Forum Acusticum 2026

8 Sept 2026, 08:00 → 12 Sept 2026, 18:00 Europe/Vienna

Tuesday 8 September

10:00 Opening Ceremony

1 Auditory modeling for realistic binaural audio

Every day, millions of people use headphones to listen to music, while watching movies or playing video games, and to communicate with others. Unfortunately, the sounds presented via headphones are usually perceived inside the listener’s head rather than at a realistic external position. This is because binaural sound reproduction via headphones typically lacks head tracking and bypasses the natural filtering of sound by the listener-specific shape of the outer ear and head. Knowing how accurately those characteristics need to be engineered in order to evoke a realistic spatial percept is of great importance for developing more efficient and effective rendering solutions. In this talk, I will review acoustical and cognitive mechanisms involved in binaural sound reproduction, and propose auditory modeling as a framework to assess the quality of binaural reproduction, showing how auditory modeling can help achieve more realistic binaural audio.

Speaker: Piotr Majdak (Acoustics Research Institute, ÖAW)

12:00 Lunch break

A02.00 Bio-acoustics: S007

Conveners: Angela Stöger-Horwath (University of Vienna, Vienna Zoo), Dick Botteldooren (U-Ghent), Michael Kerscher, Paolo Diviacco

2 Detecting Redwing (Turdus iliacus) Flight Calls in Passive Acoustic Recordings of Nocturnal Migration

Bird migration is the seasonal movement of birds between regions, often influenced by weather, food availability, and breeding needs. Passive acoustic monitoring (PAM) provides a useful way to study bird migration by recording the calls of birds passing through an area. These calls are often short and can be difficult to detect, especially when recordings contain environmental noise, overlapping sounds, or technical artefacts. In this study, we focus on the Redwing (Turdus iliacus), a migratory bird commonly detected during nocturnal migration by its characteristic flight call. We explore different approaches for automatically detecting and classifying Redwing calls in acoustic recordings. These include existing open-source bird sound recognition tools such as BirdNET, supervised methods using different preprocessing and audio representation techniques, and agile models based on Perch embeddings that support active learning. The results are discussed in terms of detection performance, practical usability, and robustness under varying recording conditions. By comparing these approaches, this study provides insight into the possibilities and limitations of automated call detection for monitoring nocturnal bird migration.

Speaker: Paul Devos (Ghent University)

3 Sonic Selva: Calibrated Ambisonic and Binaural Soundscape Baselines of Venezuelan Biotopes

The preservation of highly biodiverse acoustic environments is of critical importance given the rapid global decline of natural ecosystems. This paper presents Sonic Selva, an open-access, spatial audio dataset that captured the acoustic baselines of Venezuelan rainforests and savannas in 2024. A calibrated industrial binaural measurement system (HEAD acoustics SQobold with BHS II) and a 3rd-order Ambisonics microphone (Zylia ZM-1) were deployed. This paper details the methodological workflow across 62 measurement locations, incorporating 23 hours of 3rd-order Ambisonic and 14 hours of calibrated binaural recordings. The data is augmented by detailed measurement metadata and 3 hours of synchronized 360° video, providing a comprehensive contextual and visual ground truth for each measurement, as well as over 71 hours of continuous extended-duration Ambisonic recordings, capturing a full diel cycle as well as critical diurnal and nocturnal transitions.This paper highlights the procedures used to conduct the measurements, technical challenges encountered when operating acoustic equipment in tropical extremes as well as the overall structure of the dataset. Furthermore, the application of hardware-specific noise profiles characterized in an anechoic chamber to mitigate self-noise and restore the psychoacoustic fidelity of low-level soundscapes is examined. The resulting dataset provides a robust, open-access resource for psychoacoustic evaluation, bio-acoustic monitoring, machine learning, virtual spatial soundscape reconstruction and investigations into the effects of Neotropical soundscapes on human emotion and health.

Speaker: Marc-Laurin Schimke (TU Berlin)

Full paper (PDF)

4 Sector-based Filtering and CLEAN-SC Beamforming for Natural Soundscape Analysis : the case of the Little Owl

Acoustic monitoring has become a widely used approach to study ecosystems. However, the large amount of acquired data requires fast processing techniques to efficiently analyze soundscapes. For that, sound detection and classification algorithms, such as BirdNET, have been developed. However, the performance of BirdNET, quantified by the confidence score, is closely linked to the Signal-to-Noise Ratio (SNR) of the audio recordings. Besides, the use of Spherical Microphone Arrays (SMA), which enable rotationally invariant Spherical Harmonics (SH) beamforming, allows for directional filtering of the soundscape and sound source localization in any direction. In this work, natural soundscape analysis is performed using a SMA for recordings. In particular, the night call of the little owl (Athene noctua) was automatically searched for. Firstly, sector-based filtering is carried out to provide a set of spatially filtered signals. These signals are then analyzed using BirdNET to provide a list of sound events. Secondly, for each sound event, the CLEAN-SC deconvolution beamforming algorithm is applied to the whole 3D sound scene captured by the SMA. Directional signals are finally extracted in the direction of the sound sources and analyzed through BirdNET. An improvement of over 200% in the number of little owl detections in a natural environment has been achieved by means of sector-based filtering in comparison to a monophonic signal. After extracting directional signals, the confidence score of BirdNET has been raised by more than 50%. These results suggest that the proposed method improves the robustness of BirdNET.

Speaker: Maud Biscarat (LMFA UMR 5509 CNRS, Ecole Centrale de Lyon)

Full paper (PDF)

5 Passive acoustic filtering for automatic insect monitoring

Automated signaling of pest insects in the field could drastically improve the cost-effectiveness of integrated pest management. Acoustic monitoring of flying insects has the potential to be a low cost solution with spatio-temporal resolution for timely intervention. However, it is not used in practice due to high background noise levels, which obfuscate the insect sounds.We propose an improved trap design to enable acoustic monitoring of pest insects in practice. By tuning the acoustic filter response of the trap we aim to increase the signal-to-noise ratio to a sufficient level. We employ a multiple Helmholtz resonator configuration to attenuate noise coming from outside of the trap at the entrance. Sound recordings of flying Delia antiqua were analyzed to determine the distribution of fundamental frequencies. A side-branched Helmholtz resonator array trap was then designed to attenuate environmental noise at the wingbeat harmonics (~175 Hz multiples). The acoustic response of the full trap was tailored to the respective frequency bands with finite-element simulations, and validated experimentally. Also the position dependent response of the trap was characterized to ensure that the insect sounds produced within the trap were not deformed as much as to make recognition fail. Simulations predicted a signal-to-noise ratio improvement of 10-25 dB in the frequency bands of interest, while experiments demonstrated gains between 5-20 dB. The dependence of the microphone signal on the position of the insect in the trap was found to be sufficiently smooth to avoid that effects of the insect position would confuse recognition.

Speaker: Jef Morlion (KU Leuven)

A03.00 Building Acoustics: S460

Conveners: Teresa Carrascal García (Instituto Eduardo Torroja for Construction Sciences - IETcc-CSIC), Chiara Scrosati, Heinz Ferk, Selina Vavrik-Kirchsteiger, Catherine Guigou-Carter (CSTB)

6 Simulation and Experimental Validation of Sound Transmission Loss in Stone Wool Sandwich Panels

This paper investigates the acoustic performance of stone wool sandwich panels with steel facings at two different thicknesses. Sound transmission loss is predicted using a poroelastic model based on Biot theory, which accounts for the anisotropic structure of stone wool. Three modeling approaches are compared to assess the effect of material parameters: a simplified poroacoustic model using JCA or Delany-Bazley parameters, an isotropic Biot poroelastic model, and an orthotropic Biot poroelastic model that captures directional material properties. Finite element simulations predict sound transmission loss across different frequencies, incorporating both steel plate and stone wool poroelastic parameters. Boundary conditions are also explored to understand the best approach to predict experimental results. Experimental measurements of airborne sound transmission loss for the two sandwich panel thicknesses are compared with the numerical predictions. The study evaluates how accounting for full poroelastic parameters, boundary conditions, and material anisotropy affects prediction accuracy and identifies the influence of each modeling assumption. Results demonstrate the relevance of poroelastic modeling over simplified poroacoustic approaches and provide guidance on when material anisotropy must be considered for accurate acoustic design of sandwich panels.

Speaker: Maria Monica Ballesteros Villarreal (ROCKWOOL A/S)

7 Statistical analysis of sampling strategies for sound insulation verification: A case study

The in-situ verification of airborne and impact sound insulation in buildings relies on acoustic measurements that are necessarily limited in number. In practice, each country defines its own sampling approach based on local construction practices and economic feasibility. Typically, representative rooms are tested instead of the entire building, assuming repetition in architectural layout and construction solutions.This paper analyses the relationship between sampling strategy, number of acoustic tests, and the accuracy of sound insulation assessments in residential buildings. A reference building model is used to generate a complete set of synthetic measurements by combining calculated sound insulation values with realistic experimental uncertainty. A bootstrap resampling approach is applied to quantify the deviation introduced by reduced sampling schemes and to identify optimal testing strategies for compliance verification.The aim is to identify sampling strategies that balance measurement effort and reliability, providing quantitative criteria for minimum sampling. To the authors’ knowledge, no prior study has systematically addressed the statistical basis of acoustic sampling in buildings. The main limitation of the study is that it relies on calculated acoustic data. Nevertheless, the validity of this assumption should be confirmed through a dedicated experimental measurement campaign.

Speaker: Teresa Carrascal García (Instituto Eduardo Torroja for Construction Sciences - IETcc-CSIC)

Full paper (PDF)

8 On Statistical Sampling in Standardized Flanking Transmission Measurements

A recurring issue in standardized building acoustic measurements is determining the number of sampling positions required to achieve a prescribed confidence interval. Although measurement uncertainty can be reduced by increasing the number of samples, practical constraints limit the feasible number of excitation and measurement positions. This paper examines the measurement of the velocity level difference of rigid heavyweight junctions according to ISO 10848-1, currently under revision. Monte Carlo simulations with a wave based model are used to quantify the error and uncertainty associated with different numbers of excitation and measurement positions. While the results are influenced by junction type, transmission direction and plate properties, clear general trends emerge. Increasing the number of excitation positions is most effective in reducing uncertainty, whereas increasing the number of measurement positions beyond four provides only limited additional improvement. General formulations are proposed to determine the required sampling for a specified maximum allowable error or uncertainty. The current ISO 10848-1 requirement for Type A elements, namely a minimum of four excitation positions and three measurement positions per excitation position, appears to be an efficient compromise between accuracy and practical feasibility.

Speaker: Arne Dijckmans (Buildwise)

Full paper (PDF)

A10.06/A12.10 Acoustic wave propagation in complex media: S065

Conveners: Tristan Lawrie (University of Exeter), Malte Peter (University of Augsburg)

9 Engineering Complex Dispersion Relations with Beyond Nearest Neighbour Couplings

We present a non-Hermitian inverse-design protocol for engineering wave propagation in non-local mass–spring lattices, enabling the creation of exotic dispersion phenomena in elastic media. By extending inverse-design methods to complex frequencies and incorporating velocity-dependent, non-conservative behaviours in the governing equations, both the real band structures and their associated attenuation can be precisely tailored. Our approach demonstrates apparent wavenumber band gaps and compact Fourier representations of target functions, suitable for both passive and active systems.

Speaker: Robyn Edge (University of Exeter)

10 Natural Hyperuniform Structures: Hidden Order in Diatom Shells & Metamaterial Implications

As the most common type of phytoplankton, diatoms are unicellular microalgae responsible for producing about 20-25% of atmospheric oxygen, playing a central role in global carbon and silica cycles. Their siliceous shells, called frustules, exhibit complex porous micro and nano-architectures combining remarkable structural regularity yet with apparent biological disorder. Despite the great inter-disciplinary attention brought upon these organisms, the apparent order in frustules is lacking consistent systematic characterization. In this work, we provide a quantitative order classification of these natural microstructures using the framework of hyperuniformity. Born from statistical and material physics, the latter helps describe and characterize the orderliness of complex media at the boundaries of order and disorder; a compelling application for analyzing and classifying disordered yet regular patterns, such as those of diatoms. High-resolution scanning electron microscopy (SEM) images from 21 diatom genera spanning the 3 major taxonomic classes were analyzed through complementary real- and reciprocal-space metrics, including spectral density scaling and bond-orientational order parameters. All investigated frustules were found to display hyperuniform organization, distributed mostly as strong (Class I) and weak (Class III) hyperuniform regimes. This taxonomy reveals how biologically evolved porous architectures balance local geometric order with large-scale disorder resilience. Beyond evident statistical and biological interests, the results suggest that diatom frustules constitute a natural library of architected hyperuniform media, known for featuring exotic wave physics features found in crystals, such as isotropic Bragg bandgaps and topological mode protection. As such, diatoms can offer design inspiration for lightweight structures and disorder-based crystals and metamaterials with tailored mechanical or wave-interaction properties.

Speaker: Chiara Gazzola (Politecnico di Milano)

11 Tracking Rayleigh–Bloch Waves in Periodic Arrays of Penetrable Scatterers

Rayleigh–Bloch waves are guided modes localized around periodic arrays with unbounded unit cells, and they play an important role in resonant wave scattering by diffraction gratings. In this work, we extend a Riemann-sheet tracking framework for Rayleigh–Bloch waves to two-dimensional acoustic gratings composed of penetrable scatterers. The problem is formulated as a quasi-periodic transmission eigenvalue problem, where the exterior and interior acoustic fields are coupled through continuity conditions on the scatterer boundary. Using quasi-periodic Green’s functions, we derive a boundary-integral formulation that allows selected plane-wave components to become unbounded in the far field, thereby indexing different Riemann sheets of the Bloch wavenumber. The resulting nonlinear eigenvalue problem is solved using the block Sakurai–Sugiura method, enabling robust continuation of Rayleigh–Bloch wavenumbers as the frequency and material contrast vary. Numerical examples demonstrate how penetration into the scatterers modifies the cut-on, cut-off and interaction behaviour of Rayleigh–Bloch waves, including transitions between physical and non-physical sheets. The results reveal qualitative differences from the sound-hard case and provide a systematic way to analyse guided and near-resonant modes in periodic structures with finite material contrast.

Speaker: Kei Matsushima (Hiroshima University)

12 A Modal Analysis of Rainbow Trapping vs Rainbow Reflection

The term rainbow trapping commonly refers to using graded arrays for amplification of waves at chosen locations according to frequency. In the context of wave-energy harvesting, Chaplain et al. (NewJ. Phys. 22:063024, 2020) showed the benefits of grading such that amplifications are associated to bandcrossings within Brillouin zones rather than at their edges, referring to the latter as rainbow reflection todistinguish it from the rarer trapping cases. Here, we revisit the differences between rainbow trapping andreflection using the concepts of Rayleigh–Bloch waves and transfer matrices.

Speaker: Malte Peter (University of Augsburg)

A11.00 Musical Acoustics: S467

Conveners: Vasileios Chatziioannou (Univ. Music and Performing Arts Vienna), Charalampos Saitis, Christoph Reuter (Universität Wien)

13 Acoustical Analysis of a Hungarian Folk Flute

Sound generation by air reed instruments is an active area of research. Studies on various instruments continue to reveal new details regarding the sound production mechanism. While a substantial body of knowledge is available on classical instruments, relatively few studies have focused on folk instruments. Hungarian folk flutes exhibit remarkable design diversity and can be grouped into regional types, each associated with distinct performance practices.This study presents an acoustical analysis of a Hungarian folk flute, specifically a Transylvanian shepherd type. The investigation compares one-dimensional transmission matrix models and three-dimensional finite element simulations of the resonance properties of the air column with sound recordings obtained under laboratory conditions. In addition to documenting the acoustical features of the flute, the study also reveals interesting resonance properties of the air column.

Speaker: Péter Rucz (Budapest University of Technology)

Full paper (PDF)

14 Investigation of internal flow in a bass recorder with various tone hole configurations

Tone holes are one of the primary mechanisms for pitch control in many wind instruments, yet their acoustics are not yet fully understood. This study investigates air flow in a bass recorder with an open tone hole using time-resolved particle image velocimetry (PIV). To enable the measurements, the instrument was constructed by attaching a commercial bass recorder mouthpiece to a flue pipe with a square cross-section and transparent walls. The total length of the instrument is designed to match the lowest tone a standard bass recorder produces (note F3). Standing wave behaviour was evaluated for two tone hole positions and for five diameters, tone holes opened individually. The positions and some of the diameters were selected to produce musically relevant tones that are by design referencing fingerings used in an actual baroque bass recorder. A detailed look at the flow within these configurations allows to analyse the tone hole behaviour and establish a foundation for future research.

Speaker: Titas Lasickas (Univ. Music and Performing Arts Vienna)

15 Measurement and Analysis of Acoustic Indicators Across Trumpet Players Specialised in Different Musical Styles

This study investigates the relationship between the performance style and the acoustic features of the radiated sound in trumpet playing, with a focus on the variation of objective descriptors depending on the main musical style of the participant. To that aim, a total of 13 trumpet players, representing three primary stylistic domains (classical, jazz, and commercial), participated in controlled anechoic recordings. Each performer played the same musical excerpt using three differently style-oriented trumpets, covering the full playable range of the instrument and performing at two dynamic levels: piano and forte.The measurement setup enabled multi-positional acquisition of the radiated sound, allowing specific-point and spatial analysis of acoustic output. From the recorded signals, a set of acoustic and psychoacoustic descriptors was computed for both dynamics. These descriptors include Sound Pressure Level, spectral content, High-Frequency Energy Ratio and psychoacoustic metrics Loudness and Sharpness.The results reveal significant correlations between acoustic signal emitted by the trumpets and the performers’ stylistic background. In particular, variations in spectral distribution, specifically high-frequency content, related to harmonic enrichment. These findings contribute to a deeper understanding of how the performer musical style influences the radiated sound of brass instruments and establish a quantitative framework for comparing stylistic traits in trumpet performance. The study emphasizes the importance of style-oriented musical research in instruments and opens the door to more specific acoustic virtualizations of these musical instruments, improving the realism of virtual reality applications.

Speaker: Juan Luis de la Torre Moral (Grupo Acústica Arquitectónica, Universidad Politécnica Madrid)

Full paper (PDF)

16 Acoustic profiles of choral, popular, and amateur singing in a controlled sustained-vowel corpus

This study compares acoustic profiles of choral, popular and amateur singing using a controlled corpus of 800 sustained-vowel recordings produced by 20 singers on eight pitches of a C-major scale. Studio recordings were analysed using intonation, vibrato, perturbation, and selected formant-related and spectral features. The aim of the study was to identify interpretable acoustic features that vary across styles under controlled phonetic and pitch conditions. As each singer belonged to one style group, comparisons used singer-level median feature values. The clearest style-related difference was observed for absolute intonation error, which differed significantly between groups and was lowest in choral singing: 8.1 cents, compared with 19.7 cents in popular and 25.2 cents in amateur singing. Other descriptors (amplitude stability and vibrato-related) showed consistent but weaker tendencies. Median shimmer was lowest in choral singing, 0.127 dB, compared with 0.163 dB in amateur and 0.224 dB in popular singing. Median vibrato frequency was also higher in choral singing (5.07 Hz) than in popular (4.77 Hz) and amateur singing (4.63 Hz), but these differences did not reach statistical significance after singer-level correction. These findings support the search for interpretable acoustic features that differentiate singing styles and describe stylistic aspects of vocal technique in an interpretable way. The results suggest that controlled sustained-vowel recordings can reveal style-related tendencies, particularly in pitch accuracy, amplitude stability and vibrato-related features, with weaker evidence for spectral features, while also demonstrating that such differences should be interpreted with caution due to substantial singer-specific variability and partial overlap between style groups.

Speaker: Agnieszka Pietrzak (Warsaw University of Technology)

17 Asymmetric Pitch Control in the Vietnamese Monochord Dan Bau: Implications for Playing Technique

The Dan Bau (in Vietnamese: Đàn Bầu) is a Vietnamese monochord zither that is played by exciting the harmonics of the string by 1) dampening the string at a position of an antinode with the ulnar side of the palm, followed by 2) plucking the string in an upwards motion with a long plectrum while 3) releasing the dampening. Changing the tension of the string by inflecting the bending rod on which the string is mounted allows the Dan Bau performer to play pitches outside the harmonic series of the string and to add ornamentation. We investigated the relationship between the force applied to the bending rod, its horizontal displacement and the resulting pitch. To this end, we used a manual XY translation stage to inflect the bending rod. A load cell mounted on the translation stage measured the force applied to it. We recorded the sound of the Đàn Bầu after bending it in 5 mm steps in both directions along its axis. Excitation is accomplished using a pluck mounted on a linear guide. We found that the force needed to alter the pitch varies with the bending direction. Due to physical constraints, a player would need less force to decrease the pitch by a certain interval than to increase the pitch, which introduces an asymmetry in playing technique with the bending rod. These results may inform the development of new digital musical instruments that borrow from the gestural repertoire of the Dan Bau.

Speaker: Tim-Tarek Grund (Department of Music Acoustics – mdw)

A14.07 Computational and model-based approaches to hearing science: S104

Conveners: Helia Relaño-Iborra (Eriksholm Research Centre), Vaclav Vencovsky (Czech Technical University in Prague)

18 Computational Models of the Sub-Cortical Auditory System with Efferent Pathways

Computational models of auditory-nerve responses provide accurate predictions of neural responses to complex sounds in anesthetized animals. In these recordings, the activity of efferent neurons is assumed to be suppressed. The role of efferent control of cochlear gain in neural coding has been further explored in recent models that include inputs to medial olivocochlear efferents from brainstem and midbrain sources. This talk will provide an overview of our recent work on these models.Inclusion of efferent pathways requires a restructuring of the afferent models to allow continuous variation of cochlear gain. Our initial model (Farhadi et al., 2023) is a single-channel implementation of feedback from both brainstem and midbrain levels. This model focuses on the balance between the reflex-like, negative feedback provided by brainstem inputs and the fluctuation-driven positive feedback provided by midbrain inputs. However, the single-channel structure limits the ability of this model to explore neural representations across populations of neurons. A more recent model (Guest et al., 2026) focuses on the brainstem-level feedback, including tonotopically distributed (i.e. multi-channel) control of cochlear gain by individual medial olivocochlear neurons. This model can simulate the frequency-specific sensitivity and strength of the effects of contralateral elicitors of efferent effects that have been demonstrated in auditory-nerve recordings, including the effect of elicitors that are tonotopically distant from probe tones. The implications of these models for neural coding of complex sounds will be discussed.

Speaker: Laurel H. Carney (University of Rochester)

19 A computational framework for the simulation and interpretation of auditory evoked potentials in healthy and pathological human ears

Although auditory evoked potentials (AEP) are widely used in clinical settings, the complex relationship between AEP morphology and cochlear status limits their diagnostic specificity. Computational models of AEPs grounded in cochlear mechanics can help link AEP morphology to underlying pathology. However, no validated framework has yet demonstrated this systematically across clinically relevant stimuli and pathologies.We developed an AEP modeling framework that combines a state-of-the-art computational model of the human auditory nerve (AN) with a convolution stage using a unitary response (UR) to simulate population-level responses. Simulated AEPs in healthy ears were validated against existing experimental datasets using transient and periodic stimuli presented across multiple stimulus intensities. We then simulated AEPs to standard audiological stimuli for four isolated pathology types – inner hair cell loss, outer hair cell loss, auditory nerve neuropathy, and myelinopathy – as well as for realistic profiles derived from human post-mortem histopathological data.This framework successfully simulated a broad range of AEPs across electrode configurations and captured stimulus- and level-dependent patterns observed experimentally in healthy peripheral auditory systems, despite some limitations in reproducing latency-level functions. The modeling approach thus enables systematic investigation of how different cochlear pathologies, such as neural or outer hair cell loss, manifest as distinct AEP patterns. These large simulated datasets can, in turn, provide a foundation for training machine-learning models to predict individual cochlear damage profiles from clinical AEP recordings.Together, these tools represent a step toward model-informed, electrophysiology-based precision diagnostics for cochlear hearing loss.

Speaker: Miguel Temboury-Gutierrez (Hearing Systems section, Technical University of Denmark)

20 Approaching the Hopf bifurcation largely increases the cochlear response

This work investigates cochlear amplification in a transmission line model whose local elements are critical oscillators operating near a Hopf bifurcation. Here, starting from a physically based micromechanical description of the cochlea, the governing equations are reformulated into normal form, allowing the parameters of the reduced model to be directly related to measurable physical quantities. A transmission-line cochlear model including two-dimensional fluid coupling is solved in time domain with the state space technique and pressure Green’s functions. The analysis further demonstrates the rapid growth of the cochlear response as the system approaches the Hopf instability. The distance from the bifurcation is expressed in terms of physiologically meaningful parameters, such as the strength of the active force and the cutoff frequency associated with the outer hair cell transmembrane potential. The proposed formulation therefore establishes a direct connection between abstract nonlinear oscillator theory and the physical processes governing cochlear mechanics. Preliminary results show how approaching the critical regime it is possible to get large gain dynamics in the peak region. Interestingly, although the nonlinear force is coupled to the relative mode only, both modes show the same nonlinear behavior, due to the back-action on the pressure associated with the focusing phenomenon.

Speaker: Renata Sisto (INAIL Research)

Full paper (PDF)

21 (Non)Linear Growth and Self-Oscillation in Auditory Biophysics

It has been proposed that a "single strategy" can account for numerous characteristics of the active ear [e.g., spontaneous otoacoustic emission (SOAE), frequency-specific amplification/compression]. The basic heuristic employed essentially amounts to an individual nonlinear oscillator, describable by a single ordinary differential equation, that has the propensity to exhibit a limit cycle oscillation (i.e., self-oscillation). A common benchmark to assess the validity of such a supercritical "Hopf Oscillator" (HopfO) is to compare its level growth response with respect to a sinusoidal drive to experiment, seeking out a specific degree of compressive behavior. However, while empirical studies generally report nonlinear responses at moderate to high sound levels, motions appear linear closer to threshold. The current study examines these foundations, taking a two-fold approach combining both computational and empirical observations. First, we simulated various scenarios of a noise/sinusoidally-driven HopfO (e.g., vicinity of a bifurcation, degree of noise and detuning). Our results are broadly consistent with, but expand upon, those of O Maoileidigh & Hudspeth (2018). Further, we observed a narrow region over which entrainment occurs that creates significant intermodulation distortion. Second, we report novel vibrometry measurements from the Anolis lizard tympanic membrane that show spontaneous oscillations with displacements on the order of 10 pm and are commensurate with SOAE activity. Crucially, sound-evoked level growth near (as well as away from) those frequencies tended to be highly linear. Linking both paths raises critical questions for the validity of a "single strategy' that relies upon an individual HopfO.

Speaker: Christopher Bergevin (York University)

A15.00 Psychoacoustics: S106

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Kai Siedenburg (CvO University of Oldenburg), Jan Rennies (Fraunhofer IDMT-HSA, Oldenburg), Sarah Verhulst (Ghent University), Bernhard Laback (Austrian Academy of Sciences)

22 Association between psychophysical inferred cochlear response functions and loudness categorical data

In general, loudness growth functions of normal-hearing (NH) listeners are compressive for medium to high stimulus levels. This compressive characteristic is commonly associated with the nonlinear input-output (I/O) function of the healthy cochlea. The present study investigated the relationship between loudness growth functions and cochlear response functions in NH listeners using psychoacoustic experiments. Listeners had to have pure-tone audiogram thresholds (0.25-8 kHz), tympanometry and stapedius reflexes within normal limits. Categorical Loudness Scaling (CLS) was used to measure loudness growth. The stimulus was a one-second duration speech-shaped noise stimuli. Listeners rated the stimuli on an 11-point categorical scale. Each of these categories is assigned a numerical value referred to as a Categorical Unit (CU). Loudness in CU was converted to sone and a nonlinear function was fitted to the data. Human inferred cochlear response I/O functions were obtained using the fixed-duration masking curve (FDMC) method. The function type was the same as used for the loudness growth function. Correlation analyses were undertaken to understand the relationship between the inferred cochlear measures obtained from the I/O response functions and the loudness growth in sone from the same listeners.

Speaker: Jesko Verhey (Otto von Guericke University Magdeburg)

23 Effect of interaural delay on fundamental frequency discrimination of harmonic complex tones

Fundamental frequency difference limens (F0 DLs) for harmonic complex tones are much lower when the stimulus contains low-ranked harmonics (typically lower than 7 or 8) than when it only contains high-ranked harmonics (> 12), suggesting good performance requires the presence of spectrally resolved components. Here, we show that within a certain stimulus parameter range, introducing an interaural delay within binaural harmonic complex tones can decrease F0 DLs substantially, despite producing no change in spectral resolvability.12 normal hearing participants took part. In Experiment 1, F0 discrimination was measured for two binaural stimuli consisting of the same harmonic complex tone with 12 consecutive components presented in both ears. In one condition (SIM), the tones were presented simultaneously in both ears whereas in another condition (DELAYED), an interaural delay of half the stimulus period was imposed. F0 DLs were measured for 2 different F0s (50 and 280 Hz) and 5 different harmonic contents (rank of the first harmonic N ranging from 4 to 16). For F0=50 Hz and N<10, DELAYED provided significantly lower DLs than SIM. Experiment 2 showed that DELAYED also supported better melody recognition than SIM. Similarities of this phenomenon with other dichotic pitch phenomena will be discussed.

Speaker: Olivier Macherey (LMA-CNRS)

24 Tonal vs. Noise-Like Sharpness: An Individual Vocabulary Profiling Study

Sharpness is a well-established psychoacoustic attribute describing high-frequency spectral balance, but current objective measures treat it as a single dimension regardless of tonal content. This study was motivated by whether tonal and noise-like signals give rise to perceptually distinct types of sharpness and whether metrics should accommodate this distinction.Fourteen loudness-matched stimuli (pure tones, harmonic tone complexes, filtered noise, and tone-in-noise mixtures) span a two-dimensional space of sharpness and tonality. Selected pairs are contrasted by their tonal or sharpness content and participants are asked to generate their own descriptive vocabulary through pairwise comparison, then rate every stimulus against their elicited constructs. Constructs pooled across participants are grouped by hierarchical agglomerative clustering and related to standard psychoacoustic metrics for sharpness and tonality.This study contributes to a wider project developing an improved model of psychoacoustic sharpness. The perceptual constructs identified here will inform the model so that it better reflects how sharpness is perceived across signal types.

Speaker: Matt Torjussen (University of Salford)

Full paper (PDF)

25 Just noticeable difference and absolute thresholds of psychoacoustic roughness for simple reference and complex drone signals

The psychoacoustic metric of roughness is key to many acoustic evaluations. Although several calculation methods are available, each method yields slightly different roughness values for different signal and modulation types. In addition, the practical interpretation of roughness values is often unclear due to the limited knowledge of absolute and just noticeable difference thresholds, especially for complex technical sounds, which are typically subject of roughness analysis. To broaden the empirical basis on thresholds of roughness, a listening test was designed and conducted. Using a constant-stimuli procedure, thresholds were determined for both simple reference signals, including amplitude modulated pure tones and broadband noise, and more complex synthesized drone sounds. Thresholds were derived for three commonly used roughness calculation methods (ECMA 418-2, Zwicker and Daniel & Weber). Although threshold values depended on signal type and calculation method, low roughness values were associated with low thresholds, whereas higher roughness levels resulted in higher thresholds.

Speaker: Felix Hochbaum (TU Berlin, Department of Engineering Acoustics)

Full paper (PDF)

26 Individual acoustic preference in self-assembled sleep soundscapes: a population-scale field study

Functional sleep soundscapes are a mainstream listening category, yet which acoustic properties keep users returning is unknown. We analyze twelve months of replay data from Sonoraa: Sleep and Focus Sounds (77'408 users), reconstructing each user's self-assembled sound mixes from event logs (app-initiated audio excluded) and computing every mix's spectrum from calibrated spectra of its constituent sounds. Classical psychoacoustic descriptors (ISO-532-1 loudness, sharpness, tonality) carry little predictive power: no association with return frequency survives control for usage intensity, and they separate day-one churn only weakly (AUC 0.59). In a mel-cepstral representation, however, the content of a user's first-day mixes alone (source categories plus spectral shape, no usage information) predicts whether that user ever returns (AUC 0.69). Within-person mixed models on 2'600 multi-session users show that the coupling between a session's acoustics and its listening time is strong but individually signed: between-user variation in this coupling is seven to nine times the population-mean effect, and is organized along two largely independent axes, level response and brightness response. Opposite individual signs cancel in any population average, which is why group-level analyses of sleep soundscapes find so little. There is no universally optimal sleep soundscape, and fixed-stimulus designs in sleep-noise research may underestimate effects for the same reason. A within-user randomized field study is planned.

Speaker: Stefan Schucker (Empa Acoustics and Noise Control 509)

Full paper (PDF)

A15.05 Psychological and physiological responses to environmental and product noise: S110

Conveners: Massimiliano Masullo (Università degli Studi della Campania “Luigi Vanvitelli”), Margret Sibylle Engel (Dresden University of Technology)

27 Modeling Short-term Annoyance due to Wind Turbine Noise: Contribution of some Influential Acoustic and Non-acoustic Factors

Short-term annoyance caused by wind turbine noise (WTN) was examined under controlled laboratory conditions, focusing on some acoustic and non-acoustic factors. Forty-five participants were exposed to WTN stimuli stemming from in-situ recordings. Annoyance ratings were gathered alongside individual measures of noise sensitivity and attitude toward wind turbines. Various acoustic factors were investigated from the stimuli, such as sound intensity, amplitude fluctuation, and low-frequency content. The relative contribution to annoyance of the studied acoustic and non-acoustic factors was assessed using multilevel regression modeling. All multilevel models built included noise sensitivity and attitude toward wind turbines at the individual level. Among the acoustic and psychoacoustic indices, LAeq emerged as the dominant variable. However, metrics capturing amplitude fluctuations—such as roughness and the temporal derivative of loudness—provided additional explanatory value beyond sound level alone. In contrast, the contribution of the IOA amplitude modulation metric was minor. Low-frequency energy, highlighted by specific loudness integrated between 2 and 4 Barks, was also significantly related to annoyance, though its influence remained moderate.

Speaker: Tom COLAS (ENTPE)

28 Exploring the Influence of Spatial Attributes and Visual Cues on Psychoacoustic Annoyance Using Virtual Reality

This study builds upon a previous investigation into the effect of source azimuth on the perceived annoyance of domestic sounds. A novel experimental testing method was employed to further evaluate the observed directional effect, in addition to assessing the influence of room acoustics and visual cues. In a virtual reality (VR) listening experiment, 33 participants were presented with several stationary sound sources situated within three room simulations, each with differing acoustic characteristics. The stimuli were presented twice, once with a visual simulation of the room and source, and once without (blackout). The participants identified the most and least annoying positions independently for each stimulus and rated their perceived level of annoyance. The results not only verified the initial findings but also revealed some additional insights. For example, the directional responses were consistent with the original study, where the majority found sources around ±90° to be the most annoying, with 180° and 0° causing the least annoyance; however, some participants found that ~0° was the most annoying position, influenced by visual and acoustic factors. The difference between average perceived annoyance responses for the most and least annoying positions was significant for all stimuli, again reinforcing the previous results; although the acoustic conditions of the room also affected annoyance levels. The results validate the methodology approach for obtaining meaningful results in annoyance and soundscape quality studies. They also offer a deeper insight into the influence of spatial features and audiovisual interaction on psychoacoustic annoyance, with potential applications in annoyance modelling and environmental design.

Speaker: Samuel Dickinson (University of Huddersfield)

Full paper (PDF)

29 Effects of road traffic noise on cognitive performance in undergraduate students

Urban noise, particularly road traffic noise, negatively affects physiological and psychological health. This study provides novel evidence on how road traffic noise impacts attention, concentration, memory, and verbal fluency. Sixty-one undergraduate students (34 females, 27 males) completed a brief interview and cognitive tasks under two conditions: with experimental traffic noise (ETN) and without (NEN). To control for order effects, half began with ETN, half with NEN.Results showed males outperformed females in visual detection and auditory verbal memory. A significant main effect of treatment order was found in memory performance, with order 2 (NEN first) achieving higher scores than order 1 (ETN first). Significant sex × treatment order interactions were observed in memory (third repetition) and semantic verbal fluency, indicating that noise exposure sequence differentially affects cognition by sex. A significant main effect of treatment was also observed in semantic verbal fluency, with better performance under ETN than NEN.

Speaker: Horeb Hesiquio Sevilla Pérez (Universidad Autónoma de Tlaxcala)

30 Cognitive load under urban soundscapes: A physiological approach using digit span tasks

The effect of soundscape on cognitive performance remains an active area of inquiry in psychoacoustics, yet few studies combine behavioural measures with continuous physiological monitoring. This study is the first part of an audiovisual experiment which examines whether urban soundscape conditions modulate cognitive load during a digit span task, as reflected in performance outcomes and autonomic physiological responses. Twenty-seven participants (university students and academic staff) performed forward and backward digit span tasks while exposed to four of five urban soundscape conditions, including traffic, traffic masked with water sounds, traffic masked with simple music, traffic masked with sleep music, and traffic with dynamic loudness variations. Each participant was exposed to four of the five soundscape conditions. Electrodermal activity and heart rate were recorded continuously throughout. Task periods were preceded by two-minute silent baseline recordings, enabling baseline-corrected physiological comparison between forward and backward task conditions. Forward and backward digit span scores are compared across soundscape conditions as a measure of short-term memory retention and working memory manipulation, respectively. Skin conductance response frequency and mean heart rate, corrected to a silent baseline, are examined as indicators of autonomic arousal during each task condition. The higher cognitive demand of the backward task is expected to produce both lower performance scores and elevated physiological arousal relative to the forward condition. This methodological framework contributes to understanding how urban soundscapes interact with cognitive load, offering a combined behavioural and physiological perspective relevant to soundscape design in work and study environments.

Speaker: Margret Sibylle Engel (Dresden University of Technology)

Full paper (PDF)

31 When sounds bring back memories: A psychophysiological approach

Soundscapes influence emotional and cognitive states, yet their capacity to trigger autobiographical memory remains underexplored in psychoacoustics. This study is the second part of an audiovisual experiment investigating the relationships among soundscape perception, autobiographical memory retrieval, and soundscape arousal. Twenty-seven participants (university students and academic staff) were exposed to four of five scenes: 1. urban traffic, traffic masked 2. with water sounds, 3. with simple music, 4. with sleep music, and 5. with changes in loudness. Electrodermal activity and heart rate were recorded continuously. Following each scene, participants completed a structured memory elicitation protocol assessing whether the soundscape triggered autobiographical memory across eight associative dimensions: meaningful events, persons, places, activities, objects, time periods, thoughts, and emotions. A memory profile heatmap is proposed to visualise the proportion of participants who trigger each memory dimension in each scene, enabling direct comparison across conditions. Skin conductance response frequency and tonic electrodermal level slope are examined in relation to memory retrieval using linear mixed-effects models that account for the repeated-measures structure of the data. This multimodal framework bridges soundscape perception, autobiographical memory research, and objective physiological measurement, offering a novel perspective on how urban sound environments are experienced and remembered. Implications for soundscape design and evaluation are discussed.

Speaker: Margret Sibylle Engel (Dresden University of Technology)

Full paper (PDF)

A16.05 Performing Arts Spaces and Rehearsal Rooms: S117

Conveners: Francesco Martellotta (Politecnico di Bari, DARCOD), Jamilla Balint (Rohde Acoustics)

32 Towards Characterizing Irregular Loudspeaker Arrays in Diverse Acoustic Environments

The expansion of immersive audio into multipurpose venues, such as extended reality facilities, music studios, and cinemas, poses a challenge to the reproducibility of spatial audio content. Unlike standardized listening rooms, these environments exhibit unique geometric and acoustic constraints. As a highly transportable and effective method for describing and reproducing sound fields, Higher-Order Ambisonics is often preferred for immersive audio applications. However, multipurpose loudspeaker setups often feature non-uniform speaker distributions and vary in room acoustics. Ambisonic decoders typically do not take the room into account, and standard system calibration relies primarily on single-channel metrics. Consequently, there is a methodological gap in diagnosing and predicting how room acoustics distort spatial audio rendering. This paper is a comparative case study of six diverse venues in Switzerland. First, we compare vector measures based on an All-Round Ambisonic Decoder calculated for the selected loudspeaker layouts. Then, we analyze spatial room impulse responses measured on-site using the Spatial Decomposition Method and demonstrate how early reflections may create spatial distortions. This contribution is a foundational step toward an open-access dataset of multipurpose loudspeaker arrays. These measurements will drive future listening tests to assess the performance of the physical arrays and the reproducibility of spatial reproduction algorithms.

Speaker: Georgios Papadimitriou (ETH Zurich)

Full paper (PDF)

33 Towards an in-situ perceptual experiment in Early Reflection Levels in a Concert Hall

The perceptual characteristics of early reflections in concert halls have been extensively studied using controlled experiments based on measured or simulated spatial room impulse responses (SRIRs). In contrast, the present study adopts an in-situ approach, in which participants were invited to the Detmold Concert Hall to evaluate perceptual changes induced by controlled modifications of the sound field. The hall’s wave field synthesis (WFS) system, consisting of 300 flat-panel loudspeakers, was used to introduce first-order early reflections. It was hypothesized that increasing early reflection levels would lead to measurable enhancements in perceptual attributes. Contrary to expectation, clarity decreased with increasing reflection level, indicating that both early and late reflections were affected by the modifications. Analysis of the SRIRs using the Spatial Decomposition Method (SDM) revealed an increase in late reflection energy, attributed to loudspeaker directivity. The hall’s WFS system is primarily designed for artificial diffuse reverberation, and the results highlight its limitations when used to modify early reflections. Despite these deviations, psychometric analysis revealed clear discrimination thresholds in the similarity task that varied across reflection configurations, whereas judgments of perceptual attributes (clarity, width, envelopment, and loudness) exhibited weaker trends.

Speaker: Otavio Colella Gomes (Hochschule für Musik Detmold)

Full paper (PDF)

34 The Influence of Ceiling Shape on Spatio-Temporal Sound Energy Incidences in Concert Halls

The architectural and acoustic designs of concert halls influence their overall appearance and aesthetic appeal. However, it can be challenging to achieve both visual appeal and acoustic excellence. Concert halls are often categorized as vineyard or shoebox types, which are specified by the relationship between the orchestra and the audience, the surrounding surfaces, and the general shape of the space. In contrast, ceiling shapes can vary independently of typology and range from horizontal to curved to tilted.This paper examines how ceiling geometry affects the spatial and temporal distribution of sound energy to listeners in concert halls. A parametric study is presented in which the shape, height, and surface properties of the ceiling are varied using the ray tracing tool misuka, which was developed at TU Berlin. Substantial differences in sound energy decay were found when comparing non-flat (e.g., pyramidal) and flat ceilings, depending on their surface structure. Conversely, only minor changes were observed in established sound strength parameters. Furthermore, changes in spatio-temporal sound energy incidence are discussed.

Speaker: Linus Staubach (Müller-BBM Building Solutions GmbH)

Full paper (PDF)

35 The Impact of Large-Scale Orchestra Canopies on Spatial Sound Energy Decay

Overhead reflector (canopy) systems are critical in shaping the acoustic environment for symphony orchestras. While previous research has focused on the timing, strength and frequency content of first-order reflections and ensemble conditions, the influence of canopies on the late reverberant field remains under-explored. This study investigates how large-scale monolithic canopies affect the spatial development of the sound field, hypothesising that high stage-coverage ratios lead to a partial acoustic decoupling of the volume above the canopy from the primary hall volume. Empirical observations made during the commissioning of variable-acoustics concert halls suggest that canopy height alters the late reverberant sound, not only the early reflected sound. To quantify these effects, spatial room impulse responses (SRIRs) were captured in a concert hall across multiple canopy height configurations using a spherical microphone array. The RIRs were encoded into Higher-Order Ambisonics (HOA) to provide high-resolution spatial analysis and facilitate accurate auralisation. This work presents an investigation of the spatial energy distribution and late-part decay characteristics associated with changes to canopy height. The findings will be used to guide future studies in this area, including perceptual studies aimed at understanding how musicians and audiences perceive the orchestral sound in relation to overhead reflectors.

Speaker: Matthew Gray (Acoustics Research Centre)

36 Stage Acoustics for the Symphony Orchestra: a Laboratory Study with Auralised Rooms

A central challenge in the design of music venues concerns the architectural features of the stage. Previous field studies have shown that, provided the overall shape and size of a hall are suitable for symphonic music, appropriately designed stage enclosures can generate beneficial reflection patterns to complement the direct sound distribution within the orchestra. These reflections are essential for enabling musicians to hear themselves and one another effectively during performances. While some significant tendencies have emerged in terms of architectural measures, the numerous confounding factors inherent in field investigations have hindered the establishment of robust and generalisable relationships between architectural or acoustic parameters and perceptual qualities relevant to symphony orchestra performers. Building on a diffraction pipeline for geometrical acoustics specifically designed to accurately reproduce the sound propagation occurring within a symphony orchestra, a listening experiment was conducted in which orchestra musicians performed symphonic music in auralised rooms under controlled conditions. By systematically modifying the architecture of an archetypal concert hall, the study examined the perceptual effects of stage enclosure width and height, stage canopy structure, and exposure to the main hall volume. Musicians’ perceptions were assessed using the Stage Acoustics Quality Inventory (STAQI), a perceptual measurement instrument developed in earlier studies. Finally, correlations were calculated between recently proposed additions and alternatives to the stage acoustic parameters in ISO 3382-1 and the perceptual qualities captured.

Speaker: Emanuele Porcinai (Audio Communication Group, Technische Universität Berlin)

A18.08 Soundscape in Natural Parks: S135

Conveners: Tin Oberman (University College London), Claudio Guarnaccia (University of Salerno, Department of Civil Engineering)

37 Aircraft noise dose-response in Norwegian National Parks

Soundscape perception is an important part of the nature experience, but increasing aircraft noise events (airplanes, helicopters, etc.) are impacting the natural soundscapes in national parks. We study the influence of aircraft noise on visitor experience and how it influences visitor willingness to revisit a national park, taking into account visitors’ perceptions of “wild remote” vs. “near, peri-urban natural parksTwo Norwegian national parks were chosen based on their popular perception: Femundsmarka National Park (remote and wilderness) and Østmarka National Park (peri-urban and accessible). To study these effects dose-response methodology is used, a technique to understand the relationship between aircraft noise and visitor experience in national parks. Visitors were asked to participate in an intercept (in-situ) survey that assessed tolerance for different scenarios of aircraft noise by administering a “dose” of aircraft noise through temporarily installed speakers, and attaining their corresponding response for acceptability and emotive reactions to hearing differing aircraft levels. This study is important to inform national park management about how people perceive aircraft noise which affects soundscape management and conservation decisions. It also contributes to an international body of work mapping the effects of aircraft noise in national parks. This research also contributes to finding an acceptable threshold for aircraft noise disturbance in national parks.

Speaker: Noelien Wilsnach (Norwegian University of Life Sciences (NMBU))

38 What do Holidaymakers in an Alpine Vacation Region Want to Hear?

In its 2024 summer competition, the Tourist Organisation of the lower Engadin (TESSVM) asked holidaymakers about their favourite sound in this vacation region. It received 2,500 responses. Almost 500 of these included further details (e.g., the exact location) or a reason for their choice. The huge response made it interesting to conduct an in-depth evaluation with regard to the following questions: • Which sounds were mentioned most frequently; are they exclusively natural or also technical? Does the ranking correspond to other studies in urban or suburban or Alpine recreational areas? • Do these sounds have common acoustic characteristics (level, spectrum, time evolution)? • To what extent are they linked to a specific place, situation, or experience? • What determines the choice of favourite sound: is it the acoustic “harmony” or associations it evokes, or ...? The most popular sounds are those of water, e.g., mountain streams or village fountains. These are followed by marmots. Birds come in third place, followed by cow bells and herd bells. When it comes to technical sounds, the iconic three-tone post bus horn comes out on top. Women contributed two-thirds of the responses, men one-third. Only two sounds showed significant differences: men had a stronger affinity for stag bellowing, women for crickets chirping.The most popular sounds are completely different acoustically. One of the conclusions is that for the sound perception in such a recreational context, it is not the sound itself that matters, but its source and the associations and emotions it evokes. The presentation will be illustrated by short audio demonstrations.

Speaker: Beat W. Hohmann (SGA-SSA)

Full paper (PDF)

39 Ecoacoustic gradients and directional information flow in the soundscape of an alpine national park

High-altitude ecosystems are characterized by strong environmental gradients and increasing seasonal anthropogenic pressure, yet their acoustic structure remains comparatively understudied. In this work, the soundscape of an Alpine National Park was investigated during the peak summer period using passive acoustic monitoring across thirteen recording sites on the Nivolet plateau (Gran Paradiso National Park, Italy). A set of ecoacoustic indices was computed from the recordings and summarized through Principal Component Analysis to identify the dominant acoustic gradients. The indices with the highest loadings on the first two components, Acoustic Entropy (H) and the Normalized Difference Soundscape Heterogeneity Index (NDSHI), were retained as representative descriptors of the July soundscape. Spatial and diel patterns revealed marked heterogeneity among sites, with consistently higher nocturnal values of H and site-dependent variations of NDSHI. To investigate directional relationships among recording locations, Transfer Entropy was applied to the selected ecoacoustic descriptors. Only interactions exceeding the 95th percentile of the |ΔTE| distribution were retained, revealing a sparse but spatially structured network of directional information exchanges. These results highlight the potential of combining ecoacoustic indices with information-theoretic metrics to explore the spatial organization of soundscapes in fragile mountain environments.

Speaker: Giorgia Guagliumi (University of Milano-Bicocca)

Full paper (PDF)

40 Sound Source Dominance in Mountain Soundscapes: Machine Listening, Human Perception, and Memory

This paper evaluates sound source dominance inAlpine mountain soundscapes by comparing threedistinct methodologies: human perception in situ,automatic sound event detection (SED), and retrospectivejudgement. Findings indicate that while SEDand human perception ratings are generally consistentfor categories such as human sounds, traffic, and water,retrospective recollections can be influenced bynon-acoustical factors and prior knowledge. Notably,traffic noise is often overestimated in memory, whilenatural sounds like water and wildlife are retrospectivelyassociated with an overall experience of silence.These results suggest that integrating quantitative andqualitative data is essential for a comprehensive understandingof mountain soundscapes to support effectiveenvironmental monitoring and protective measures.

Speaker: Sara Lenzi (University of Deusto, Ikerbasque Basque Foundation for Science)

Full paper (PDF)

41 Investigating Associations Between Human Perception and Ecoacoustic Indices: Short Term Measurements and Soundwalks in Natural or Protected Areas

Within the field of soundscape studies, two distinct approaches have consolidated: a human-centric one, focused on people’s perception; one related to soundscape ecology, which assesses biodiversity via bio- and ecoacoustic indices (EIs). These two branches frequently operate in silos, with minimal cross-disciplinary communication or data integration, even when collected from the same areas, as in the case of mountainous ones. To address this gap, the present study cross-references ISO 12913 standardized perceptual data collected from 219 participants during the "Silenzi in Quota" soundwalks, conducted in 10 European natural or protected areas, with EIs computed and bird species identified from 71 binaural recordings in 68 listening stops. Given the beneficial effects of natural sounds on humans, and the role of EIs as proxies for their presence, the goals are: 1) to identify associations between human soundscape perception and EIs, 2) as well as between these perceptual dimensions and bird species richness. Results from a preliminary correlational analysis revealed that more pleasant soundscapes are associated with sites characterized by higher total entropy (H), lower Acoustic Evenness Index (AEI), and higher number of species present, while more eventful soundscapes are related to higher Acoustic Biodiversity Index (BIO) and lower Spectral Entropy (sh) values.

Speaker: Giacomo Gozzi (University of Trento, Dept. of Civil, Env., Mech. Eng.)

Full paper (PDF)

A20.01 Pathological Speech: S145

Conveners: Péter Mihajlik (Budapest University of Technology and Economics), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Juan Ignacio Godino Llorente, Melanie Jouaiti, Julián David Arias Londoño

42 Hydra VI-LoRA: Speaker-Aware Mixture-of-Experts for Personalized Recognition of Atypical Speech

Automatic speech recognition remains unreliable for individuals with impaired speech, in part because a single adapter must absorb the high inter-speaker acoustic variability that characterizes dysarthric, apraxic, and neurodegenerative populations. Parameter-efficient fine-tuning with Low-Rank Adaptation (LoRA), and its Bayesian extension via variational inference (VI-LoRA), can personalize foundation models such as Whisper to atypical speech, but a monolithic adapter still forces averaging across speakers whose articulation, prosody, and phonemic realization differ markedly. We extend VI-LoRA into a Hydra-style Mixture-of-Experts framework, Hydra VI-LoRA, in which multiple low-rank expert branches share a single down-projection but maintain independent stochastic up-projections, dispatched by a learned gating network. The router is trained without explicit speaker labels, allowing it to discover an implicit clustering of speakers and to adapt expert assignment dynamically at inference. We evaluate the framework on the English UA-Speech dysarthric corpus and the larger multi-etiology SAP-2024 dataset, with Whisper-Large V3 as the backbone. A per-layer top-k routing variant reduces character error rate on UA-Speech from 18% (pure LoRA) to 14%, and lowers WER on SAP to approximately 9% from a 10.5% LoRA baseline. Heatmap analyses reveal speaker- and etiology-correlated expert preferences that emerge without supervision, suggesting that the router captures meaningful subgroup structure. These preliminary results indicate that combining variational regularization with sparse expert routing is a promising path toward scalable, label-free personalization of ASR for atypical speech. We outline ongoing work on expert scaling, cross-lingual transfer, and interpretability of the discovered speaker clusters.

Speaker: Niclas Pokel (ETH Zürich)

Full paper (PDF)

43 Text-to-Speech Data Augmentation for Dysarthric Child Speech Reconstruction

The goal of Dysarthric Speech Reconstruction is to improve intelligibility by converting dysarthric into healthy speech. In the future, this technology could become a valuable communication aid for people with this speech impairment. The development of systems performing complex speech tasks typically requires large amounts of data, which is especially sparse for pathological child speech. This paper explores the potential of Text-to-Speech data augmentation to improve performance in resource constraint scenarios. In this augmentation technique, synthetic dysarthric child speech is generated from text. To evaluate the effectiveness, a Voice Conversion model is trained with just 2.37 hours of real speech in two configurations: (1) only real speech (2) pre-trained with synthetic speech and fine-tuned on real speech. While the overall intelligibility remains low, data augmentation leads to noticeable improvements, which encourages further research. Future work should focus on optimizing both Text-to-Speech and Voice Conversion models for this task.

Speaker: Moritz Pfeiler (Signal Processing and Speech Communication Laboratory)

Full paper (PDF)

44 Multi-Teacher Distillation for Domain-Robust, Real-Time Representations of Electro-laryngeal Speech

Recent advances in self-supervised learning (SSL) have significantly improved speech representations, yet their performance degrades in pathological domains such as electrolaryngeal (EL) speech. Additionally, the large computational footprint of state-of-the-art SSL models limits their applicability in real-time, on-device voice rehabilitation systems. We propose a multi-teacher knowledge distillation framework to train a lightweight, fully causal, and streaming-compatible content encoder that generalizes across healthy (HE) and EL speech. Our approach leverages two complementary teachers: (1) a frozen SSL model that provides phonetic cluster targets derived from healthy speech, and (2) an EL-adapted ASR model that supplies bottleneck feature regression targets to anchor representations in the pathological domain. Experimental results, evaluated via downstream ASR using word error rate (WER) and character error rate (CER), show that the proposed method substantially reduces error rates on EL speech compared to zero-shot SSL baselines, while maintaining competitive performance on HE data. We further analyze the impact of temporal modeling by comparing causal CNN, Transformer, Conformer, and Mamba-based student architectures, showing that effective temporal context modeling is a key factor for cross-domain generalization. The final model operates in real time on a single CPU core, providing a compact and practical representation backbone for cross-domain voice conversion.

Speaker: Benedikt Mayrhofer (Signal Processing and Speech Communication Laboratory)

45 Exploring voice analysis for Turner syndrome characterisation

Turner syndrome is a rare chromosomal condition affecting females, characterized by high phenotypic heterogeneity and frequent delays in diagnosis. In recent years, digital health approaches have highlighted the potential of non-invasive biomarkers to support earlier and more accessible screening strategies. Within this context, the BeNeXT project aims to develop scalable diagnostic and prognostic tools by integrating multimodal phenotypic information (such as facial, body, gait and voice data) with machine-learning techniques.This contribution focuses on the exploration of voice as a digital biomarker for Turner syndrome. Speech signals were collected using different microphone types and a smartphone, reflecting realistic, low-cost acquisition scenarios. Multiple phonatory tasks, ranging from highly controlled productions to more complex speech activities, were considered in order to capture complementary aspects of voice production. Acoustic features related to phonation stability, resonance, prosody and temporal organization were extracted, and statistical analysis together with data-driven models were employed to assess their descriptive and discriminative potential.Preliminary analyses suggest that voice contains systematic acoustic patterns that may be associated with Turner syndrome. Beyond classification performance, key open challenges include ensuring robustness across recording conditions and improving the interpretability of the observed acoustic differences.Overall, this work supports the relevance of voice analysis as a promising component within multimodal pipelines for rare-disease characterization. Ongoing and future work within BeNeXT will extend these analyses to larger samples, additional speech tasks and integrated models, with the goal of advancing reliable, explainable and clinically meaningful voice-based biomarkers.

Speaker: Marc Freixes (La Salle, Universitat Ramón Llull)

A24.01 Auralization of complex environments: S168

Conveners: Josep Llorca-Bofi (Fraunhofer Institute for Building Physics), Jonas Heck (Institute for Hearing Technology and Acoustics)

46 Influence of wind noise on a record-based simulation approach of vehicle driving by scenarios

In order to investigate pedestrian safety, it is important to understand how they behave in different traffic scenarios. For reasons such as cost efficiency, consistent experimental conditions, and increased safety, these investigations are transferred to virtual environments. Recent studies have used single-source recordings of vehicles to simulate driving by scenarios. A main concern at this point was the interfering noise caused by wind, which increased with vehicle speed. This study reviews previous literature and develops five different windscreen concepts. These concepts were tested and compared in anechoic conditions with a fan that could control wind speeds. Next, one concept was selected, and a transfer function and additional low-shelf filters for wind noise reduction were investigated. Driving scenes were simulated with and without pre-editing through filtering. These simulations were recorded with a dummy head in an anechoic chamber and compared to determine the influence of the improvements introduced in this paper on physical values.

Speaker: Jonas Krautwurm (Professur für Akustik und Haptik)

Full paper (PDF)

47 Influence of the Curb Shape on Road Traffic Auralization

In outdoor sound propagation simulations, geometric modeling often focuses on large-scale features, while small roadside structures are neglected or simplified. For vehicle noise sources located close to the ground, roadside structures of similar height can modify the sound field and become perceptually relevant. Therefore, it can be important to consider in auralization, as it allows for listening to simulated scenarios by filtering an audio source signal through sound propagation effects. This contribution investigates the effects, different curb shapes introduce to the sound field emitted by road vehicles. The results show that curb geometry can have perceptual relevance for the auralization filters, which is discussed based on the psychoacoustic metrics loudness and sharpness. The effect size depends not only on curb shape but also on source–receiver configuration. Curb modeling may therefore improve perceptual realism for low-height traffic noise sources.

Speaker: Jonas Heck (Institute for Hearing Technology and Acoustics)

Full paper (PDF)

48 Interacting with soundscapes in Virtual Reality: desktop vs. immersive VR

Interactive soundscape tools for urban design remain underdeveloped, both in adoption and in their fundamental form. City Ditty, an interactive soundscape simulator, enables city makers to engage with soundscape in virtual reality. Developed through a user-centered design process, it provides core functionalities supporting accessible learning and rapid prototyping of soundscapes by non-sound experts. But how does immersion affect user experience? This study examines how immersion and interaction shapes user experience and soundscape engagement using City Ditty. It uses identical functionality with different modalities of immersion and interaction: a desktop version of City Ditty (controlled with a keyboard and mouse on a 24-inch monitor) and immersive VR version (using a head-mounted device (Varjo XR-3) and controllers (HTC Vive). The same binaural audio rendering, tracking the head position and orientation, was presented across conditions over headphones (AKG K702).33 participants completed a within-subjects evaluation with counterbalanced ordering. Each participant completed 13 structured tasks per modality, introducing controls, exploring and listening to the space, finding, adding, and manipulating different sound sources, and redesigning a space within two comparable virtual park environments. Distinct but functionally equivalent task sets per condition reduced repetition while preserving comparability. Post-evaluation questionnaires captured user preferences and engagement, with exit interviews providing qualitative description of their experience.Preliminary results indicate a strong preference for immersive VR over desktop in terms of visual, auditory, and overall enjoyment, despite identical binaural audio across conditions. This suggests that immersion and interaction shape not only the overall experience but also specifically the listening experience.

Speaker: Catherine Guastavino (McGill University)

Full paper (PDF)

49 Real-time modelling of sound transmission through apertures

The hybrid approach of geometrical acoustics (GA) for early reflections and a simplified late reverberation model such as feedback delay networks (FDN) offers a high computational efficiency for virtual acoustics. Their combination has been used in recent time-domain approaches for real-time simulation of sound propagation and reverberation. In complex environments, the requirement for multiple coupled volumes (e.g., rooms) with individual reverberation times and connecting apertures (e.g., doors) arises. However, such geometries typically increase the computational complexity by far, necessitating simplified solutions for diffraction and diffuse sound transmission through apertures. For the GA early reflections, we suggest a physics-based model to approximate the diffraction effects caused by the aperture’s edges without increasing the number of sound propagation paths to be calculated. For the late reverberation, we suggest a statistical-acoustics-based model for the transmission of diffuse sound through apertures by using multiple FDNs. In combination, a computationally efficient acoustic portal model for realistic simulation and spatial rendering of sound propagation through apertures is suggested. The model ensures continuous early reflections and a continuous late reverberant field when the listener or sound source moves through the portal.

Speaker: Kevin de Haas (Medizinische Physik, Carl von Ossietzky Universität Oldenburg)

A24.04 Audio for Augmented Reality: S171

Conveners: Annika Neidhardt (Audio Engineering, Faculty of Media, HS Mittweida), Nils Meyer-Kahlen (Aalto University), Johannes M. Arend (Aalto University)

50 A Densely Sampled Dataset of Room Impulse Responses for Sound- Field Reconstruction

Reconstructing sound-fields from a limited number of room impulse responses is a key challenge in spatial audio. Existing datasets are often spatially sparse or limited to a few rooms, hindering systematic benchmarking and training of sound-field reconstruction methods. This paper introduces Spatially Dense Room Impulse Responses (SpaDenRIR), a simulated dataset designed for developing and benchmarking sound-field reconstruction methods for early reflections. The dataset comprises 100 generated rooms, each containing a cubic microphone array of 22×22×22 microphones with 21.4 mm spacing, allowing spatial interpolation without aliasing up to 8 kHz. In every room, five sources are placed in distinct regions relative to the array. Room impulse responses are simulated using the image-source method and stored only for the first 100 ms, capturing the direct sound and early reflections. We describe the stochastic generation of room geometries, materials, source configurations, and array placement, and outline several application scenarios enabled by SpaDenRIR, including moving microphones and arbitrary array geometries. The dataset is intended as a shared training and evaluation resource for both model-based and machine-learning approaches to sound-field reconstruction.

Speaker: Mads Lang Matthesen (Aalborg University and GN Advanced Science)

Full paper (PDF)

51 Accessible 6DoF rendering with MPEG-I Immersive Audio

A growing cohort of multimedia consumers requires advanced accessibility solutions due to diverse sensory and cognitive needs such as age-related hearing loss, noise-induced hearing impairment, or congenital auditory conditions. Ensuring that emerging media technologies such as virtual-reality and augmented-reality (VR/AR) remain inclusive has become a critical requirement for broad-scale technological adoption.Concurrently, the development of immersive Six Degrees of Freedom (6DoF) audio technologies represents a shift in media consumption. Unlike traditional static media playback, 6DoF audio gives listeners the freedom to navigate within a virtual acoustic scene. However, this interactivity introduces substantial challenges for the traditional authoring of accessible audio experiences during content production. The low-latency requirements of real-time 6DoF audio systems preclude complex post-processing approaches, e.g., clarity enhancement in reverberant scenes containing multiple sound sources.To enable accessible immersive audio while satisfying latency constraints, the recently finalized ISO MPEG-I Immersive Audio Standard (ISO/IEC 23090-4) integrates several accessibility tools directly within its 6DoF audio renderer.This contribution examines these integrated accessibility features within MPEG-I and demonstrates how 6DoF audio rendering can be individualized to foster more equitable experiences for individuals with diverse sensory and cognitive needs. Results from a listening test alongside feedback from a usability study demonstrate the effectiveness of the proposed solution.

Speaker: Nils Peters (Trinity College, The University of Dublin)

Full paper (PDF)

52 A Statistically Validated Bayesian Listener Model for Spatial Audio Personalisation in Augmented Reality

Accurate sound localisation in augmented reality depends also on employing head-related transfer functions (HRTFs) that match the listener's anatomy. Yet, predicting how a specific listener will perceive a given spatial audio rendering remains an open problem. Bayesian observer models can, in principle, predict individual localisation behaviour from acoustic measurements, but their parameters have traditionally been fitted by minimising ad hoc summary metrics, such as polar error and quadrant error rate. This metric-matching approach lacks a probabilistic foundation, prevents formal hypothesis testing, and reduces the full localisation response distribution to scalar summaries that discard the bimodal structure which distinguishes polar uncertainty from front-back confusions. We address this by reformulating parameter estimation as maximum-likelihood optimisation over full listener response distributions, yielding a model that is individually fitted and capable of tracking listener-specific localisation behaviour within a full-sphere HRTF framework. Applied to the SONICOM HRTF dataset across 33 listeners, the model reliably recovers its parameters and accurately predicts individual localisation performance, as validated through behavioural experiments. These results establish a statistically rigorous, individually calibrated listener model that opens the door to principled spatial audio personalisation in AR systems, from automatic HRTF selection to perceptual quality evaluation.

Speaker: Roberto Barumerli (Imperial College London)

53 Comparative Analysis of Sound Localization Accuracy in Real and Virtual Environments Using a Head-Tracked Dynamic Beacon Source

With the advancement and implementation of Virtual Reality (VR) and Augmented Reality (AR) technologies, there is an increasing demand for improvements in spatial audio that accurately mimics real-world perception. Regarding human auditory perception, sound localization is a crucial element. Previous research has shown that localization accuracy often differs between physical and virtual environments when comparing horizontal (azimuth) and vertical (elevation) planes, with the latter exhibiting more pronounced discrepancies. This study aims to examine the differences in localization accuracy by utilizing a moving sound source (dynamic beacon) rather than traditional static sources. The experimental setup involves participants placed in an acoustically controlled environment surrounded by a loudspeaker configuration, where signals are routed to simulate a continuous trajectory. Participants are instructed to track the moving source with their head, while the movement is measured with a head tracker. This trajectory is subsequently reproduced binaurally using generic Head-Related Transfer Functions (HRTFs), and the tracking task is repeated in the virtual domain. By comparing head-tracking data from both experiments, the study analyzes the precision of binaural synthesis relative to loudspeaker-based Ambisonics reproduction serving as a physical acoustic reference. Preliminary expectations suggest superior localization performance in the loudspeaker-based measurements, particularly within the horizontal plane. The implications and results of these findings will be discussed further, along with the potential uses of individualized HRTFs.

Speaker: Ivan Ucović (University of Zagreb Faculty of EE and Computing)

Full paper (PDF)

A02.01 Machine Learning in Bioacoustics: S008

Conveners: Daniel Haider (Acoustics Research Institute, Austrian Academy of Sciences), Jure Zeleznik (Acoustics Research Institute, Austrian Academy of Sciences)

54 F0 Contour Extraction and Pattern Analysis of Rumbles from African Savanna Elephants

We present a pipeline for extracting and analyzing the fundamental frequency (F0) contour patterns in the rumbles of African savanna elephants (Loxodonta africana). Rumble F0s can go as low as 15 Hz and are often masked by infrasonic background noise or overlap with other calls, making standard contour extraction methods perform poorly. To address this, we introduce a custom infrasonic-adapted time-frequency representation and a three-stage extraction procedure: (i) tracking spectral peaks of the first harmonic, (ii) manual inspection and correction in a GUI, and (iii) refinement of the final F0 contour by maximizing harmonic coherence at a higher time resolution. Once extracted, we analyze the F0 contours of approximately 1,600 rumble recordings using unsupervised methods to identify groups based on different contour features, with particular emphasis on amplitude and frequency modulation patterns. Matching these patterns with behavioral observations and contextual data will provide a more in-depth view of the use of modulation in elephant acoustic communication.

Speaker: Daniel Haider (Acoustics Research Institute, Austrian Academy of Sciences)

55 On the Effects of Tightness and Frequency Resolution in Bioacoustic Classification

We investigate whether the frame-theoretic properties of filterbanks can predict downstream neural network classification performance. Using a dataset of 3,080 African savanna elephant rumble vocalisations classified into four age groups, we train a convolutional neural network on magnitude spectrograms derived from 15 filterbank configurations spanning equivalent rectangular bandwidth (ERB), constant-Q, gammatone, and mel families, including their tight-frame variants.Two findings emerge: First, the condition number kappa = B/A is not a statistically significant predictor of CNN performance within the controlled ERB family (r = 0.738, p = 0.155, n = 5). Second, the considered invalid-frame configurations significantly outperform the valid-frame configurations (p = 0.006). An apparent overall correlation between condition number and F1 of r = 0.954 is identified as a filter-family confound. In summary, we find that in our setting tightness does not benefit CNN classification performance but that the filter family design - specifically frequency resolution and spacing - is the dominant factor.

Speaker: Clara Hollomey (Univ. of Appl. Sciences St. Pölte)

Full paper (PDF)

56 Mitigating Domain Shift in Bird Call Recognition

Automated bird call recognition systems trained on one recording environment often produce Automated bird call recognition systems trained on one recording environment often produce worse results when deployed to others due to differences in background noise and recording equipment, a problem known as domain shift. To isolate this issue we process data from two different recording sources into comparable datasets with matched marginal label distributions and similar durations. We fine-tune a pre-trained bird recognition CNN on each dataset and find substantial domain shift in both directions, with 18--84\% performance loss depending on preprocessing method. We test different methods for addressing this problem, including normalization of the spectrogram, adding noise, and domain adversarial training. Of the methods tested, adaptive background subtraction helps models generalize the most, reducing performance loss to as low as 18--32\%. Moderate noise augmentation can help at the cost of reduced in-domain performance, while noise diversity shows mixed effects.

Speaker: Stephen Marsland (Victoria University of Wellington)

Full paper (PDF)

57 Machine Learning to Study Individual Budgerigar Vocalizations in a Social Setting

Artificial intelligence holds powerful potential for studying animal vocal communication in naturalistic group environments, where manual annotation of audiovisual data is often prohibitively time-consuming. We are developing a pipeline to separate individual calls within a group of budgerigars (Melopsittacus undulatus) to investigate potential language-like structure in their vocalizations. Our approach includes a computer vision system using YOLO11 and MobileNetV2 models fine-tuned with minimal annotation effort, which reliably detects (98.05% mAP) and identifies (up to 94.38% accuracy) individuals within the group. The video-derived bird positions are combined with audio recordings from a planar microphone array to isolate sound sources and attribute vocalizations to individuals. Currently, source separation in our pipeline is achieved through beamforming based on the birds’ locations as candidate source positions. However, deep learning methods may later be considered to enhance separation performance. We are developing these techniques using a synthetic corpus of group budgerigar recordings, which is generated by combining individually recorded song segments with the room’s measured acoustic parameters, so that a group of birds can be simulated to be singing from random but ecologically relevant positions within the aviary. This enables the creation of an arbitrarily large, fully labelled acoustic dataset, where the number of individuals and the level of complexity (e.g., spatial proximity) can be freely selected. Overall, our pipeline combines machine learning and signal processing techniques to support analysis of non-human social vocalizations at scale, including call-coordination and low-level structures. Such an automated system is adaptable across species and offers a tool for comparative research on animal communication and language-like systems.

Speaker: Zsofia Katona (Acoustics Research Institute, OEAW)

A13.03 Ultrasound for Medical and Biomedical applications: S498

Conveners: Lynda CHEHAMI (UPHF), Robert Nuster (University of Graz), Markus Saurer (University of Graz)

58 Exploring methods for ultrafast ultrasound computed tomography

Ultrasound Computed Tomography (USCT) is a promising complement to traditional mammography for breast cancer detection. 3D USCT systems, which allow for good illumination of the breasts, are currently limited by the long data acquisition (DAQ) times. Conventionally, sources are activated sequentially, requiring the signal to dissipate before the next wavelet is emitted. These extended DAQ times cause lower image resolution due to patient motion. Here we present an exploratory study on diverse methods to reduce DAQ time, such as source encoding and random wavefield tomography.Regarding source encoding, we consider source apparition, vortex encoding, and compressive sensing. For source apparition and compressive sensing, we encode multiple simultaneous sources with specific time shifts to allow the isolation of the contributions of the individual sources from the mixed signal. Alternatively, with vortex encoding, all sources emit simultaneously. Using full-waveform inversion, we directly invert the blended signals without prior source separation by modifying the misfit accordingly.Random wavefield tomography mimics the concept of ambient noise tomography from geophysics. We actively create the noise by activating all the sources simultaneously, each with a random source time function. By cross-correlating the signals from different receiver pairs, we obtain deterministic travel time information that we can use for inversion. If the duration of simultaneous random sources is shorter than the time needed for sequential emission, we get a reduced DAQ time that is independent of the number of transducers, enabling the scaling to denser measuring systems.We provide a comparison of the different approaches and determine their advantages for later implementation in actual USCT systems.

Speaker: Jade van Leersum (ETH Zurich, Department of Earth and Planetary Sciences)

59 High-Resolution Amplitude-Based Reflection-Mode Sound Speed Mapping

The tissue’s speed of sound (SoS) is a promising marker for a non-invasive diagnosis of diseases such as cancer or liver steatosis. In addition, knowledge of the spatial distribution of SoS allows for aberration correction which is key for high-resolution and high-contrast photoacoustic and echo ultrasound imaging. Computed ultrasound tomography in echo mode (CUTE) retrieves spatially resolved SoS using pulse-echo ultrasound, by analysing the variation of echoes when probed with varying transmit and receive beam steering angles and can be readily integrated with standard ultrasound probes. It conventionally uses echo phase shift, which is interpreted as differential time-of-flight (dToF) and related to SoS. While promising results have been obtained, the trade-off between beamwidth and angular aperture not only limits spatial resolution but also leads to artefactual noise in media with short-scale SoS variations. To overcome this limitation, we propose an alternative approach: for each echo, instead of reducing the angle-dependent variations to a single parameter (the dToF), we retrieve aberration profiles with sub-beamwidth resolution by inverting a model that relates these profiles to the echo amplitudes detected for all angle pairs. We demonstrate that this amplitude-based (a-CUTE) approach results in substantially improved spatial resolution and reduced artifacts compared to dToF-CUTE.

Speaker: Michael Jaeger (Institute of Applied Physics, University of Bern)

60 Pulse-echo waveform tomography for speed-of-sound imaging

Speed-of-sound (SoS) inhomogeneities in tissue induce aberrations that degrade the quality of conventional ultrasound images. These aberrations, however, can be exploited to map the SoS, the key parameter governing wave propagation and a promising biomarker for breast cancer detection. Inspired by seismic full-waveform inversion, we present a pulse-echo SoS tomography approach with three main components. First, it builds on reflection matrix imaging, which is a redatuming operation that yields the impulse responses between virtually relocated transducers at arbitrary depths inside the medium. Correlations between these responses are then used to quantify the phase distortions experienced by the incident and reflected wavefronts across the entire field of view. Second, wave propagation in heterogeneous media is modeled by numerically solving a paraxial form of the wave equation via the Fourier split‑step technique. This accounts for complex wave phenomena, including diffraction and refraction, during matrix imaging while remaining computationally efficient. Third, the SoS distribution is estimated by minimizing the measured phase aberrations, whose exact sensitivities to SoS are computed using the adjoint-state method. This involves an adjoint matrix-imaging step in which aberrations are propagated upward in depth and correlated with the downward-propagated wavefields. The resulting inverse problem is solved iteratively using a BFGS quasi-Newton algorithm. We validate the method on experimental tissue-mimicking phantoms and demonstrate its performance in vivo for breast cancer detection. Importantly, the method naturally yields aberration-corrected images as part of the iterative process.

Speaker: Naiara Korta Martiartu (EPFL)

A02.00 Bio-acoustics: P421

Conveners: Angela Stöger-Horwath (University of Vienna, Vienna Zoo), Dick Botteldooren (U-Ghent), Michael Kerscher, Paolo Diviacco

61 bacpipe: a Python package to make bioacoustic deep learning models accessible

Natural sounds have been recorded for millions of hours over the previous decades using passive acoustic monitoring. Improvements in deep learning models have vastly accelerated the analysis of large portions of this data. While new models advance the state-of-the-art, accessing them using tools to harness their full potential is not always straightforward. Here we present bacpipe, a collection of bioacoustic deep learning models and evaluation pipelines accessible through a graphical and programming interface, designed for both ecologists and computer scientists. Bacpipe is a modular software package intended as a point of convergence for bioacoustic models. Bacpipe streamlines the usage of state-of-the-art models on custom audio datasets, generating acoustic feature vectors (embeddings) and classifier predictions. A modular design allows evaluation and benchmarking of models through interactive visualizations, clustering and probing. We believe that access to new deep learning models is important. By designing bacpipe to target a wide audience, researchers will be enabled to answer new ecological and evolutionary questions in bioacoustics. In conclusion, we believe accessibility to developments in deep learning to a wider audience benefits the ecological questions we are trying to answer.

Speaker: Vincent S. Kather (Naturalis Biodiversity Center)

62 Characterisation of Asian hornet flight sounds

The invasive Asian hornet (Vespa velutina) poses a threat to insect biodiversity, pollination services, and honey production. Current control strategies rely on locating and destroying nests. One method to locate nests is by using a wick bait station and recording return times of hornets, which allows an estimation of nest distance to the station. This approach could potentially be automated using acoustic sensing at bait stations. To achieve this, it is necessary to distinguish individuals based on their acoustic signatures.As a first step, this paper presents a characterisation of hornet flight sounds under low-noise conditions. Flight sound recordings were collected from 68 Asian hornets and 9 European hornets (Vespa crabro). Each hornet was recorded individually inside a mesh enclosure within an anechoic chamber using a calibrated microphone. The resulting dataset contains 2.5 hours of isolated flight sound from Asian hornets and half an hour from European hornets.The analysis focuses on the fundamental wingbeat frequency (f₀) and its variability within and between individuals. Among hornets exhibiting >20 s of flight events (51 Asian, 9 European), the mean individual f₀ was 110 Hz for Asian hornets and 115 Hz for European hornets, with average within-individual standard deviations of 3.8 Hz and 4.5 Hz, respectively. Initial results indicate noticeable inter-individual variability in fundamental frequency, suggesting that hornet flight sounds may carry information useful for discriminating individuals.Although this study is limited to low-noise measurements and restricted acoustic features, it establishes a reference dataset and methodology for future work. The findings support future work involving additional features, machine learning, and outdoor validation.

Speaker: Jan Vandoorne (KU Leuven)

A02.01 Machine Learning in Bioacoustics: P423

Conveners: Daniel Haider (Acoustics Research Institute, Austrian Academy of Sciences), Jure Zeleznik (Acoustics Research Institute, Austrian Academy of Sciences)

63 A Lightweight MiniROCKET-Based Pipeline for Bat Echolocation Call Detection and Classification with Anomaly Filtering

Automatic detection and classification of bat echolocation calls is essential for large-scale biodiversity monitoring, yet remains challenging under real-world acoustic conditions and constrained hardware environments. This work presents a lightweight and deployable pipeline for bat call analysis that combines simple signal processing with efficient time-series machine learning. An adaptive amplitude-based segmentation using the Otsu algorithm is applied to extract dominant acoustic events from continuous recordings. These candidate events, primarily representing bat echolocation chirps, are transformed into feature representations using MiniROCKET. The resulting feature vectors are classified by a Multi-Layered-Perceptron trained on a reference dataset of bat taxa common in Germany. To improve robustness, an auxiliary anomaly detection stage is introduced. An Isolation Forest model is trained on the same feature space to identify atypical events. During inference, detected segments are first evaluated for similarity to known bat calls before being passed to the classifier. This cascaded strategy reduces false detections in acoustically challenging environments.The proposed approach provides a practical balance between performance, robustness, and deployability for scalable acoustic wildlife monitoring.

Speaker: Sercan Alipek (University of Siegen)

Full paper (PDF)

64 Activity-Aware Denoising for Weakly Labelled Insect Sound Classification

Large-scale bioacoustic datasets are typically weakly labelled, assigning species labels to entire recordings without temporal annotations. When recordings are divided into short audio chunks for training, many chunks inherit the recording-level label despite containing no target species, introducing weak-label noise. In this work, we quantify this source of noise in the InsectSet459 dataset and investigate activity-aware denoising strategies for insect sound classification. We evaluate several approaches for incorporating activity information, including activity-based loss weighting, background class modelling, activity-aware pooling, and Multiple Instance Learning (MIL). Our results suggest that explicitly modelling inactive chunks consistently improves classification performance, with a macro F1-score improvement of 3.4 percentage points over the baseline. The findings demonstrate that inactive audio segments represent a measurable source of weak-label noise and that explicitly modelling them provides a simple and effective way to improve large-scale insect sound classification.

Speaker: Florian Krebs (Joanneum Research Forschungsgesellschaft mbH)

Full paper (PDF)

65 CowTalk: Deep Learning-Based Acoustic Monitoring of Cattle for Vocal Activity Detection

Effective management of respiratory diseases in cattle (Bos taurus), particularly within the context of Bovine Respiratory Disease Complex (BRDC), relies on early, non-invasive monitoring strategies to safeguard animal health and welfare. Among the earliest observable indicators, vocal and respiratory-related sounds such as coughs and vocalizations provide valuable insight into physiological and behavioural states. These acoustic cues offer strong potential for continuous, automated surveillance in precision livestock farming systems. In this study, we present a deep learning-based framework for the detection of cow vocal activity using environmental audio recordings collected from multiple farm settings. The proposed system employs convolutional neural networks operating in real time to identify vocal events. Unseen recordings were used to evaluate the robustness and generalization capabilities of the approach.Results indicate that incorporating data from heterogeneous farm environments enhances model generalization, achieving an F1-score of 57.40% and a recall of 74.05%. While models trained on single-farm data can reach higher peak performance under matched conditions, cross-farm training yields more stable behaviour across varying acoustic contexts. Additionally, detection performance is shown to depend strongly on temporal segmentation strategies, highlighting the importance of parameter selection for deployment. This work demonstrates the feasibility of scalable, non-invasive acoustic monitoring systems for livestock, supporting early detection of health and behavioural changes. Such approaches contribute to the development of autonomous tools for improving animal welfare and farm management practices.

Speaker: Ignasi Nou-Plana (La Salle - URL)

Full paper (PDF)

A03.00 Building Acoustics: P459

Conveners: Teresa Carrascal García (Instituto Eduardo Torroja for Construction Sciences - IETcc-CSIC), Chiara Scrosati, Heinz Ferk, Selina Vavrik-Kirchsteiger, Catherine Guigou-Carter (CSTB)

66 Commissioning of Rockwool's Vertical Transmission Facility According to ISO 10140

In August 2024, construction of Rockwool’s Acoustic Laboratory in Hedehusene, Denmark was completed. The laboratory includes a reverberation chamber and vertical, horizontal, and flanking transmission facilities. The vertical transmission facility is being commissioned to comply with the requirements of ISO 10140. This paper presents the commissioning process, interpretation of standard’s requirements, and the challenges encountered. In particular, the relationship between reverberation time, diffusivity, and loudspeaker optimization is discussed.Selected measurement results are presented together with the testing framework used to evaluate the facility. Based on this experience, a proposed sequence of commissioning activities is also outlined for future laboratories or companies planning and commissioning similar facilities.

Speaker: Daniel Nuñez Solano (ROCKWOOL A/S)

Full paper (PDF)

67 Sound attenuation performance of compact multi-resonator silencers for natural ventilation ducts

Although natural ventilation apertures are installed on building façades to improve indoor air quality, they reduce the sound insulation performance of the exterior walls. One countermeasure to address this issue is installation of silencers inside the apertures. Such silencers require a compact design that is thinner than the wall thickness. In addition, the silencer should have a simple, easy to manufacture, and cost-effective structure for practical implementation. This study presents a compact silencer for natural ventilation apertures in building façades using a multi-resonator structure. Assuming an aperture with an inner diameter of 100 mm, incorporating a multi-resonator structure with a total thickness of 100 mm and an outer diameter of 200 mm achieves an insertion loss of 20 dB over a broad bandwidth around 1 kHz. This compact and simple silencer could improve sound insulation performance in the critical frequency range of human hearing. The silencer design offers a practical solution for improving indoor acoustic environments.

Speaker: Takumasa Tsuruha (Takenaka R&D Institute)

A10.06/A12.10 Acoustic wave propagation in complex media: P526

Conveners: Tristan Lawrie (University of Exeter), Malte Peter (University of Augsburg)

68 FDTD Simulation of Acoustic Absorption in Multilayer Textile Structures

A 1D finite-difference time-domain (FDTD) simulation is proposed for the acoustic characterisation of multilayer textile absorbers. Dissipative mechanisms are derived directly from five measurable microstructural parameters per layer, open porosity, airflow resistivity, tortuosity, and viscous and thermal characteristic lengths, through the Johnson–Champoux–Allard (JCA) equivalent-fluid model. Frequency-dependent viscous and thermal losses are introduced into the time-domain solver via two Auxiliary Differential Equations (ADE). Reflection, transmission, and absorption coefficients are extracted. The extracted absorption spectrum is compared against the Transfer Matrix Method (TMM),and the two results are in close agreement across the frequency band of interest.

Speaker: Ahmed Mehrem (German Institutes of Textile and Fiber)

A13.00 Physical Acoustics and Ultrasound: P487

Conveners: Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe), Daniel A. Kiefer (Institut Langevin, ESPCI Paris, Université PSL, CNRS), Lynda CHEHAMI (UPHF), Esma Tuzovic (George W. Woodruff School of Mechanical Engineering)

69 Characterization of acoustic wave propagation in granular media for an earthquake analogue model

Seismogenic analogue fault models offer a novel approach for studying earthquake rupture physics at laboratory scale, enabling simulation of hundreds of seismic cycles and acquisition of data on surface deformations and radiated wavefields. Repeated rupture events occur spontaneously in a multi-layered granular medium under quasi-static shear loading, generating wavefields that carry information about the spatiotemporal evolution of the fault and associated earthquakes. A full characterization of the granular medium is therefore essential for implementing signal processing and inversion techniques.This work investigates elastic wave propagation in unconsolidated granular media under controlled conditions. Acoustic wave velocities were measured through three materials — sand, rice, and rubber pellets — in the 0.1–10 kHz frequency range. The setup consists of a sandbox with two embedded piezoelectric transducers on opposite faces, excited by a Gaussian pulse sweep, and a scanning laser vibrometer recording surface particle velocity. Wave velocity was estimated from signal time-of-flight. Results confirm that coherent wave velocity can be measured in all three materials. A dispersive regime is observed in the lower frequency band, while strong attenuation dominates the upper half due to scattering and diffusion.

Speaker: Louis Demange (CY Cergy Paris Université)

70 Pressure field calculation using the Born series method for an array of point sources in inhomogeneous fractional viscoacoustic medium

Steady-state pressure field simulation, for a ultrasound transducer owing to time harmonic excitation, remains an active area of research because it provides important information regarding beam profile along the axial and lateral directions, and its convergence at the focal region and modification in the field due to scatterers etc. In this work, traditional Born series (TBS) approach has been applied to calculate the pressure field for a linear array of point sources. A total of 501 points was considered. The simulation was performed in a square computational domain of size 4096x4096 grid points with spatial resolution of 20 µm and the medium was considered to be absorbing, dispersive and inhomogeneous. The numerical code was written in CUDA C and was executed in a GPU architecture. The estimated field at ≈ 1 MHz demonstrates excellent agreement with that of a standard time-domain method. This study reveals that the TBS procedure has the potential for rapid computation of pressure fields for real transducers.

Speaker: Ratan K Saha (IIIT Allahabad)

Full paper (PDF)

71 Ultrasonic handling of microplastics in water

Microplastic management using acoustic waves, especially ultrasound, is an emerging and effective technique that utilizes radiation forces to aggregate, separate, and extract particles present in water samples. By applying controlled sound waves in microfluidic systems, the particles are entrained toward equilibrium points due to a radiation force induced by the ultrasonic waves at certain acoustic conditions, facilitating their concentration and subsequent removal separated from their host samples. This enrichment-based approach avoids the use of chemicals and reduces reliance on physical filters that can easily become clogged. Furthermore, it allows for precise process control, improving energy efficiency and the sustainability of water treatment. In practical applications, this technology can be integrated into treatment plants or portable devices, offering an innovative solution to the growing problem of microplastic pollution in fresh water or sea samples. Its implementation also opens up opportunities for real-time water quality monitoring and the adaptation of parameters such as frequency, amplitude, and channel geometry according to the type of particle present. This results in a flexible, scalable, and environmentally friendly system for various industrial applications.This paper presents some strategies to optimize microplastics detection and separation with focus on different conditions of the suspensions: different chip materials, particle concentration or viscous properties of the water samples define the ultrasonic work conditions to achieve highly efficiencies in the microplastic collection/separation.

Speaker: Itziar Gonzalez (National Research Council of Spain CSIC)

A14.00 Physiological Acoustics and Audiology: P512

Conveners: Bernhard Laback (Austrian Academy of Sciences), Sarah Verhulst (Ghent University)

72 Towards the Application of Photoacoustics for Bone Conduction Hearing

The work aims to generate audible sound by photoacoustic (PA) excitation of an optically absorbing target, with a future goal to develop a photoacoustics-based bone conduction hearing device. Very few studies have explored photoacoustic bone conduction hearing, existing work either fails to reach clinically relevant hearing levels (i.e. a minimum of 60 dB HL) or requires unsafe power levels. This study aims to estimate achievable hearing levels using skin safe PA excitation power (≤300 mW/cm2) in the near infrared (808 nm). We present experimental measurements on a range of PA targets with high optical absorption, high thermal expansion coefficient and low specific heat capacity. The amplitude of a laser was modulated with sinusoidal tones ranging from 250 Hz to 20 kHz and directed to the targets. Experimental measurements on phantoms like a skull simulator and a healthy human volunteer demonstrated the generation of audible sound in mid to high frequencies, although the observed amplitude was rarely above the hearing threshold. While we show that increasing the laser power increase sound pressure linearly, reaching, i.e., 60 dB HL with this approach would require excitation power on the order of 100 W, which is far beyond safety limits. Therefore, ongoing work focuses on optimizing PA target materials and excitation to improve photoacoustic conversion efficiency to safely achieve relevant sound levels.

Speaker: Rajalakshmi Sivarajan (Martin Luther University Halle)

A15.00 Psychoacoustics: P496

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Kai Siedenburg (CvO University of Oldenburg), Jan Rennies (Fraunhofer IDMT-HSA, Oldenburg), Sarah Verhulst (Ghent University), Bernhard Laback (Austrian Academy of Sciences)

73 Experimental Study on Impulse Characteristics Affecting Perceived Impulsiveness and Annoyance

Impulsive components can substantially increase noise annoyance, motivating the inclusion of impulse penalties in environmental noise assessment. While existing research and impulse penalty procedures mainly focus on level- and onset-related characteristics, the roles of the spectral content of impulses and temporal irregularity of impulse occurrences have received less attention. To explore these aspects further, this study investigated how selected impulse characteristics affect perceived impulsiveness and annoyance in a controlled laboratory experiment. Thirty participants rated synthetically generated stimuli with impulses systematically varied in level difference, onset rate, irregularity, and spectral content of the impulse. Ratings were compared with model predictions to assess the explanatory power of different approaches. For the stimuli considered in this study, a more complex method incorporating onset-related characteristics provided the most accurate prediction of perceived impulsiveness, whereas a loudness-based metric predicted annoyance well. A simple level percentile-based approach predicted both well. Level difference and the spectral content had statistically significant effects on both rating criteria, while onset rate and regular vs. irregular impulse occurrences showed no significant influence. However, methodological constraints, particularly difficulties in impulse-level scaling and loudness differences between the impulses due to the different spectral content, limit generalizability of the observed effects. The study thus provides relevant considerations for future experiments and highlights the potential roles of psychoacoustic loudness and onset detection in modelling perceived impulsiveness.

Speaker: Felix Hochbaum (TU Berlin, Department of Engineering Acoustics)

Full paper (PDF)

A15.05 Psychological and physiological responses to environmental and product noise: P510

Conveners: Massimiliano Masullo (Università degli Studi della Campania “Luigi Vanvitelli”), Margret Sibylle Engel (Dresden University of Technology)

74 Application of Psychoacoustic Metrics in Environmental Noise: A review

Environmental noise assessment relays on energy-equivalent sound pressure level metrics that quantify the physical properties of sound, while human perception is influenced by perceptual attributes. Psychoacoustic metrics characterise these attributes including loudness, sharpness, roughness, and fluctuation strength, and provide a complementary perspective for the analysis of environmental noise. This work presents an overview of the state-of-the-art of psychoacoustic metrics in environmental noise studies. The analysis considers their application across different sources including road traffic, railway, aviation, and industrial noise, and examines to which extent they can be linked to human responses such as annoyance, disturbance, stress, and cognitive load. The findings highlight both the potential and current limitations of psychoacoustic approaches in this environmental noise assessment. While psychoacoustic metrics can improve the interpretation of perceived sound quality, their relationship with health-related outcomes is not yet fully established and remains context-dependent. This study aims to support the integration of perception-based metrics into environmental noise assessment frameworks.

Speaker: Giada Cardellino (Graz University of Technology)

A20.01 Pathological Speech: P439

Conveners: Péter Mihajlik (Budapest University of Technology and Economics), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Juan Ignacio Godino Llorente, Melanie Jouaiti, Julián David Arias Londoño

75 A multidimensional speech profile in early multiple sclerosis: exploratory associations with mood and cognition

Background: Speech production is sensitive to cognitive and affective states. In people with multiple sclerosis (pwMS) with no-to-minimal disability, multidimensional acoustic measures may help characterize subtle variability beyond overt speech disorders.Aim: To explore psychologically related variability in speech measures in pwMS and healthy subjects(HS).Methods: Thirteen pwMS (9females; median age 44years[IQR38–52]; EDSS 1.50points[IQR1-2]) and 23 matched HS (13 females; median age 41years[IQR32–51]) completed reading and sustained-vowel tasks. Reading-derived measures included speech rate, articulation rate, pause ratio, pause duration, AVQI, and cepstral peak prominence smooth (CPPS). Vowel space area (VSA) and formant centralization ratio (FCR) were calculated from sustained vowels. Associations with Symbol Digit Modalities Test (SDMT), Trail Making Test B-A (TMT B-A), State-Trait Anxiety Inventory (STAI-Y1), and Beck Depression Inventory-II (BDI-II) were explored using Spearman correlations.Results: In pwMS, higher state anxiety (STAI-Y1) was strongly associated with values of VSA (ρ=0.95, p=0.001). In HS, higher anxiety was associated with lower speech (ρ=−0.83, p<0.001) and articulation rate (ρ=−0.62, p=0.031), and lower vowels centralization (FCRρ=−0.56, p=0.045). Better speed process(SDMT) in HS was associated with higher voice quality (CPPS ρ=0.76, p=0.002; AVQI ρ=−0.58, p=0.029). Mood disorders(BDI-II) in HS were associated with smaller VSA (ρ=−0.59, p=0.034).Conclusions: Multidimensional acoustic profiling may capture subtle psychologically related speech variability in this population. Preliminary findings suggest different patterns of association between acoustic measures, mood, and cognition in pwMS and HS. Larger studies are needed to better explore group-specific patterns.

Speaker: Giulia Fusari (University of Milan)

76 An Evaluation of Hypernetwork-Based Parameter-Efficient Fine- Tuning for Hungarian and English Dysarthric ASR

In this study, we evaluate a recently proposed hypernetwork‑based, parameter‑efficient fine‑tuning approach for dysarthric speech recognition in Hungarian and English, using the Whisper ASR (Automatic Speech Recognition) model. We compare its performance to Low‑Rank Adaptation (LoRA) demonstrating the promise of hypernetwork‑driven personalization for atypical ASR. To better understand speaker‑level variability, we analyze the encoder’s latent‑space representations and examine the presence of intra‑cohort structure. Our results show that, even under severe data limitations, the hypernetwork‑based method consistently improves recognition accuracy for previously unseen speakers while requiring substantially fewer trainable parameters than LoRA, yet achieving competitive performance. These findings highlight the potential of hypernetwork‑based adaptation as an efficient and effective strategy for dysarthric ASR in low‑resource scenarios.

Speaker: Péter Mihajlik (Budapest University of Technology and Economics)

Full paper (PDF)

77 Dysarthric Speech Intelligibility Evaluation – A Comparison Between Human Agents and ASR Systems

Motor speech disorders can often lead to reduced intelligibility, communication difficulties, ultimately affecting the person’s quality of life. In recent years, several initiatives have been launched to improve the processing of disordered speech [1]. However, these initiatives differ in their methodology and their degree of success.We are working on a collaborative project called the ‘Dysarthria’ project at the University of Szeged with two main goals in mind. First, we conduct detailed analysis of dysarthric speech while considering etiological aspects [2, 3]. Secondly, we aim to find solutions which intend to improve the intelligibility of dysarthric speech using automatic procedures [4, 5]. This study focuses on the intelligibility of Hungarian dysarthric speakers, comparing four perceptual metrics: speech intelligibility, speech severity, speech naturalness, and listener effort. We invited trained and untrained listeners in speech analysis to rate the level of intelligibility of dysarthric speech samples. We compared their ratings with the error rates of automatic speech recognition (ASR) systems. Previously [5], we discussed automatic speech evaluation methods using four ASR systems. We selected 15 dysarthric speech samples from the Hungarian Dysarthria Database [6]. Our dataset contained recordings from 7 mild, 5 moderate, and 3 severe dysarthric speakers, and 5 healthy controls [7]. Speech samples were evaluated based on the four metrics using a visual analog scale. The judgements of human listeners were compared with the error rates of ASR systems. Findings showed great variation in the evaluation of severity, the degree of severity correlates with intelligibility in both listener groups.

Speaker: Bernadett Dam (University of Szeged)

Full paper (PDF)

78 Effect of Headphones and Hoarseness-Induced Auditory Feedback on Voice Quality: A Multiple Case Study

Real-time auditory feedback modulation (AFM) alters how speakers perceive their own voice during speech, enabling the study of vocal motor control. This multiple-case study specifically examines how auditory feedback with artificially increased hoarseness (i.e., hoarseness-induced feedback) affects acoustic voice quality in healthy speakers, relative to two control conditions: unaltered headphone feedback and no headphones. Eleven participants completed a total of three sustained-vowel production sessions – one per condition (AFM and controls). Acoustic voice quality was assessed using smoothed cepstral peak prominence (CPPS). Results show that hoarseness-induced AFM elicited a small but significant improvement in voice quality. In the control conditions, voice quality slightly declined with unaltered headphone feedback over the course of the task, while it improved without headphones. These patterns were closely linked to changes in voice intensity, with higher RMS values associated with higher CPPS. Overall, our findings indicate that hoarseness-induced AFM may enhance acoustic voice quality and suggest that including no-headphone controls may provide a more thorough picture of factors influencing vocal motor control.

Speaker: Isabel S. Schiller (Work and Engineering Psychology, RWTH Aachen University)

Full paper (PDF)

79 Error patterns in ASR for elderly Austrian German speakers with dementia

This paper focuses on the automatic speech recognition (ASR) of speech produced by elderly speakers diagnosed with dementia. In clinical contexts, reliable transcriptions can support the work of clinical linguists and serve as a basis for (automatic) analyses to identify linguistic biomarkers for dementia. ASR is further also increasingly used in voice-based assistive technologies supporting elderly people in everyday life. Since ASR systems are usually optimized for fluent, standard-language, they perform significantly worse on spontaneous speech from atypical speakers. This becomes even more challenging when dialectal variation and disfluencies are present. We investigate ASR methods for elderly Austrian German with a focus on faithful transcription rather than normalized text output. We compare four ASR architectures: zero-shot Whisper, fine-tuned Whisper model and two encoder-based models (wav2vec2 with greedy and with language-model decoding), which are first fine-tuned on adult Austrian German and then adapted to elderly speech. Besides standard evaluation metrics on word and character level, we present a detailed error analysis showing how well the models preserve disfluencies or hesitations and analyze transcription errors. Beyond Austrian German, this work provides methodological insights into ASR for atypical, disfluent speech in general, informing the development of more robust and clinically applicable speech technologies.

Speaker: Barbara Schuppler (Signal Processing and Speech Communication Laboratory)

80 Multi-Modal Voice Conversion for Pathological Speech: Leveraging Visual Articulation to Restore a Healthy Voice

Pathological speakers whose intelligibility is caused by impairment of vocal fold movement retain largely intact lip and mouth movements, making the visual modality a reliable signal even when the acoustic channel is severely degraded. We propose a multimodal end-to-end voice conversion system that exploits this preserved articulatory information to convert pathological speech into a healthy-sounding waveform of the same utterance, without relying on a separate target speaker encoder or conversion module. Audio is encoded with Whisper and visual articulatory features with AV-HuBERT; a speech rate predictor dynamically adjusts the number of Q-Former query tokens, making the system robust to the reduced and variable speaking rates characteristic of pathological speakers. A Conformer processes the fused representation while injecting target speaker identity through cross-attention at every layer, and a HiFi-GAN vocoder synthesises the output waveform. Adversarial fine-tuning with a Constant-Q transform discriminator further improves perceptual quality. We hypothesize that the visual stream, capturing not only lip movements but also micro-expressions and head movements, improves the expressiveness of converted speech compared to audio-only approaches. Since no public parallel audio-visual dataset exists for voice conversion, training data is constructed synthetically at each stage using a combination of publicly available datasets, web-sourced material, and private pathological speech data recorded under controlled conditions. The current work targets pathological-to-healthy voice conversion in German, a task underrepresented in voice conversion research, nevertheless, the methodology is language-agnostic and transferable to other languages. Evaluation is across subjective naturalness and speaker similarity scores alongside objective speaker similarity metrics.

Speaker: Enrique Orozco Olivares (Signal Processing and Speech Communication Laboratory)

81 Temporal structure of creak in pathological read speech

Creak is a voice quality frequently occuring in healthy speech, sometimes carrying linguistic or paralinguistic information. Due to the lack of precision in voice control, pathological speakers may produce creak independently from its usual functions. This paper investigates the similarities and differences in temporal characteristics of creak occurences using recordings of German read speech. Based on automatically annotated and manually corrected creak intervals, our results show that pathological speakers creak longer and speak slower than healthy speakers, while the creak rate and the gap between two subsequent creaks remain similar. Analysis shows a loss of correlation between articulation rate and mean creak length as well as articulation rate and creak rate in pathological speech. Furthermore, we find that creak occurs at phrase and sentence boundaries in both groups. The pathological dataset shows a less distinct and more dispersed structure, which suggests reduced targeted voice control and more irregular creak onsets. The study provides new insights into the temporal characteristics of creak in pathological speech and how they reflect impaired voice control.

Speaker: Anna Viehhauser (Graz University of Technology)

Full paper (PDF)

A24.04 Audio for Augmented Reality: P449

Conveners: Annika Neidhardt (Audio Engineering, Faculty of Media, HS Mittweida), Nils Meyer-Kahlen (Aalto University), Johannes M. Arend (Aalto University)

82 Benchmarking the Accuracy of Consumer-Grade Camera-Based Head Tracking Systems

Deep learning and facial landmarking have enabled the development of consumer-grade camera-based head tracking systems for binaural rendering that do not require specialized hardware. Instead, standard webcams or smartphone cameras provide sufficient input for real-time head pose estimation on desktop or mobile devices. The estimated head orientation is typically transmitted via network protocols, such as Open Sound Control, to drive a binaural renderer. However, precise tracking with low latency is required for high quality binaural rendering, and it has not been investigated if landmarking based systems meet these strict requirements. In this study, we benchmarked multiple such systems against a high-precision reference tracking system with respect to accuracy and latency under static and dynamic tracking conditions, as well as varying illumination conditions and camera setups. Results show that three out of four tested systems reach an accuracy that may be sufficient for many applications, but only one system also achieved a sufficiently low latency.

Speaker: Christian Scheer (Audio Communication Group, TU Berlin)

Full paper (PDF)

83 SuRIR: A Dataset of Successively Measured Room Impulse Responses for Localizing Silent Humans

Acoustic localization of humans traditionally focuses on scenarios in which they emit sound. In certain settings, however, a human may remain silent while another sound source is active. Recent work demonstrated that silent humans can be localized by analyzing the subtle perturbations their bodies introduce into room impulse responses (RIRs). The existing SoundCam dataset includes RIRs estimated with a silent human present, but provides only single-snapshot RIRs per configuration. In contrast, successively measured RIR estimates enable the analysis of subtle temporal variations caused by involuntary human micromotion, e.g., breathing, which can lead to more robust localization performance than single-snapshot human‑present vs. human-absent comparisons.To advance research in this field, we introduce SuRIR, the first dataset that provides four successive RIR estimates for each configuration, measured a few seconds apart. SuRIR comprises 14,400 RIR estimates, acquired at 48 kHz using exponential sine sweeps. Measurements were conducted using six microphones and two loudspeakers across two rooms, with three participants performing five distinct movement patterns. To provide a static reference condition, SuRIR additionally includes 960 RIRs collected using a mannequin. Besides silent human localization, SuRIR provides a controlled testbed for evaluating RIR-based algorithms under subtle, real-world environmental variations.

Speaker: Jeremy Lawrence (International Audio Laboratories Erlangen)

Full paper (PDF)

14:40 Coffee break

14:40 Coffee break

14:40 Coffee break

14:40 Coffee break

14:40 Coffee break

15:00 Coffee break

15:00 Coffee break

15:00 Coffee break

A10.06/A12.10 Acoustic wave propagation in complex media: S301

Conveners: Tristan Lawrie (University of Exeter), Malte Peter (University of Augsburg)

84 A numerical study of the sound insulation properties of segmented vibro-acoustic metamaterial plates

Vibro-acoustic metamaterial plates (VAMPs) consist of a thin baseplate with periodically distributed structural resonators. Through band gaps generated by localised resonances, VAMPs can achieve high sound transmission loss (STL) over targeted frequency ranges, with the potential to outperform mass-equivalent homogeneous plates. This makes VAMPs promising for low-frequency noise control using thin and lightweight treatments.Previous studies have demonstrated the performance of finite-sized VAMPs, bandwidth broadening using multi-resonant or multi-modal resonators, and multi-layered VAMP partitions. Alongside these advances, practical implementation at larger scales remains an important area for further investigation. Manufacturing methods such as additive manufacturing, thermoforming, and injection moulding have shown potential for producing VAMPs with complex resonator designs. However, extending these methods to panels of several square metres size can introduce constraints related to manufacturing complexity, cost, and scalability.This contribution investigates a modular approach in which multiple small VAMP segments, each easier to manufacture individually, are mechanically connected, for example using tape, to form a larger metamaterial plate. This strategy may support more efficient manufacturing and enable application to complex structures, including those with obstacles or curvature. A finite element model of a segmented VAMP, coupled to a Rayleigh integral formulation for efficient STL prediction, is used to examine how segment size, coupling stiffness, and related design parameters affect wave propagation and sound insulation. The results aim to assess the potential of segmented VAMP assemblies as a practical route towards large-scale and geometrically adaptable metamaterial noise-control treatments.

Speaker: Felix Langfeldt (University of Southampton)

85 Controlled Wave Trapping in an Acoustic Waveguide via Localized Non-Hermitian Feedback Interactions

We study the problem of trapping acoustic waves in a waveguide using a local realization of flatband-supporting non-Hermitian interactions. These interactions are implemented via active feedback control, using loudspeakers driven by non-collocated microphone measurements to synthesize the desired coupling within a confined region of the waveguide. The design is inspired by PT-symmetric non-Hermitian lattice models that exhibit flatbands in periodic configurations. Here, we translate this concept to a hybrid continuous-discrete acoustic waveguide system and show that the flatband can still be re-produced. This enables trapping of incoming sound waves within the controlled region, preventing both further transmission and backscattering, while the rest of the waveguide remains uncontrolled. The model is developed analytically and numerically, and demonstrated experimentally in a one-dimensional waveguide.

Speaker: Gal Ben Yelid (Tel Aviv University)

86 Permeo-Elastic Effects in Thin 3D-Printed Sound-Absorbing Metamaterials with High Tortuosity

Thin layers of acoustic metamaterials can exhibit remarkable low-frequency sound absorption when their microstructures contain long, tortuous channels or coiled resonators. This is because the coiled, spiral or labyrinthine channels require a relatively small material thickness to significantly extend the path of oscillatory viscous flows induced by airborne acoustic waves that penetrate the air-saturated permeable material. From a homogenization perspective, these materials are characterised by exceptionally high tortuosity, which slows down acoustic waves. Test prototypes of such metamaterials with excellent acoustic properties can be easily produced using low-cost additive manufacturing techniques. However, experimental validation can exhibit deviations from model predictions, arising from elastic vibrations of the 3D-printed metamaterial structure. In this work, we use direct numerical simulations to show in which situations and how these effects occur. Furthermore, we demonstrate that the theory of permeo-elasticity provides a framework to capture these fluid-structure interaction effects and to determine the acoustic material performance by using a computationally more efficient approach based on unit-cell calculations. The proposed method based on this approach enables the informed and very efficient design of highly-tortuous permeo-elastic metamaterials, in which permeo-elastic effects enhance the desired low-frequency sound absorption. This method essentially involves designing a (printable) metamaterial structure that forms the walls of a tortuous channel and is flexible enough to exhibit elastic resonances in a desired frequency range between simultaneously designed quarter-wavelength resonances.

Speaker: Tomasz G. Zieliński (Institute of Fundamental Technological Research, PAS)

87 A discontinuous Galerkin approach for the radiative transfer approximation of high-frequency wave energy in complex media

The accurate prediction of high-frequency wave behaviour is essential for engineering applications related to noise and vibration control, but traditional finite- and boundary-element approaches typically become computationally impractical at these frequencies. We will present a transport-based framework for high frequency waves, formulated using the stationary Radiative Transfer Equation (RTE) in complex media such as piecewise homogeneous acoustic cavities and multi-plate configurations attached at line joints. Building on earlier developments in high-frequency wave energy modelling -most notably Dynamical Energy Analysis (DEA)- we retain DEA’s improved fidelity and applicability over Statistical Energy Analysis. Furthermore, the RTE is spatially discretised using a discontinuous Galerkin scheme that can be implemented efficiently for any polynomial degree basis by making use of analytic integration techniques based on Stokes’ theorem. Our method supports general finite element–style unstructured meshes with arbitrary polygonal elements and delivers competitive computation times suitable for large-scale applications

Speaker: David Chappell (Nottingham Trent University)

Full paper (PDF)

A13.00 Physical Acoustics and Ultrasound: S089

Conveners: Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe), Esma Tuzovic (George W. Woodruff School of Mechanical Engineering)

88 Diffraction of sound waves by a periodic façade displaying impedance boundary conditions

Our study aims to develop an analytical and numerical model to quantify the influence of a building's façade geometry and all its components (windows, balconies, etc.) on scattered and reflected sound waves. We have therefore chosen to adapt a diffraction model originally used in underwater acoustics and elastodynamics for ultrasonic waves and to apply it to airborne acoustics for a frequency range corresponding to urban noise. This formalism, based on the expansion of the potential solutions of the Helmholtz equation into infinite series of plane waves, allows calculating the intensity maxima of far-field diffraction lobes when the studied grating is periodic. Impedance boundary conditions are applied to the surface to take account of the physical properties of the building materials used in the façade. We also provide finite element simulations and compare our model with experimental measurements taken on a 1:20 scale model under the same assumptions for a Galois diffraction grating (GF(2³)).This profile, derived from number theory and widely used in room acoustics, attenuates the intensity of the specular lobe at a characteristic frequency. Numerical results are compared with experimental data and finite element simulations, showing good agreement and demonstrating the model’s ability to predict diffraction patterns for building façades.

Speaker: François-Marie Lucchetti (Università di Corsica)

Full paper (PDF)

89 A combined acoustical and visual analysis of Surface-Acoustic-Wave-induced ice melting

Ice removal and the prevention of ice buildup is a hot topic in industry and academia due to its implications, e.g. on sensors, aircrafts and wind turbines. We recently researched de-icing and anti-icing methods using Surface Acoustic Waves (SAW) to transfer energy into the ice or water to melt it or protect it from freezing respectively. It has been shown that SAWs are a very energy efficient method of ice removal compared to caloric heating and that the interdigital transducer exciting the SAW can be used for ice buildup detection at the same time. In this study the SAW-induced melting process of ice is further investigated using an acoustic analysis combined with microscopic camera footage. We identify different de-icing and anti-icing mechanism during early, middle and late surface activation times. Those results are useful to understand the underlying physics of the energy transfer from the SAW into the ice and the water.

Speaker: Steffen Kempen (German National Metrology Institute)

90 Ultrasonic Evaluation of Plies’ Interface Strength in AFP Thermoplastic Composites

Composites, specifically fibre-reinforced polymers, are increasingly utilised in aerospace industry particularly as they are lightweight, and possess respectable mechanical properties capable of sustaining environmental constraints. Automated Fibre Placement (AFP) has emerged as a manufacturing solution in the composites industry over the past few years. It has improved the production process, and delivers better quality parts. Mechanical tests show that the adhesion quality between plies is strongly depending on the manufacturing process (placement speed, consolidation). The aim of this work is to give an ultrasonic non-destructive evaluation (NDE) method of the adhesion quality in such structure.In this study, the sample under investigation is a representative plate of an AFP-manufactured part provided by Safran. It consists of 8-plies carbon fibre-reinforced Low-Melt Polyaryl-ether-ketone (LM-PAEK) composite, with a total thickness of 1.47 mm, resulting in an average ply thickness of approximately 183 µm. To evaluate the integrity of the ply interfaces, a high-frequency non-destructive ultrasonic method is employed using the Scanning Acoustic Microscope (SAM) PVA TEPLA 301 in pulse-echo method. A numerical model based on the Debye Series Method (DSM), developed by J.M Conoir and implemented by P. Marechal is used. It expresses the reflection and transmission coefficients at each interface and, recursively, calculates the global reflection and transmission coefficient. The latter enables signal reconstruction resulting in echoes associated with each layer of the structure. Interphase conditions can be fed to the model in order to simulate its nature and the potential defects. Optimisation methods were used in order to estimate the interphase properties and, consequently, the adhesion quality of the composite.

Speaker: Soufiane BELMEKKI (Université Le Havre Normandie)

Full paper (PDF)

91 Active control of ESI aerosols by airborne ultrasound for mass spectrometry applications

Electrospray ionization (ESI) is the most widely used ionization technique in liquid chromatography-mass spectrometry (LC-MS), generating charged microdroplets that undergo evaporation and Coulomb fission to produce gas-phase ions. However, in high-flow ESI regimes, a significant fraction of large, highly charged droplets (up to 100μm) can enter the MS inlet. These droplets contribute to contamination of ion optics, increased chemical noise, and degradation of detector performance. In this work, we introduce a new approach for active control of ESI-generated aerosols, enabling steering of the aerosol stream using acoustic radiation forces induced by airborne ultrasound. Acoustic fields are generated by a phased array of 240 ultrasonic emitters operating at 40 kHz, with independent amplitude and phase control. This enables dynamic shaping of the acoustic field, allowing the formation of pressure distributions such as extended focal regions and acoustic vortices between the ESI source and the MS inlet. The ultrasonic system is integrated with a Bruker amaZon ETD quadrupole ion trap mass spectrometer, allowing direct assessment of the influence of the acoustic field on mass spectral signals. Aerosol imaging experiments demonstrate controlled steering of ESI aerosol flows under acoustic excitation. The application of ultrasound altered the measured mass spectra, shifting components with high m/z values toward lower values. We further investigate the underlying mechanisms of acoustic–aerosol interactions using various mass spectrometer operating modes and evaluate the impact of ultrasonic field configurations on analyte transmission and mass spectral measurements. This approach demonstrates a contactless, reconfigurable method for aerosol control in LC–MS, opening new possibilities for improving measurement stability and reducing device contamination.

Speaker: Dmitrii Nikolaev (University of Helsinki)

92 Ultrasonic transmission through a water-saturated porous medium with an embedded PVC interlayer

Internal erosion can produce localized losses of material inside fluid-saturated porous structures and detecting such hidden defects from outside the structure is difficult. This work studies, theoretically and experimentally, the transmission of normally-incident ultrasonic waves through a fluid saturated porous medium in which a thin solid layer is embedded at the midplane to stand in for the missing material. The thickness of the solid layer is treated as a control parameter. Each porous layer is described either within Biot's single-porosity framework or within the Berryman–Wang double-porosity extension, and the multilayer boundary-value problem is solved using a global matrix method. Transmission spectra are computed as functions of frequency over a range of solid-layer thicknesses.Inserting the solid layer changes the transmission response in two ways: the frequencies of the transmission extrema are shifted and the amplitudes across the band are modified. A peak-frequency sensitivity is introduced to quantify the frequency shift and the amplitude change gives a complementary indication of the internal discontinuity. Measurements on water-saturated glass-bead (single-porosity) and Robu sintered borosilicate glass (double-porosity) samples, with and without a PVC plate inserted at mid-height of the column, show good agreement with the model within the transducer bandwidth. The combined dependence of peak positions and transmission amplitudes on the internal-layer thickness suggests that such measurements can be used as a practical indicator of internal discontinuities in saturated porous structures.

Speaker: Mehdi Abdu Mohammed (Le Havre Normandy University)

A13.03 Ultrasound for Medical and Biomedical applications: S519

Conveners: Lynda CHEHAMI (UPHF), Robert Nuster (University of Graz), Markus Saurer (University of Graz)

93 Investigating Transcranial Ultrasound Transmission and Wavefield Dynamics

Transcranial focused ultrasound (tFUS) is advancing as a non-surgical technique for neuromodulation and tissue ablation by transmitting ultrasound waves through the skull and focusing them inside the brain. Most studies characterize ultrasound waves passing through the skull at normal incidence and for focusing at deep brain targets, while focusing on shallow brain targets and application of the guided waves remains largely unexplored. The skull bone causes major distortion to the ultrasound waves because of the high acoustic impedance, PS mode conversions, and heavy scattering in the diploë layer. Our goal in this study is to explore how we may stimulate parts of the brain with guided waves that might be otherwise hard to target, and to explore the application of guided waves for skull characterization. We developed a numerical simulation framework for tFUS to simulate the acoustic-viscoelastic wave equation using the spectral-element method. We employ an ex-vivo human skull model in the numerical simulation framework and place a focused transducer bowl source at some incidence angle to the skull. We analyze the ultrasound waves transmitted through the skull into the shallow areas close to the skull layer. Incident P waves convert to S waves at the outer skull, form guided waves via internal reflections, and reconvert to P waves upon exiting the inner skull layer. Thus, when we place incident waves at certain angles to the skull, the setup introduces shear wave effects in the transmitted wavefield. We investigate these complex ultrasound wavefields as they pass through the skull bone to observe their behavior in shallow regions.

Speaker: Isha Lohan (ETH Zurich)

94 Motion Artifact Mitigation in Audioplethysmography: Signal Processing and ML-Based Approaches

Audioplethysmography (APG) enables cardiovascular monitoring through ear-worn devices by detecting cyclical changes to the acoustic properties of the ear canal. However, body-motion-induced artifacts significantly degrade signal quality in real-world applications. This paper investigates the impact of body motion on the accuracy of APG-derived heart-rate estimates and compares a digital signal reconstruction method and three machine-learning-based estimators against an unmitigated baseline. The candidate-ranking model achieved the best overall performance, with a mean absolute heart rate error of 5.36 bpm and a mean absolute percentage error (MAPE) of 6.35%. Its MAPE remained below 5% in all conditions except cycling, where it reached 15.72%. These results show that machine-learning-based candidate selection provides more robust motion-artifact rejection than the signal reconstruction method, although severe body motion remains the primary obstacle to reliable cardiovascular monitoring using APG in everyday scenarios.

Speaker: Dževad Ćoralić (USound GmBH)

Full paper (PDF)

95 Predicting the Acoustic Emissions from Clinical Proton Beams: a Step Towards Ionacoustic Dosimetry in Proton Therapy

Rapidly producing accurate 3D spatial maps of the radiation dose distribution delivered to patients during proton therapy treatment would reduce existing beam delivery uncertainties and improve patient outcomes. Current dosimetry methods that use ionisation chambers have low spatial resolution and long acquisition times. An alternative to this would be the use of ionacoustic imaging to produce the required dose maps. This methodology utilises the back propagation of acoustic waves generated by the thermoelastic expansion of material where the proton energy is deposited. Ionacoustic imaging would be able to provide fast and accurate 3D dose maps non-invasively. In the short term, ionacoustic imaging could provide essential beam quality assurance prior to treatment and it is hoped that in the future it could be used for online dose monitoring during patient irradiations. The latter would help to mitigate against uncertainties arising from anatomical changes and internal organ movement. Some ionacoustic measurements have been documented in the literature. However, accurately predicting these signals to the degree necessary for optimising detection systems and enabling precise 3D map reconstruction remains challenging. As a first step towards experimentally measuring a 3D spatial map of radiation dose using ionacoustic imaging, we will present recent work into a modelling framework. The model demonstrates the ionacoustic signals expected from clinically relevant beam energies and geometries. These modelling results will serve as the foundation for designing future experimental setups and determining equipment requirements, including identifying the transducer frequency band response and minimum signal to noise ratios for ionacoustic signal detection.

Speaker: Catherine Burne (University of Birmingham)

A14.07 Computational and model-based approaches to hearing science: S248

Conveners: Helia Relaño-Iborra (Eriksholm Research Centre), Vaclav Vencovsky (Czech Technical University in Prague)

96 Perceptual Sensitivity is Explained by Optimization for Ecological Tasks

When asked to discriminate between two stimuli that differ along some dimension, such as intensity, frequency, or orientation, humans exhibit a smallest stimulus difference, or threshold, below which discrimination becomes impossible. These limits on perceptual sensitivity have traditionally been explained by internal noise. Contemporary theories instead link perceptual sensitivity to efficient coding of the natural environment, but have not been comprehensively tested. Here, we measured the psychoacoustic sensitivity of a computational model optimized for everyday sound localization and recognition tasks, estimating thresholds using linear classifiers operating on the model’s internal representations. Thresholds derived from learned model representations exhibited qualitative and quantitative matches to human sensitivity across a large set of perceptual dimensions. The results show that many classical perceptual phenomena (such as spectrotemporal modulation sensitivity and binaural unmasking) are side effects of representations optimized for natural listening behavior, and raise the possibility that psychophysical thresholds are determined by linear separability in these optimized representations, rather than intrinsic neural noise.

Speaker: Mark Saddler (Technical University of Denmark)

97 Temporal Modulation Detection in a Model of Auditory Neuropathy Spectrum Disorder

Auditory Neuropathy Spectrum Disorder (ANSD) is a hearing impairment characterized by abnormal or absent auditory brainstem responses and difficulty understanding speech, especially in noise, while maintaining normal outer hair cell function. Despite this definition, clinical presentation of ANSD is highly variable in practice. Additionally, ANSD has a wide range of underlying pathologies, making it difficult to diagnose without performing electrophysiology. ANSD listeners have worse performance than normal hearing (NH) listeners on psychophysical tasks that require precise temporal resolution. Therefore, using a model featuring physiologically accurate impairments, a test battery of tasks can be identified that distinguishes ANSD from normal hearing, as well as specific ANSD pathologies from each other.Using the Bruce et al. (2018) auditory nerve (AN) model, we proposed that perturbations of vesicle release dynamics are sufficient to accurately portray ANSD resulting from synaptic disorders. Model performance was quantified for modulation detection using a noise carrier, by calculating thresholds over multiple modulation frequencies.Modulation detection thresholds were significantly degraded in the ANSD model compared to NH, mirroring the trend seen in human subjects. Furthermore, threshold details differ between perturbation types, demonstrating that ANSD subtypes with different mechanisms can be distinguished even within the same psychophysical paradigm. Finally, we show that overall thresholds and impact of ANSD perturbation depend on model parameters like spontaneous rate, suggesting that certain types of AN fibers may be resistant to ANSD impairment.These results not only demonstrate the feasibility of modeling ANSD mechanisms and their perceptual consequences, but also suggest that similar methods can be applied to diagnose human patients in the future.

Speaker: Paul Mitchell (CeRIAH, Institut Pasteur)

98 Low and High Frequency Sounds Are Processed Differently in the Cochlea

Healthy hearing relies on the careful orchestration and strong coupling of the mechanical, electrical, and fluidic domains in the mammalian cochlea. There is a dearth of data and simulation results for a broad swath of the cochlea including the apex, a region responsible for processing speech frequencies. With the advent of vibrometry based on optical coherence tomography and signal processing techniques, it is now possible to measure the sound-evoked response of intracochlear structures at the apex without opening the cochlear capsule. The experimental data available to date suggest that the apex functions differently from the oft-studied cochlear base, but there is significant disagreement amongst different groups’ measurements. We analyze the cochlear response to sound using a combination of time-domain and frequency-domain algorithms for predicting the nonlinear electro-fluidic-structural response providing mechanistic insights not possible through experiments. We show how different optical axes of measurement can give rise to dramatically different results, explaining some variations in measured responses. We find that the transition to the broadly tuned low-frequency response region arises due to the combination of geometric changes, frequency regimes, and viscous fluid-structure interaction at the apex. Moreover, experiments and theory for transient responses at the lowest frequencies sensed in the cochlea show peak locations that do not correlate with the standard place-frequency mapping. These results have practical implications for the design of auditory prosthesis and algorithms for hearing restoration.

Speaker: Karl Grosh (University of Michigan)

99 Viva la Transmission Line!

Transmission-line models have a unique place in the cochlear-mechanics field. Because they are the simplest model that includes an explicit representation of the traveling wave’s physics, they have been instrumental in developing and popularizing several cochlear mechanical theories that are now well-established (e.g., coherent reflection and amplification theories).Because transmission-line models are computationally light enough to perform time-domain simulation in a reasonable time, they occupy a special niche at the interface between cochlear mechanics, auditory modeling, and, more recently, AI-powered auditory signal processing.Despite having been at the forefront of hearing science for more than a century, transmission-line models are not without flaws. Because they neglect two-dimensional hydrodynamics effects that largely contribute to the cochlear response, this class of models struggles to replicate well-known features of the experimental data (e.g., the full dynamic range of cochlear amplification), and it is inadequate to study the fine details of cochlear amplification. Transmission-line fans shall not despair: here we show how to modify the transmission line—at no extra computational cost—to incorporate a physically accurate account of two-dimensional cochlear hydrodynamics. With the proposed modification, the transmission line performance becomes virtually identical to that of a complete two-dimensional model. Viva la Transmission Line!

Speaker: Alessandro Altoe (University of Southern California)

100 An End-To-End Tunable Cochlear Model With Instantaneous Nonlinearity For Simulating Otoacoustic Emissions

Computational cochlear models are essential for auditory research. However, personalization to individual audiological data remains a challenge. While previous work achieved personalization using transient-evoked otoacoustic emissions (TEOAEs), simulating distortion product OAEs (DPOAEs) requires a more explicit bidirectional framework.We introduce TDLM-SNLv3, a fully differentiable model implemented in JAX. Unlike traditional models using time-varying parameters, TDLM-SNLv3 employs instantaneous nonlinearity to represents the rapid mechano-electrical transduction of outer hair cells. The primary innovation of this work is an extended backward path that couples nonlinear distortion products directly into a filter chain while maintaining linear reflection mechanisms for TEOAEs. Experimental results demonstrate that TDLM-SNLv3 enables simultaneous TEOAE fitting and DPOAE simulation, providing an end-to-end tunable framework for personalized hearing applications.

Speaker: Yong-Yue Xu (National Tsing Hua University)

Full paper (PDF)

A18.02 Soundscape practice and interventions: S129

Conveners: Enkela Alimadhi (AARE-SA), Francesco Aletta (University College London), Siebein Keely (Siebein Acoustic)

101 Perspective on Soundscape Practices – Public and Occupational Health Benefits

Soundscape approaches, as formalized in the ISO 12913 series, have introduced perceptual dimensions into environmental acoustics, complementing traditional noise metrics. Recent developments, including the Soundscape Perception Index (SPI) from the ERC-funded Soundscape Indices (SSID) project, provide quantitative tools for describing and comparing sound environments. However, their uptake in professional practice remains limited. This perspective paper argues that current methodological and data-related advances enable a transition towards implementation. A strategy is proposed to support the integration of soundscape indices into existing design, engineering, and policy processes. The approach focuses on identifying feasible entry points within current workflows and evaluating application scenarios through simulated and case-informed studies. Particular attention is given to public and occupational health contexts, where soundscape considerations can complement existing noise management practices. The proposed framework seeks to accelerate the uptake of soundscape practices by producing transferable workflows, practical recommendations, and an open-access multimedia evidence base. Ultimately, this work contributes to bridging the gap between research and application, supporting the integration of soundscape thinking into mainstream environmental design and policy.

Speaker: Tin Oberman (University College London)

Full paper (PDF)

102 Exploring the Discipline of Soundscape Architecture: Lessons from Landscape

The etymology and history of landscape and soundscape are explored to identify opportunities for enriching the concepts of architectural soundscape design as it is applied in buildings and environments. Stillgoe (2018) developed an entire text to explore the concepts underlying the word landscape and the ways in which landscape has been defined in theory and in practice. Southworth (1967), Schafer (1977), and Truax (2000) described early efforts in parallel studies of the concept of soundscape. There has been much work done in the development of standards and some design applications of soundscape in recent years to the point that there is the potential for a new field of Soundscape Architecture. There is much potential to be gained from exploring the transformation of landscape architecture as a discipline that could shed light on defining a discipline of soundscape architecture particularly as it applies to sounds in cities, towns, urban centers, and buildings. Siebein and Siebein (2023) have identified five levels and nineteen elements of soundscape theory that can serve as an initial outline for the conceptual expansion of this field. Case studies from the theory and practice of landscape and soundscape architecture illustrate potentials for enriching the conceptual basis for this evolving field of applied soundscape study and practice.

Speaker: Gary Siebein (Siebein Associates, Inc.)

Full paper (PDF)

103 An Exploration of Acoustic Diversity Across Areas with Varying Levels of Tourism Intensity on a Mediterranean Island

This study investigates the acoustic diversity observed on the Naxos island in Greece, through the combined use of eco-acoustic indices and manual sound-source annotations. Recordings from 84 locations were grouped into six regions based on tourism intensity. Sound events were organized and statistically analyzed according to a proposed taxonomy, and Acoustic Entropy (H), Normalized Difference Soundscape Index (NDSI), and Dynamic Spectral Centroid (DSC) data were used to characterize soundscape structure. Results show coherent patterns across regions: high-tourism areas exhibit increased anthropogenic acoustic presence, while low-tourism ones maintain strong biophonic content. The complexity of effective soundscape interpretation and the benefits of using combined analysis approaches which work in a complementary manner is also discussed.

Speaker: Emmanouil Lianis (NKUA)

Full paper (PDF)

104 Augmenting the Sonic Environment in Dementia Care Through Clinically Validated Design Protocols

Translating soundscape research into clinical practice requires empirical evidence of efficacy. It also needs structured, operationalizable guidance for implementation. This presentation addresses that translational gap through a case study, focusing on soundscape augmentation in a specialized dementia unit serving individuals aged 65 and older with moderate to severe dementia. In this setting, complex acoustic dynamics frequently exacerbate agitation and impede care delivery.Drawing on findings from a pilot randomized controlled trial that demonstrated a reduction in resistance-to-care behaviours through targeted soundscape augmentation, the presentation outlines an implementation protocol. The protocol begins with a soundscape investigation following ISO/TS 12913-2 to characterize the ambient acoustic environment. Sound selection prioritizes higher sharpness values and natural vocalizations, delivering gentle alertness cues during morning care and energetic masking during evening routines, at levels slightly above ambient background (40–60 dBA). The system operates autonomously within existing care workflows, supports immediate muting, and accommodates age-related hearing impairment. A real-time feedback mechanism enables ongoing personalization without requiring additional workforce, leveraging existing clinical observation competencies.This implementation model is presented as a replicable protocol with direct relevance to the ongoing development of ISO/CD TS 12913 series, addressing a population whose perceptual characteristics are largely absent from existing soundscape design standards.

Speaker: Arezoo Talebzadeh (Ghent University)

105 Bringing sound to the design table: A guide for urban planners

Sound is a key but overlooked component in how we experience the city. Professionals of the built environment are aware of the importance of sound, yet rarely consider it proactively in their practice. They feel ill-equipped to think about sound early in their projects, and when asked, they request a best practice guide for decision-makers. This article presents a sound design guide co-designed with urban experts to look like the urban design and planning guides they know and use in their practice. Specifically, it provides guidance on how different urban design decisions may impact the sound environment. Developed through an iterative process of document analysis, interviews and work sessions with urban experts, the guide is organized into general guidance providing contextual information and methods for readers to adapt to their own projects, as well as sample exercises of realistic situations to illustrate that information and to help readers practice thinking through the sound component of an urban project. We aimed to be more descriptive than prescriptive or exhaustive, as readers know best what their own projects will look like. The guide is meant as an opportunity for practitioners to think through the sound components of their project to strengthen them.

Speaker: Cynthia Tarlao (Université du Québec à Montréal)

Full paper (PDF)

106 From Soundwalks to Auralisation: A Participatory Workflow for Biophilic Soundscape Design in Urban Green Spaces Using City Ditty

Soundscape design is increasingly recognised in urban planning, yet its integration into design processes remains limited by methodological and technological gaps. This study presents the advancement of a participatory workflow for urban soundscape design, integrating field data, simulation, and perceptual evaluation.The work builds on two complementary contributions previously presented at Internoise 2025: the development and application of the City Ditty soundscape tool, and a methodological framework for biophilic soundscape design in urban green spaces. Here, both approaches are integrated into a unified process and applied to the redevelopment of a large urban park in Santiago, Chile.The workflow combines soundwalks, multidimensional data collection, interdisciplinary design, and predictive auralisation. Soundwalks were conducted with community members and design teams to characterise the existing environment and inform proposals. Prior to the design phase, participants engaged in a workshop with invited experts covering urban planning, sustainability, social dimensions, soundscape, and green space design. Four interdisciplinary teams—comprising students and professionals from acoustic engineering, environmental sciences, social sciences, and planning and design disciplines—then developed design proposals through a competition.These proposals were auralised using City Ditty, an interactive soundscape simulation tool developed at McGill University and adapted to the local context. For each team, five listening positions were auralised, resulting in 20 auralisations. These are currently being evaluated through 45-minute controlled listening tests in an acoustically treated room, using circumaural headphones and randomised playback. Evaluation includes ISO 12913-based perceptual assessment, overall soundscape assessment, and the Green Soundscape Index (GSI), alongside assessment by an interdisciplinary jury.This study demonstrates the integration of participatory processes, auralisation tools, and perceptual evaluation within a coherent workflow for sound-informed urban design.

Speaker: Pablo Kogan (ChucaoLab, Dept. of Sound, Faculty of Arts, University of Chile)

Full paper (PDF)

A19.02 Listening experience in sound-driven design: S140

Conveners: Georgios Marentakis (Østfold University of Applied Sciences), Sandra Pauletto (KTH Royal Institute of Technology), Elif Özcan (Erasmus MC / TU Delft), Simone Spagnol (Iuav University of Venice, Dept. of Architecture and Arts)

107 The SoundRise Alarm: The Impact of a Gradually Rising Alarm on Morning Mood and Waking-up Behavior

This study examined whether a gradually rising pre-wake alarm affects the morning awakening experience differently from a sustained version of the same alarm. In a 14-night home-based within-subjects study, 18 participants experienced both alarm conditions and reported morning states and waking-up behavior. The gradually rising alarm was associated with significantly higher morning positive affect and longer alarm stop time, and with no significant differences in sleepiness, perceived sleep quality, snooze use, or waking-up delay. Associations between pre-sleep states and morning mood were also generally weaker under the gradually rising condition. These findings suggest that a gradually rising alarm may improve the sleep-to-wake transition and support a smoother, more positive awakening.

Speaker: Ke Yang (Royal Institute of Technology)

Full paper (PDF)

108 Bubbling: imagining breath as a sonic interaction technique

Breathing is an increasingly being explored as an alternative input technique and is intrinsically linked to sound production. The sound of breath is directly affected by our emotional and physiological state while fine conscious control of the breath-sound loop is possible, as demonstrated by wind instrument players. Despite this, breath as an input technique has not received much attention in sonic interaction design. To understand how to design using breath as an input modality for sonic interactions, we designed a technology probe that can facilitate the playful exploration of breath as a sonic interaction technique and elicit early qualitative feedback. To establish a tight coupling between breath and sonic feedback, weimplemented an interaction based on the metaphor of creating bubbles in a liquid by blowing into a straw. The user is asked to breathe into the microphone and breathing intensity is sensed based on the devicemicrophone signal. This is used to simulate air flow through the straw and into the liquid. A simplified parametric real-time model simulates bubble formation and synthesizes the resulting bubble sounds. Thefinal prototype is a mobile phone app that can sense breath, create bubbles of different sizes, simulate their sound, and provide visual feedback on the bubble life-cycle. We present early reflections andpreliminary results from a small-scale qualitative evaluation. Participants naturally experimented with breathing intensity and described the interaction as engaging and playful. Based on the comments, we discuss how to approach the intricacies of coupling breath to sound in digital interfaces in future research.

Speaker: Georgios Marentakis (Østfold University of Applied Sciences)

109 Crossmodal Correspondences Between User Interface Sounds and Material Sensations: A Cross-Cultural Investigation

This study investigated crossmodal correspondences across cultures between acoustic properties of user interface sounds and five sensations: finish, weight, opacity, texture and physical state. Correspondences between these sensations and mean fundamental frequency (f0), melodic and loudness contour, and dissonance in 30 sounds were compared across three countries: France, China, and the USA. Results showed links between mean f0 and three sensations, finish, weight and opacity, in France and the USA but not China. Similarly, many links were shown between sensations and melodic and loudness contour and dissonance for France and the USA, whereas correspondences were rarer for China. Potential mechanisms for this cross-cultural difference include linguistic and cultural associations. These crossmodal associations (where they exist) have important implications for product sound design.

Speaker: Anjali Bhatara Morard (Ircam Amplify)

Full paper (PDF)

110 Exploring Sonic Feedback in a Truck Driving Simulator

The quiet sound profile of electric trucks can reduce auditory cues related to speed and acceleration, potentially affecting the driver´s sense of feedback andcontrol. Therefore, it is relevant to examine whether active sound design can optimize the driver experience. A simulator study was conducted to explorewhether adding complex tonal sounds can improve the driving experience and change drivers´ sense of feedback, and whether biometric measurements can capture human responses to the sounds. Sixteen participants experienced three different sound treatments during simulated driving. Two treatments provided enhanced sound feedback depending on vehicle speed and accelerator pedal position. Besides biometric measurements, subjective ratings and written comments were collected. Qualitative analyses of the results indicated that enhanced sound feedback can increase perceived sense of feedback and control for some drivers. Approximately half of the subjects appeared to ap preciate the added sound feedback. Galvanic skin response revealed a significant difference between the treatments and is recommended for further investigations of human response to sound feedback. Although attitudes toward and perceived need for active sound design vary considerably, the demand is still sufficiently pronounced to justify further research.

Speaker: Birgitta Nyman (Luleå University of Technology)

Full paper (PDF)

111 Evaluation of Psychoacoustic Tonality Models for Cabin Noise Across Vehicle Powertrains

Tonality strongly affects the perceived quality and annoyance of vehicle interior noise. This study evaluates two widely used tonality metrics—the Aures tonality metric and the Sottek psychoacoustic Hearing Model tonality metric (ECMA-418-2)—against subjective ratings collected in two complementary listening experiments. Stimuli were drawn from chassis-dynamometer recordings of four vehicles representing different propulsion technologies (two electric, one hybrid-electric, one diesel), captured at the driver and rear-passenger head positions under full and partial throttle. Study 1 used 112 unprocessed 3 s segments rated by 33 listeners. Study 2 used 48 order-tracking resynthesised stimuli, which preserved the spectral and tonal structure while suppressing temporal fluctuations and equalising loudness across stimuli; each of 19 listeners rated every stimulus three times in randomised order. Hierarchical clustering of Study 1 ratings revealed two listener subgroups with distinct evaluation strategies: one showed a moderate association with the Sottek metric (r = 0.59) but none with Aures, whereas the other showed only weak associations with either metric; ratings in both clusters also correlated with loudness. In Study 2, with loudness and temporal fluctuations controlled, both tonality metrics were strong predictors of subjective ratings (Sottek: r = 0.77; Aures: r = 0.72). Overall, the Sottek metric outperforms the Aures metric for complex, realistic stimuli, while both metrics perform comparably well under controlled conditions. Although tonality alone is a difficult perceptual attribute for non-expert listeners to evaluate, both metrics remain valuable as input features for higher-level perceptual models such as annoyance prediction.

Speaker: Zhenxian Li (INSA Lyon, LVA UR677)

Full paper (PDF)

A24.04 Audio for Augmented Reality: S446

Conveners: Annika Neidhardt (Audio Engineering, Faculty of Media, HS Mittweida), Nils Meyer-Kahlen (Aalto University), Johannes M. Arend (Aalto University)

112 Baseline localisation biases in audio-only virtual reality prior to investigating floor reflections

In real environments, direct sound from a source is typically followed by a strong floor reflection whose level and time delay depend on the location of the source and the listener, and may potentially provide additional localisation cues. However, prior findings on reflection effects are mixed, and evidence for virtual reality (VR) applications and environments remain limited. This question is especially relevant for blind and visually impaired (BVI) users, who may rely heavily on accurate auditory spatial cues in audio-only VR for tasks such as orientation and locomotion involving footsteps, cane-impact and echolocation. We developed a dynamic VR sound localisation platform with spatialised rendering of direct sound and options for reflections. This paper reports baseline results from initial experiments using direct sound only. Across stimuli, responses showed significant compression toward the horizontal plane (systematic elevation bias) for controller responses and head pointing. Elevation was a strong predictor of signed localisation error, and sound stimulus significantly modulated the elevation bias slope. In contrast, no robust effect of HRTF was observed on the elevation-bias function, and no sound and HRTF interaction was found. These baseline results provide a reference for subsequent experiments testing whether adding floor reflections reduces elevation-bias compression in VR. We hypothesise that, relative to this baseline, adding floor reflections will reduce elevation-bias compression.

Speaker: Max Væhrens (Department of Electronic Systems, Aalborg University)

Full paper (PDF)

113 Plausibility of Sound-Source Directivity Resynthesis in a Contemporary Edison Test

At the beginning of the 20th century, the Edison company presented so-called "tone tests" to promote their new phonograph. To demonstrate its quality, listeners were presented with live performances and reproductions of different sound sources. Here, we perform a contemporary Edison test that assesses the influence of different methods for resynthesizing sound-source directivity on the plausibility of musical reproduction in a chamber music hall. Performances of violin, oboe and trumpet and reproductions thereof were played behind an acoustically transparent curtain. The participants then had to decide if it was a live performance or a loudspeaker reproduction. The reproduction methods included a studio loudspeaker and a twelve-channel dodecahedral speaker that was operated either with a one-channel recording fed to all drivers or with a twelve-channel recording from a dodecahedral array surrounding a performing musician. Additionally, beamforming was applied in one condition to enhance low-frequency control. In about 48% of all trials, participants accepted the reproduction as a live performance. There were differences between the reproduction methods, indicating that reproductions using the twelve-channel array, where every driver is fed its own signal, might be more plausible than using omnidirectional reproduction or a studio speaker.

Speaker: Otavio Colella Gomes (Hochschule für Musik Detmold)

Full paper (PDF)

114 Perceptual Relevance of Source Width in Binaural Rendering of Human Speakers

In recent years, numerous studies have investigated the plausibility of binaural rendering. Typically, the auralized sound sources are assumed to have point-source characteristics and static directivity. While this can be considered a good representation of loudspeakers, which typically have a well-defined acoustic center and time-invariant directivity, other natural sound sources, such as the human voice, may have different properties.This study presents the results of an initial listening experiment that determined the relevance of the ’source width’ aspect when considering the sound radiation of human speakers. The experiment assessed static binaural recordings of frontally positioned human speakers in an anechoic environment using several attributes from the Spatial Audio Quality Inventory (SAQI). Two human speakers each articulated two different sentences at distances of 1.5 m and 3 m to the listener. The stimuli were compared to the reference of the same human speakers themselves articulating the same sentence. Visual cues were removed by placing a curtain between the speakers and the listeners. Two variants were analyzed: in the first, the binaural recording was presented dichotically; in the second, the left-ear signal was presented diotically to both ears. This diotic presentation of equal signals to both ears can be considered comparable to the perception of a frontal, horizontally non-distributed sound source.The results of the listening experiment revealed no significant differences in the considered SAQI attributes between the two variants, suggesting that representing a human speaker by a point-like sound source with a defined acoustic center is sufficient for plausible binaural rendering.

Speaker: Tobias Weber (TH Köln - Institute of Computer and Communication Technology)

Full paper (PDF)

115 Speech Recognition and Listening Effort with Real and Virtual Speech in Variable Acoustics

Augmented reality (AR) telepresence systems often aim to enhance immersion by adding room-adapted reverberation to remote speech, potentially at the cost of reduced intelligibility and increased listening effort. In this study, we investigated speech recognition and subjective listening effort in a speech-in-noise task across real and virtual acoustic environments with varying degrees of reverberation. Twenty participants performed a sentence recognition task in babble noise under seven conditions, including dry, reverberant, and highly reverberant scenarios, as well as mixed conditions in which the virtualized target remained dry rather than being matched to the acoustic environment in which the noise was placed. Results confirmed that increased reverberation reduces word recognition and increases listening effort. Notable discrepancies were observed between real and virtual presentation, particularly in dry conditions. Importantly, omitting room-adapted reverberation for the virtualized target speech provided only limited benefits for intelligibility and effort. These findings highlight trade-offs in AR audio design and suggest that the acoustics in which the background noise is placed may be more critical than that applied to the target speech in determining listening effort.

Speaker: Nils Meyer-Kahlen (Aalto University)

Full paper (PDF)

A02.01 Machine Learning in Bioacoustics: S252

Conveners: Daniel Haider (Acoustics Research Institute, Austrian Academy of Sciences), Jure Zeleznik (Acoustics Research Institute, Austrian Academy of Sciences)

116 Deep-Learning-Based Detectors For the Calls of the Shallow Water Common Spadefoot Toad (Pelobates fuscus)

In the context of the decline of the European common spadefoot toad (Pelobates fuscus), long term bioacoustic monitoring of this secretive species that vocalizes underwater is highly relevant. In this study, we present a software detector that features superior performance compared to literature [1]. For this study, hydrophone data was collected in Denmark and Poland with a different recorder than in [1]. Two candidate detectors were designed: a spectrogram-based EfficientNet-B0 classifier and a waveform-based CNN-BiGRU one. Both were trained on existing French hydrophone data and the new Danish Hydromoth data with call-aware losses and post-processing. The spectrogram-based model reached TPR 84.5% at FPR 0.11% with 94 % accuracy. The raw-waveform model reached TPR 75% at FPR 0.02% with 93 % accuracy. The new detectors were used to turn long recordings into phenologies.

Speaker: Guillaume Dutilleux (NTNU/IE/IES/Acoustics)

Full paper (PDF)

117 Sound Synthesis and Physics-Based Data Augmentation for the Detection of Eurasian Lynx Calls

In Scandinavia, the Eurasian lynx (\textit{Lynx lynx}) is a hunted species whose populations must be monitored. We present a deep-learning-based software detector of E. lynx calls. It can be applied to long-term audio recordings carried out during the breeding season where the species produce loud and far-reaching calls. To our knowledge, this is the first published software detector for this species. Considering the large data sets required by deep learning, to compensate for the dearth of E. lynx recordings, we harnessed sound synthesis and data augmentation. Starting from a statistical analysis of the handful of available recordings of lynx calls, we designed a spectrogram synthesizer based on additive sound synthesis. The diversity in the synthetic data set was further extended using a subset of an engineering outdoor noise prediction model to simulate the received spectrograms as a function of topography, source-receiver geometry and other site-specific parameters. A one-channel neural network inspired from LeNet-5 was trained on this synthetic data. The E. lynx detector trained from scratch achieved 95.4 % accuracy. This novel approach of repurposing noise prediction schemes should be valuable in other applications of supervised machine learning for a wide range of outdoor sounds.

Speaker: Guillaume Dutilleux (NTNU/IE/IES/Acoustics)

Full paper (PDF)

118 Evaluating unsupervised clustering of deep embeddings for unknown species identification in bioacoustics

Natural soundscapes, especially in hyper-diverse environments often contain unknown sound events. Strategies to identify unknown sound events, especially those produced by species not included in training sets have traditionally been infeasible to manually identify. Recently, the classification performance of bioacoustic deep learning models is becoming competitive with human annotators. Yet, classifier predictions are often unreliable for hyper-diverse soundscapes and under-represented regions or taxa, and for unknown species they are inherently ill-defined. Recent research suggests that by using the feature extraction capability of state-of-the-art bioacoustic deep learning models, similar sounds can be clustered together in the feature space despite the model never having encountered these sounds during training. However, current methods to evaluate the ability of bioacoustic feature extractors to specifically identify unknown sounds are limited. Therefore, in this study we curate a dataset of under-represented species vocalizations from diverse taxonomic groups (i.e. birds, mammals, amphibians and insects) and superimpose background noise from different habitats (i.e. tropical and temperate forests). By doing so we are able to vary signal-to-noise ratio (SNR) and background noise while mimicking real recorded soundscapes. We evaluate a wide variety of acoustic deep learning models and clustering algorithms. Preliminary results show that models trained on a large variety of species are capable of clustering sounds from species they have not been shown during training over different noise environments for high SNR values. This evaluation method provides a fine-grained understanding of the limitations of current deep learning models and, when paired with interactive visualizations, can be used to inspect how models handle unknown sounds.

Speaker: Vincent S. Kather (Naturalis Biodiversity Center)

119 Evaluating Noise Reduction for Underwater Bioacoustic Event Detection Using Synthesised Acoustic Recordings

Marine soundscapes contain valuable information about local ecosystems, making comprehensive analysis of their constituent events essential for accurately assessing ecosystem health. However, they are subject to high levels of continuous noise, hindering our ability to disentangle these events and analyse the soundscape. One potential approach to mitigating noise is the use of a noise-reduction front end. However, because noise-reduction algorithms alter the input distribution seen by the pre-trained embedding networks used in event detection, they must be applied carefully. At the same time, appropriate tools for synthesising marine acoustic recordings at controlled signal-to-noise ratios (SNRs) are currently not available. In this work, we introduce an open toolbox for synthesising controlled recordings aimed at evaluating noise-reduction front ends. Users can select the number of events, the SNR, and the length of the recorded files to simulate underwater recordings containing acoustic events originating from biological, anthropological, or geological processes from the Marine Soundlib. We introduce binary time-frequency masks indicating where event energy is located. These masks are used in the mixing process and in evaluating noise reduction, allowing us to assess performance in the time-frequency bins where event energy is active. We then mix the extracted events with noise recordings from different locations at different SNRs to assess noise-reduction effectiveness. Using this toolbox, we show the relation between event detectability and SNR, and how different pre-trained embedding networks react to noise reduction using a simple Wiener filter; some models show improved event-detection performance, whereas others degrade. This illustrates that the effects of noise-reduction methods should be studied carefully within a detection pipeline.

Speaker: Bram Cuyx (Flanders Marine Institute (VLIZ))

A02.02 Effects of noise and light on terrestrial and aquatic biodiversity: S009

Convener: Dick Botteldooren (U-Ghent)

120 Disentangling Animal Vocalisations and Anthropogenic Sound to Assess Noise Effects on Biodiversity

Passive acoustic monitoring (PAM) has become a widely used tool for biodiversity assessment, particularly for monitoring vocalising animal species. However, in soundscapes dominated by natural or urban background noise, or in the presence of human voices, reliable detection and classification of animal vocalisations becomes challenging for widely used pretrained classifiers, such as BirdNET, and even expert auditory analysis. Conversely, when assessing the impact of anthropogenic noise such as the noise of high-altitude aircraft, intense biological activity—such as bird dawn chorus—can interfere with the monitoring.To address these challenges, we investigate artificial intelligence–based sound separation as a preprocessing step for both biodiversity monitoring and environmental noise assessment. Two approaches are explored. The first approach, inspired by speech enhancement techniques, employs complex ideal ratio masks trained with a U Net architecture to separate bird vocalisations from natural or urban background sounds. Results show that this effectively enhances bird sound separation and improves the accuracy of automatic classification.The second approach is based on a recently proposed transformer based Task-aware Unified Source Separation (TUSS), which allows token controlled sound extraction. By fine tuning the pretrained model for task specific targets, such as aircraft noise or animal vocalisations, the system accuracy increased. When applied to aircraft noise, the separated sound can be directly used to estimate its contribution to overall noise levels. When applied to animal vocalisations, it can serve as an enhanced input for subsequent PAM.These results demonstrate that AI based sound separation can support a more robust assessment of the effect of anthropogenic noise on biodiversity in complex real world soundscapes.

Speaker: Dick Botteldooren (U-Ghent)

Full paper (PDF)

121 When Mountains Meet the City: Hybrid Soundscape Monitoring in Andorra la Vella with Sensors, AI and Citizen Science (ECHO-POLIS)

This paper presents ECHO-POLIS, a 24-month project to characterise and interpret the soundscape of a mountain city—Andorra la Vella and Escaldes-Engordany—through a hybrid approach that integrates objective acoustic monitoring, AI-based source analysis, and citizen perception. The experimental design combines a fixed sensor network deployed across contrasting urban typologies (pedestrian commercial areas, traffic corridors, green spaces, mixed plazas, and quieter residential pockets) with mountain and transition points located along plausible “sound pathways” towards the city. This dual-layer configuration supports a central question: whether, when, and under which meteorological and activity conditions natural mountain sounds (water, wind, fauna) measurably influence the urban sound environment and its perceived quality. Continuous time series are collected (e.g., LAeq profiles and temporal variability descriptors) and complemented by synchronised seasonal campaigns (summer and winter) to capture tourism-driven dynamics and environmental variability. Acoustic data are processed using sound event detection and AI-based source classification to derive indicators of source composition, dominance, intermittency, and time-of-day patterns, consistently computed across urban and mountain sites. In parallel, participatory soundwalks, short ISO 12913-aligned perception surveys, and mobile contributions provide contextual and perceptual descriptors across key target groups (schools, residents, shopkeepers, and tourists). The project delivers physically grounded yet user-relevant hybrid indicators, interactive maps and dashboards for non-expert stakeholders, and a transferable methodological toolkit to support replication in other urban–mountain territories. A final co-validation phase with local authorities links indicators to planning decisions, producing two adoptable action plans (urban soundscape management and mountain-sound integration) and guidance for standardised deployment and reporting across seasons.

Speaker: Rosa Ma Alsina-Pagès (La Salle, Universitat Ramón Llull)

122 Assessing the impact of outdoor music festivals on biodiversity

Outdoor music festivals are often organized outside urban areas to minimize nuisances for city dwellers. This results in possible proximity of festivals to natural areas and potential impact on biodiversity in these areas. However, assessments of festival impact often use anthropocentric modelling or monitoring, approaches. Music is targeted at human experience and therefore its frequency range and spectral temporal modulations aim at creating an emotional response in humans. Other organisms may exhibit different responses to humans. Hence, it is proposed to focus assessment on the relevant taxa by including their hearing thresholds. Moreover, we propose combining impact monitoring with passive acoustic monitoring (PAM) of biodiversity. The proposed approach is tested at a medium-sized outdoor music festival in Flanders. During successive editions, directivity of the loudspeaker arrangement was suggested as a mitigation measure. To validate the directivity of the loudspeaker arrangement, drone-based mobile acoustic measurements were performed.

Speaker: Marcel Kok (U-Ghent)

Full paper (PDF)

123 Towards biodiversity-aware road traffic noise mitigation measures

Environmental noise of anthropogenic origin is increasingly recognized as a major threat to biodiversity in natural areas. These areas are often intersected by major highways, exposing large adjacent zones to road traffic noise. Current knowledge and guidelines on traffic noise mitigation are largely focused on human listeners. However, road traffic noise, outdoor sound propagation, mitigation measures, and auditory sensitivity of terrestrial animals are all strongly frequency dependent. As a result, their combined effect on animal sound exposure is far from straightforward. In this simulation study, the Harmonoise/Imagine sound propagation model—well suited to capturing the complex physics of sound propagation in natural environments—is used to predict sound exposure levels for a wide range of terrestrial animals. The simulations account for species-specific receiver heights and measured hearing threshold spectra. This approach allows a systematic evaluation of the effectiveness of various road traffic noise mitigation measures, including noise screens, natural berms, depressed road segments, local landscape depressions, and traffic management strategies such as vehicle speed reduction or bans on heavy vehicles, in reducing sound exposure for wildlife.

Speaker: Timothy Van Renterghem (Ghent University)

Full paper (PDF)

124 AI-supported Large-Scale Passive Acoustic Biodiversity Monitoring in German National Parks

Large-scale, standardized biodiversity monitoring is critical for assessing ecosystem responses to climate change and increasing anthropogenic pressures in protected areas. The KI-Nationalpark project establishes a nationwide biodiversity monitoring network across 13 national parks and two wilderness areas in Germany. Combining camera trapping with passive acoustic monitoring over a 13-month period starting in October 2025, the project enables standardized assessments of biodiversity and human-induced pressures to support adaptive management.This contribution focuses on the acoustic monitoring component, comprising 304 recording positions across 11 study regions, each equipped with paired AudioMoth devices for simultaneous recording of bird vocalizations and bat echolocation. Within the first three months of monitoring, more than 38 TB of acoustic data were collected, highlighting the scale of the approach.Acoustic data are processed with AI-based pipelines to classify species and anthropogenic noise. For bats, Kaleidoscope Pro, a non-AI template-matching tool, provided a first pass to flag files with bat calls and classifications; validated detections will train a dedicated AI classifier. We show spatial and temporal activity patterns: detection rates declined steadily from about 1.7% of recordings in October to below 0.1% in January, reflecting hibernation phenology. At a conservative match-ratio threshold (>=0.95), 21 species were detected, though one (Nyctalus lasiopterus) is not established in Germany and likely reflects call confusion, requiring expert verification. Notably, eight of eleven regions already hold winter recordings, with confirmed activity of nine species despite very low call rates.This project is funded by the BMUKN as part of the Action Programme for Natural Climate Protection under the AI Lighthouses funding scheme.

Speaker: Ursula Verfuss (NA Nature Analytics GmbH)

A15.00 Psychoacoustics: S268

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Kai Siedenburg (CvO University of Oldenburg), Jan Rennies (Fraunhofer IDMT-HSA, Oldenburg), Sarah Verhulst (Ghent University), Bernhard Laback (Austrian Academy of Sciences)

125 Psychoacoustic-based acoustic analysis recommendation based on the symbiose of domain knowledge and AI models

Loudness, sharpness, tonality, roughness, fluctuation strength, relative approach, Fast Fourier Transform, modulation spectrum (and many others...). The number of acoustic analyses can be daunting for inexperienced users. One consequence of this is that even when there may be a perfect match between an industrial problem and an acoustic solution, this link can be obfuscated by the lack of user expertise. We propose a new solution paradigm for this problem: we train an AI model able to perceive the prominent psychoacoustic characteristics of sounds and combine it with domain knowledge for improved analysis parameterization. This allows us to tailor analysis recommendations based on the symbiosis of how the sound is perceived by people together with expert domain knowledge. This is achieved by a psychoacoustic-based network pre-training coupled with a fine-tuning step using expert-generated listening test results on strongly diverse industrial sounds. The results of the model are then piped into a rule-based model extracted from expert knowledge. We show that the model can achieve good results in predicting expert ratings and can produce helpful recommendations. This is a first step in the direction of fully automated acoustic analysis recommendations, which we believe can empower a whole class of non-expert users.

Speaker: Andreas Herweg (HEAD acoustics GmbH)

Full paper (PDF)

126 Exploring Binaural Beat Delivery via Silent Disco Headphones for Large Studies: A Dataset and Analysis

This paper presents proof of concept of studying the psychological effects of binaural beat (BB) stimulation with silent disco headphones as means of delivery. It describes a comprehensive dataset and analysis on the effects of BB embedded in music on working memory and self-reported relaxation. The experiment employed a between-subjects design with three conditions (10-Hz BB, 40-Hz BB, and control) and involved 317 participants. Participants were exposed to a 13-minute multimedia art performance in a planetarium theatre using wireless silent disco headphones as means of delivery of the auditory stimulus. Data were collected via smartphone-based Qualtrics surveys, capturing pre-performance demographics, including hormonal fluctuations, sound sensitivity and personality traits, as well as post-performance self-reported relaxation and working memory performance (n-back task). The dataset includes raw and processed data, alongside supporting materials such as survey instruments, experiment code, a codebook, and informed consent documents. This resource enables the exploration of individual differences and their potential influence on BB efficacy. Analysis confirmed the potential of 10-Hz BB for enhancing relaxation, however 40-Hz BB effects on working memory and mediating effects of personality and assigned sex were not observed.

Speaker: Kirsten van den Bosch (University of Groningen)

127 Auditory Function as a Potential Early Marker of Alzheimer’s Disease: Preliminary Study

Background. More than a century after the first description of Alzheimer’s disease (AD), its etiology remains poorly understood and currently available treatments offer limited efficacy. The socioeconomic burden is substantial, with care costs projected to exceed €250 billion annually in Europe by 2030. Aging is the strongest risk factor, making prevention and early diagnosis key public health priorities. Importantly, AD pathology develops silently over many years, highlighting the need for early detection strategies.Objective. This study aims to investigate auditory function in individuals at prodromal or an early stage of AD, and to compare the results with those obtained from a healthy control group.Methods. The study included a set of audiological and psychoacoustic assessments. Basic audiological tests—pure-tone audiometry, impedance audiometry, and distortion product otoacoustic emissions (DPOAE)—were conducted to exclude peripheral hearing impairment. Central auditory processing was evaluated using auditory brainstem responses (ABR). Binaural processing abilities were assessed using three tests: speech intelligibility in noise, dichotic listening with non-directed attention, and binaural masking level difference (BMLD).Results. Preliminary results indicate differences in auditory processing between AD individuals and healthy controls, particularly at higher levels of the auditory pathway and in binaural processing tasks.Conclusions. The findings suggest that auditory processing measures, particularly those reflecting central and binaural mechanisms, may serve as promising early indicators of AD. Nevertheless, further research is needed to confirm these observations and explore their clinical applicability.

Speaker: Julia Wesołowska (Adam Mickiewicz University)

128 Listening effort in audiovisual matrix sentence tests: Evaluating the effect of virtual avatars

This study investigates listening effort in audiovisual speech perception, with a focus on the use of virtual avatars as an alternative to real speakers. An audiovisual version of the Italian Matrix Sentence Test was developed, together with a dedicated toolbox (SPiNE - SPeech in Noise & Listening Effort Assessment) for administering speech intelligibility and listening effort measurements. The audiovisual material was created using recordings of a real speaker and a corresponding avatar generated through audio-driven facial animation within Unreal Engine. Two experiments were designed to compare speech intelligibility and listening effort between real and avatar-based stimuli under controlled conditions. The experiments included different types of background noise and signal-to-noise ratios. Listening effort was assessed through both behavioral (i.e., response time) and subjective measures, alongside speech intelligibility metrics. The study aims to provide an ecologically valid methodology for audiovisual speech testing. Indeed, by enabling precise control over visual speech cues, virtual avatars represent a promising tool for advancing research on listening effort and for supporting future clinical applications.

Speaker: Chiara Visentin (University of Ferrara)

Full paper (PDF)

129 Investigation of methods for assessing the audibility of warning sounds in train cabs: comparative analysis of perceptual testing and predictive indicators

This study was carried out within CEN TC256 WG3 as part of ongoing work towards a new European standard for evaluating audible warnings in train drivers’ cabs. The purpose was to investigate and compare methods of assessing audibility to inform the proposal for a standardised methodology.Controlled subjective tests were conducted in a mock-up train cab, where ambient noise recordings from several train cabs were replayed at realistic levels and eight representative warning sounds were superimposed. Participants rated the audibility of each sound presented at each of seven signal-to-noise ratios. These subjective results were compared to the TSI LOC&PAS homologation requirement (+6 dB signal-to-noise ratio) and to sound levels computed using ‘Detectsound’, a tool developed by the University of Ottawa to predict the audibility of sounds in noise from the one-third octave band spectra of the signal and noise. Detectsound models the auditory filter characteristics of the ear and takes into account the hearing abilities of different listener populations.Initial findings confirm that the +6 dB signal-to-noise criterion aligns with perception for speech signals but tends to be conservative for tonal warning sounds, resulting in warning sounds that are sometime louder than desirable. Detectsound calculations show closer agreement with perceived audibility, although discrepancies remain. Detectsound also highlights the substantial influence of hearing ability. Warning sound levels need to be increased by 5 dB to 7 dB for persons just meeting the minimum hearing requirement for train drivers in Europe compared with persons with normal hearing.

Speaker: Fabrice Aubin (SNCF Voyageurs)

Full paper (PDF)

A16.05 Performing Arts Spaces and Rehearsal Rooms: S250

Conveners: Francesco Martellotta (Politecnico di Bari, DARCOD), Jamilla Balint (Rohde Acoustics)

130 Acoustic Design of Orchestra Rehearsal Hall Renovation of the Finnish National Opera and Ballet

The orchestra rehearsal hall of the Finnish National Opera and Ballet is situated below the stalls of the main hall of the Opera. This space was originally intended to be use as a musical library, but it was realized in the late stage of design, that the orchestra also needed a rehearsal room. Because of this, the finished room was oddly shaped and had less than ideal height. This caused several problems and the hall was never considered to be well suited for its purpose. In this paper we describe the acoustic renovation of the rehearsal hall, and the way an electroacoustic system was integrated to the overall acoustical design. In spite off the challenging starting point, the project was a success, and the orchestra now has a rehearsal hall with lower sound level, longer reverberation time and overall, more favourable acoustic conditions.

Speaker: Anssi Ruusuvuori (Akukon Oy)

131 Balancing Voices and Ensembles: Methods for Tuning and Assessing Adaptable Performance and Rehearsal Spaces

Acoustically adaptable performance spaces offer the possibility to actively shape their sonic response, allowing them to be “tuned” in relation to specific artistic requirements. In this sense, the space itself can be approached as an instrument, whose parameters—such as reverberation, clarity, and spatial distribution—are adjusted to support different ensembles and repertoires. A central challenge lies in defining appropriate strategies for tuning these environments, particularly with regard to perceptual aspects such as the audibility between musician groups and the balance between singers, orchestra, and choir.This contribution examines methods for assessing and adjusting such spaces, with a focus on capturing interaction within and across performer sections. While established room acoustic metrics provide a useful foundation, they often fall short in describing the nuanced relationships that emerge in complex performance settings. The paper therefore addresses the limitations of current measurement approaches and explores complementary strategies that better reflect ensemble conditions.In addition, the role of participative design processes is highlighted. Involving musicians and other stakeholders in the tuning and evaluation of the space can support more informed adjustments and tends to increase acceptance of the resulting acoustic conditions. The study aims to contribute to an ongoing discussion on how architectural design, acoustic adaptability, and user involvement can be integrated, and how such processes can be supported by appropriate measurement and evaluation methods.

Speaker: Jamilla Balint (Rohde Acoustics)

Full paper (PDF)

132 Vibro-Acoustic characterisation of a historic theatre ceiling: structure-hall interaction

In the context of a research project conducted at the Teatro Comunale di Ferrara concerning the assessment of vibrations induced by amplified music events, particularly their impact on delicate architectural elements, this study investigates the dynamic response of the hall’s ceiling, assessing its contribution to the hall's overall acoustics. An experimental vibro-acoustic analysis was conducted by measuring sound pressure levels in the hall and vibration levels on the ceiling, induced by either an airborne sound source or impact excitation on the ceiling structure. Furthermore, acoustic and structural reverberation times were measured respectively in the room and on the ceiling. Experimental observations indicate that the influence of the ceiling on the hall's acoustics may be only substantial at very low frequencies. For a deeper investigation, a hybrid model was implemented to analyse this vibro-acoustic interaction. This approach utilised numerical Finite Element (FE) analysis for low frequencies and Statistical Energy Analysis (SEA) for mid-to-high frequencies. The model results were finally validated against a different set of experimental data.

Speaker: Andrea Santoni (University of Ferrara)

133 Analysis of Changes to the Early and Late Response in an Existing Orchestral Rehearsal Studio

Successful rooms for orchestral and chamber music allow the musicians to balance their own playing as part of the full ensemble. Work by Marshall in 1978 inspired Gade to propose measures in 1989 that addressed these two factors. Recent work in this field has verified that both the early and late room response have a significant effect on what musicians hear and can respond to while playing.Tuning a 7,500m3 orchestral rehearsal studio in Brisbane in late 2025 provided an opportunity to analyse the subjective impressions of 70 musicians with repertoire ranging from a Shostakovich symphony to a wind quintet and a string music quartet.The early reflection strength and late room response were independently varied by changing the height and orientation of an existing large array of overstage reflectors and varying the extent of operable absorption banners.This paper describes the creation of a loosely coupled reverberant volume in the upper part of the room following the design theory of reflection sequence halls.This was activated by directing some of the early energy from the reflectors. Subjective responses and objective data have been analysed to provide insight into preferred conditions.

Speaker: Peter Exton (Creative Acoustics)

Full paper (PDF)

134 Acoustic Characterization of the Theatre Diogo Bernardes (Ponte de Lima)

This paper provides an acoustic characterization of the Theater Diogo Bernardes (Ponte de Lima, Portugal). The main results of the measured acoustic parameters were: LAeq(HVAC) at 26 dB, Speech Transmission Index ranging from 0.56 to 0.71, Reverberation Time (avg 500-1k Hz) of 0.99 s, C80 (avg 500-1k Hz) of 4.7 dB, D50 avg 500-1k Hz) of 0.59, and Center Time (avg 500-1k Hz) of 63 ms. A comparison is made with 19 Portuguese halls and five international horseshoe-type theaters.

Speaker: Afonso Pires (Faculty of Enginnering U. of Porto)

Full paper (PDF)

A22.06 Passive Acoustic Monitoring (PAM) for Biodiversity, Ecosystem Assessment and Ship noise: S473

Conveners: Paolo Diviacco, María Campo-Valera (Universidad Internacional de la Rioja)

135 In-Field Setup to study the Effects of Underwater Noise on Marine Fouling Comunity

Maritime traffic represents a growing contributor to underwater acoustic pollution. Laboratory studies on noise impacts benefit from tightly controlled conditions, making it possible to reliably compare organisms exposed to sound with those kept in near-silent environments. By contrast, field-based experiments are often complex, particularly in confined and shallow systems such as the Venetian Lagoon, where preliminary surveys have revealed widespread background noise and a lack of quiet reference areas suitable for biological comparisons. Here we describe an experimental setup designed to evaluate the influence of noise on marine fouling communities. Two prototype structures (50×50×80 cm) were constructed: one made of stainless steel (internally lined with polyurethane) and the other of black nylon. Their capacity to attenuate a synthetic sound signal, generated by a submerged loudspeaker positioned 1 m from a hydrophone, was then assessed. Results indicated that the coated steel structure reduced sound levels by approximately 27 dB, while the nylon one achieved a reduction of about 11 dB. Based on these findings, we plan to build replicas incorporating a plexiglass panel (25×25 cm) to support fouling settlement and development. This system will enable comparative biological studies between environments with different levels of noise pollution.

Speaker: Giovanni Zambon (University of Milano-Bicocca)

Full paper (PDF)

136 Long-term acoustic monitoring in the deep Mediterranean: the INFN-LNS cabled infrastructure and the multidisciplinary observatories

The deep-sea cabled infrastructure operated by INFN-LNS provides a platform for continuous acoustic monitoring in the Ionian Sea, southeast of Sicily, Italy. In recent years, it has expanded significantly, integrating different observatories and sensing technologies for long-term monitoring of the underwater acoustic environment. The system includes the Test Site offshore Catania and the deep-sea observatory at Capo Passero, where the KM3NeT/ARCA telescope is located. These sites are connected to shore laboratories through electro-optical cables for power supply and real-time data transmission, with termination frames (junction boxes) that allow the connection of different sensors. A central element of this system is the ITINERIS long-baseline hydrophone array deployed at the Capo Passero site. The array is composed of three synchronized broadband receivers spaced by approximately 300 m and operating continuously. The hydrophones are calibrated over a frequency range from a few Hz up to about 80 kHz. This provides the opportunity to analyze temporal variability and spatial coherence of ambient noise, as well as to identify and localize acoustic sources. The system is expected to be further expanded with the deployment of additional hydrophones. In this work, an overview of the infrastructure and preliminary results from the hydrophone data analysis are presented, along with the approach adopted for long-term data processing and archiving.

Speaker: Abdelghani Idrissi (INFN - LNS, Unict-DFA)

137 Temporal and Environmental Drivers of Soundscape Composition in Galway Bay: An Ecoacoustic Analysis of Passive Acoustic Monitoring Data

The underwater soundscape off the west coast of Ireland is a dynamic acoustic environment shaped by a complex interplay of natural sounds and anthropogenic noises. This region, exposed directly to the north Atlantic, experiences pronounced seasonal and inter-annual fluctuations driven by meteorological forces, oceanographic processes, and biological rhythms.This study investigated the impact of environmental conditions on the underwater soundscape of Galway Bay using passive acoustic monitoring data collected from SmartBay, an integrated marine observatory located off the coast of Spiddal, Co. Galway. The platform collects simultaneous hydrophone and environmental data, facilitating direct comparison between acoustic and environmental records. Data were gathered at one-minute intervals over a week for each season from February 2023 to December 2024, with supplementary recordings obtained for notable weather events. Acoustic indices (acoustic complexity index, normalized difference soundscape index, acoustic diversity index, and acoustic evenness index) were used to characterise the underwater soundscape, and were compared against environmental factors including tidal state, current speed, and calendar season.The soundscape was dominated by geophony year-round, with this dominance being noticeably more intense during the winter months. Sea temperature and wave height emerged as the proximal (environmental) drivers across seasons, with tidal phase having a statistically significant influence on key indices (Kruskal-Wallis, p<0.001). Ebb tide was consistently associated with elevated acoustic complexity, while diel patterns in biological activity indices were most pronounced during calmer seasonal periods.

Speaker: Conor Byrne (The Galway Sound Lab, School of Engineering,University of Galway)

Full paper (PDF)

138 Multi-probe Analysis of the Marine Soundscape: Approaches and Results from the VONGOLA Project

The VONGOLA project (Visual and noise-eNhanced AI Analysis for Marine Biodiversity Monitoring, Observation and LeArning), led by CSFNSM (Centro Siciliano di Fisica Nucleare e di Struttura delle Materia), developed innovative approaches for monitoring marine biodiversity and evaluating the impact of human interference on marine ecosystems. To achieve these goals, an heterogeneous monitoring network was established, integrating different acoustic technologies to investigate both deep-sea and shallow-water environments. One task enabled the first permanent Distributed Acoustic Sensing (DAS) system in the Mediterranean Sea. Operated offshore Catania and connected to an existing submarine electro-optical cable managed by INFN-LNS, the DAS interrogator transformed 41 km of optical fiber, stretching from the coastline down to the deep sea, into a dense array of virtual acoustic sensors. By the end of 2025, the system was moved to the INFN infrastructure in Portopalo di Capo Passero where it identified acoustic signals from fin whales. Another task was dedicated to shallow-water multiparametric recordings using multi-probe observatories. This includes a real-time cabled station at the NOEL laboratory, featuring the first cabled station in the Messina Strait, and an autonomous station in the Plemmirio Marine Protected Area. This contribution presents a preliminary analysis of the marine soundscape, demonstrating the successful detection and classification of geophysical, biological and anthropogenic sources. These observations were conducted in real-time and on a continuous basis, following FAIR principles to ensure data interoperability. The results were obtained by combining standard analysis with Machine Learning algorithms

Speaker: C. De Maria (UniCT -DFA, CSFNSM)

Full paper (PDF)

139 Underwater radiated noise level database collected in shallow water in Baltic Sea

Underwater noise emission of vessels is usually declared by radiated noise level (LRN). LRN must be tested in Finnish seas using a method which is suitable for shallow water (30150 m). There is not much public data about the LRNs of vessels sailing in Finnish seas nor vessels built in Finnish shipyards. Knowledge of LRN of current vessels is needed in maritime industry since IMO has encouraged all member countries to reduce LRN in the future. Our purpose was to describe the method which was recently applied in Finland to collect LRN data, and to demonstrate the data obtained so far. All measurements were undertaken within 10  20 000 Hz using hydrophone recorders and applying “DNV shallow” test method. Measurement stations were placed along such shipping lanes, where the traffic is frequent. The station consisted of two hydrophone recorders. The movements of vessels were obtained from AIS system. Vessel passes which took place sufficiently close to the station were accepted and analyzed. So far, 645 passes have been analyzed. The database consists of 164 different vessels. The method will be applied in 6 more shipping lanes during 2026. The final database will be available in 2027.

Speaker: Valtteri Hongisto (Turku University of Applied Sciences)

Full paper (PDF)

A24.01 Auralization of complex environments: S448

Conveners: Josep Llorca-Bofi (Fraunhofer Institute for Building Physics), Jonas Heck (Institute for Hearing Technology and Acoustics)

140 Room acoustic 3D models from automatically meshed point clouds

This study addresses the challenge of generating reliable three-dimensional models for room acoustic simulations from massive data capture techniques, such as photogrammetry and terrestrial laser scanning. Although these methods provide high metric precision, their direct use in acoustic simulation remains limited by geometric and topological issues arising from automatic meshing, including excessive complexity, inconsistent surfaces, and insufficient control over mesh simplification.We propose a parameter-controlled meshing methodology for deriving acoustic simulation models from captured spatial data while preserving the essential geometrical features of complex architectural spaces. The novelty of the proposed methodology lies in the combination of Poisson Surface Reconstruction, Alpha Wrapping, and controlled quadric-based mesh simplification, with parameters defined in relation to the input point cloud density. The approach seeks to reduce operator subjectivity and improve the standardization of model generation by balancing geometric accuracy, computational efficiency, and acoustic relevance.The methodology is assessed through historically significant architectural case studies, showing that geometric simplification can produce acoustically relevant deviations when not properly controlled. The proposed workflow therefore provides a framework for generating more consistent and reliable 3D models for room acoustic simulation from reality-based data.

Speaker: David Infantes-López (Universitat Politècnica de Catalunya)

Full paper (PDF)

141 How does the room's modelled level of detail influence the choice of scattering for auralization?

Architectural room models used for auralization are typically created with different geometric levels of detail (LOD), which directly affect both the surface areas and the effective scattering behaviour, and thus introduce uncertainty into room acoustic simulations. This study reviews the influence of geometric LOD on the choice of scattering and proposes a compensatory modelling strategy for auralization. Two auditoria are investigated as case studies: the Gasteig philharmonic hall in Munich and the chamber music hall of the Konzerthaus in Berlin. An overview of the modelling challenges and the auralization opportunities will be given.

Speaker: Josep Llorca-Bofi (Fraunhofer Institute for Building Physics)

Full paper (PDF)

142 Simulation-based prediction of RIR measurement locations

The advent of robotic-aided acoustic measurements has enabled the precise acquisition of Room Impulse Responses (RIRs) across dense sampling grids, which is essential, e.g., for documenting the unique acoustics of cultural heritage sites and music venues. Despite these technological advances, capturing a high volume of RIRs in expansive spaces remains a time-intensive process, and determining the most salient measurement points would be ideal to optimize this process. This study investigates the use of basic room simulations to predict the necessary spatial density of RIR measurements. Our approach is perceptually driven, incorporating the Just Noticeable Differences (JNDs) of objective acoustic parameters as defined by ISO 3382-1, such as Early Decay Time (EDT) or Clarity (C80).Our predictions are grounded in physical measurements from a high-density RIR dataset and benchmarked against a basic geometric-acoustic simulation of the same environment. Results from the case study show that such simulations can effectively estimate the required measurement resolution. Moreover, the optimal resolution depends on the chosen objective metrics and their associated JNDs.Simulation-based prediction thus provides a practical framework for optimizing measurement strategies and reducing the time required for dense RIR acquisition. In addition, these predictions enable the detection of anomalies and potential measurement errors during the acquisition process.

Speaker: Nils Peters (Trinity College, The University of Dublin)

Full paper (PDF)

17:00 Welcome Reception

Wednesday 9 September

A03.02 Measurements and experimental analysis in building acoustics: S014

Conveners: Martin Schneider (Hochschule für Technik Stuttgart), Selina Vavrik-Kirchsteiger

143 Sound insulation of modular constructions

As part of a publicly funded interdisciplinary research project at the Stuttgart University of Applied Sciences, the development of an innovative timber-hybrid modular construction system for the creation of affordable housing is being supported in cooperation with the Ed. Züblin AG. The research focuses on logistics, the application of BIM for sustainable constructions, climate control systems, building physics and acoustics. The modular construction system is characterised by the separation into a primary and a secondary structure. The primary load-bearing structure consists of reinforced concrete columns, in which prefabricated modules are suspended as the secondary structure. Each module consists of a reinforced concrete ribbed floor as well as walls and ceilings manufactured using a novel timber plug-in system. The individual components of this system are cut from plywood sheets using a CNC machining centre and joined together with precision via milled tenons and mortises. As the project progresses, a demonstrator consisting of five modules across three storeys will be constructed on the university campus for use as a real-world laboratory for the various research disciplines. Preliminary extensive investigations of the airborne and impact sound insulation as well as some additional structure-borne sound measurements, e.g. vibration reduction index, were carried out on a mock-up consisting of two modules arranged both side by side and one above the other. Furthermore, the use of elastic interlayers to decouple the modules of the support brackets was investigated to the reduce structure-borne sound transmission. The article presents the current status of the building acoustic measurements.

Speaker: Andreas Ruff (HFT Stuttgart)

Full paper (PDF)

144 In-situ Measurement of the Sound Insulation of Building Elements

The verification of the sound insulation of façades and individual building elements, such as windows, is typically carried out in accordance with ÖNORM EN ISO 16283-3. However, in the presence of traffic noise or in the case of complex façade structures, such measurements are often difficult and require considerable effort.This paper therefore presents an adapted measurement method based on sound intensity, following the principles of ÖNORM EN ISO 15186-2. The two methods are compared in terms of their applicability and accuracy.Furthermore, specific challenges associated with large rooms and with building elements exhibiting high sound insulation values are discussed.

Speaker: Gustav Luckinger (Akustik Engineering Luckinger GmbH)

Full paper (PDF)

145 Improvement on Nighttime Sound Insulation of Windows by external Shutters- Part 1: Measurement Results

Window elements are often equipped with external shutters for the purpose of shading a room especially during nighttime. Although it is known that a shading device can have an effect on sound insulation of the overall window construction it is usually not taken into account for a prognosis of sound insulation performance. In terms of enhanced sound protection requirements during nighttime it is desirable to incorporate this effect in design and planning procedures.To investigate this effect a joint research project by ift Rosenheim, TH Rosenheim and HfT Stuttgart was started with support by German „Bundesministerium für Wirtschaft und Klimaschutz“ within the funding program „WIPANO – Normung und Standardisierung“.The project includes experimental, numerical and theoretical methods for assessment of the improvement on sound insulation of window elements by external shutters. Measurements of airborne sound insulation of window elements with and without external shutters were done in the acoustic test stand of ift Rosenheim. A variety of roller shutters with different mass per unit area, different cross sections and materials varying from plastic and aluminum to wood were combined with window elements featuring different glazing configurations.Within this paper first results and analysis of experimental investigations are reported. Measurements show a strong dependence of sound insulation of a combined element on the distance between roller shutter and glazing. Additionally, the influence of other constructive parameters (e.g. glazing configuration, mass per area of roller shutter and sealing of the shutter to the guide rail) on sound insulation were investigated and first results are presented within this paper.

Speaker: Joachim Hessinger (ift Rosenheim GmbH)

Full paper (PDF)

146 Improvement on Nighttime Sound Insulation of Windows by External Shutters – Part 2: Calculation of Weighted Sound Reduction Index

The current version of the German standard for sound insulation in buildings, DIN 4109, explicitly takes into account the external noise level at night. If the external noise level at night is no more than 10 dB lower than during the day, the required sound insulation is based on the relevant external noise level at night. In this case, a surcharge in sound insulation of 10 dB must be taken into account in bedrooms due to the increased disturbance at night. For such external noise situations, temporary sound insulation in form of an external enclosure can be an economical solution if the existing roller shutters provide an improvement instead of expensive soundproof windows.As part of a research project conducted jointly with Rosenheim University of Applied Sciences (HS Rosenheim) and the Rosenheim Institute for Window Technology, (ift Rosenheim) the sound reduction of various combinations of glazings and shutters was determined in the laboratory of ift Rosenheim. Several parameters e.g. the distance between the glazing and the shutter, the surface-related mass of the shutter, the type and the sound reduction of the glazing and the airtightness of the shutter were identified as most important for the airborne sound improvement. To implement the experimental findings in the German standard DIN 4109 a simple calculation procedure with single number quantities has to be developed.

Speaker: Martin Schneider (Hochschule für Technik Stuttgart)

Full paper (PDF)

147 Experimental Investigation of Spatial Sampling Requirements for Vibration-Based Sound Power Estimation of Lightweight Façade Systems

Accurate estimation of radiated sound power from lightweight facade systems is challenging due to their complex vibro-acoustic behavior, particularly at higher frequencies. Vibration-based approaches, such as those based on the Acoustic Radiation Matrix (ARM), offer an efficient framework for predicting sound power from measured surface velocities, but discrepancies are often observed when compared with standardized measurements.This study investigates the influence of spatial discretization on vibration-based sound power estimation for an ETFE cushion system using scanning laser Doppler vibrometry. A grid refinement analysis is performed to assess the effect of measurement resolution on the predicted sound power. The results show that insufficient spatial sampling leads to systematic underestimation at higher frequencies due to unresolved spatial variations in the vibration field. Based on these observations, a spatial sampling criterion is proposed to ensure reliable sound power estimation.

Speaker: Abhishek Singh (Czech Technical University in Prague)

A06.01 Microphone and MEMS transducers: S037

Conveners: Dominik Mayrhofer (Graz University of Technoloy), Petr Honzík (Czech Technical University in Prague)

148 Diaphragm Sealing Strategies for MEMS Loudspeakers

This paper presents a comparative analysis of diaphragm sealing strategies for piezoelectric MEMS loudspeakers, focusing on their impact on sound pressure level (SPL) across the audible frequency range. Effects on actuator design, manufacturing complexity, and reliability are also evaluated. An analytical model is introduced to estimate acoustic leakage, along with a new criterion for defining leakage.

Speaker: Nils Wittek (University of Stuttgart, Institute for Micro Integration (IFM))

Full paper (PDF)

149 Time-Domain Modeling of Distortion Reduction in In-Ear MEMS Loudspeakers via Feedback Linearization

MEMS loudspeakers are promising for next generation in ear headphones and hearing aids due to their compact size, low power consumption, and CMOS compatibility. However, their miniaturized structure requires large membrane displacements to achieve sufficient sound pressure levels. In piezoelectric actuators, these displacements often require higher input voltages, and both factors increase nonlinear distortion. While conventional loudspeakers mitigate distortion through structural design, active control offers a more adaptable solution by using a sensor to monitor membrane motion and adjusting the drive signal in a closed loop. In this work, we present a simulation of an active control scheme that integrates a collocated sensor and actuator within a MEMS loudspeaker.The loudspeaker considered is a clamped guided cantilever based structure, divided around the inflection point. One side is covered by a piezoelectric actuator and the other side is used as a collocated sensor. The voltage generated by the sensor provides a measure of the displacement, is amplified by a feedback gain Kp, and subtracted from the input voltage. The input signal is amplified with a feedforward gain Kff. The loudspeaker operates in an ear occluded coupler represented by an equivalent electrical circuit. The governing time domain equations are written in matrix form and discretized using a bilinear scheme, including linear parameters and nonlinear effects such as geometric nonlinearities, ferroelectric transduction, and varying capacitance.Simulations show that combining feedforward and feedback reduces total harmonic distortion by about 50 percent in the quasi static range, while preserving the frequency response.

Speaker: Romain Liechti (CEA-Leti)

Full paper (PDF)

150 Low-Power Push–Pull Actuated Piezoelectric MEMS Loudspeakers

Piezoelectric MEMS speakers are emerging as promising microscale loudspeaker solutions, driven by the increasing demand for smaller, lighter, and more power-efficient audio devices. However, their integration into mass-market products remains challenging, as it requires simultaneously achieving compact size, high acoustic performance, and compatibility with standard fabrication processes.The push–pull actuation scheme has recently been proposed as an effective strategy to enhance performance compared to single-patch actuation. It employs two independent piezoelectric patches on a silicon diaphragm, driven in anti-phase to increase diaphragm displacement and acoustic output.In this work, a numerical investigation of four design iterations with progressively improved acoustic output is presented. Each design explores the interplay between structural compliance, tuned through the number and pattern of 5 μm slits, and the area and placement of the piezoelectric patches. The design variations include outer actuator reshaping, the introduction of curved beam elements, and the progressive extension of the central diaphragm hub.The best-performing architecture achieves a sound pressure level (SPL) exceeding 113 dB SPL in an IEC 60318-4 coupler from 100 Hz onward, with a piezoelectric capacitance of 15 nF. Overall, this study contributes to the development of design guidelines for low-power, high-performance piezoelectric MEMS loudspeakers compatible with standard industrial processes.

Speaker: Filippo Pietro Perli (Politecnico di Milano)

151 Modulated ultrasound MEMS loudspeakers – a paradigm shift

In recent years, loudspeakers that generate audible sound by demodulating an ultrasonic carrier have gained increasing attention. For MEMS applications in particular, this approach offers the advantage of achieving higher sound pressure levels at low frequencies, a regime in which conventional MEMS loudspeakers typically struggle to meet industry requirements. This advantage stems from a fundamental change in the loudspeaker’s operating principle: Classical loudspeakers act as a volume source, whereas modulated ultrasound loudspeakers act as a volume flow source, enabling a +20 dB per decade increase in sound pressure for decreasing frequencies. This fundamentally different operation is achieved by introducing a shutter that demodulates the ultrasound carrier signal in order to create an audible signal. In this contribution, we review the current status of modulated ultrasound principles and outline their challenges in both technological and medical contexts. Besides fabrication, the topics of ultrasound tolerance in humans and animals, simulation methods, design optimization, and sound quality are tackled. Additionally, we compare modulated ultrasound principles with current MEMS loudspeaker designs and identify opportunities and application examples enabled by this new MEMS loudspeaker technology.

Speaker: Dominik Mayrhofer (Graz University of Technoloy)

Full paper (PDF)

A10.07 Micro-perforated and Helmholtz-type resonant absorbers: modelling and characterisation: S066

Conveners: Teresa Bravo (Spanish National Research Council), Cedric Maury (Laboratory of Mechanics and Acoustics)

152 Slow-sound optimization of rainbow-trapping silencers

Designing new kind of absorbers to avoid the use of porous or fibrous materials due to the presence of mean flow in a duct is a challenging topic in the field of noise control. Lately, the development of metamaterials has provided new noise control perspectives for this issue. Traditional Acoustic Black Holes (ABHs) have been derived with closed-ended configurations to achieve broadband absorption. These solutions are appropriate as anechoic termination but they cannot be used as silencers as they require a mean flow going through the axis of the acoustic treatment. The use of open ABH, also known as Rainbow-Trapping Silencers (RTS) has been proposed to tackle this issue. RTS are fully-opened in-duct metamaterials, of rectangular or cylindrical cross-sections, traversed by a low-speed flow and whose walls are lined by graded cavity depths. In order to reduce reflection and transmission and to obtain full dissipation of the acoustic disturbance over a broad bandwidth, a strategy is to enhance its slow sound properties. They are obtained when the axial phase speed gets closer to the bulk flow velocity. Parametric and optimization studies from theoretical (transfer matrix method) and numerical (finite element model) approaches show how a suitable choice of converging RTS parameters maximize slow sound effects. Such compact lightweight devices could be used to reduce broadband sound emissions in transport systems.

Speaker: Teresa Bravo (Spanish National Research Council)

Full paper (PDF)

153 Robustness of Rainbow Trapping Silencers to Low-Speed Flow Effects

This study investigates the aero-acoustic robustness of rainbow trapping structures (RTSs) with straight and coiled cavities in fully opened and flow-compliant convergent configurations under a low-speed grazing flow. A transfer matrix model is developed to predict dissipation, reflection, and transmission loss (TL), and validated against finite element simulations of the Linearized Navier–Stokes equations in the no-flow case. In the absence of flow, increasing the acoustic path length, particularly in convergent designs, broadens and downshifts the high-dissipation plateau through merging of quarter-wave resonances and improved impedance matching. The convergent straight (CS) and coiled (CC) RTSs exhibit the widest efficiency bandwidths, while the CS configuration provides the highest TL peak. Under grazing flow, all configurations remain robust, with reflection coefficients below 1.5% above 400 Hz. Upstream flow conditions (UPC) enhance attenuation relative to downstream flow conditions (DPC), yielding broader dissipation bandwidths and higher TLs, whereas DPC induce over-damping at the cavity mouths and reduced performance. In convergent designs, geometric acceleration of the flow increases the flow-induced modification of the cavity impedances. It triggers a low-frequency axial half-wavelength resonance that scales on the reduced sound speed. Overall, RTS performance is maximized in suction mode, confirming their robustness for flow-integrated noise control.

Speaker: Cedric Maury (Laboratory of Mechanics and Acoustics)

Full paper (PDF)

154 Broadband Noise Attenuation in Ventilated Ducts using an Acoustic Black Hole-Micro-perforated Panel Structure

Broadband noise attenuation in ventilated duct systems remains challenging in the low-to-mid frequency range. This study proposes a ventilated acoustic black hole-micro-perforated panel (ABH-MPP) structure that combines axial geometric variation with sidewall micro-perforations to enhance thermoviscous dissipation. A one-dimensional transfer matrix model is developed and validated against finite element simulations and available data from the literature. Results show broadband sound power dissipation and transmission loss without relying on conventional resonant mechanisms. Introducing internal partitions further enhances dissipation by extending thermoviscous losses from the sidewalls to the partition surfaces, leading to stronger distributed attenuation along the propagation path. Time-domain analysis indicates that the MPP does not weaken the intrinsic slow-wave effect of the ABH; instead, it increases the residence time of acoustic energy and promotes sustained dissipation. Parametric studies identify the influence of key geometric parameters and support a mechanism-based design approach. The ventilation capability of the proposed structure is also evaluated under a typical flow condition, demonstrating that a continuous airflow passage can be maintained.

Speaker: Yuchen Zhao (School of Architecture, Southeast Univ)

Full paper (PDF)

155 Strong fluid-structure interaction effects in subwavelength Helmholtz resonators for underwater acoustics: a comparative study for vibroacoustic design applications

Low-frequency sound absorption in underwater environments remains challenging due to the large acoustic wavelengths and strong fluid–structure interaction effects. While Helmholtz-type resonators generally offer a promising route toward compact subwavelength solutions, achieving efficient dissipation mechanisms in water is non-trivial because of the suppressed role of visco-thermal losses relative to air. This often implies moving beyond ideal rigid-wall assumptions toward more realistic fluid-structure interactions.In this work, a comparative study of 16 water-filled resonator configurations is presented, examining the influence of the wall material across analytical, rigid, aluminum, and compliant-polymer models. Results show that analytical and rigid-wall models remain closely aligned (frequency deviations <1%) and achieve near-perfect absorption. Aluminum implementations preserve high absorption while introducing moderate and predictable resonance downshifts (≈5–13%). In contrast, compliant polymeric walls lead to strong detuning and reduced peak absorption, highlighting that structural compliance must be treated as an integral part of the resonant mechanism rather than as a perturbation.These findings provide practical guidelines for the modeling and design of compact underwater acoustic absorbers and metamaterials, and clarify the conditions under which rigid-wall approximations remain valid and when they must be replaced by compliance-aware approaches.

Speaker: Dario Magliacano (Politecnico di Torino)

Full paper (PDF)

A11.03 Physical modelling and simulations of musical instruments: S076

Conveners: Péter Rucz (Budapest University of Technology), Juliette Chabassier

156 Physics-based Modelling of Mechanical Impact Noise in Piano Acoustics: Hammer-string Acceleration Noise

This study examines a rarely considered acoustic consequence of the piano hammer–string collision: the direct sound produced by the hammer’s rapid acceleration at impact. Using analytical methods and 2D finite element modelling, the resulting acoustic pressure is simulated and combined with the synthesised responses of a stiff string and soundboard radiation. The composite signal is analysed to assess the relative relevance of the impact-generated sound across the piano range. Results show that its contribution is generally minor, particularly for lower notes, and while typically masked under normal playing conditions, it may become noticeable in recordings with microphones placed near the action.

Speaker: Pablo Miranda Valiente (University of Southampton)

Full paper (PDF)

157 Equivalent Soundboard for Piano Model

This paper investigates the substitution of traditional ribbed piano soundboards with an equivalent single-ply structure while preserving key acoustical and mechanical characteristics. Piano soundboards play a crucial role in sound radiation, inter-string coupling, and overall tonal quality, all of which are strongly influenced by geometry and material properties. With the decreasing availability of high-quality tonewood and increasing interest in sustainable and cost-effective alternatives, engineered materials and optimized design methods are gaining importance.Building on prior findings that sound quality can be described through quantitative indicators linked to soundboard admittance and modal behavior, this work focuses on replicating these properties in a simplified structure. A two-step estimation approach is applied: first, ribs are replaced by an equivalent transversely isotropic layer, and second, the effective properties of the resulting laminate are determined.The proposed method systematically derives a one-ply equivalent model and evaluates its performance through modal analysis and admittance comparison. Although the resulting parameters are hypothetical and not yet validated for physical realization, the study yields to offer a conceptual framework for future development of alternative piano soundboard designs.

Speaker: Dóra Jenei-Kulcsár (Budapest University of Technology)

Full paper (PDF)

158 Low-frequency dissipation of a musical string in a viscous fluid: numerical study

The distinct sound of a string instrument depends on harmonic attenuation. While past studies have mostly focused on high-frequency thermoelastic damping, we investigate here low-frequency dissipation in a real (viscous) fluid. Todo so, we solve the incompressible Navier-Stokes equations using Basilisk, a finite volume solver capable of handling adaptive octree meshes around moving bodies with immersed boundaries. Assuming a long coherence length, we model the string as a 2D oscillating cylinder. Accordingly, flow regimes are characterized via the Reynolds (Re) and Keulegan-Carpenter (KC) numbers. We first obtain reference time-resolved data via prescribed cylinder motion. The Valette & Cuesta model, partially based on the potential flow theory, reveals limited validity within the (Re, KC) space. We propose Morison’s equation as an improvement: a semi-empirical, physically consistent model that extends the predictions validity by extracting parameters from the simulation data. Finally, a two-way coupled simulation dynamically injects the computed drag forces at each time step, demonstrating Basilisk’s ability to capture the coupling between variable oscillation amplitude and viscosity. This last method, although computationally heavier, provides a versatile method valid for any Re, KC, paving the way for comparison with experiments described in a companion paper.

Speaker: Antoine Hajczak (Sorbonne Université, CNRS, Institut Jean Le Rond d'Alembert)

159 Explicit and Stable Pseudospectral Time-Domain Method for the Föppl–von Kármán Equations

Modal synthesis is a widely used technique for simulation of musical instruments. In the linear case, a modal decomposition leads to an uncoupled system of damped and forced harmonic oscillators which can be efficiently solved by regular time-stepping methods. However, extensions to nonlinear problems are challenging due to the presence of products of modal expansions in the governing equations. In the case of the Föppl–von Kármán plate, the nonlinear coupling between the modes is described by a fourth-order tensor and is computationally expensive to evaluate in the modal domain. In this work, we propose a pseudospectral method in which the products are evaluated on a grid in the spatial domain while spatial derivatives are computed exactly in the modal domain. Discrete sine and cosine transforms between the modal and spatial domains are used to impose simply supported boundary conditions for the plate. Finally, we prove non-negativity of the nonlinear potential energy of the system and employ a scalar auxiliary variable technique for explicit and stable time integration in the modal domain. As a result, we reduce the computational cost of modal synthesis while preserving its advantages like a precise control over the simulated frequency range. Sound examples are presented.

Speaker: Victor Zheleznov (University of Edinburgh)

Full paper (PDF)

A14.04 Unraveling the auditory periphery with physiological measures: S101

Conveners: Gerard Encina-Llamas (University of Vic - Central University of Catalonia), Bastian Epp

160 Active hearing in different human populations

Most studies of hearing are based on Europeans or Asians or their descendants, in, e.g., N. America. There is little data on other human populations. The San people of southern Africa represent one of the oldest distinct human populations, being almost completely separate from all other groups for > 100k years. We report a study in young adults belonging to the Ju/’huansi of the Nyae Nyae San community in Namibia compared to West European students in Groningen. The studies were approved by relevant review boards, the local San community and the Namibian government. A closed insert microphone system (Etymotic ER10B) was used for measuring audiograms, spontaneous otoacoustic emissions (SOAE), suppression tuning of SOAE, and stimulus frequency emissions (SFOAE). Audiograms were indistinguishable in the Namibian and European groups. SOAE peak values were higher in Namibian subjects, both in the highest peak frequencies (13.7 kHz in in Namibians vs. 9.6 kHz in Europeans) and highest levels (19.0 vs. 16.4 dB SPL, respectively). Suppression tuning curves had a V- shape with additional side-lobes on the high-frequency slope. While the main V-shape was similar in both groups, sidelobes were more numerous and deeper in Namibians. SFOAE amplitudes and phase delays were similar between the groups. However, SFOAE amplitudes were usually lower in the Europeans at higher frequencies.In the absence of differences in the audiogram, the reduced number SOAE and lower amplitudes of SFOAE at higher frequencies, suggest a previously unknown form of hidden hearing loss in the Europeans. The pronounced sidelobes in the suppression tuning curves of Namibians suggest enhanced coupling along the cochlear partition.

Speaker: Pim Van Dijk (University Medical Center Groningen)

161 Simulation-based validation of the distortion invariance hypothesis for distortion-product otoacoustic emissions

Distortion-product otoacoustic emission (DPOAE) level maps measured across audiological frequencies differ. Keefe JASA 111:249–260 (2002) suggested that if the distortion characteristics of the basilar membrane (BM) are uniform along a large part of its length, the differences in DPOAE level maps across frequencies from 0.5 kHz to 8 kHz should reflect the forward and reverse transmission of the middle ear. We use a cochlear model that accounts for longitudinal hydrodynamic coupling via cochlear fluids between individual BM segments. At each segment, a memory less spatially uniform nonlinearity is applied to the feedback force to generate DPOAEs. The model was designed to produce roughly uniform amplification of approximately 50 dB of transversal BM motion at characteristic frequencies from 1 to 6 kHz. DPOAE level maps in this frequency range derived from this model were aligned by applying level shifts comparable to shifts between BM input/output functions. Therefore, the model supports the distortion invariance hypothesis. We took average DPOAE growth functions from Abdala et al. Ear Hear. 42:832–845 (2021) and fitted them into the DPOAE level map simulated at 2 kHz. The smallest fitting error was obtained for DPOAE growth functions measured for young, normally hearing subjects. The error increased for middle-aged and older subject groups in which the cochlear nonlinearity is altered in comparison to the young group. The fitting procedure can be used to estimate the onset of compression in the BM input/output functions across frequencies.

Speaker: Vaclav Vencovsky (Czech Technical University in Prague)

162 Posture-Driven Intracranial Pressure Changes Revealed by Otoacoustic Emissions: Toward Noninvasive Monitoring in Microgravity

It is well established that middle ear transmission is influenced by posture. This phenomenon is associated with variations in intracranial pressure (ICP) when a subject transitions from an upright to a supine position. Typically, this transition leads to an increase in ICP. The resulting rise in endocranial fluid pressure alters middle ear (ME) impedance, generally producing an increase in stiffness, which is linked to the imaginary component of the impedance. These impedance variations also affect distortion product otoacoustic emission (DPOAE) phase. This study is aimed at optimizing joint OAE- an ME-based techniques to assess ICP increases under microgravity conditions. DPOAEs were recorded in both upright and supine positions. Additionally, ME energy reflectance, along with the real and imaginary components of impedance, was measured. Typically, the results align with theoretical expectations, but several counterexamples are observed, which need explanation to enhance the diagnostic power of the proposed technique.Overall, joint OAE-based and ME-based ICP estimation demonstrate potential as a benchmark technique for monitoring intracranial pressure changes in space environments, where such non-invasive and reliable physiological monitoring is essential during prolonged missions in microgravity conditions. Further investigations are needed to refine methodological accuracy and broaden clinical applicability across diverse populations and conditions.

Speaker: Arturo Moleti (University of Rome Tor Vergata, Dept of Physics)

Full paper (PDF)

163 Assessment of auditory efferent strength using click-evoked otoacoustic emissions in listeners with and without hearing loss

The ability to understand speech in adverse listening conditions, such as in the presence of masking noise, is severely degraded in people with hearing loss. The mechanisms that underly these speech deficits are still not fully understood. A healthy auditory efferent system has been argued to facilitate understanding speech in noise. Evidence from computational auditory models that include the efferent system suggests that efferent suppression might play a significant role in the representation of speech sounds in the auditory pathway. However, the relationship between efferent strength, hearing loss, and reduced speech understanding is still debated, partly because measuring efferent suppression is not straightforward in participants with hearing loss. In this contribution we apply a new method for measuring efferent strength using click-evoked otoacoustic emissions (CEOAEs). Our approach is to compare the magnitude of long-latency CEOAEs for pairs of clicks, before and after an efferent-eliciting tone. In a cohort of participants with a diverse range of hearing status, we studied the influence of click level and elicitor level on the measured efferent strength. Different descriptors of efferent strength in terms of their on-frequency magnitude as well as the frequency range of the suppressing effect were considered. Future studies will extend this approach to study the relationship between hearing loss and efferent strength and its potential relationship with speech intelligibility.

Speaker: Helia Relaño-Iborra (Eriksholm Research Centre)

A14.06/A15.07 Music perception and consequences of hearing impairment: S103

Conveners: Alinka Greasley, Kai Siedenburg (CvO University of Oldenburg)

164 A non-randomized controlled study for older adults with hearing loss: investigating cumulative effects of choir and sign language on music- and speech perception

Older adults are at risk for both mild cognitive decline and feeling socially isolated, and hearing loss increases these risks. While previous studies report older adults to experience improved aspects of cognition and social wellbeing after both language learning and musical activities, the role of hearing loss in the success of these outcomes remains unclear. Moreover, despite the clear advantages of randomized controlled trials in diminishing potential bias effects, we previously experienced participant attrition when participants were not allocated to their preferred intervention. Therefore, this project aims to address these gaps by focusing on older adults with hearing loss while using non-randomized, self-selected training and control groups. Over one year, participants will successively engage in a 3-month choir and 3-month sign language lessons. Before-, after, and follow-up testing for each activity will be compared with a do-nothing control group. Primary outcomes assess cognitive function and social wellbeing, and secondary outcomes assess speech- and music-perception metrics. Baseline metrics include the Cognitive Reserve Index Questionnaires (CRIq – a proxy for resilience against cognitive decline) and the online digits-in-noise test (DIN). Preliminary analysis of baseline metrics in cohort 1 (n= 14, preliminary best ear PTA4 range: 20 – 55 dB) show a moderate association (R = -.646, p = .013) between CRIq and DIN scores, already suggesting a relationship between auditory processing and cognitive resilience. Ultimately, we hope to provide valuable insight into how language learning and musical training might prospectively contribute to healthy aging, accounting for the complex relationships among hearing loss, cognition and wellbeing, and speech- and music perception.

Speaker: Eleanor Harding (University of Groningen)

165 Beyond ageing: Effects of hearing loss and musical training on music perception

Individual differences in music perception are shaped by peripheral sensory, cognitive, and experiential factors. The roles of hearing loss (HL) and ageing are particularly difficult to estimate because both are typically confounded in adult samples, leaving it unclear whether reduced musical perceptual performance reflects elevated hearing thresholds, age-related changes, or both. To disentangle HL from ageing, this ongoing study includes four listener groups: older normal-hearing (oNH; n = 31, M = 62.2 years), older hearing-impaired (oHI; n = 34, M = 67.8 years), younger normal-hearing (yNH; n = 26, M = 25.3 years), and younger hearing-impaired listeners (yHI; n = 14, M = 36.0 years). Participants completed a battery assessing Musical Scene Analysis, Melody Discrimination, Beat Alignment, Mistuning Perception, and Timbre Perception, alongside pure-tone audiometry and the Goldsmiths Musical Sophistication Index (GMS). Confirmatory factor analysis yielded a unidimensional General Musical Perception Score (GMPS), with the latent construct accounting for 40% of task variance. Ridge regression predicting GMPS from hearing thresholds, age, and GMS revealed that pure-tone average robustly predicted lower GMPS (β ≈ −.30, p < .001), while age contributed negligibly once hearing was controlled (β ≈ −.05, p = .583). GMS showed the largest effect (β ≈ .40, p < .001). Together, predictors explained ~45% of individual variance. These findings suggest that reduced musical perception in older adults is better explained by elevated hearing thresholds than by ageing per se. The strong contribution of musical training further highlights its potential role in supporting perceptual performance and compensating for HL-related constraints.

Speaker: Kai Siedenburg (CvO University of Oldenburg)

166 Effect of Hearing Aid Output Limiter on Sound Quality Prediction for Loud Music

Hearing aid requirements extend beyond restoring audibility to ensure user comfort, a critical determinant of overall satisfaction. Comfort is largely managed by controlling maximum power output, typically via fast-acting compression in output limiters. However, a significant trade-off exists between comfort and sound quality, particularly for signals like music; literature consistently reports negative user experiences during loud music exposure due to these processing constraints. Objective sound quality models offer a promising way to reduce listener exposure to loud situations and expand testing conditions during development.This presentation reviews existing metrics and evaluates their performance in a specific use case: predicting the impact of output limiters on the sound quality of loud music. The relationship between predicted sound quality and perceptual ratings was examined across different limiter implementations. Group analysis indicates that concordance between perceptual and predictive methods is not systematic. While agreement is high when perceptual differences are large, objective metrics can yield qualitatively different conclusions when perceptual differences are negligible. Furthermore, we discuss how reference condition definition, rater selection, and sample size influence the analysis. These results highlight the need for caution when interpreting sound quality predictions outside of their original application domains.

Speaker: Christophe Lesimple (Sonova AG)

167 Effects of HRTF Augmentation on Predicted Spatial Release from Masking in Music

Separating individual musical instruments within a complex mixture of sounds poses a persistent challenge for listeners with hearing loss. Although spatial separation of sources improves speech recognition in this population, the potential benefits of spatial cue enhancement for music perception remain largely unexplored. This paper introduces a method to increase spatial cue salience through the augmentation of individual head-related transfer functions (HRTFs). Auditory model analyses indicate that augmented HRTFs may enhance the separability of musical instruments relative to individual HRTFs. Predicted benefits persist when moderate sensorineural hearing loss is modelled, though they are substantially reduced. Simulated hearing aid processing does not restore these benefits to normal-hearing levels.

Speaker: Jack Webb (Imperial College London)

Full paper (PDF)

A15.01 Spatial Hearing: Modeling and Applications: S107

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Ville Pulkki (Aalto University)

168 Building Internal Models of Auditory Space in Elevation

Spatial hearing depends on internal representations of acoustic cues that vary across space. Elevation perception relies on learned spectral cues from pinna filtering, making it uniquely dependent on prior knowledge rather than instantaneous acoustic information alone. It remains unclear whether listeners track statistical properties of elevation consistent with Bayesian inference and whether this differs across regions of auditory space.To test this, listeners monitored sequences of noise bursts with elevations sampled from a distribution defined by a mean and variance. Each sequence either remained constant or changed in mean and/or variance (mean shifts: 30° or 60°; standard deviations: 15° or 25°). Participants reported when they detected a change.Preliminary results (N = 6) show reliable detection of changes in elevation distributions, with sensitivity increasing for larger mean shifts (d' = 1.4 vs 2.2 for 30° vs 60°, p < .001). Detection was more accurate for sequences with initial distributions in the front than for those starting above (p = .025) or behind (p = .034). This spatial asymmetry may reflect stronger or more reliable prior representations for frontal elevations. The influence of variance is still under investigation, with data collection in progress. Ongoing work will fit behaviour with Bayesian change-detection models to test consistency with Bayesian inference and estimate how latent variables, such as temporal integration, vary across space.

Speaker: Katarina C. Poole (Imperial College London)

169 A Probabilistic Mixture Model for Evaluating Sound Localisation Performance in the Median Plane

Accurately evaluating sound localisation performance is essential for developing and validating methods to personalise head-related transfer functions (HRTFs). Localisation in the median plane is typically summarised with complementary metrics: polar error for local precision, polar gain for spectral decoding ability, and quadrant error rate for front-back confusions. These metrics rely on hard thresholds applied to subsets of trials, which limits statistical robustness and compresses individual differences (for instance, polar error excludes responses beyond 90° of the target, capping polar error at 52° for listeners responding at chance). Here, we propose a probabilistic model of the complete polar response distribution, based on a four-component von Mises mixture, that requires no data partitioning and negligible computation. The model free parameters are jointly identifiable, as confirmed by a synthetic recovery study, and can be estimated from as few as 45 trials per listener. Fitted to 33 listeners, each parameter correlates with a classical metric: confusion rate onto quadrant error rate, concentration onto polar error, and a spatial prior parameter onto polar gain. Model comparison further supports including the spatial prior for the large majority of listeners. The framework lends itself naturally to Bayesian extensions, offering a principled basis for more data-efficient individualised HRTF evaluation.

Speaker: Roberto Barumerli (Imperial College London)

Full paper (PDF)

170 Deep neural networks as integrative models of binaural hearing: linking neural tuning and spatial perception

Human spatial hearing relies on the integration of multiple binaural cues, including interaural time differences (ITDs), interaural level differences (ILDs), and spectral information. While psychophysical experiments have characterized how these cues are combined across a wide range of conditions, linking perceptual behavior to underlying neural mechanisms remains a central challenge. Deep neural networks (DNNs) trained on sound localization tasks provide a promising framework for bridging this gap.Here, we evaluate a DNN model of spatial hearing with a focus on its ability to reproduce human psychophysical data across diverse and often highly artificial stimulus conditions. The model captures key aspects of human perception, including envelope ITD–based lateralization, interactions between ITD and ILD, robustness to strong reflections, and stimulus-dependent spectral cue weighting. While these properties were not directly imposed by design, input characteristics shaped by cochlear filtering and hair-cell nonlinearities constrain the functional space of both artificial and biological binaural processing.At the same time, systematic deviations from human performance are observed. The model fails in conditions that rely on weak or ambiguous cues, such as broadband noise with supranatural ITDs, and does not reproduce known perceptual biases in pure-tone localization. These discrepancies point to differences in cue integration strategies and suggest missing biological constraints, such as frequency-dependent priors or limitations in combining binaural information.By directly comparing model behavior with established psychophysical paradigms, this work demonstrates how DNNs can be used not only to reproduce but also to interrogate human spatial hearing, providing a framework for generating targeted experimental predictions and refining theories of binaural perception.

Speaker: Mathias Dietz (Universität Oldenburg)

171 Do Spectral Localization Cues Mitigate Azimuthal Localization Bias by a Precursor?

The perceived location of a sound is influenced by preceding (precursor) sounds. In the horizontal plane, lateral precursors shift the perceived location of a subsequent probe sound toward the midline, whereas central precursors push perceived locations further toward the sides. These effects are typically stronger for headphone presentation than for loudspeaker playback in anechoic environments. We hypothesized that accurate, listener-specific spectral cues provided by loudspeaker presentation enhance spatial precision and thus reduce precursor effects. To that end, we manipulated spectral cues by progressively smearing head-related transfer functions (HRTFs) using cepstral smoothing, thereby reducing spectral detail while preserving binaural localization cues. Participants performed a headphone-based localization task in a virtual audiovisual environment using a VR headset in an anechoic room. Visual markers at 0°, 45°, and 90° allowed participants to realign their position after each trial. Probe sounds were presented from −90° to 90° in 10° increments. Precursors were delivered at −70°, 0°, or 70°, or omitted in a control condition without precursors. Probe and precursor sounds were pink noises with durations of 600 and 300 ms, respectively, separated by 10 ms of silence. Listeners were instructed to focus on the probe. Preliminary results show small but reliable precursor effects within listeners, with notable variability across individuals. Crucially, reducing spectral cues did not significantly change the magnitude of the precursor effect. Consistent with this finding, spectral cues also did not affect localization accuracy in the absence of precursors. Overall, the results do not support the hypothesis that spectral cues mitigate spatial precision and precursor-induced spatial biases in the frontal horizontal plane.

Speaker: Bernhard Laback (Austrian Academy of Sciences)

A16.04/A24.06 Acoustics for learning environments across physical and virtual realities: S116

Conveners: Arianna Astolfi (Politecnico di Torino), Janina Fels (IHTA, RWTH Aachen University)

172 Long-Term Attentional Processing Under Reverberant Conditions: A Room Impulse Response–Based Controlled Listening Study

This study investigates how reverberation time influences late-stage attentional control during extended acoustic exposure. Room impulse responses were generated from systematically modified architectural-acoustic models and convolved with intelligible speech recordings to produce controlled auralized background stimuli. In a laboratory-based listening experiment employing a between-group design, participants were exposed to a single reverberation condition while completing three sequential attention tasks—Sustained Attention-to-Cue Task (SACT), Stroop Adaptive Response Deadline, and Selective Visual Arrays. The sequential administration of multiple tasks enabled examination of attentional performance across successive phases of sustained engagement, allowing assessment of potential cumulative and time-dependent effects under reverberant conditions. By focusing on late-stage attention rather than isolated short-term performance, the study aims to contribute to a deeper understanding of how room acoustic parameters may interact with prolonged cognitive demands. The findings are expected to inform reverberation-sensitive design considerations in acoustically complex learning environments

Speaker: Zekiye Sahin Ozturk (Bilkent University)

173 Noise Perception and Cognitive Load in a Simulated Reverberant Library under Full Occupancy

In recent years, libraries have evolved from silent study spaces into multifunctional environments hosting diverse activities. This shift has resulted in increasingly complex acoustic conditions. This study investigates noise perception and cognitive performance in the simulated environment of the New Central Civic Library of Torino during full occupancy. The library consists of a large space with a volume of approximately 160,000 m³ and a mid-frequency reverberation time of around 5 s. A three-dimensional geometrical acoustics model was developed to reproduce the sound environment, including typical library sources such as traffic, HVAC systems, speech and user-generated noise. Simulated sound pressure level was in the range 50.2 – 52.0 dB(A) across two receivers, consistent with values reported for occupied libraries. An experimental session involving 20 participants was conducted in an ambisonics laboratory to assess noise sensitivity and annoyance, together with a reading-aloud cognitive task, conducted both in noise and in quiet. Results showed no statistically significant differences across receivers for subjective evaluations overall. A comparison with a previous study conducted with headphones under low occupancy, within a sound pressure level range of 45 - 47 dB(A), showed different outcomes, with a significant increase in reading errors under noise conditions compared to quiet. The different submission methods and the type of noise can explain the results of the comparison.

Speaker: Ioana Tsankova Grozeva (Politecnico di Torino)

Full paper (PDF)

174 Cognitive, Auditory, and Room Acoustic Effects on the Irrelevant Speech Effect in Primary School Children: A Structural Equation Modelling Approach

Children learn in complex environments, often marked by high levels of background noise that is largely generated by students themselves and inadequate acoustic conditions. Young primary school children are particularly vulnerable to these conditions due to immature selective attention abilities, with even greater risk for those with additional learning needs or hearing difficulties. Over the past decades, models have been proposed to explain how auditory, cognitive, and acoustical factors interact in speech perception. However, many classroom activities affected by noise do not rely on listening. To address this issue, a structural equation model was developed to simultaneously examine how individual (audiological and cognitive) and environmental factors (noise exposure and room acoustics) jointly contribute to children’s susceptibility to noise interference, as reflected in the irrelevant speech effect (ISE). Data were collected from a stratified sample of 465 children aged 7–10 years. Preliminary findings indicate that ISE is primarily driven by acoustic conditions, particularly speech clarity (C50). Chronic noise exposure also contributes, whereas audiological factors show minimal influence. The pattern of relationships varies with students’ age. The findings underscore the need to consider individual and environmental factors jointly, adopting an integrated approach to better understand and mitigate the impact of noise on children’s learning.

Speaker: Chiara Visentin (University of Ferrara)

175 Evaluating Speech Intelligibility across Seating Zones in a Naturally Ventilated Classroom

Acoustic conditions in naturally ventilated classrooms are influenced by open windows, ceiling-fan operation, and external noise sources. The resulting spatial heterogeneity cannot be adequately captured by conventional single-point measurements or energy-averaged acoustic metrics. This limitation constrains the ability to assess and design for speech intelligibility in real-world educational environments. Our study examines speech intelligibility across classroom spaces under typical operating conditions during summer.The study was conducted in a secondary school classroom in Roorkee, India. Word recognition test was administered to students seated across multiple spatial zones under baseline conditions (fan on, windows and doors open). Concurrent acoustic measurements (LAeq, L90) were collated and RT60; SNR and STIPA computed across the classroom. The results indicate marked seat-wise variation in perceived speech intelligibility. The observed patterns suggest that proximity to openings and external noise sources may influence word-recognition performance. Temporal variation between morning and afternoon sessions and linguistic variation between Hindi and English word tests were also examined. Overall, the acoustic environment allowed speech to remain generally audible, but word-level intelligibility varied across classroom locations.The study foregrounds spatial variability as a critical dimension of classroom acoustics. It proposes a field-based protocol for evaluating speech intelligibility in naturally ventilated classrooms.

Speaker: Syed Azhar Akhtar (Indian Institute of Technology Roorkee)

176 Psychological Outcomes Among Teachers in School Acoustic Contexts: A Preliminary Descriptive Study Within the INAIL BRiC Project

Background: Teachers work in environments characterized by multiple occupational demands, while anxiety represents an important psychological outcome. However, evidence linking classroom acoustics to psychological outcomes remains limited. This preliminary study provides a descriptive overview of teachers' anxiety and the acoustic characteristics of the school environments in which they work.Methods: Twenty-six teachers from three public schools (preschool, primary, and lower secondary) participated. State and trait anxiety were assessed using the State–Trait Anxiety Inventory Form Y (STAI-Y). Standardized acoustic measurements included Background Noise (NC), Reverberation Time (RT) and Speech Intelligibility (STI). Descriptive statistics and internal consistency analyses were performed.Results: Complete anxiety data were available for 23 teachers. Mean STAI-State and STAI-Trait scores were 42.57 (SD = 9.20) and 42.04 (SD = 8.96), respectively. Descriptive statistics suggested some variability in anxiety levels across schools, while acoustic measurements reflected heterogeneous environmental characteristics. No inferential analyses were performed to examine associations between anxiety and acoustic variables.Conclusions: These preliminary findings show that standardized psychological assessments and objective acoustic evaluations can be conducted in educational settings and provide a basis for larger multidisciplinary studies.

Speaker: Veronica Raspa (Department of Medicine and Surgery, University of Perugia)

Full paper (PDF)

A16.09 Acoustic Comfort in Hospitals: S120

Conveners: Simone Secchi (Department of Architecture, University of Florence), Andrea Giglio (European Acoustics Association Young Acousticians Network (YAN)), Juan Negreira, Maria Quinn

177 Acoustic Characterization and Noise Assessment in Italian Hospitals: a Preliminary Study

Acoustic comfort in hospitals is vital for staff performance and patient safety, yet it remains a critical challenge in healthcare design. This paper presents the preliminary results of the BRIC INAIL 2024 project (ID07), aimed at defining a standardized protocol to assess the extra-auditory effects of noise on healthcare workers. The study characterizes the acoustic environment of selected Italian public and private hospitals through an observational approach involving three levels of depth: preliminary noise monitoring, room acoustic characterization, and integrated environmental-biomedical assessment. Preliminary measurements reveal that sound pressure levels significantly exceed WHO recommendations, with high sound dynamics (ΔL ≈ 13 dBA) and excessive reverberation times. Based on these findings, specific intervention strategies—including the installation of high-efficiency sound-absorbing materials—were designed and simulated. The results demonstrate a substantial predicted reduction in reverberation.

Speaker: Veronica Amodeo (Department of Architecture - University of Florence)

Full paper (PDF)

178 Soundscape Quality Across Hospital Departments: Psychoacoustics and ISO 12913 Assessment

Hospital noise is a clinical concern, but average A weighted sound pressure levels (LAeq) do not explain how people experience sound in different departments. This study examined whether department type drives hospital soundscape quality by combining psychoacoustic measures with soundscape questionnaires based on ISO 12913. We collected 124 acoustic measurements and 86 questionnaire responses across four departments: Emergency, Intensive Care Unit (ICU), Oncology, and Hematology, in four Dutch tertiary hospitals. Psychoacoustic annoyance (PA) was computed using the Zwicker-Fastl model alongside Loudness (N5), sharpness, roughness, and fluctuation strength. Results showed clear department-specific profiles. ICU and Oncology had nearly identical LAeq values (54.15 vs. 54.17 dB(A)) but differed substantially in psychoacoustic profile and perceived sound quality, showing that equal dB levels can mask meaningful experiential differences. Emergency had the highest annoyance (PA =15.80) and lowest perceived sound quality, while Hematology had the lowest annoyance (PA = 9.13) and highest appropriateness. PA correlated positively with ISO Eventfulness (ρ = 0.41) and negatively with ISO Pleasantness (ρ = −0.22). Roughness and Loudness (N5) showed the strongest negative associations with perceived pleasantness. Staff and ICU patients rated the same environment markedly differently, highlighting the role of listener context. These findings support department-specific interventions and show the value of integrating psychoacoustics with ISO 12913 soundscape assessment.

Speaker: Pasupathan Chithra Barani (Delft University of Technology)

Full paper (PDF)

179 Acoustic Design Recommendations for Intensive Care Units

The acoustic design of an intensive care unit should not only ensure low noise levels and short reverberation times to support effective communication and, as far as possible, reduce noise related stress for both patients and staff. In multi-bed patient rooms, it should also provide adequate acoustic privacy between patients. As these requirements can be partly contradictory from an acoustical perspective, a study was conducted in which twelve different room configurations were evaluated using a full-scale mockup of an intensive care unit. Three different types of screens between bedside areas were assessed, together with the effect of adding sound-absorbing wall panels, as well as the combined effect of screens and wall panels. The results of the study provide tangible outcomes that can be used as input for acoustic design recommendations for intensive care environments.

Speaker: Per Hiselius (Ecophon)

Full paper (PDF)

180 Noise assessment in healthcare facilities: a preliminary approach for estimating annoyance based on source intrusiveness

Hospital noise severely impacts patient recovery and staff well-being. Traditional metrics fail to capture true perceived annoyance, while accurate psychoacoustic models require continuous audio recordings, violating clinical privacy constraints. Conversely, current intrusiveness standards (UNI/TS 11844) use privacy-compliant sound level meters but only evaluate energetic emergence, ignoring the tonal and fluctuating characteristics of sound. This study compares the psychoacoustic annoyance and intrusiveness of isolated noise events (alarms, speech, equipment) in an Italian hospital ward. To bridge the gap between existing models, a data-driven methodology is proposed. The aim is to introduce a weighted Intrusiveness Index. The model extracts temporal variance and spectral prominence from the acoustic time-history, generating objective penalties that simulate fluctuation and tonality. This approach aims to approximate psychoacoustic annoyance solely from sound-level data. It provides researchers and designers with a robust, scalable, and privacy-compliant tool to assess acoustic comfort in healthcare environments without compromising patient confidentiality.

Speaker: Domenico De Salvio (University of Bologna)

A17.07 Source directivity and sound source identification: S126

Conveners: Matthieu Hartenstein (L2S, CentraleSupélec), Samuel D. Bellows (University of Utah Asia Campus)

181 Improving Independent Component Analysis-Based Acoustic Blind Source Separation with Beamforming

Blind Source Separation (BSS) is a long-standing signal processing problem where individual signals cannot be observed separately, only as noisy mixtures produced by an unknown mixing system. The objective is to recover the original sources as efficiently as possible. This problem is highly relevant in telecommunications, medical signal processing, and acoustics, and has been widely studied in the literature.A related problem in acoustic and radio signal processing is beamforming, which aims to determine the spatial orientation of sound sources using a microphone array of known geometry and to separate them by focusing on their directions. Given the conceptual similarity between BSS and beamforming, this paper investigates how classical BSS techniques can be adapted for acoustic beamforming.We explore frequency-domain independent component analysis (FD-ICA), a widely used BSS method for convolutive mixtures in real acoustic environments. The paper reviews Complex-ICA methods underlying FD-ICA and examines how microphone array properties --- known sensor positions and a larger number of microphones than sources --- provide additional information that can mitigate limitations of classical BSS approaches. The applicability of traditional beamforming techniques within the FD-ICA framework is also analyzed.Based on these insights, we propose a new FD-ICA-based source separation algorithm tailored to microphone arrays. The method is evaluated on real measurements, with performance compared to classical beamforming techniques and the impact of using multiple microphones assessed.

Speaker: Péter Fiala (Budapest University of Technology and Economics)

Full paper (PDF)

182 High-Performance Stochastic Optimization of Phased Microphone Arrays for Acoustic Imaging of UAV Propulsion Systems

The rapid proliferation of Unmanned Aerial Vehicles (UAVs) demands advanced diagnostic tools to analyze rotor-generated noise. While acoustic imaging via phased arrays is widely used, standard static geometries fail to suppress spatial aliasing across varying UAV rotational speeds and complex spectral signatures. Building upon wavelet-based beamforming methodologies for high-speed rotating sources, this paper proposes a paradigm shift: a quasi-real-time variable-geometry microphone array. Instead of altering the topology class, the array dynamically morphs the mutual distance between sensors to adapt to the instantaneous noise spectrum of the target drone. This dynamic shape-shifting is driven by a stochastic Monte Carlo optimization algorithm designed for embedded electronics. To demonstrate the computational feasibility of this approach, Monte Carlo runs are benchmarked across three geometries using single and dual Intel Xeon E5-2697 V4 CPUs, and an NVIDIA Tesla P100 GPU. The results deduce the necessary TFLOPS required for edge-computing implementations. By coupling stochastic geometric adaptation with exact Doppler inversion via generalized Morse wavelets, the proposed framework drastically minimizes the Maximum Sidelobe Level (MSL), delivering an adaptive tool for next-generation computational aeroacoustics.

Speaker: Alessandro Di Marco (Universita degli Studi Roma Tre)

Full paper (PDF)

183 Limitations of the array-source distance estimation using planar arrays

Most studies on acoustical source localization focus on estimating directions of arrival (DOAs) of far-field source, or localization of sources in a plane parallel to the array. Nevertheless, in several applications, an estimate of the source-array distance is desirable, such as the localization of drones. Numerical and experimental studies on source localization have evidenced that distance estimation fails for distant sources. The critical distance above which the distance estimation fails is generally considered to be the Rayleigh distance. This study provides a theoretical analysis of the Cramer-Rao bound (CRB) for the distance estimation problem using planar microphone arrays. The CRB provides a lower bound for the variance of the estimation of the parameters of the sources (e.g. position, power), for any unbiased estimator. Asymptotic approximations of the CRB under the conditional (time-harmonic source signal) and unconditional (random source signal) hypotheses are derived. In both cases, a high increase rate of the CRB in function of the distance is found. It is also shown that the Rayleigh distance does not predict the distance at which source localization fails. The results of numerical simulations and of an experiment strongly support the theoretical analysis.

Speaker: Matthieu Hartenstein (L2S, CentraleSupélec)

184 Cepstral Decomposition of Microphone Directivity in Arrays for Acoustic Direction Finding

Many acoustic direction-of-arrival (DoA) algorithms rely on digital beamforming with omnidirectional receivers, which demands the highest computational load per method when applied to a full-sphere focus grid. Conventional methods often neglect the spatial filter characteristics inherent in passive acoustic gain patterns, which are standard in studio microphones. This work introduces a phase-less DoA estimation method for a full-sphere focus grid, leveraging the comb-like filter characteristics of non-omnidirectional microphones when excited by a noise source. Through cepstral decomposition, informed by domain knowledge of the spatial filter, we derive an analytical relationship between the rotation-ambiguous angle of arrival and the filter parameters observed in the spectrum. For an array of orthogonally positioned non-omnidirectional microphones, superposing relative direction estimators yields one or more sound source directions in 3D space. We first simulate scenarios for far-field noise sources and then experimentally validate them under free-space conditions using real noise sources. Classification metrics and spatial error for source detection are presented, with comparisons across different array geometries. As the proposed method employs cepstral analysis per channel rather than per discrete focus point, it reduces computational load, making it suitable for autonomous low-power sensor networks in noise monitoring applications.

Speaker: Jakob Tschavoll (Technische Universität Berlin)

Full paper (PDF)

185 Impact of Model Complexity on Acoustic Source Localization Accuracy

This paper investigates the limits of acoustic source localization when using multipole-based models in a two-dimensional cylindrical-wave framework. A small displacement of a monopolar sound source generates additional spatial patterns beyond the basic monopole field. These patterns carry essential information about the source position. When estimating the source location from noisy measurements, we show that including more multipole components in the model can unintentionally remove this information. For example, adding components such as dipoles and quadrupoles to the estimation model can absorb the effects of source displacement, making the position increasingly difficult to identify. This leads to a loss of identifiability and a significant degradation in estimation accuracy. More generally, we show that the estimation behavior is governed by the lowest-order remaining term in the field expansion with respect to the source displacement. These results are derived analytically and confirmed through numerical simulations, showing strong agreement between theory and practice.

Speaker: Muhammad N. Albezzawy (DAAA, ONERA)

Full paper (PDF)

A18.03 Soundscape methods, monitoring and metrics: S130

Convener: Rosa Ma Alsina-Pagès (La Salle, Universitat Ramón Llull)

186 Translating Soundscapes into Practice using Virtual Reality: city makers’ insights on Interactive Soundscape Simulation

Despite advances in soundscape research, its application to urban planning and design practice remains limited. In many professional contexts, engagement with sound is still reduced to simplified metrics, if considered at all, leaving practitioners without the conceptual grounding or tools required to consider soundscapes. While raising sound-awareness remains an ongoing challenge, awareness alone is insufficient without accessible means to translate these concepts into practice. This is further constrained by the difficulty city makers face in imagining how soundscape tools might be used before they exist.To support this translation, we evaluated the interactive soundscape simulator City Ditty with 15 city makers from diverse professional backgrounds, including urban planners, architects, designers, and acousticians across public and private sectors. Participants used City Ditty’s immersive VR version (using head-mounted display) to engage in interactive soundwalks that introduced key soundscape concepts while allowing free manipulation of sound sources and contextual factors, such as time of day and weather, to explore the sonic implications of their actions. Participants then took part in semi-structured interviews examining how such tools might integrate into existing workflows, where they could provide the most value, and what barriers might hinder adoption.Results highlight both the potential and limitations of immersive, interactive tools in supporting sound-aware design practices. Participants identified opportunities for improved early-stage planning, training, workload support, and stakeholder engagement. Challenges were also observed across user, technological, and systemic contexts. Together, these findings contribute to ongoing efforts to translate soundscape approaches into adoptable, practice-oriented tools for city-making.

Speaker: Catherine Guastavino (McGill University)

Full paper (PDF)

187 From noise reduction to soundscape design: systematic methods for improving the integration of soundscape thinking in urban and (landscape) architectural design

This paper reviews urban and (landscape-) architectural design approaches that aim to better integrate sound mitigation and soundscape enhancement within spatial planning and design. While conventional noise control strategies often rely on late-stage technical fixes, a growing body of work advocates for embedding acoustic considerations earlier in the design process. Central to this review is the “pattern language” approach, originally developed as a generative design method to link spatial configurations with human experience. Its emphasis on reusable, context-sensitive design patterns offers a promising framework for incorporating acoustic quality as a fundamental spatial parameter. The paper examines how pattern language principles have been adapted in urban design, landscape architecture, and environmental acoustics to address both noise reduction and the promotion of positive soundscapes. These adaptations include the development of sound-oriented pattern libraries, participatory design tools, and multi-sensory planning strategies that align acoustic performance with social and ecological values. Key benefits identified include improved interdisciplinary collaboration, earlier integration of acoustic thinking in design workflows, and the ability to balance quantitative noise control with qualitative soundscape goals. However, challenges remain in translating abstract patterns into measurable outcomes and in bridging gaps between design intent and regulatory frameworks.The paper concludes by outlining directions for further research, including the operationalization of soundscape patterns in digital design tools, the validation of pattern-based approaches through empirical studies, and the expansion of spatial design methodologies that foreground auditory experience. These directions aim to support more holistic, human-centered approaches to acoustic design in the built environment.

Speaker: Martijn Lugten (TU Delft)

188 SonoRezé: Lessons from a Co-Constructed Action Research Framework on Urban Sound Environments

SonoRezé project (Rezé, France) explores how participatory action research can bridge the gap between expert knowledge and lived experience in urban sound environments through a residents / researchers / elected officials collaboration framework. Conducted in two phases (2021–2025), the project combines citizen-based noise measurements, qualitative investigations, and co-design workshops involving residents, researchers, and public authorities (elected officials and technical city services). A first phase mobilized 130 inhabitants to produce participatory noise maps of sound levels and source typologies, and to identify perceived sound quality. The second phase focused on aircraft noise, leading to the co-production of analyses and awareness-raising indicators highlighting trajectory changes, sleep disturbance, and the underestimation of annoyance in regulatory frameworks. The work also included a reflection on how results should be communicated and disseminated. Complementary surveys and experience-based maps further documented residents’ experiences. Beyond knowledge production, SonoRezé generated tangible outcomes, including new forms of citizen involvement in decision-making and strengthened dialogue with institutions. The project demonstrates that integrating experiential knowledge can reorient acoustic assessment and support more inclusive public action, offering a transferable framework for addressing other sound-related challenges.

Speaker: Arnaud CAN (UMRAE)

189 Structural Validity of the ISO 12913 Soundscape Circumplex: A Bayesian Predictor-Driven Assessment

Recent soundscape research commonly adopts the ISO 12913 circumplex to evaluate and predict perceptions of acoustic environments. This psychometric model represents soundscape perception in a circular space defined by the latent dimensions Pleasantness and Eventfulness, with eight Perceived Affective Quality items (pleasant, annoying, calm, chaotic, vibrant, monotonous, eventful, and uneventful) uniformly distributed around its perimeter. However, the structural validity of this circumplex has rarely been examined within predictor-driven modeling pipelines. This study addresses that gap by estimating PAQ item angles, with uncertainty, from ordinal questionnaire ratings conditioned on psychoacoustic and acoustic predictors. We also quantify the relative contributions of Pleasantness and Eventfulness to each PAQ item in order to identify dominant dimensional loadings. Using Bayesian multivariate hierarchical models, we evaluate the circumplex structure across multiple soundscape datasets while accounting for person- and scene-specific effects. The results show that Eventfulness-related PAQ items deviate systematically from their ISO reference positions, while the circumplex geometry remains largely preserved. Eventfulness appears less clearly separable from Pleasantness-related variation than assumed by the ISO model. This suggests that prediction performance based on dimension scores should be interpreted cautiously, as high accuracy may partly reflect coupling between dimensions rather than clean recovery of independent perceptual axes. Practically, the findings suggest that applications of the ISO 12913 circumplex in prediction pipelines could be refined by explicitly assessing PAQ item positions and circumplex geometry, enabling more robust evaluation of whether predicted dimension scores preserve the intended perceptual structure.

Speaker: Patrick Blättermann (Hochschule Düsseldorf)

190 Towards the Evaluation of Complex Acoustic Environments Using Sound Source Separation Methods

Perception of a soundscape is shaped significantly by the dominant sound sources present in an acoustic environment. Spatial characteristics have also been suggested to influence how a soundscape is perceived. Soundscape recordings are often analysed holistically rather than at the level of individual sources; the ability to separate and analyse sources individually could enable deeper insight into how specific sources drive overall perception.Spatial audio formats are commonly used in soundscape assessment, yet a source stripped of its directional character cannot be used to investigate spatial contributions to perception, or re-rendered for listening experiments. The preservation of spatial characteristics, however, appears to be relatively unexplored for the task of universal sound separation.This paper presents a minimal modification of Conv-TasNet to accept and output first-order ambisonic (FOA) signals for universal sound separation (separating sources of any class), comparing three loss functions (SNR, SI-SDR, and a multichannel SI-SDR variant) for FOA-to-FOA separation. Results indicate that the multichannel SI-SDR loss substantially improves both separation performance and preservation of source direction compared with conventional SI-SDR and SNR losses. These early-stage findings suggest spatially-preserving source separation may hold promise as a tool for future soundscape studies.

Speaker: Sean Kenji Cragg (Acoustics Innovation Institute)

Full paper (PDF)

A20.00 Speech: S144

Conveners: Philipp Aichinger (Medical University of Vienna), Barbara Schuppler (Signal Processing and Speech Communication Laboratory), Oliver Niebuhr (University of Southern Denmark), Franz Pernkopf (Signal Processing and Speech Communication Laboratory), Peter Balazs (Acoustics Research Institute, Austrian Academy of Sciences)

191 Assessing the Efficacy of Markov Chain Monte Carlo with People to Estimate the Mental Representation of Vowels

Vowel recognition relies on specific physical properties of sound, known as ``acoustic cues'', that listeners use to identify phonemes. The first two formant frequency values (F1 and F2) have been repeatedly shown to be the primary acoustic cues for vowel categorisation. However, identifying the exact distribution underlying their mental representation remains a difficult task. Current methods, such as reverse correlation, are very time intensive as they typically require thousands of trials. In the present study, we tested whether the method Markov Chain Monte Carlo with people (MCMCp) with adaptive Metropolis-Hastings sampling could provide a more efficient alternative to probe the mental representation of vowels. In a perceptual two-interval, two-alternative forced choice task with synthetic vowels, we showed that vowel-specific regions in F1-F2 space were more precisely estimated using the MCMCp approach than with the reverse correlation method, given the same number of trials. These regions were also congruent with those reported in the production literature. Furthermore, these results do not critically depend on the number of features considered. The use of the MCMCp may therefore prove highly useful for experimental phonetics.

Speaker: Mauro Manucci (ENS Paris)

Full paper (PDF)

192 Evaluation of AI-based Text-to-Speech Generators

Text-to-speech systems are an essential building block in our modern AI-driven world. Various models have been established and are often also provided as open source. In this study, we evaluate ten different AI-based text-to-speech systems, with eight of them se- lected from the publicly available toolbox Coqui.ai TTS. In addition, with OpenAI and Google translate, two proprietary models are included for a wider evaluation. We conducted a perceptual evaluation based on ten pre-defined text-prompts. In an online study, participants rated audio quality, clarity, and naturalness. Our results indicate that the generators VITS and Google translate are the best regarding audio quality and clarity, while ChatGPT achieves the highest score in naturalness. In addition, a high correlation between quality, clarity, and naturalness could be observed. We further carried out an objective evaluation with low- and high-level audio features, where the Meta Audiobox Aesthetics shows the high- est correlation with all three dimensions. In the future, more generators and text-prompts should be included in similarly designed evaluation studies. Further, a comparison to real speech will be a valuable next step.

Speaker: Steve Göring (Audiovisual Technology Group, TU Ilmenau)

Full paper (PDF)

193 Keeping Up With The Rhythm: Temporal Dynamics Of Speech Rate Normalization

Speech comprehension relies on normalization mechanisms that help listeners cope with the natural variability in speech production. In speech rate normalization, for instance, the perception of an ambiguous target phoneme can be influenced by the rate of the preceding context. While previous research has demonstrated that both proximal and distal contexts contribute to the effect, they provide only limited insight into the the temporal structure of the rate integration window. This study uses a prosodic reverse correlation approach to assess how 100-millisecond segments within the context sentence contribute to the speech normalization effect. Eight pairs of ambiguous sentences were used, each participant categorizing one pair across 300 trials. On each trial, every segment was randomly time-compressed or stretched. Segment-based correlations between compression factors and responses reveal how speech rate across different segments influences perception. The results demonstrate a broad temporal integration window, with significant effects observed up to 1 second before the target. A detailed analysis of the correlation patterns across the eight sentences challenges the idea that loudness, lexical stress, or phonetic similarity alone drive speech rate integration. The presence of negative correlation coefficients further suggests that more complex mechanisms are involved in the temporal integration of speech rate.

Speaker: Azal LE BAGOUSSE (Laboratoire des Systèmes Perceptifs, ENS-PSL)

Full paper (PDF)

194 Can Algorithms Be Illuded? Machine Learning and the Speech-to- Song Illusion

This study explores the 'Speech-to-Song Illusion' discovered by Diana Deutsch. Specifically, it questions whether a correlation can be established between subjective human perception of the illusion and objective analyses performed by machine learning tools within the open-source Essentia library and others. To identify and examine the specific region, where machine learning algorithms yield ambiguous classifications (speech - song) for speech recordings, this study utilizes the spectro-temporal degradation techniques developed by Philippe Albouy in 2023. Can the model’s predictions for the same speech recordings, evaluated by human listeners, be correlated with the results of a subjective experiment examining their ability to evoke the speech-to-song illusion?

Speaker: Robert Gogol (Adam Mickiewicz University in Poznan)

A06.02 Electroacoustics and transducers arrays: P431

Conveners: Manuel Melon (Université du Mans - LAUM - UMR 6613 CNRS), Thomas Gmeiner, Marko Horvat (University of Zagreb Faculty of EE and Computing)

195 An Open-Ear Clip-On Earbud Research Platform for Perceptual Studies

Open-ear clip-on earbuds, a growing consumer product category, place a transducer against the pinna without occluding the ear canal, preserving natural ambient awareness and improving wearing comfort over traditional headphones. Their low-occlusion design would make them attractive for research applications requiring simultaneous acoustic interaction with the real environment, such as spatial hearing or augmented listening. Yet consumer products are ill-suited for controlled experiments: their inaccessible DSP signal chains prevent full control over the audio chain, their frequency responses deviate substantially from established headphone target curves, and the absence of an acoustic seal fundamentally limits low-frequency output.This paper presents a research platform addressing these limitations. A wired 3D-printed dual-transducer open-ear clip-on earbud prototype was developed through finite element simulations and empirical design. The dual-transducer configuration provides substantially greater headroom for bass extension via equalization. Target frequency responses are achieved using an extension of the established virtual headphone equalization method, adapted to account for elevated harmonic distortion inherent to small drivers at low frequencies. The equalized response to a flat target was validated as robust to earbud repositioning.

Speaker: Baptiste FOURRIER (IKT - Leibniz Universität Hannover)

Full paper (PDF)

196 Development and Numerical Simulation of Ultra-Open Multi-Driver Headphones for Individualized Binaural Reproduction

Current headphone-based binaural audio playback systems rely on conventional stereo headphones with a single transducer per ear. Their frequency response, shaped by interactions between the sound field produced by the transducer, the headphone cup, and the pinna, is typically equalized first to enable hardware-independent playback. Individual adaptation to the user is then achieved using direction-dependent head-related transfer functions. In contrast, this paper presents a multi-driver headphone featuring a total number of 30 dynamic micro-speakers, along with an associated virtual digital twin, with which a fundamentally different approach can be pursued: The ultra-open design minimizes structural influence on the generated sound field, so transfer functions are primarily influenced by the enclosed pinna. The large number of transducers enables excitation of individual, direction-dependent pinna cues that may be used to customize the auditory impression. The numerical simulation model provides highly precise sound field data for developing advanced signal processing under ideal, disturbance-free conditions, while the physical headphones enable subsequent metrological and psychoacoustic evaluation. Design and manufacturing aspects are discussed, and the simulation model is validated in the frequency domain using measurements of sound pressure and particle velocity. Measured and simulated data show good agreement across the investigated frequency range.

Speaker: Benjamin Pries (IKT - Leibniz Universität Hannover)

Full paper (PDF)

197 LoudPy : An Open-Source Python-Based Finite Element Tool for Electro-Vibroacoustic Loudspeaker Simulation

An important aspect of loudspeaker development is understanding how structural geometry and material properties influence the small-signal frequency response and directivity. This question can be answered by the use of numerical methods such as FEA, which approximate the solution of the governing partial differential equations by solving their weak variational form over a discretised computational domain. However, commercial FEM software remains expensive and offers limited flexibility for customization. This paper presents LoudPy, an open-source Python-based simulation framework that models asymmetrical loudspeakers by coupling a FEM formulation for the acoustic and mechanical domains with a lumped-element model for the electrical circuit. Validation against COMSOL simulations demonstrates agreement within <1 % of relative error across the studied frequency range. Key results include small-signal pressure frequency responses, pressure field maps, and a complex eigenvalue analysis, establishing LoudPy as an accessible and reliable alternative for loudspeaker acoustic modeling.

Speaker: Romain Degraeve (Le Mans Universite)

Full paper (PDF)

A10.05/A16.11 Advanced Characterization of Acoustic Materials: S064

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Luc Jaouen

198 Experimental Analysis of Leading-Edge Diffraction from Finite-Sized Rigid and Porous Plates under Grazing Incidence

Leading-edge diffraction and early scattering from finite-sized rigid, porous, and perforated plates under grazing incidence are investigated using dense planar measurements obtained with a moving-microphone array. A signal-subtraction approach is employed to isolate sample-induced contributions in the upstream region, where specular reflections are negligible and time-domain windowing is not applicable. The resulting diffraction-induced level distributions are analysed across multiple grazing angles. A rigid plate serves as a reference condition, enabling consistent characterization of the measured fields. The responses of porous and perforated samples are evaluated both independently and relative to this reference, allowing systematic identification of material-dependent modifications in the early diffraction field. The proposed approach provides a controlled framework for assessing edge-dominated scattering from finite-sized samples under grazing-incidence conditions.

Speaker: Arif Onur Yurek (Aalto University)

Full paper (PDF)

199 Experimental Determination of Dynamic Viscous Resistance and Thermal Relaxation Conductance in Porous Materials

Porous materials are widely used in acoustic and thermoacoustic applications due to their ability to dissipate acoustic energy through viscous and thermal interactions between the fluid and the solid skeleton. These mechanisms are commonly described by visco-thermal parameters such as the viscous resistance and the thermal relaxation conductance, which play a central role in equivalent fluid models of porous media. However, their direct experimental determination remains challenging.This work presents two experimental methodologies enabling the estimation in the low frequency regime of viscous resistance and thermal relaxation conductance in porous materials subjected to oscillating flow. The techniques rely on acoustic pressure measurements performed in a modified two-microphones technique in a standing wave tube. Two different measurement configurations allow the viscous and thermal contributions to be independently evaluated.The proposed approach is tested experimentally on a polyester fibers sample. The results show good agreement between experimental estimations and theoretical predictions, demonstrating that the proposed methods allow direct estimation of visco-thermal parameters of porous materials. The presented techniques provide a useful experimental tool for the characterization of porous materials in acoustic and thermoacoustic applications.

Speaker: Elio Di Giulio (University of Naples Federico II)

Full paper (PDF)

200 Characterization of Blown-in Insulation Dynamic Stiffness Using a 2-DOF Measurement System

This contribution summarizes the results of multi-year research focused on the integration of biobased blown-in insulation materials, such as straw and cellulose, into building structures, with an emphasis on their acoustic properties. Although initial studies confirmed the environmental sustainability and satisfactory sound reduction index of these constructions, they simultaneously revealed critical limitations of the current ISO 9052-1 normative methodology. The standardized load (4kPa) applied during dynamic stiffness (s’) measurements causes excessive compression of these materials, leading to a significant overestimation of stiffness and subsequent errors in acoustic predictions. The core of this work presents a methodology utilizing a two-degree-of-freedom (2-DOF) system and a lightweight load plate. To atenuate nonlinear response of the lightweight plate during top-down excitation, a baseplate excitation system (bottom-up) was implemented. The mechanical behavior of the porous structure under low loading conditions was analyzed using Kraak's Extended Model (KEM) and validated through numerical simulations (FEM and Simscape).

Speaker: Daniel Urbán (Slovak University of Technology)

A11.03 Physical modelling and simulations of musical instruments: P469

Conveners: Péter Rucz (Budapest University of Technology), Juliette Chabassier

201 Aerodynamic Mechanism of Accordion Pitch Bending: Simulation and Theoretical Analysis

The accordion is a type of free-reed instrument. In this instrument, a pitch bending technique can be performed where a player achieves a decrease in pitch up to a semitone or so by partially pressing a key while operating the bellows forcefully. The mechanism of accordion pitch bending is fundamentally different from that of the harmonica. While the latter is caused by the player's vocal tract resonance, the former occurs due to an aerodynamic effect. When a key is pressed partially, the pallet only partially opens the vent hole of the reed chamber. This restriction induces a significant aerodynamic pressure loss, causing the pressure within the chamber to fluctuate in synchronization with the vibration of the reed. This pressure acts to close the reed strongly when it is near its closed position and weakly when it is open, thereby partially offsetting the restoring force of the reed. This reduction in the effective spring constant leads to the pitch bend.In this presentation, we first provide a qualitative explanation of this mechanism using an equivalent circuit. We then present a physical model of accordion sound production. Through numerical simulations and theoretical analysis of nonlinear vibrations, we reproduce the experimental results of pitch bending reported by Okada et al. [J. Acoust. Soc. Jpn. 81(11), 739--742 (2025)]. Our findings clarify how this unique musical effect arises from the interaction between aerodynamic pressure loss and reed dynamics.

Speaker: Seiji Adachi (School of Marine Science and Technology, Tianjin University)

Full paper (PDF)

A13.10 Acousto-Optics and Distributed Acoustic Sensing: P485

Conveners: Emmanuel Dekemper (BIRA-IASB), Samuel Dupont (Université Polytechnique Hauts-de-France), Ireneusz Grulkowski (Nicolaus Copernicus University)

202 Acousto-optic Tuneable Filter design for marine plastic debris detection

Marine plastic pollution represents one of the most pressing environmental challenges of the contemporary era. Despite extensive research, the formation, transformation, and short-term dynamics of floating plastic debris remain inadequately understood, primarily due to insufficient high-resolution spatiotemporal observations. Furthermore, conventional space-based multispectral sensors, while valuable for ocean monitoring, lack the spectral and polarimetric capabilities required for effective plasticdetection.To address these limitations, this study investigates the potential of hyperspectral imaging in the Short-Wave InfraRed range, focusing on the development of a tuneable acousto-optic filter optimized to exploit plastic-specific spectral signatures. The proposed approach integrates advanced spectral indices and polarimetric measurement requirements to enhance discrimination between plastics and their marine surroundings under complex environmental conditions. This article outlines the theoretical framework for filter design and evaluates its expected performance, offering a promising pathway toward improved remote sensing of marine plastic debris.

Speaker: Emmanuel Dekemper (BIRA-IASB)

Full paper (PDF)

203 Multi-Fiber Distributed Acoustic Sensing for Urban Seismic Monitoring and Subsurface Imaging

Dense instrumentation is essential for urban seismic monitoring, yet deploying traditional seismometer networks within city environments poses significant logistical and financial challenges. Distributed Acoustic Sensing (DAS) offers an alternative by repurposing existing telecommunication infrastructure as continuous arrays of seismic receivers. Athens presents a particularly suitable study site: the Greek capital lies within a tectonically active sedimentary basin where more accurate subsurface velocity models are needed to better characterize local seismic hazard. We present a multi-fiber DAS experiment conducted in Athens in partnership with the Hellenic Telecommunications Organization (OTE), in which four fibers spanning 24 to 48 km from a central location were simultaneously interrogated using two units, achieving wide azimuthal coverage across the metropolitan area. This constitutes one of the most spatially extensive urban DAS deployments to date. Two sets of acquisition parameters were used, optimized for earthquake detection and ambient noise interferometry workflows. Event detection via phase-weighted stacking and STA/LTA triggering identified 548 National Observatory of Athens (NOA) catalogue events on at least one fiber, as well as numerous sub-catalogue events. P- and S-wave arrival times from ~29 NOA events are used as input to 3D Eikonal traveltime tomography, employing a second-order fast marching method with L-BFGS inversion and L2 damping regularization, yielding a refined velocity model of the Athens basin. This model is used to relocate NOA catalogue events, enabling an assessment of hypocenter constraint quality relative to catalogue solutions, and to locate DAS-only events. The newly located events are then incorporated into an iterative inversion to further refine the velocity model and improve tomographic resolution.

Speaker: Isha Lohan (ETH Zurich)

A14.05 Neural substrates of complex auditory perception and auditory scene analysis: P491

Conveners: Katrin Krumbholz (University of Nottingham), Joseph Sollini (University of Nottingham)

204 The influence of masker type and intelligibility level on listening effort in adolescents

Understanding speech in background noise is a complex task, and adolescents persistently perform worse than adults in situations in which the competing noise is composed of interfering talkers (informational masking). The underlying mechanisms explaining this protracted development of performance are still not completely understood although they could be related to the recruitement of cognitive resources. The aim of this study is to investigate the potential differences in listening effort between adolescents and adults, in different types of maskers and different intelligibility levels. Considering that listening effort does not reflect intelligibility scores, investigating listening effort might provide more insights to the differences observed between adolescents and adults in intelligibility measures. Pupil dilation was measured in children (12-13 years old) and young adults with typical hearing during a speech intelligibility task against two types of maskers (speech-shaped noise and 2-talker speech) and at two intelligibility levels (50 and 84% correct responses). Working memory was evaluated using the backward digit span. Data collection is still ongoing. A repeated measures ANOVA will assess the effects of masker type, intelligibility level and age group on pupil dilation. In adults, results show larger peak pupil dilation in the 2-talker condition compared to the SSN condition and larger peak pupil dilation in the 50% compared to the 84% intelligibility level. For speech-in-noise performance, we expect a group difference only in the 2-talker condition. We expect to observe an interaction between group and condition on pupil dilation. Those results could help understand whether performance in children is limited by their cognitive ressources or their ability to allocate those.

Speaker: Luna Prud'homme (Université libre de Bruxelles)

A16.00 Room Acoustics: P478

Conveners: Jamilla Balint (Rohde Acoustics), Francesco Martellotta (Politecnico di Bari, DARCOD), Mélanie Nolan (Universidad Politécnica de Madrid)

205 Exploration of Acoustic Applications of Air-Inflated Latex Balloons Beyond Impulsive Sources for Reverberation Time Measurement

Air-filled balloons are commonly used for party celebrations and, in acoustics, as simple impulsive sources (when burst) to measure sound decay within a room. However, their potential application in other areas, particularly their contribution to sound absorption, has not been extensively investigated, as addressed in this paper. Acoustic measurements were carried out on air-inflated latex balloons with a diameter of 100 mm, arranged either as single or double layers; in the latter case, the external layer was left unsealed at the neck. Two measurement methods were compared: results obtained using an impedance tube and those measured within a reverberation chamber. The results show that the absorption coefficients obtained from the reverberation chamber are reasonably consistent, especially at high frequencies, whereas the absorption performance measured with the impedance tube is relatively limited. This research suggests that such configurations could be applied as rapid, low-cost mitigation measures in situations where acoustic correction is required.

Speaker: Gino Iannace (Università degli Studi della Campania)

Full paper (PDF)

206 Room Acoustic Challenges of Heritage Spaces: Case Studies from the Budapest University of Technology and Economics

World Heritage sites often have high reverberation times and low speech intelligibility, which are rarely suitable for their new, modern‑day functions. This paper shows how the acoustic refurbishment of two rooms can be achieved while preserving their architectural and historical character.Both rooms are located in the Central building (Building K) of the Budapest University of Technology and Economics: the Main Hall (Aula), which is often used for ceremonies and musical performances, and a regular classroom, one of the largest lecture rooms in the building.Initial measurements using interrupted noise and impulse sources showed extremely high reverberation‑time values in both rooms, which explain the generally negative subjective opinions about their overall acoustical quality.To obtain a more detailed evaluation, we performed impulse‑response measurements using a swept‑sine signal and also created and circulated a subjective questionnaire among the audience and musicians of the Main Hall and the students using the lecture hall. In addition, the rooms were analysed using CATT‑Acoustic software.Based on these observations and analyses, we determined where sound‑absorbing materials could be placed while respecting the visual integrity of these historic spaces. The effects of the proposed solutions were also analysed and are presented here.

Speaker: Judy Housam Mohamed Fathi Abdellatif (Budapest University of Technology and Economics)

207 STI Measurement and Evaluation of Directional Loudspeakers for Emergency Broadcasting in Noise Barrier Tunnels

This study investigates the measurement and evaluation of the Speech Transmission Index (STI) of emergency broadcasting speaker systems in noise barrier tunnels under emergency conditions such as fire incidents. STI measurements were conducted for two types of directional loudspeakers in both a semi-anechoic chamber and an actual noise barrier tunnel. The evaluation was conducted under various conditions, including speaker mounting height, speaker orientation, and the distance between the sound source and the receiver.In addition to STI evaluation, sound pressure levels (SPL) were measured and analyzed under each test condition. STI measurements were performed using two methods: a pink noise signal radiated from the loudspeakers and a speech-based method utilizing a talk box. The SPL results confirmed that the target sound pressure level of 80 dB or higher was achieved under the tested conditions. Furthermore, the STI evaluation results demonstrated that the target STI value of 0.60 or higher was successfully satisfied. [ This work was supported by the Technology Innovation Program [(RS-2025-05492970, Development of a Disaster-Responsive Noise Barrier System for Ensuring Evacuation Safety in Tunnel Fires) funded By the Ministry of Interior and Safety)(MOIS, Korea)]

Speaker: JUNOH YEON (Korea Marine Equipment Research Institute)

208 Using feature maps to classify room impulse responses

Room impulse responses (RIRs) are used to characterize rooms, their acoustic behavior and the resulting coloration of the sound. They are widely used in applications for room acoustic design, but also in speech recognition and spatial audio. To describe room impulse responses, well defined features according to ISO 3382 are typically applied, such as e.g. reverberation time, clarity, strength or the direct-to-reverberant energy ratio. However, are these features sufficient to capture and objectively describe the overall sound quality of a room?In this article, we approached this research question by analyzing the contribution of several acoustic features in a multidimensional feature space for a large data set of RIRs. In addition to classical room acoustic features as defined by ISO 3382, new features adapted from statistics, signal processing and speech analysis, were explored. The room impulse responses under test were represented in this feature space and were clustered based on their distribution in this space. It was found that classical acoustic features play a major role in the clustering structure, while additional descriptors, like statistical moments, provide complementary information, particularly in characterizing the reverberant part of the impulse response. Future work will conduct listening tests to clarify whether RIRs that are grouped together sound alike. Such knowledge is relevant for data-driven acoustics or machine learning applications involving room impulse responses.

Speaker: Chloé Balmes (HARMAN International Industries, DTU)

A16.02 Reliable Characterization of Acoustic Absorption, Reflection, and Scattering: P482

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Marco Berzborn (Eindhoven University of Technology), Jonathan A. Hargreaves (Acoustics Innovation Institute)

209 Exploring practical approaches to sound absorption measurement: validating a repurposed space as a small-scale reverberation room

While normal-incidence absorption can be measured with a simple impedance tube (ISO 10534-2), its random-incidence counterpart (ISO 354) assumes a level of infrastructure most university labs simply don't have. This paper first compares these two standardized methods with several non-standardized alternatives, weighing their accuracy, practicality, and cost in a summary table intended as a practical guide. It then addresses the specific case of small-scale reverberation rooms, evaluating a small, unmodified bathroom as a low-cost alternative to a dedicated reverberation chamber. The validity of this approach is assessed by comparing the absorption coefficients obtained in the adapted room with those measured in an impedance tube, across three sample configurations of increasing exposed area. The results show that compatibility between the two methods depends on two independent conditions: measuring above the room's Schroeder frequency, and using a sufficiently large sample area, neither being sufficient on its own. These findings provide a practical guideline for researchers with limited resources seeking to validate small-scale reverberation room measurements in non-dedicated spaces.

Speaker: David Bosonin (Politecnico di Torino)

Full paper (PDF)

A16.04/A24.06 Acoustics for learning environments across physical and virtual realities: P492

Conveners: Arianna Astolfi (Politecnico di Torino), Janina Fels (IHTA, RWTH Aachen University)

210 Room Acoustics and Compliance of Austrian Classrooms

A dataset of 416 reverberation time measurements, collected in rooms of Austrian educational institutions enable a quantitative characterization of current room acoustics in education. 144 classrooms in schools are looked at to assess compliance with national requirements for reverberation time. Occupied conditions are calculated using ÖNORM B 8115‑3 (2023) with a fixed per‑pupil absorption. Results are classified and interpreted according to the building regulation OIB-5 (2023) as compliant, compliant within tolerance (±20% in all octave bands), or non‑compliant. Overall, 27% of all surveyed classrooms meet requirements. Restricting to new/renovated rooms since 2007 (n=97), compliance is at 34%. For the whole classroom dataset (N=144), technical absorption is present in 96 rooms (mostly ceilings), most frequently used materials are stated. Frequent causes of non‑compliance are insufficient effective absorption, over‑absorbing rooms (too short reverberation times), and exceedances of reverberation times at 250 Hz. Limitations include potential selection bias (e.g. measurements followed by complaints), fixed person‑absorption assumptions, and uneven distribution above the nine Austrian states. Findings indicate that roughly one third of Austrian classrooms in this dataset achieve current targets.

Speaker: Sarah Ambros (Universität Wien)

Full paper (PDF)

A17.07 Source directivity and sound source identification: P495

Conveners: Matthieu Hartenstein (L2S, CentraleSupélec), Samuel D. Bellows (University of Utah Asia Campus)

211 Acoular: Acoustic testing and source mapping software

Acoular is an open-source Python toolkit for analyzing multichannel recordings from microphone and sensor arrays. Widely used by researchers and engineers, it supports beamforming, deconvolution, source localization and general signal analysis across aeroacoustics, automotive, environmental and bioacoustic applications. Its modular, extensible design lets users adapt and extend functionality, and the project is actively maintained. Extensions include SpectAcoular, a browser-based GUI for interactive visualization, and AcouPipe, a pipeline for reproducible data generation aimed at machine-learning workflows.

Speaker: Art J.R. Pelling (Technische Universität Berlin)

212 SpectAcoular & AcouPipe: Extensions for Acoular

This poster presents two complementary extensions to Acoular that streamline interactive exploration and reproducible dataset creation for acoustic research.SpectAcoular is a Bokeh-based GUI layer for Acoular that exposes interactive widgets and prebuilt applications via the web browser. By delegating rendering and event handling to a high-level JavaScript stack, SpectAcoular enables cross-platform, low-latency visualizations and simple construction of custom interfaces for real-time source mapping and signal inspection—without requiring platform-specific GUI toolkits.AcouPipe is a pipeline for reproducible acoustic data generation targeted at machine-learning workflows. It provides tools for statistically controlled random sampling of virtual measurement scenarios, scalable distributed computation across machines. Users can generate large datasets while storing only derived features needed for training and share the containerized code required to reproduce results rather than the raw data itself, avoiding large transfers and improving reproducibility.Together, SpectAcoular and AcouPipe extend Acoular’s modular architecture: SpectAcoular lowers the barrier to interactive analysis and dissemination, while AcouPipe accelerates and documents data production for ML-driven studies. The poster summarizes design, example applications, and resources for adoption.

Speaker: Adam Kujawski (Technische Universität Berlin)

A18.03 Soundscape methods, monitoring and metrics: P508

Convener: Rosa Ma Alsina-Pagès (La Salle, Universitat Ramón Llull)

213 Development and Validation of a Common Data Model for Sound Level Data from Music Venues and Events

Measurement of sound levels at music venues and events is important, and sometimes mandated, to protect audience members’ hearing, limit annoyance to neighbours, and ensure compliance with regulatory requirements and/or best-practice guidelines. However, ongoing logging of sound levels generates a vast amount of data that varies greatly by format depending on the equipment and methods used. This makes it difficult to extract relevant insights, particularly across different datasets. To address this issue, we propose DB-SPL (“Database for Sound Pressure Levels”), which consists of i) a common data model designed to standardise the organisation of sound level data; and ii) a standardised vocabulary of relevant key terms pertaining to sound level measurement and monitoring. Our goal is to provide a common framework for storing, describing, and analysing sound level data from music venues and events, facilitating improved data sharing and data re-use for research.

Speaker: Oluwademilade Oshin (University of Nottingham)

A21.04 Railway noise and vibration: P521

Conveners: Karoline Alten (AIT Austrian Institute of Technology), Christopher Knuth (Empa), Slimane Ouakka (Empa)

214 A Study of the Relationship Between Train Speed and Acceleration Amplitude Based on ABA Measurements

This paper investigates the relationship between train speed and acceleration amplitude measured through axle‑box acceleration (ABA) at insulated rail joints (IRJs) on the Norwegian railway network, using the track‑recording vehicle Roger 1000. A total of 56 passages (from two IRJs) recorded with a tri‑axial piezoelectric accelerometer mounted near the axle box on the leading wheelset were analyzed. For each passage, a +-2.5 s window around the IRJ was extracted to identify the exact peak where the train is assumed to have passed the IRJ, and a RMS-value of the absolute amplitudes was computed in a +-1 ms gliding window to show the local vibration energy. Second-order regression models were fitted for both vertical and longitudinal directions that show a near-linear increase in amplitude with speed, with the vertical direction exhibiting the highest sensitivity. Further, normalized RMS-values with speed based on the second order regressions help reveal outliers from a perfect amplitude-speed-relationship, indicating additional influences. The longitudinal (X) direction exhibits larger relative variability than the vertical response after speed normalization, and the changes following documented rail-grinding events were more pronounced in X-direction, suggesting that longitudinal accelerations are more sensitive to subtle changes in track condition.

Speaker: Helena Steine Tysland (NTNU)

Full paper (PDF)

A24.01 Auralization of complex environments: P444

Conveners: Josep Llorca-Bofi (Fraunhofer Institute for Building Physics), Jonas Heck (Institute for Hearing Technology and Acoustics)

215 Rapid background sound generator for urban soundscapes studies in Stuttgart

European cities are adapting soundscape-based urban planning to create better living spaces for residents. Auralization acts as one of the possible tools that provides immersive auditory experiences for urban planners, architects and the community to participate in the design process without acoustical expert. Exisiting auralization tools suffer from long computation time. This paper explores using surrogate background sound generators. By combining Text-to-Audio models (TTA) and BRIR-based models, the paper highlights the potential of surrogate audio renderers to support more perceptually grounded, design‑oriented decision‑making in urban environments.

Speaker: Hou Hin Au Yeung (Fraunhofer Institute for Building Physics)

Full paper (PDF)

216 The Acoustic Signature of Awe in Sacred Architecture

While the psychology of awe has been extensively studied through visual stimuli, the role of acoustic environments as a driver of self-transcendent experience remains underexplored. This study investigates how structural properties of room acoustics—such as reverberation, spectral balance, and clarity—influence perception, behavior, and auditory awe.Drawing on frameworks of awe (perceived vastness and need for accommodation) and predictive processing, we examine whether variations in acoustic environments induce measurable changes in the action–perception loop. Participants are immersed in a controlled multichannel listening environment and assigned to one of three conditions: a minimally reverberant control space, a moderately reverberant “Warm” environment, and a highly reverberant “Vast” environment inspired by monumental architecture.Participants engage in sound-making and listening tasks, including vocalization, speech, rhythmic interaction, and instrumental exploration. Behavioral responses are quantified using vocal sound pressure level, inter-onset interval variability, and interaction patterns, while subjective experience is assessed through standardized affective scales and self-report measures of awe and self-perception.We hypothesize that highly reverberant environments will promote exploratory interaction, alter temporal perception and rhythmic stability, and influence vocal production and speech clarity. These acoustic features are proposed to form an “acoustic signature of awe,” enhancing perceived vastness, immersion, and awe-related experiences.This work links room acoustics to measurable behavioral and emotional responses and provides a framework for studying the sonic dimensions of awe in immersive environments.

Speaker: Nima Farzaneh (Stanford University)

A21.04 Railway noise and vibration: S154

Conveners: Karoline Alten (AIT Austrian Institute of Technology), Christopher Knuth (Empa), Slimane Ouakka (Empa)

217 A blocked force-based approach for railway-induced vibration simulation considering strongly coupled track-soil-building systems

Ground-borne vibration induced by railway traffic is a growing concern in densely populated urban environments. Railway-induced ground-borne noise and vibration prediction tools are valuable assets when implementing or upgrading railway systems, as well as when constructing new civil structures near existing operational lines. Railway-induced vibration simulations are typically computationally expensive due to several factors, including the unbounded and heterogeneous nature of the soil, the moving dynamic loads generated by wheel-rail interaction, and the dynamic coupling between the different subsystems involved. To reduce computational effort, it is commonly assumed that no significant dynamic coupling exists between the track (or track–tunnel) system and the building of interest. However, in situations where the proximity between the tracks and buildings is small, this assumption may no longer be valid. To address this issue, this paper proposes an efficient methodology to simulate the response of strongly coupled track-soil-building or track-tunnel-soil-building systems based on the blocked force concept. The proposed approach consists of first determining the blocked forces at the building-soil interface using a model of the track-soil or track-tunnel-soil system excited by the moving train-track dynamic and quasi-static loads. These forces are then applied to the corresponding track-soil-building or track-tunnel-soil-building system to compute the building vibration response. By avoiding the explicit modelling of moving loads in this second stage of the methodology, the computational procedure is significantly simplified. The methodology is verified with other numerical schemes to ensure its correctness and a measure of its computational efficiency benefits is given.

Speaker: Robert Arcos (Universitat Politècnica de Catalunya)

Full paper (PDF)

218 On-Site Measurements of Transfer Functions and Prediction of Vibration Velocity along Lisbon Subway Lines

The vibration transfer function method employed by the Portuguese National Laboratory of Civil Engineering (LNEC), based on the FTA methodology, is a predictive and diagnostic approach used to evaluate the propagation of vibrations from railway traffic through the ground and into buildings, causing structural vibrations and re-radiated noise.Approximately 30 surface points representative of sensitive receivers, corresponding to 1.8 km of Lisbon new subway line, were evaluated using measurement and calculation procedures based on this method. In this paper, results of measurements of the transfer functions and prediction of vibration velocity are presented. These results were used to define anti-vibration solutions that allow compliance with the established criteria which are intended to ensure human comfort and prevent damage to buildings. The results illustrate the need to apply anti-vibration measures of different types and degrees of attenuation that vary considerably from one another, depending on the geological characteristics and buildings structure.

Speaker: Tatiana Teixeira (InAcoustics)

219 Exploring Graz by Tram: Monitoring the Network using Vibro-Acoustic Onboard Sensors

Onboard measurement systems installed on rail-bound vehicles have emerged as an effective means of monitoring the interaction between rolling stock and track infrastructure under real operating conditions. The systems can be deployed either on in-service passenger vehicles or on dedicated inspection units and typically integrate a range of sensing technologies. The present study investigates the application of vibro-acoustic sensors mounted below the primary suspension of tram vehicles, with the aim of characterising emission behaviour over a monitoring period of over one year. The analysis is based on data collected during a transnational research initiative conducted across Austria, Germany and Switzerland and focuses on a subset of this dataset obtained from three in-service tram vehicles operating in Graz on a standard-gauge network. The sensor configuration comprised microphones positioned with direct line-of-sight to the wheel–rail contact point, as well as accelerometers mounted on the axle boxes of both wheels of the central, unpowered bogie. The analysis demonstrates characteristic vibro-acoustic signatures associated with different wheel–rail contact conditions, including faults and regular rail components such as crossings, switches or expansion joints, and highlights statistically significant trends observable across the network. Particular emphasis is placed on the challenges of processing measurements obtained under highly variable operating conditions, in contrast to the more homogeneous conditions typically found in heavy-rail environments. The study underlines the potential of advanced signal processing techniques to extract meaningful information on infrastructure condition from large-scale, long-term datasets, thereby contributing to the advancement of condition monitoring strategies for urban rail systems.

Speaker: Karoline Alten (AIT Austrian Institute of Technology)

Full paper (PDF)

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:20 Coffee break

10:20 Coffee break

10:20 Coffee break

10:20 Coffee break

10:20 Coffee break

A06.01 Microphone and MEMS transducers: S316

Conveners: Dominik Mayrhofer (Graz University of Technoloy), Petr Honzík (Czech Technical University in Prague)

220 Numerical Simulation and Evaluation of Viscous Effects in a Coupled Modulated Ultrasound Microspeaker

Research on modulated ultrasound microspeakers has been increasing steadily in recent years, with new modulation and actuation techniques being proposed. Such devices are highly integrated, often consisting of multiple moving parts and varying channel widths that require complex simulation to accurately model their behavior. This makes accurate simulations very challenging. In this contribution, we present a fully coupled multiphysics finite element model of a microspeaker comprising two moving membranes, a pressure chamber between them, and air channels with varying acoustic impedance. Incorporating electrostatic actuation, nonlinear membrane mechanics and a viscous air domain, the model aims to accurately capture surface traction and squeeze-film damping effects. Using a frequency-modulated excitation, the modulation behavior is investigated. In the spectrum of the resulting modulated ultrasound pressure signal, the low-frequency audio signal components are present, confirming the expected modulation behavior. Initial results indicate an increase in total harmonic distortion when nonlinear and viscous effects are included compared to a linear model. This provides crucial insight into the device's development and demonstrates the importance of a fully coupled simulation for predicting the modulation behavior. All steps, from modeling and meshing to the coupled simulation, were done using only open-source software (Gmsh, openCFS), providing a reproducible framework for simulation of highly coupled systems.

Speaker: Tobias Wilczacki (Graz University of Technoloy)

Full paper (PDF)

221 Eigenfrequencies of Perforated Plates in MEMS Transducers: an Analytical Approximation

Modern MEMS transducers frequently employ electrostatic transduction with perforated plates. Despite the prevalence of these devices, 3D numerical simulations of their structural vibrations remain computationally expensive, and a clear analytical understanding of the underlying physics is often lacking. This paper addresses these challenges by providing analytical approximations for the eigenfrequencies of clamped non-perforated and perforated square plates.Using symbolic regression, we provide analytical approximations while avoiding the instabilities associated with high-degree polynomials. For perforated plates, the eigenfrequency is modeled as a normalized discrepancy to the solid case using a rational function characterized by only two geometry-dependent parameters. This sparse, interpretable form facilitates the subsequent derivation of the effective Young’s modulus and effective plate density. To the best of our knowledge, this is the first work to employ symbolic regression for the analytical approximation of perforated plate eigenfrequencies.

Speaker: Aneta Furmanová (Czech Technical University in Prague)

Full paper (PDF)

222 Segmented-backplate condenser microphone with controlled directivity: advanced analytical modeling

This contribution presents a condenser microphone concept based on a segmented backplate, where directional characteristics are obtained from differences in the mean membrane displacement between electrode segments. The approach enables the realization of bidirectional (figure-of-eight) responses without the need for multiple acoustic paths and provides a flexible basis for synthesizing other directivity patterns through signal combination. The segmented-backplate concept and its analytical description build on previously published work and are here extended and further developed.To support the design and analysis of such microphones, an advanced analytical model is introduced. The model describes the coupled behavior of a clamped circular membrane and the thermoviscous acoustic field in the air gap, while explicitly accounting for arbitrary backplate perforation patterns and plane-wave excitation from an arbitrary direction. In contrast to simplified or symmetry-restricted approaches, the formulation captures non-axisymmetric membrane motion and spatially varying pressure fields across a broad frequency range, including frequencies above the first membrane resonance.Analytical predictions are validated against three-dimensional finite-element simulations, showing very good agreement in membrane displacement, air-gap pressure, and frequency-dependent mean displacement. The model further enables accurate prediction of directional characteristics derived from segmented outputs, including their frequency-dependent deviations.The presented framework provides an efficient tool for the design and optimization of condenser microphones with tailored directivity, including extensions toward cardioid and other directional patterns.

Speaker: Petr Honzík (Czech Technical University in Prague)

223 Miniaturization of bio-inspired MEMS sensors for next-generation hearing aids

Bio-inspired acoustic sensors, which integrate functionalities of the human cochlea like frequency decomposition and non-linear amplification directly in the sensor itself, offer a great potential to improve speech-in-noise processing and energy-efficiency for embedded systems and hearing aids. Particularly, the dynamic MEMS cochlea demonstrated advantages regarding robustness of sound recognition in noisy conditions and enables adaptation of sensing and processing capabilities to the current hearing situation. It is based on micro-electromechanical silicon cantilevers in combination with electronic feedback. To cover the auditory frequency range, an array of MEMS cochlea sensors is required due to their frequency filtering properties. This requires miniaturization of the sensor system and electronics to take advantage of its positive properties in next-generation hearing aids.Here, we present the design and fabrication of an array of 8 sensors. Each of these sensors responds to a different frequency, dependent on their geometry. However, due to the mechanical and acoustical coupling between them, artefacts can occur. To optimise the design regarding acoustic sensitivity, detect coupling artefacts and derive sensor parameters for simplified descriptions of the system, FEM simulations are performed and compared to measurements of different arrays, using different actuation schemes for the array, as well as results from the simplified single mode ODE model. Based on these results, the combination of multiple 8-sensor arrays can be designed for covering the auditory range and improve speech processing and sound quality.

Speaker: Claudia Lenk (Universität Ulm)

224 Design, Fabrication, and Packaging of Multi-Band Piezoelectric MEMS Accelerometers for Auditory Prostheses

Hearing aids and cochlear implants restore auditory perception for many patients but remain limited by environmental susceptibility and restricted daily use. Completely implantable auditory prostheses promise continuous, unobtrusive operation. However, their widespread adoption is constrained by the lack of implantable acoustic sensors that are simultaneously low-noise, mechanically robust, and suitable for long-term implantation. We have designed and fabricated ultra-miniature piezoelectric microelectromechanical systems (MEMS) accelerometers for middle-ear sound sensing. Methods for meeting process flow and manufacturing challenges to build the muti-band devices for speech-relevant noise performance over the 0.1–8 kHz bandwidth are presented. Building on this validated MEMS sensor design, we also address implantable packaging as the next-in-queue critical challenge. Titanium-based package architectures are developed using precision titanium stamping and additive manufacturing, integrated with alumina ceramic substrates and metallized electrical feedthroughs. The resulting packages aim to meet the unique challenges of sub-millimeter form factors and sub-20 milligram total mass while maintaining mechanical robustness and hermetic encapsulation. Ongoing and planned evaluations consider seal integrity, material biocompatibility, and the influence on sensing performance in a temporal bone configuration, guided by established standards for implantable medical devices.

Speaker: Karl Grosh (University of Michigan)

A10.05/A16.11 Advanced Characterization of Acoustic Materials: S356

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Luc Jaouen

225 Variability of the physical and acoustic characteristics of polyurethane foams

Open-cell polyurethane (PU) foams are widely utilised for acoustic insulation in the automotive and industrial sectors; however, the foaming process often introduces significant spatial non-homogeneity and anisotropy. This variability complicates the prediction of macroscopic performance via standard semi-phenomenological models, such as the Johnson-Champoux-Allard (JCA) framework. This study presents a systematic experimental investigation into the spatial variability of PU foams within a single production batch, focusing on samples extracted from varying depths of the same injected block. The acoustic properties, specifically normal and random incidence sound absorption, were measured alongside an experimental characterisation of the five non-acoustic transport parameters. To bridge the gap between microscopic morphology and macroscopic behaviour, a semi-empirical multiscale approach was employed to evaluate transport parameters from structural characteristics. The novelty of this work lies in the implementation of an inversion technique designed to determine the reticulation index directly, rather than relying on estimates from microscopic images. Results demonstrate a significant reduction in the reticulation index as extraction depth increases. Validation against experimental data confirms that this approach provides a robust tool for the industrial design and optimisation of acoustic treatments with inherent density and morphological gradients.

Speaker: Andrea Santoni (University of Ferrara)

226 Nonlinear characterization of vibro-acoustic absorbers

This study is concerned with the nonlinear characterization of thin membranes undergoing large vibrations under high acoustic loads, designed for noise mitigation application based on the nonlinear energy sink principle. An experimental vibroacoustic setup is equipped to prescribe high pressure levels to circular thin samples and accurately measure their amplitude-dependent behaviors. This is made possible by monitoring in real-time acoustic and vibration measurements and implementing an experimental continuation procedure based on a phase-locked loop (PLL) control. In particular, amplitude continuation with prescribed phase resonance allows for obtaining the so-called backbone curves and analyzing separately the conservative and dissipative contributions in nonlinear regimes. The signal monitoring and recording as well as the PLL are applied through a software interface devoted to audio applications in which the main functions of analogue electronics are digitally mimicked. The increases of stiffness and damping with oscillation amplitude are characterized using this experimental approach. Finally, a theoretical model including the main nonlinear mechanisms is developed: it allows to fit the measured backbone curves, thus opening the possibility to perform a parameter identification.

Speaker: Arthur Givois (Université de technologie de Compiègne)

227 Indirect Characterization of Fibrous Materials Using In-Situ Acoustic Impedance and JCAL-Based Parameter Optimization

This paper presents a novel indirect methodology for the characterization of fibrous sound-absorbing materials based on in-situ acoustic impedance measurements. The proposed approach relies on the combined measurement of sound pressure and particle velocity to estimate the surface acoustic impedance of the sample under test, enabling material characterization directly in realistic measurement conditions. An inverse identification framework based on the Johnson–Champoux–Allard–Lafarge (JCAL) model is employed to optimize the material parameters that exhibit the strongest influence on the acoustic response, namely airflow resistivity, effective sample thickness, and viscous characteristic length. Other parameters are kept fixed, as a sensitivity analysis showed within a physical feasible range their impact on the measured impedance to be negligible within the investigated frequency range. The optimization procedure is designed to be robust and computationally efficient, making it suitable for automated measurement scenarios. The method is evaluated on a wide variety of fibrous samples, covering different thicknesses and densities, and the extracted parameters are systematically compared against results obtained from standardized laboratory measurement techniques. Experimental results demonstrate good agreement between the proposed indirect approach and reference methods, confirming the validity and reliability of in-situ impedance-based parameter estimation. The study highlights the potential of this technique for fast, non-destructive material characterization and paves the way for its application in industrial and end-of-line quality control environments.

Speaker: Juan Díaz Carabias (Microflown Technologies)

Full paper (PDF)

228 Effects of Water Content on the Acoustic Properties of Porous Media

Porous materials are widely used for sound absorption, yet their acoustic performance can be significantly altered by the presence of liquid water. The effect of overall moisture content on the acoustic properties of porous and granular media has already been investigated, almost through experimental approaches. However, the spatial distribution of water introduced by capillary rise altering the acoustic properties of porous materials remains comparatively unexplored. This work is a first attempt to quantitatively explain how capillary-driven partial water saturation modifies the acoustic properties of a porous material with straight cylindrical pores. The porous structure considered here is fabricated using additive manufacturing and characterized by X-ray tomography and contact-angle measurements. The acoustic transmission and reflection properties of samples with pore diameters from 0.5 mm to 1 mm are measured in an impedance tube for several values of water content in a frequency range from 500 Hz to 3 kHz. The results highlight that capillary rise creates a heterogeneous wetted sublayer that reduces the acoustically active thickness and modifies the acoustic response in a frequency-dependent manner. These findings suggest that the spatial distribution of capillary water, rather than total water content alone, governs the acoustic response of partially saturated porous media.

Speaker: mohammad morovati (Université Le Havre Normandie)

A10.07 Micro-perforated and Helmholtz-type resonant absorbers: modelling and characterisation: S312

Conveners: Teresa Bravo (Spanish National Research Council), Cedric Maury (Laboratory of Mechanics and Acoustics)

229 Added damping and vibration behavior of a microperforated plate in nonlinear regimes

Microperforated plates (MPP) are simple structures that can operate in hostile conditions and withstand high-amplitude mechanical excitations. MPPs have been shown to dissipate energy through thermoviscous interactions between shear fluid layers adjacent to solid perforation walls. In linear operation, the added damping is maximal at a characteristic frequency determined by the perforation parameters. However, its behaviour as amplitude of mechanical excitation levels increases remains less well understood. Since MPPs consist of both a fluid and a solid part, two types of nonlinearities gradually appear as the excitation level increases: nonlinearities induced by the fluid inside the perforations; nonlinearities induced by large deformations of the solid part. An experimentally validated analytical model based on the linear MPP model and Forchheimer's law is proposed to capture fluid-induced nonlinearities. Results show that nonlinear added damping achieves its maximum at a nonlinear characteristic frequency depending on the relative fluid-solid velocity and perforation parameters. It is shown that the nonlinear added damping is a variable in space on the plate. It is therefore maximum locally in areas where the relative fluid-solid velocity is equal to the critical velocity. The results further highlight a transition zone between the two nonlinear regimes: at low excitation amplitudes, acoustic nonlinearities dominate the system response, while at high amplitudes, geometric nonlinearities associated with large solid deformations become prevailing. The relative fluid-solid velocity is identified as the key quantity governing the balance between these two competing nonlinearities, providing a physically meaningful threshold that delineates the transition between both regimes.

Speaker: Lucie Gallerand (École de technologie supérieure)

230 Characterization of micro-perforated materials

Microperforated plates are thin plates containing perforations in the submillimeter range. As the incident sound wave penetrates the material, energy is dissipated by friction in the perforations. In this study, the acoustic mechanisms in microperforated plates and foils are investigated. Due to different manufacturing techniques, the investigated materials exhibit different perforation geometries: circular and slitted perforations that form the microperforated absorber. To characterize the materials' absorption behavior, measurements in the impedance tube are performed. Using optimization techniques, the parameters of theoretical models, such as the Maa model, can be obtained. Here, a cost function is defined to optimize the absorption in a specified frequency range using the material parameters and the geometry as design variables. Since the investigations cover circular and slitted plates, the latter require a more in-depth analysis since the Maa model assumes a circular geometry. Here, numerical simulations using the finite element method are employed to accurately model the absorption via a unit cell of the microperforated plate. The numerical computations are based on the linearized compressible flow equations to investigate the viscous dissipation mechanism in depth. Comparisons with measurements suggest that the unit cell simulations can predict the absorption behavior of the microperforated plates well and efficiently, also for more complex slit geometries.

Speaker: Felicitas Spörk (Graz University of Technology)

Full paper (PDF)

231 Modeling and Characterization of Micro-perforated Panel Absorbers with Parallel and Coiled Sub-cavities for Multifunctional Building Acoustic Applications

Achieving high-quality building acoustic environments requires absorbers that combine frequency-tailored absorption, multifunctional performance, and compact form factors, while remaining amenable to predictive modeling and reliable characterization. Within this scope, this contribution reviews recent advances developed by the authors on micro-perforated panel absorbers (MPPAs) with parallel-arranged or coiled-up sub-cavities of different depths (PCD-MPAs). Analytical prediction models that combine MPP impedance theory with multi-cavity coupling are established for normal, oblique and random incidence conditions, and validated against finite element simulations and impedance tube measurements. Characterization is further extended to thermo-viscous regimes inside ultra-compact coiled cavities, where viscous and thermal boundary layers govern absorption peak shifts and absorption bandwidth. Beyond pure absorption, periodic sub-cavity arrangements derived from quadratic residue sequences are shown to provide simultaneous sound absorption and diffusion, validated through semi-anechoic measurements that comply with AES-4id-2001. Engineering verification is demonstrated by a PCD-MPA muffler whose sound transmission loss exceeds 20 dB across 700–1600 Hz, and by an in-situ traffic tunnel application where the mean reverberation time at 250 Hz decreases by 66% and noise attenuation at 100 m source–receiver distance improves by 17 dB. Building on these results, ongoing work on a ventilated acoustic black hole–micro-perforated panel (ABH–MPP) hybrid structure further extends broadband attenuation through coupled viscothermal dissipation and slow-wave residence effects. Together, these studies form a unified modeling, characterization and application framework that supports reversible, space-efficient design of next-generation multifunctional acoustic materials.

Speaker: Hequn Min (School of Architecture, Southeast Univ)

232 Material Design From Psychoacoustic Targets: Application to Micro-perforated Panels for Sound Absorption

Conventional acoustic material design typically optimizes for physical parameters such as absorption coefficients or transmission loss, which do not always correlate with human perception of sound quality. This research proposes an alternative design methodology consisting of identifying material designs which satisfy a set of target psychoacoustic metrics, and is here applied to micro-perforated panel (MPP) designs for sound absorption. that starts with target psychoacoustic metrics and identifies compatible micro-perforated panels (MPPs) designs for sound absorption applications. To enable this inverse design, global sensitivity analysis is used to quantify the relationship between MPP design parameters (perforation diameter, perforation rate, cavity depth, and panel thickness) and the resulting psychoacoustic metrics (loudness, sharpness, and tonality), from a series of noise stimuli according to ISO, ECMA, and DIN standards. A backward search is then proposed, incorporating just noticeable differences on the target psychoacoustic metrics, resulting in manufacturing tolerances on the design parameters. Optimized MPP designs based on psychoacoustic targets were manufactured and experimentally validated in an impedance tube setup following ISO and ASTM standards. Reflection coefficients were measured and converted to impulse responses for convolution with sound stimuli, enabling calculation of the above-mentioned metrics to verify the psychoacoustic optimization approach.

Speaker: Jiahua Zhang (Test Division, Siemens Industry Software NV)

A11.03 Physical modelling and simulations of musical instruments: S265

Conveners: Péter Rucz (Budapest University of Technology), Juliette Chabassier

233 Towards Co-Prediction of Steady-State and Transient Behaviour in Bowed-String Instruments: Initial Findings with the Thermal Elasto-Plastic Friction Model

The function of bowed string instruments has been a subject of research since the times of Helmholtz and Raman. However, the characteristics of the frictional interaction at the bow-string interface have evaded researchers for several decades. Taking the rapid temperature variations during stick-slip motion into account was deemed necessary in order to obtain an accurate model of bow-string interaction. The developed thermal friction models did offer an improvement in comparison to previous friction-curve models, but are still struggling to predict both transient and steady-state behaviour in comparison to experimental measurements. This presentation outlines a model that, by incorporating thermal effects to an elasto-plastic friction modelling framework, has managed to improve the above comparison [M. van Walstijn, V. Chatziioannou, A. Lampis, and E. Matusiak. A thermal elasto-plastic friction model for bowed-string simulation. Acta Acustica, 2026, doi:10.1051/aacus/2026042]. It offers a satisfactory qualitative match between experimentally measured and numerically simulated playability diagrams regarding both transient (Guettler diagrams) and steady-state (Schelleng diagrams) behaviour. Comparisons of individual waveforms obtained from measurements and simulations show that, while in many cases model predictions are very accurate, improvements are still forthcoming in the formulation of the thermal model and in the extraction of the friction model parameters. This points towards the need for experimental measurements that are capable of tracking rapid temperature variations at the bow-string interface.

Speaker: Vasileios Chatziioannou (Univ. Music and Performing Arts Vienna)

234 Non-iterative energy balanced scheme for vocal and reed instrument synthesis

An energy-balanced linearly implicit scheme for aclass of systems including nonlinear dissipation lawsis proposed. It is based on previous work on virtualanalog simulations and bowed string simulations, andallows an efficient update. Specifically, the scheme isgeared towards real-time synthesis of reed instrumentsand voice.In order to include conservative nonlinearities (derivedfrom a non-quadratic potential energy), the proposedmethod is built as an extension of an already existingscalar auxiliary variable scheme, handling the nonlin-earity in the conservative part of the dynamics.The scheme is framed in terms of the port-Hamiltonianframework. It is shown that the method satisfies adiscrete equivalent to a continuous-time power-balancelaw, with positive dissipated power, thus ensuringstability.Simplified models are used, focusing on the nonlinear-ities of interest. Simulation codes are available in apublic repository.

Speaker: Thomas Risse (STMS lab - IRCAM)

Full paper (PDF)

235 Computational Study of Sound Production in a Flue Instrument using Helmholtz Decomposition

A direct numerical solution of the Navier-Stokes equations was used to obtain the time dependence of the pressure and air velocity in and around a recorder as it produced a musical tone. A Helmholtz decomposition was then applied to extract the irrotational and solenoidal components of the velocity. The simulations also yielded the Lamb vector, allowing a comparison with predictions for the generation of acoustic energy based on Howe's energy corollary.

Speaker: Nicholas Giordano (Auburn University)

Full paper (PDF)

236 Numerical characterization of Cretan lute sound-hole rosettes

Rosettes covering the sound hole are characteristic of traditional Greek long-necked string instruments such as the Cretan lute. Although rosettes are known to influence the resonance of musical instrument cavity, their vibrational and airflow effects remain insufficiently studied. This work numerically examines 27 typical Cretan lute rosette geometries using finite element modal and frequency-response analyses combined with steady-state computational fluid dynamics. All rosettes were modeled with identical material properties and dimensions to isolate geometric effects. FEM eigenfrequencies were verified against physics-based models, presenting a good agreement per mode (MAPE < 5%). Clustering analysis showed that vibrational behavior was primarily governed by perforation rates and ligament patterns which strongly controls airflow. Overall, rosette geometries tend to act as independent tuning factors for soundboard behavior.

Speaker: Michail Evangelos Terzakis (Eindhoven University of Technology)

Full paper (PDF)

A12.09 Hybrid Modelling Approaches for Efficient and Accurate Simulations: S086

Conveners: Albert Prinn (International Audio Laboratories Erlangen), Jan Wouter Smits (University of Edinburgh)

237 A Neural Network Approach to Spherical Wave Reflections in the Image Source Method

The standard image source method approximates surface reflections using plane wave reflection coefficients, which can introduce errors at low frequencies. The complex image source method addresses this issue by using reflection coefficients derived from the Sommerfeld integral solution for a point source above a sound-absorbing plane of infinite extent. However, this approach is more computationally expensive than the standard image source method because the spherical wave reflection coefficient must be computed for each image source, which limits its practical applications. In this study, we propose a multilayer perceptron that models the reflection of spherical waves at the surfaces of a reverberant room. This surrogate model is trained on solutions to the Sommerfeld integral. In an image source implementation, for each image, the surrogate model receives the distance between the image and the receiver, and the angle of reflection. The model outputs a corresponding spherical wave reflection coefficient. The resulting modified image source method is more accurate than the standard and complex image source methods when compared to finite element solutions of the room transfer functions.

Speaker: Muyue Xi (International Audio Laboratories Erlangen)

238 MIMO Room Impulse Response Analysis and Synthesis through a Modal-Domain Approach: Application to a Reverberation Chamber

A computational pipeline for modal analysis and synthesis of room impulse responses (RIRs) is presented, with application to a reverberation chamber. Modal frequencies and damping coefficients are estimated from measured RIRs using a sub-band least-squares complex frequency-domain (LSCF) algorithm. Synthesis is performed as a direct modal sum in the time domain. The method extends previous single-channel work to a multiple-input multiple-output (MIMO) setting comprising 36 source-receiver pairs. Poles extracted independently for each measurement are consolidated into a common frequency basis via a voting procedure across spatial positions. Residual amplitudes are then obtained for each measurement by banded least-squares fitting. The resulting parametric model yields per-band $T_{60}$ statistics with inter-measurement uncertainty. Reconstructed RIRs are validated against measurements in both frequency and time domains.

Speaker: Giulia Fratoni (University of Bologna)

Full paper (PDF)

239 Towards a Hybrid Numerical Wave Method for Rooms with Extended Reaction Materials

Sound fields in rooms containing extended reaction materials can be predicted using numerical wave methods, such as finite difference or finite element methods. To describe the material's acoustic behaviour, an equivalent fluid model is often employed. This model typically includes complex-valued, frequency-dependent descriptions of the material's effective density and effective sound speed. The effective sound speed is typically lower than that of the surrounding air, and because mesh resolution depends on the wavelength of the propagating sound wave, the mesh in the fluid region needs to be much finer than in the air-filled region. This requirement for a refined mesh in the fluid region results in models that can be costly to solve. More efficient models may be achieved through hybridization of methods, in which the air-filled region is modelled using the finite element method while the equivalent fluid region is modelled using an alternative, more efficient method. In this study, a range of hybrid models is compared in terms of their efficiency and accuracy. A simple 2D test case is used to evaluate under which conditions a hybrid model outperforms a purely finite element model for achieving the same level of accuracy.

Speaker: Albert Prinn (International Audio Laboratories Erlangen)

240 Secondary source method for polygonal scattering objects with cavities

A tube ending in an infinite baffle is a classical radiation problem and has been studied extensively, with closed-form or semi-analytical results for radiation impedance, directivity, and end correction. In contrast, an unflanged tube introduces additional physics, such as edge diffraction, making analytical treatment more challenging. This problem is highly relevant for simulating musical instruments based on resonances in tubes.Here, secondary sources are used to couple the interior and exterior domains of a partially open shoebox. Green’s functions between all secondary-source pairs are needed for both domains under rigid-walled boundary conditions. For the interior, a modal solution is used, while the exterior domain is modeled with an edge-diffraction integral-equation formulation.A rectangular tube with one open end and inner dimensions 10 cm × 10 cm × 60 cm was investigated. Measurements were made in an anechoic chamber, with a microphone placed near the bottom corner inside the tube and a tweeter loudspeaker positioned 2 m from the opening. The ratio of microphone pressure amplitudes was measured with the box present and in an otherwise identical free-field setup. The same scenario was simulated in MATLAB using a modal sum for the inside domain and an edge-diffraction-based toolbox for the outside domain.Measurements and simulations agree for the first six resonance frequencies, with a maximum deviation of 3.5 %. For these modes, the largest Q-factor discrepancy occurs for the first resonance and decreases for higher modes, likely due to physical losses not fully captured in the simulation.

Speaker: Peter Svensson

Full paper (PDF)

241 JAX-BEM: Gradient-Based Acoustic Shape Optimisation via a Differentiable Boundary Element Method

Engineering structures are increasingly designed using numerical optimisation. However, traditional optimisation methods can be challenging with multiple objectives and many parameters. In machine learning, stable training of artificial neural networks with millions or billions of parameters is achieved using automatic differentiation frameworks such as JAX and Pytorch. Because these frameworks provide accelerated numerical linear algebra with automatic gradient tracking, they also enable differentiable implementations of numerical methods to be built. This facilitates faster gradient-based optimisation of geometry and materials, as well as solution of inverse problems. We demonstrate JAX-BEM, a differentiable Boundary Element Method (BEM) solver, showing that it matches the error of existing BEM codes for a benchmark problem and enables gradient-based geometry optimisation. Although the demonstrated examples are for acoustic simulations, the concept could be readily extended to electromagnetic waves.

Speaker: James Hipperson (Acoustics Research Centre)

Full paper (PDF)

A14.03 Hidden hearing loss & synaptopathy: Diagnostics, treatment & functional consequences: S174

Conveners: Lukas Rüttiger (University of Tübingen), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

242 Electrophysiological markers of auditory nerve integrity in humans: insights from recordings and computational modeling

Age-related auditory nerve (AN) degeneration has been observed in post-mortem human and animal studies and proposed to underlie suprathreshold perceptual difficulties despite normal audiometric thresholds. However, this link remains unconfirmed in vivo as reliable electrophysiological markers of AN integrity, independent of hair-cell status, are currently lacking.Brainstem-generated responses to pure tones – frequency-following responses (FFR) – are reduced in older normal-hearing listeners, though whether this reflects central or peripheral degeneration has been debated. To address this question directly, we simultaneously recorded peripheral electrocochleographic responses and brainstem FFRs in young and older normal-hearing listeners. Both measures were reduced in the older group, supporting a peripheral origin. In contrast, although compound action potential (CAP) N1 amplitudes were reduced at the periphery, auditory brainstem response (ABR) wave V amplitudes were preserved, consistent with central gain compensation.Computational modeling confirmed the mechanistic basis of this differential sensitivity, showing that suprathreshold responses to periodic stimulation are sensitive to AN degeneration while remaining robust to hair-cell damage. The modeling further indicated that pure tones and amplitude-modulated (AM) tones excite partly distinct AN fiber populations, suggesting that these stimuli could, in combination, probe apical versus basal neural integrity.Building on these findings, we developed a multi-stimulus paradigm presenting pure and AM tones separately and in combination with normal-hearing listeners. Parallel simulations in healthy and synaptopathic cochleae are used to predict the experimental observations and to assess whether combined stimulation can improve diagnostic sensitivity while reducing test time. Together, these findings represent a step toward a model-informed electrophysiological assessment of cochlear nerve integrity in humans.

Speaker: Miguel Temboury-Gutierrez (Hearing Systems section, Technical University of Denmark)

243 Age-Related Declines in Subcortical Sound Coding Are Not Associated with Perceptual Discrimination Deficits

Age-related cochlear synaptopathy (CS) degrades neural coding of sound, as reflected in envelope and frequency following responses (EFR/FFR). However, its perceptual consequences remain unclear. This study examined whether age-related declines in subcortical neural sound representation are reflected in deficits in perceptual discrimination of temporal and spectral features. Two groups of young (yNH; n = 20) and older normal-hearing adults (oNH; n = 20) completed three perceptual tasks in quiet and noise: (1) formant-frequency and (2) formant-bandwidth discrimination tasks, both targeting sensitivity to spectral information at low frequencies; and (3) a temporal envelope discrimination task assessing sensitivity to the shape of temporal envelopes at high frequencies. Stimuli consisted of single-formant harmonic complexes (113-Hz fundamental) and amplitude-modulated tones (4-kHz carrier, 113-Hz modulation). FFRs were recorded using dynamic versions of these stimuli, with parameters varying continuously over 1 second. The envelope component of the FFR was reduced in oNH compared to yNH, whereas the temporal fine-structure component was preserved. However, perceptual performance did not differ significantly between groups in any task, either in quiet or in noise, and did not correlate with global FFR metrics. This dissociation suggests that age-related declines in subcortical neural encoding are not directly reflected in basic perceptual discrimination. These findings are consistent with the possibility that the perceptual impact of CS is subtle or mitigated by central compensatory mechanisms. Our study provides a detailed characterization of how neural representations of dynamic sounds are reshaped with age, which is critical for refining computational models of CS.

Speaker: David López-Ramos (STMS (Ircam-CNRS-SU))

244 Neurophysiological and Behavioral Consequences of Cochlear Synaptopathy in Chinchillas

Moderate-level noise exposure and/or aging can eliminate cochlear synapses without permanently damaging hair cells or elevating thresholds in animals. Cochlear synaptopathy (CS) has been hypothesized to contribute to human perceptual difficulties in noise even with normal audiograms. However, testing this hypothesis is difficult because 1) ethical limits exist in measuring human synaptopathy directly, and 2) synaptopathy has been most completely characterized in rodents for which behavioral measures at speech frequencies are challenging. We established a relevant mammalian behavioral model by showing that chinchillas have corresponding neural and behavioral amplitude-modulation (AM) detection thresholds comparable to humans. Furthermore, immunofluorescence histology confirmed synaptopathy occurs in chinchillas across a broad frequency range, including speech frequencies, following a lower-frequency noise exposure that avoids permanent ABR-threshold and DPOAE-amplitude changes. Auditory-nerve-fiber responses showed that low-SR fibers were reduced in percentage (but not eliminated) following noise exposure, as in guinea pigs. Wideband middle‐ear muscle-reflexes in awake chinchillas showed large and consistent reductions in suprathreshold amplitudes following noise exposure, whereas suprathreshold ABR wave-1 amplitude reductions were less consistent, in line with findings from parallel studies in middle-aged and noise-exposed humans. Behavioral assays of tonal-carrier AM detection in chinchillas before and after noise-induced TTS found subtle, but significant modulation-frequency dependent performance effects, suggesting limited impact of CS on this relatively simple task. More complex stimuli that better challenge population neural coding may be required to show similar effects as experienced by humans in noisy environments. These anatomical, physiological, and behavioral data illustrate a valuable animal model for linking physiological and perceptual effects of hearing loss.

Speaker: Michael Heinz (Purdue University)

245 Neurophysiological underpinnings of perceptual changes with cochlear synaptopathy: insight from animal models

Cochlear synaptopathy is a common pathology involving damage to and loss of auditory-nerve synapses with inner hair cells. Cochlear synaptopathy is undetectable with the clinical audiogram and widely thought to cause hearing-in-noise problems known as hidden hearing loss. Animal models are advantageous for studying links between cochlear synaptopathy and hidden hearing loss because synaptic injury can be induced and readily quantified. We induced excitotoxic cochlear synaptopathy in budgerigars, a parakeet, using intracochlear kainic-acid infusions. Budgerigars were selected based on extensive prior research showing human-like perception of complex sounds. Effects of synaptopathy on auditory perception and neural processing were evaluated with operant-conditioning behavioral experiments and single-unit auditory-nerve recordings, respectively. Many aspects of auditory perception were normal following kainic-acid infusions including the audiogram, temporal integration, and sensitivity to differences in envelope statistics. Behavioral impairment was observed for detection of tone pips at the onset of a simultaneous noise burst. The result suggests selective impairment of acoustic onset-cue perception. Single-unit auditory-nerve recordings showed that perceptual changes were associated with increased spontaneous activity and onset excitability of surviving fibers. Increased excitability of auditory-nerve responses following synaptic injury may contribute to impaired onset-cue perception with cochlear synaptopathy.

Speaker: Kenneth Henry (University of Rochester)

246 Development of an Awake Animal Model for Hyperacusis Screening

Hyperacusis affects 2-15% of the population, causing painful sensitivity to everyday sounds, yet its pathophysiology remains obscure due to limited animal models and frequent tinnitus comorbidity. This study aimed to identify neural biomarkers and behavioral changes reliably characterizing hyperacusis.Mice were exposed to acoustic trauma (2h, 95 dB SPL) inducing temporary hearing loss. Behavioral and electrophysiological assessments were conducted in awake animals pre- and post-trauma. Noise exposure increased startle reflex amplitude (hyperacusis correlate) and decreased gap inhibition (tinnitus correlate) in subsets of animals. Phenotypic scales revealed positive correlations between hyperacusis severity and evoked response amplitude in the inferior colliculus and auditory cortex 5-10 weeks post-exposure. No such correlation existed for tinnitus, and spontaneous EEG remained unchanged. Crucially, startle reflex amplitude correlated with central auditory evoked responses despite recovered peripheral thresholds.We successfully identified a hyperacusis phenotype based on simultaneous behavioral and neural response enhancements. These findings elucidate pathophysiological mechanisms impacting the central auditory system following noise overexposure, independent of permanent hearing loss, paving the way for targeted treatments.

Speaker: Boris Gourevitch (Institut de l'Audition, Institut Pasteur)

A14.04 Unraveling the auditory periphery with physiological measures: S260

Conveners: Gerard Encina-Llamas (University of Vic - Central University of Catalonia), Bastian Epp

247 Functional Characterization of Human Auditory Nerve

Researchers and clinicians measured the gross action potentials from the round window or the scalp to investigate the physiological state of the auditory nerve. However, the generators underlying these responses are not fully understood. By combining intraoperative auditory nerve recordings with computational modeling, we show that the unitary action potential recorded directly on the nerve is markedly shorter than the responses measured at the round window or tympanic membrane. This likely reflects the heavier myelination of auditory nerve fibers (ANFs) near their exit from the internal auditory meatus, which shortens the extracellular action potential waveform and produces a spectral peak around 1300 Hz. Our findings are consistent with previous work demonstrating that sound activates a broad population of ANFs, whereas the compound action potential (CAP) reflects only a small, highly synchronized subset of these fibers. The onset response is dominated by high CF fibers—even when the stimulus frequency is lower—because basal fibers respond first and with the greatest synchrony. In contrast, offset responses arise primarily from fibers with lower characteristic frequencies, confirming that they are not simply the mirror image of onset responses. We also show that the human auditory nerve exhibits robust neurophonic phase locking to acoustic frequencies up to at least 2–3 kHz. Because the simulated single fiber action potentials closely match the simulated neurophonic responses, the human auditory nerve neurophonic appears to be a relevant physiological marker for defining the upper limits of phase locking in the human auditory nerve.

Speaker: Jérôme Bourien (University of Montpellier)

248 Renewing the classics: amplitude- and latency-growth functions of human CAP and ABR to improve auditory nerve models towards precision audiology

Amplitude- and latency-growth functions of auditory evoked potentials are classical measures of peripheral auditory function, yet existing datasets suffer from insufficient age control, small sample sizes, and insufficiently detailed stimulus calibration — all of which limit their precision for validating computational models. This study provides a rigorously controlled reference dataset by simultaneously recording compound action potentials (CAPs) using tympanic membrane electrodes and auditory brainstem responses (ABRs) in twenty young, normal-hearing adults across a wide stimulus intensity range, using both clicks and level-specific chirps. Chirp stimuli produced consistently larger amplitudes and improved peak detectability compared to clicks, particularly at low-to-moderate intensities, while ABR wave-V latencies remained stable across intensities as intended by design. Robust, level-dependent amplitude- and latency-growth functions were obtained for both cochlear and brainstem responses, with notably lower inter-subject variability in latency than previously reported. Interestingly, ABR wave V can be well approximated as a linear transform of CAP N1, supporting the validity of convolution-based frameworks for modelling retrocochlear signal processing in normal-hearing ears. Individualized unitary responses derived from the data reproduced key waveform features, but systematically underestimated latency slopes — pointing to specific limitations in current auditory nerve model front ends that this dataset can help address. These results feed directly into ongoing work on the auditory evoked potential modelling framework developed by Temboury-Gutierrez (also presented at FA 2026), and can be related to the auditory phenotyping (AudPhen-DK) project (presented at FA 2026 by Encina-Llamas), which aim to develop tools towards precision audiological diagnostics.

Speaker: Gerard Encina-Llamas (University of Vic - Central University of Catalonia)

249 Multiway Canonical Correlation Analysis for Enhanced Signal Quality in Subcortical EEG Recordings

Frequency-following responses (FFR) may be sensitive to auditory nerve degeneration and auditory processing deficits beyond what is captured by the clinical audiogram. The clinical potential of this measure relies on robust, quick and reliable acquisition in the clinic. Multiway canonical component analysis (MCCA) presents a way of utilizing common response patterns across multiple channels and subjects to boost the fidelity and reliability of such responses. Here, we demonstrate the potential of MCCA to substantially improve the signal-to-noise ratio (SNR) of FFRs in a large dataset of EEG recordings (16+2 channels) obtained from normal-hearing adults (n = 110) in response to 326 Hz pure tones. When applied as a denoising method, MCCA increased the proportion of participants with significant FFRs by 17%, allowing for robust responses in 97% of the sample. Additionally, acquisition time was reduced, with FFRs of equivalent significance using only 20% of the full dataset (approximately 2 minutes of recording time), consistent with clinical constraints. Potential overfitting was addressed through cross-validation using unseen EEG noise. These results demonstrate that MCCA can enhance the reliability, efficiency, and clinical feasibility of FFR measurements by aligning patient data to a multi-subject reference space.

Speaker: Jonatan Märcher-Rørsted (Technical University of Denmark)

A15.01 Spatial Hearing: Modeling and Applications: S296

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Ville Pulkki (Aalto University)

250 Auditory-Model Predictions of Quality of Experience in Immersive Music Concerts

Immersive live music productions promise three key benefits for audiences: audio‑visual consistency (you hear sound where you see it), improved source separation and intelligibility through spatial unmasking, and new creative possibilities for mixing engineers. Yet it remains unclear how to assess whether a given venue is suitable for this type of production, based on its measurable acoustic characteristics. Using virtual acoustics and binaural auditory models, we analyze how the presence of diffuse reverberation, specified by a Reverberation Time (RT), direct-to-diffuse ratio (DRR) and Initial Time Delay Gap (ITDG) govern three perceptual predictors of Quality of Experience: source‑localization accuracy, localization blur, and binaural advantage. In our simulation results, we observed that the DRR was the primary determinant of all three predictors, while the RT played a secondary role. ITDG primarily reduced localization blur, with minor effects on accuracy and spatial unmasking. Its benefits were most apparent at moderate DRRs. These results help define what makes a venue good or poor for immersive live events from a QoE standpoint and provide actionable criteria to secure the “hear‑what‑you‑see” promise and enhanced spatial unmasking in production.

Speaker: Nicolas Epain (L-Acoustics)

Full paper (PDF)

251 Comparison of binaural models for the localization of virtual sound images using a novel two-listener virtual imaging system

Virtual sound imaging systems, such as those which work on the principle of crosstalk cancellation (CTC), are often evaluated by means of localization accuracy. For crosstalk cancellation systems, high localization accuracy requires that the actual listener position is the same as the target listener position. In this paper, the inputs to a binaural localization model are varied based on the simulated effect of deviation from target position on a CTC system. The output of the model is compared to localization results from a listening test using a CTC system, revealing that the model is capable of predicting various features of listener responses, including greater variance in prediction for more lateral angles, and asymmetry of variance in left compared to right target azimuths.

Speaker: Isaac Lambert (University of Southampton)

Full paper (PDF)

252 Influence of Reverberation on the Horizontal Localization of Sound Sources Auralized in an Audiovisual Immersive Room

Audiovisual immersive rooms represent emerging virtualization spaces that address the individual isolation inherent to head-tracked binaural Virtual Reality (VR). However, auralizing highly reverberant environments remains a challenge, as reverberant energy degrades acoustic cues essential for human spatial localization. This study evaluates sound source localization in reverberant fields within a 16.2.10 audiovisual immersive room, comparing results to traditional head-tracked binaural VR. A 4-Alternative Forced Choice (4-AFC) horizontal localization task was conducted with 26 participants. Stimuli were auralized using pink noise bursts and Third-Order Ambisonics impulse responses generated in ODEON for the same reverberation room, with absorption progressively increased to produce reverberation conditions ranging from highly reverberant to nearly anechoic. As expected, the results demonstrate that localization accuracy decreased with increasing reverberation. However, across all conditions, the immersive room achieved higher accuracy and lower task response latency, exhibiting fewer localization confusions and lateral mislocalizations than binaural VR.

Speaker: Jesus Vaquerizo-Serrano (La Salle, Universitat Ramón Llull)

Full paper (PDF)

253 Modeling Binaural Speech Intelligibility in Symmetrical Listening Conditions using Envelope Fluctuation Cues

Modeling speech intelligibility in spatially separated versus co-located masker conditions can reveal how listeners use spatial cues, such as interaural time and level differences, to separate speech from noise. Symmetric listening conditions, with maskers symmetrically positioned around a frontal target are particularly relevant because they approximate realistic situations with interfering sources evenly distributed around the listener and therefore do not provide a long-term signal-to-noise advantage at either ear. Binaural benefits arise solely from binaural unmasking (BU) and short-term better-ear (BE) listening.We used a modified binaural speech-based envelope power spectrum model to evaluate speech reception thresholds (SRT) and spatial release from masking (SRM) in symmetric two-masker conditions. The model analyzes envelope fluctuations of noise and noisy speech in short-time windows using three parallel processing pathways: two monaural pathways – one per ear – to simulate BE listening, and one binaural pathway that integrates information across ears via an equalization-cancellation process to simulate BU.Model predictions were compared with behavioral data across different spectro-temporal maskers in anechoic and reverberant symmetrical listening conditions. The model captured masker-dependent SRM effects for non-speech maskers in both anechoic and reverberant conditions. For speech-like maskers, it underpredicted SRTs and SRM in anechoic conditions and failed to capture SRM effects in reverberation, despite reproducing the SRT trends. We analyzed independent BE and BU simulations, evaluated their contributions, and compared the model’s predictions to those of other binaural models. Our findings underscore the importance of refining binaural models to account for realistic listening conditions, including reverberation and complex maskers, thereby improving our understanding of auditory speech processing in real-world acoustic environments.

Speaker: Cathrina Veigel (Technical University of Denmark)

A20.00 Speech: S288

Conveners: Philipp Aichinger (Medical University of Vienna), Barbara Schuppler (Signal Processing and Speech Communication Laboratory), Oliver Niebuhr (University of Southern Denmark), Franz Pernkopf (Signal Processing and Speech Communication Laboratory), Peter Balazs (Acoustics Research Institute, Austrian Academy of Sciences)

254 Spatial Release from Making in Turkish Speech Intelligibility

This study investigated the effects of Spatial Release from Masking (SRM) on speech intelligibility in normal-hearing native Turkish speakers using the Turkish Modified Rhyme Test (MRT). Thirty participants (15 females, 15 males; mean age: 23.0 years, SD = 2.3) completed the MRT under nine spatial conditions defined by the combination of target word azimuth angle (0°, -50°, +50°) and speech-shaped noise (SSN) azimuth angle (0°, -50°, +50°) at a fixed 0 dB signal-to-noise ratio (SNR). Accuracy and reaction time of the participants were also recorded. A two-way repeated-measures ANOVA revealed a significant main effect of word angle (p = .016) but not SSN angle (p = .206). No significant effect of spatial separation between target and masker was observed. Among the five distinctive features, graveness yielded the lowest accuracy and sibilation the highest (p = .022), with no significant interaction between distinctive features and SSN angle. The data establish a normative baseline for Turkish MRT and draw attention to spatial configuration and distinctive features as variables that deserve consideration in multilingual speech intelligibility research.

Speaker: Kivanc Kitapci (TOBB ETU)

Full paper (PDF)

255 How Clear Is Clear Speech? Effects on Speech Intelligibility, Speaking Effort, and Listening Effort During Conversations

When conducting conversations in adverse acoustic conditions (e.g., in background noise), talkers tend to alter their speech production to increase speech clarity for their interlocutors. Compared to casual speech, such “clear speech” has been reported to facilitate speech intelligibility in challenging listening situations, but this evidence has been obtained through passive listening tests where listeners attended to stimuli that did not meet the requirements of standard speech-intelligibility test materials. Furthermore, the physical and cognitive effort required by the talker to produce clear speech, as well as the cognitive listening effort required by the listener to attend it, are currently underexplored. This study introduces a novel conversation-based testing paradigm combining measures of speech intelligibility, speaking effort and listening effort. Pairs of interlocutors conducted a conversation elicited through a visual task, during which the participants were alternatingly required to read sentences from standard speech-intelligibility test material and their interlocutor had to repeat them, allowing to extract a measure of speech intelligibility from a conversational setting. During the conversation, a set of voice features were measured to quantify the physical component of speaking effort, while pupil dilation measured during speaking and listening activity provided proxy measures of cognitive speaking and listening effort, respectively. Through separate acoustic environments, an independent manipulation of noise exposure allowed to disentangle how speaking and listening effort were affected by the talker and listener acoustic condition. The experimental method will be presented and discussed in terms of the outcome measures of speech intelligibility, voice features, and pupil dilation.

Speaker: Paolo Mesiano (Eriksholm Research Centre)

256 Spectral, temporal, and intensity patterns in sibilants associated with malocclusion

Malocclusion impacts articulation of speech sounds, in particular fricatives. The present study compares spectral and prosodic patterns between adults with and without jaw disharmony during production of the English sibilants /s/ and /ʃ/. 36 adult participants were recruited and categorized using the Angle Classification into control (Class I) and experimental (Class II and Class III) groups. Speech recordings were obtained using an iPhone in a clinical setting. Speech recordings were analyzed for first and second spectral moment (i.e., center of gravity (COG), and spectral dispersion), duration and intensity of the English sibilants /s/ and /ʃ/ using a linear mixed effects model. Significant differences in each acoustic variable were found between the malocclusion classes including prominent interactions with sex. Class III malocclusion exhibited more prominent acoustic spectral distortions followed by Class II. Class III females had the highest COG for /s/ and /ʃ/ along with longer duration. Class III males showed reduced acoustic distinction between the two sounds along with intensity changes. Compensatory speech strategies involving duration and intensity changes might be learned to mitigate the impact of malocclusion on speech sound distinctions. These findings underscore intricate relationships between jaw alignment and speech articulation.

Speaker: Daniel Aalto (University of Alberta)

Full paper (PDF)

257 Transmission model of aircraft audio chain to generate non-sensitive CVR-like speech

Aircraft Cockpit Voice Recorders (CVRs) contain sensitive data which are incompatible with open database to develop specific speech processing tools. The access to aircraft is also difficult and expensive. For these reasons, no realistic CVR database exists to be shared with the scientific community to train their algorithms. Acoustic and electronic measurement campaigns were conducted in aircraft cockpits and manufacturer audio test benches to identify impulse responses of the transmission chain from the pilot's voice to the CVR audio files, as well as recordings of flight noises at different flight phases. This led to a model to transform raw speech into CVR-like speech, shared to the community via a Python program and open-source database of 1) impulse responses of cockpits, aircraft microphones, aeronautical headset microphones, aircraft audio processors and CVRs, 2) in-flight noises from different flight phases and 3) raw speech corpus typical of cockpit conversations to demonstrate our model. The corpus generated from this model will support the development and evaluation of speech processing tools tailored to aviation safety investigations.

Speaker: Lionel Feugère (Bureau d'Enquêtes et d'Analyses (BEA))

Full paper (PDF)

258 Could current spread and device coding explain vocal-tract length perception difficulty in cochlear implant users?

Previous studies have demonstrated that vocal tract length (VTL) perception is strongly limited in most cochlear-implant users, with post-lingually deaf adults showing greater deficits than those implanted in early childhood. Three potential factors may underlie this difficulty, as well as this difference between groups: (1) the implant’s coding strategy, which may distort VTL cues as it is not designed to directly encode them; (2) current spread between the stimulating electrode and the activated neural terminals, which reduces the effective spectral resolution and may disproportionately affect VTL perception, as it relies on fine spectral contrasts; and (3) the listener’s ability to relearn associations between acoustic cues and speaker size, influenced by their age at implantation.To test the first two hypotheses, we recorded the electrical outputs of a Cochlear Nucleus 7 processor using the MP3000 strategy with a standard map, and resonified them using a noise vocoder simulating different degrees of current spread. By comparing normal-hearing participants’ VTL discrimination performance with the original and vocoded stimuli, we isolated the contributions of the coding strategy alone and the electrode-neuron interface. The results indicate that the implant coding strategy alone can account for the VTL discrimination thresholds observed in paediatric implant users. However, realistic levels of current spread substantially degraded performances, bringing them to the level of adult cochlear-implant users.

Speaker: Etienne Gaudrain (CRNL, CNRS UMR5292, Inserm U1028, Université Lyon 1)

A03.02 Measurements and experimental analysis in building acoustics: S309

Conveners: Martin Schneider (Hochschule für Technik Stuttgart), Selina Vavrik-Kirchsteiger

259 On the use of dodecahedral sound sources for the field measurement of façade sound insulation according to the ISO 16283-3:2016 standard

Façade sound insulation is a key feature to reduce sound immission in different kinds of buildings, ranging from standard dwellings to spaces dedicated to entertainment like music clubs, among other activities. In the last decade, since the publication of the standard ISO 16283-3:2016, an important claim has been highlighted: a dodecahedral sound source, as usually employed to measure airborne sound insulation according to the ISO 16283-1 standard, can be used to generate the excitation in the measurement of a whole façade (global method). However, the employment of such source introduces a new degree of freedom concerning the influence of its directivity pattern in high frequencies. It is hence expected that these directional properties could, for example, make us derive different sound insulation values only rotating the source on its axis, keeping its position constant with respect to that of the external microphone. In this paper, to corroborate this hypothesis, we present the results and conclusions arose from an experiment specially designed to evaluate the influence of the rotation of a dodecahedral source in the sound insulation, in a scenario where all the other parameters remain unmodified. The conclusions of our study indicate that, although the influence of the spinning of the source is moderated in terms of Single-Number Quantities (i.e., Dls2m,nT,w), the frequency-dependent sound insulation shows an important variation with respect to the rotation of the source, mostly in frequency bands around 2 kHz. These findings would suggest a limitation of the method if additional conditions are not incorporated.

Speaker: Luis Antonio Azpicueta-Ruiz (Universidad Carlos III de Madrid)

260 Influence of Fiber Diameter measurement techniques on Airflow Resistivity predicted by the Bies–Hansen Model

In acoustics, the Bies-Hansen model is a well-established empirical correlation for estimating the airflow resistivity (AFR), supporting the design and evaluation of fibrous acoustic materials such as mineral wool. It provides a mathematical relationship between the physical properties of material—including density and fibre diameter—to AFR. The model’s empirical constants were originally derived for fiberglass materials with negligible binder content, under the assumption of uniform fiber diameters below 15 µm. In the original correlation, fibre diameter was obtained by optical microscopy and represented by a mean value used to fit the constants. In practice, multiple techniques are available for fibre diameter determination, each based on different measurement principles and data interpretation. Moreover, it is questionable if the mean value fully captures the complexity of non-uniform fiber diameter distribution typical for fibrous materials.This study compares several fiber diameter measurement approaches and assesses the agreement between measured AFR and Bies–Hansen predictions when each method’s diameter is used in the model. It outlines how measurement technique selection affects the derived fiber diameter and therefore the predicted AFR.The study highlights that variations in fiber diameter determination methods affect AFR predictions and underscores the need to recalculate the original empirical constants for better alignment with measured values. Recalibration improves model fidelity across materials with different properties and non-uniform diameters, enabling more reliable AFR prediction beyond the original scope of the Bies–Hansen correlation.

Speaker: Urska Kovacic (Knauf Insulation d.o.o.)

Full paper (PDF)

261 Floor Impact Sound Reduction Structures for Multi-Family Housing Renovation

The residential remodeling market in South Korea is experiencing a compound annual growth rate of 5.4%, yet antiquated multi-family dwellings often feature slender concrete slabs (120–180 ㎜) that restrict the thickness and loading of retrofit systems. This study develops a wet-type floating floor system by jointly optimizing impact sound insulation and the long-term dimensional stability of resilient layers. Structural stability was evaluated via residual deformation (KS F 2873) and time-dependent deflection monitored using an 18 kg steel load plate (approximately 1.96 kPa), where displacement stabilized within 144 hours with deflection below 0.2 ㎜. To suppress low-frequency resonance-induced amplification around 50 Hz, a design target for dynamic stiffness (KS F 2868) not exceeding 5 MN/㎥ was adopted while maintaining residual deformation within 3 ㎜. The proposed multi-layered system—consisting of embossed ethylene vinyl acetate (22–25 ㎜), a sound-absorbing interlayer (12–15 ㎜), and a 3 ㎜ polypropylene board for load distribution—was validated through field tests on a 150 ㎜ thin slab. The optimized configuration achieved standardized impact sound levels of 38 dB for light impact and 47 dB for heavy impact, successfully meeting the performance targets of Grade 2 and Grade 4, respectively. These results validate that the developed system provides superior acoustic comfort and structural durability, offering a robust engineering solution for environmental noise mitigation in renovated dwellings. Future investigations will explore point-support mechanisms and long-term creep tests exceeding one year to further enhance low-frequency damping performance.

Speaker: Won-Hak Lee (Korea Conformity of Lobratories)

Full paper (PDF)

262 Comparison of bending wave speed of viscoelastic panels using two approaches

This work employs a laser Doppler vibrometer (LDV) as a sensing modality for experimental measurements of bending wave properties in viscoelastic panels. A recently developed transfer function method is applied to characterize these properties, with a focus on experimentally determining the bending wave propagation coefficient—a key quantity from which other bending wave properties can be derived. The LDV measures the bending velocities at two surface points on the panel, enabling the determination of the propagation coefficient as a primary parameter. From this, other properties such as the bending phase speed, Young’s modulus, and the loss factor are obtained. Two approaches for determining the phase speed are examined: one based on the propagation coefficient, and another that calculates the phase spectrum of the transfer function. Following a brief introduction of the theoretical foundation, this paper discusses validation efforts for the experimentally measured bending phase speed using both approaches, with the ultimate goal of deriving additional parameters from propagation coefficient measurements.

Speaker: Ning Xiang (Rensselaer Polytechnic Institute)

Full paper (PDF)

A16.04/A24.06 Acoustics for learning environments across physical and virtual realities: S329

Conveners: Arianna Astolfi (Politecnico di Torino), Janina Fels (IHTA, RWTH Aachen University)

263 Real vs. Virtual Speech Intelligibility in a Lecture Room

Improving acoustic quality in learning environments is important for students of different ages, especially vulnerable groups, such as hearing aid users. When field assessments are not feasible, virtual reality (VR) provides a valuable alternative for controlled laboratory-based subjective evaluations. In this study, thirteen normal-hearing participants performed speech intelligibility tests in a university lecture room at Politecnico di Torino and repeated the same tests about one month later in a VR system based on Fifth-Order Ambisonics (5OA). The aim was to perceptually validate the VR system, previously assessed using objective measures. Five audio-visual scenarios were selected for systematic comparison, varying in target speech azimuth, distance from the listener, and the presence or absence of frontally localised masking noise. These conditions enabled the evaluation of the spatial performance of the VR system in relation to speech intelligibility. The results showed that spatial performance was preserved; although a small difference in Speech Reception Thresholds (SRTs) of 1 dB was statistically significant, it was likely not perceptible. This suggests that the VR system can realistically approximate real-world listening conditions with good perceptual accuracy.

Speaker: Arianna Astolfi (Politecnico di Torino)

Full paper (PDF)

264 Spatial decay of speech has an important role in open and enclosed learning spaces

Open learning spaces, OLSs, which enable teaching more than one group of students at a time, have become increasingly popular. Enclosed learning space, ELS, (conventional classroom) is still the most popular learning space type. Our purpose was to report on the room acoustic properties of both learning space types using measurements according to ISO 3382-3. This method was justified since Finnish regulations set distance-dependent target values for STI in OLS and distance-independent target values for ELSs. Measurements concerned speech transmission index, STI, background noise level, LAeq, and reverberation time, T. Variation in results in both OLSs and ELSs was rather large. In OLSs STI was 0.47–0.91, LAeq 29–44 dB, and T 0.44–0.72 s. In ELSs STI was 0.64–0.83, LAeq 25–47 dB, and T 0.34–0.82 s. The results of LAeq and T in OLSs and ELSs were surprisingly similar. STI values were within target values in most ELSs, unlike in most OLSs, where the target values were violated at longer distances, because STI did not decrease as steeply as required. Our finding suggests that measurement of spatial decay is justified in both learning space types because it reveals how STI behaves at different distances from the speaker.

Speaker: Jukka Keränen (Turku University of Applied Sciences)

Full paper (PDF)

265 The relationship between listening effort and speakers' comfort in quiet and noisy teaching environments: an experimental study

Room acoustics play a central role in speech communication by influencing both speaker behavior, speech intelligibility and listening effort. While the effects of acoustics and background noise on listening effort are well established, the relationship between speaker comfort and listening effort remains largely unexplored. This study examines how room acoustic conditions, background noise, and sound field amplification (SFA) influence listening effort, with a specific focus on its relationship to speaker’s comfort and listener motivation.An experimental within-subject design was conducted during an acoustic refurbishment of a university lecture hall. Data were collected across four stages (pre- and post-intervention) with 17 listeners and 15 speakers. Self-reported measures of listening effort, speaker comfort, and motivation were obtained under controlled conditions with and without background noise, and with and without SFA.The results showed that improved room acoustics significantly reduced listening effort, confirming the importance of acoustic treatment. SFA further decreased listening effort in background noise conditions. Importantly, the relationship between speaker comfort and listening effort was weak and context-dependent: no significant association was found overall, although a small positive correlation emerged in quiet conditions. This suggests that increased speaker comfort does not necessarily translate into reduced listening effort, and that the two may be influenced differently by acoustic parameters.Listener motivation was negatively associated with listening effort, indicating an interaction between cognitive and acoustic factors.These findings highlight that in order to optimize classroom acoustics, both speaker comfort and listening effort should be considered to achieve effective communication environments

Speaker: Viveka Lyberg-Aahlander (Abo Akademi University)

266 Effects of room acoustics and background noise on vocal behavior in speakers with and without voice problems

Speakers continuously adapt their voice to the acoustic environment, and prolonged speaking under adverse conditions may increase vocal load and contribute to voice problems. Classrooms constitute acoustically demanding environments, where both background noise and room acoustics influence vocal behavior through mechanisms such as the Lombard effect and altered auditory feedback. While these factors have been studied separately, their combined influence, and potential differences between speakers with and without voice problems, remain insufficiently understood.This study investigated the effects of room acoustics and background noise on vocal behavior, perceived vocal effort, and voice quality in a classroom-like environment. Nine female speakers, with and without self-reported voice problems, performed speaking tasks under four acoustic conditions and three noise levels (≈28, 65, and 75 dBA). The acoustic conditions differed in reverberation time (T20), clarity (C50), sound strength (G), voice support ($ST_V$), and mouth-to-ear decay time (DT40). Vocal sound pressure level (SPL), fundamental frequency (f0), perceptual voice quality, and self-assessments were analyzed using linear mixed-effects models.Background noise was the dominant factor, with increasing noise leading to higher SPL and $f_0$, increased dysphonia, and reduced speaker comfort. Speakers with voice problems produced lower SPL but showed reduced comfort and altered voice quality. Room acoustics showed limited main effects but moderated noise-induced vocal changes. Improved acoustic conditions were associated with increased comfort and audibility, although these effects diminished at higher noise levels.The findings indicate that background noise is the primary driver of vocal adaptation, while room acoustics exert more subtle, interaction-dependent effects. Differences between speaker groups suggest distinct vocal strategies in acoustically challenging environments.

Speaker: Viveka Lyberg-Aahlander (Abo Akademi University)

A17.07 Source directivity and sound source identification: S396

Conveners: Matthieu Hartenstein (L2S, CentraleSupélec), Samuel D. Bellows (University of Utah Asia Campus)

267 Optimal choice of parameters for the identification of the spherical wave decomposition of the field radiated by a sound source

When available, the spherical wave decomposition (SWD) of a sound source can be used to reconstruct the acoustic field radiated by this source. SWD coefficients are usually determined by fitting distributed pressure measurements taken around the source using a regularized least-squares inversion. The performance of this method critically depends on the adequate tuning of the SWD truncation order, and of the regularization parameter. The smoothness of the radiated field causes high order SWD coefficients to decay. Based on these considerations, early studies on sound radiation proposed truncating the SWD at order kd (the kd rule), where k is the acoustic wavelength and d is the extent of the source. However, in the case of complex radiation phenomena, defining the extent of the source is difficult. Furthermore, at higher frequency, the kd rule yields excessively high truncation orders.In this study, we propose a cross-validation-based framework in order to optimally define the SWD estimation parameters. Numerical simulations show that the regularization parameter can be sought for prior to the SWD truncation order, saving computational time. The optimal truncation orders obtained using both simulated and experimental data show good agreement with the kd rule at lower frequency, and comply with the order limitation of the array used for the measurements at higher frequency.

Speaker: Matthieu Hartenstein (L2S, CentraleSupélec)

268 Dynamic Voice Directivity Prediction from Facial Geometry via Feed-Forward Neural Networks

Quantifying voice radiation is essential for applications such as telecommunications, room acoustic modeling, and virtual reality. However, the precise directional characteristics of speech can be complex due to phoneme-dependent variations and movements of the head and torso. This work proposes a feed-forward neural network to predict dynamic voice directivity patterns in the horizontal plane based on extracted facial features from a simultaneous video capture of a singer’s face. A static time and phoneme-averaged directivity pattern serves as the baseline for evaluating the model’s performance. While the model improves predictions across the majority of frequency bands between 80 Hz and 8 kHz relative to the baseline, the neural network performed best in the mid-frequency bands. The results demonstrate that facial landmarks such as the mouth aperture surface area can provide relevant information in predicting voice radiation characteristics, providing an approach for rendering real-time, dynamic voice directivity.

Speaker: Samuel D. Bellows (University of Utah Asia Campus)

Full paper (PDF)

269 Directivity of individual phonemes and running speech, how do they differ?

Speech directivity has been methodically studied in theatrical design for centuries, and more precisely characterised using electroacoustic measurements for nearly a century. More recently, measurement systems comprising several dozen microphones have enabled detailed characterization of individual phonemes across different languages. It nonetheless remains unclear how the directivity of individual phonemes relates to that of running speech. To investigate such differences, 26 participants produced 9 vowels, 12 sustained consonants, and 6 phonetically balanced sentences in French. Measurements were carried out using a circular array of 180 MEMS microphones at an angular resolution of 2°, enabling high spatial resolution investigation of directivity patterns in the horizontal plane. Resulting directivity patterns are compared as a function of frequency band in terms of inter-channel level differences, as well as through circular harmonics decomposition. This preliminary work will be followed by subsequent studies comparing with other languages. In the context of immersive audio simulations, results of this study can be used for perceptual studies examining if a single general directivity pattern for running speech is sufficient, or whether phoneme-dependent directivity variations would improve rendering quality.

Speaker: Paul Luizard (Audio Communication Group, Technische Universität Berlin)

270 The Influence of Speech Directivity on Speech Recognition and Scene Analysis in Multi-talker Scenes

Sound source directivity plays a role in perception, including for scene analysis and speech recognition in complex acoustic environments. For example, speech directivity cues enhance target speech recognition when the target faces the listener and masker talkers face away from the listener. We examined this effect of speech directivity on target speech recognition as a function of masker head orientation angle. We also tested listeners’ ability to use speech directivity cues to discriminate facing vs. non-facing talkers in a multi-talker scene and to identify the spatial location of the target talker facing the listener. Speech recognition showed significant improvement when maskers faced >45 degrees away from the listener, with additional incremental improvements as masker head angle increased to 180 degrees. Scene analysis (i.e., target talker location identification) accuracy and speed likewise improved as the head angles of non-facing talkers increased. Performance in all tasks was poorer when extended high frequencies (>8 kHz) were filtered out, likely due to the loss of directivity cues associated with these highly directional frequencies. Directivity cues may play an important role for speech perception in complex acoustic environments.

Speaker: Allison Trine (University of Illinois Urbana-Champaign)

A18.03 Soundscape methods, monitoring and metrics: S254

Convener: Rosa Ma Alsina-Pagès (La Salle, Universitat Ramón Llull)

271 Exploring Urban Soundscape Diversity in Large Metropolitan Cities: A Case Study of Athens, Greece

This study investigates the acoustic characteristics of urban soundscapes in Athens, Greece, through the application of four complementary acoustic indices: Acoustic Entropy (H), Bioacoustic Index (BI), Dynamic Spectral Centroid (DSC), and Normalized Difference Soundscape Index (NDSI). A dataset of 36 recordings was collected across diverse urban environments, including traffic-dominated streets, pedestrian zones, and urban parks, and processed under standardized conditions. The analysis revealed distinct spectro-temporal patterns across categories, highlighting variations in soundscape complexity, spectral distribution, and the balance between anthropogenic and biophonic components. Results indicate the combined use of acoustic indices provides a robust framework for the quantitative assessment of complex urban soundscapes and supports a deeper understanding of their ecological and perceptual dimensions.

Speaker: Emmanouil Lianis (NKUA)

Full paper (PDF)

272 Annoyance in Soundscape and Noise Research: Towards a Comparison of ISO/TS 12913-2 and ISO/TS 15666

The acoustic environment can be assessed using two related but conceptually different frameworks: the soundscape approach defined in ISO/TS 12913-2:2018 and the noise annoyance approach described in ISO/TS 15666:2021. Their comparison is particularly important in relation to annoyance. In the soundscape framework, “annoying” is one of eight perceptual attributes, whereas in ISO/TS 15666 annoyance is generally understood as a long-term response to noise exposure. Despite the shared terminology, these concepts may refer to different perceptual phenomena. This study examines the relationship between ratings of the eight soundscape attributes and overall annoyance ratings, and explores the conceptual compatibility of the two frameworks. The analysis is based on 27 audio samples from the Soundscape Attributes Translation Project dataset recorded in London. Statistical compatibility between the two approaches is assessed using mixed-effects modelling. Particular attention is given to whether the “annoying” attribute in the Soundscape Circumplex Model can be treated as equivalent to annoyance evaluated according to ISO/TS 15666, or whether it reflects a distinct, more affective dimension of soundscape perception. The findings are discussed in terms of whether noise annoyance ratings can be modeled using the circumplex structure and the full set of eight soundscape attributes.

Speaker: Jakub Dumanowski (Adam Mickiewicz University in Poznan)

273 Echoes of renewal: the soundscape of a maldivian coral reef

Coral reefs are increasingly threatened by climate change and human pressures, highlighting the need for effective monitoring tools. This study assesses whether underwater soundscape features reflect ecological conditions in a lagoon reef (Magoodhoo, Maldives) using Passive Acoustic Monitoring at a control site (coral rubble) and a treated site with a restoration structure; acoustic metrics (PSD, L5, L95) and ecoacoustic indices (ACI, ADI, DSC) were analyzed and compared across sites. The results show greater acoustic variability and more transient events (higher L5, ADI, DSC) at the treated site, while the control site exhibited more stable background noise (higher L95) and an overall similar ACI. These patterns align with ecological differences, indicating higher structural complexity and biological activity at the treated site, and more stable, less dynamic conditions at the rubble-dominated control site.

Speaker: Valentina Zaffaroni-Caorsi (University of Milano-Bicocca)

Full paper (PDF)

274 A portable measurement system to categorise local, low- frequency noise immission

Infrasound, referring to acoustic sound waves between 0 and 20Hz, is becoming more prevalent in our daily lives as large energy converters such as wind-turbines and biogas power plants are commissioned with an increase in scale and in closer proximity to residential areas. The ability to capture the extent of this noise exposure currently relies on localised measurement campaigns with a narrow field of view. To measure infrasound exposure more reliably, we developed a portable infrasound dosimeter that continuously measures both infrasound, and other low-frequency noise (20-85Hz) in the immediate environment during normal daily life. The gadget measures the immission, meaning the exposure of pollutants, in this case noise in the surrounding environment. It fills a missing middle-ground between long term, fixed acoustic arrays, and instantaneous measurement sensors. The devices are calibrated in the frequency domain against primary calibrated reference microphones at PTB. While carrying the device during pilot studies, we were able to measure infrasound in a multitude of scenarios and environments, such as train journeys and flights, bus routes, walking, cycling, and in the vicinity of wind turbines. In the future, our device will be useful in a range of scientific and industrial scenarios, such as investigating the effects of low-frequency sound on humans, predictive maintenance, environmental monitoring, and the commissioning and operation of large industrial sites and power converters.

Speaker: Criostoir Gerasch (Physikalisch-Technische Bundesanstalt)

A21.04 Railway noise and vibration: S321

Conveners: Karoline Alten (AIT Austrian Institute of Technology), Christopher Knuth (Empa), Slimane Ouakka (Empa)

275 ROLLAND: A New Framework for Realistic and Computationally Efficient Rolling Noise Modeling in the Time Domain

Accurate simulation of noise generation and propagation during train pass-bys remains a persistent challenge in railway acoustics. Most current rolling noise models are forced into a compromise, suffering from either oversimplified assumptions or expensive computational demands. To address this, the present work introduces the fundamental framework for ROLLAND, a novel time-domain model. By applying an explicit finite difference scheme, the approach achieves detailed simulations while maintaining exceptional computational efficiency. The developed time-domain approach allows for the integration of spatially varying track properties and the implementation of multiple, truly moving sound sources. This provides a comprehensive basis for simulating realistic train pass-bys on tracks.At its core, the model utilizes Timoshenko beam theory to characterize bending wave propagation. Moving beyond the limitations of conventional models that often isolate vertical dynamics, ROLLAND also integrates lateral bending, torsional waves, and rail warping, thereby capturing the complex behavior of coupled waves. The framework also accounts for eccentricities in both rail head excitation and rail foot support. In this initial phase, the study focuses on a continuous slab track, modeled as an infinite structure through the application of complex-frequency-shifted perfectly matched layers (CFS-PML). Initial results, focused on the track's impulse response under vertical and lateral loads, demonstrate excellent agreement with an analytical reference model.

Speaker: Maximilian Mantel (Technische Universität Berlin)

Full paper (PDF)

276 Influence of Lateral Wheel-Rail Contact Position on Rolling Noise

Railway vehicle hunting motion is associated with lateral and yaw oscillations of the wheelset on the rails, induced by the conical shape of the wheel tread.During this motion, the wheel-rail contact moves laterally on the wheel tread and rail head, affecting the dynamics of both structures. This study investigates the influence of the lateral contact position on wheel and rail vibrations and sound radiation. 3D Finite Element (FE) models of the track and wheelset are combined with 3D Boundary Element (BE) models for noise radiation, coupled by an analytical wheel-rail interaction model. This framework is used to assess the effect of changing the lateral contact position on the wheel and track responses. The contact is shifted laterally by approximately ±1 cm from the nominal positions on the rail head and wheel tread. The results show that the lateral contact position has a clear influence on the dynamic and acoustic responses. On a structural vibration level, the lateral shift changes the coupling between vertical and lateral vibration components, thus affecting the simulated mobilities and contact forces. These changes lead to higher radiated sound pressure levels. For the considered cases, differences in the one-third octave bands spectra reach 10 dB, while the single value of the total noise differs by about 2 dB. The results indicate that even small lateral shifts of the contact point are relevant for wheel-rail vibration and rolling noise prediction.

Speaker: Chakib Drias (Empa)

277 Numerical assessment of the contribution of railway sleepers to pass-by noise

Railway rolling noise is generated by the vibrations of wheel and track components induced by the wheel-rail roughness interaction. Most existing modelling approaches focus on the wheel and rail as the dominant sources, while sleepers are generally simplified. This study comprehensively evaluates the sleeper contribution by combining a reduced-order multi-sleeper track model with a wheelset model, both based on three-dimensional Finite Element (FE) formulations, to compute the structural responses generated by wheel/rail interaction. The resulting surface velocity fields of the wheel, rail, and sleepers are incorporated into a separate three-dimensional Boundary Element (BE) calculation to compute the radiated noise. This yields both the typically investigated sound power and sound pressure levels at any receiver position.The model is validated against field measurements from the Swiss Railway Field Laboratory (RFL). First, sleeper vibrations are analysed under controlled excitation of the free track and during pass-by to assess the accuracy of the structural model. Predicted rolling noise levels, with and without sleeper contributions, are then compared with RFL data.The validated results demonstrate that sleepers contribute significantly to rolling noise at low frequencies, highlighting the importance of including sleeper radiation in railway noise prediction models.

Speaker: Slimane Ouakka (Empa)

278 Interactive Virtual Reality for Hearing Research: Opportunities, Limitations, and New Insights

Understanding everyday hearing requires approaches that go beyond traditional laboratory experiments with isolated sounds. This plenary introduces interactive audiovisual virtual reality (VR) as a methodology to “bring real life into the lab” and to study perception, attention, memory, and communication in complex, yet experimentally controlled environments. VR can recreate realistic acoustic scenes while allowing for systematic manipulation. However, it also has limitations, such as reduced social and situational context, which must be carefully considered.This talk will highlight the general principles that emerge when auditory and visual information are combined in such environments. Visual cues have been found not to automatically improve listening; their benefit appears to depend on task demands, the difficulty of the acoustic scene, and the need to integrate information across modalities. Overall, VR will be discussed as a tool for investigating these often nonintuitive interactions between seeing and hearing and for linking controlled experimentation with ecologically meaningful listening situations.These developments point toward a future in which hearing is studied within interdisciplinary frameworks that embed auditory perception and processing in rich, audiovisual VR environments. These approaches draw on expertise from acoustics, psychology, and computer science to achieve a more comprehensive understanding of human multimodal perception.

Speaker: Janina Fels (IHTA, RWTH Aachen University)

13:00 Lunch break

A03.02 Measurements and experimental analysis in building acoustics: S310

Conveners: Martin Schneider (Hochschule für Technik Stuttgart), Selina Vavrik-Kirchsteiger

279 Natural Rain and Droplet Impact Sound on ETFE Membrane and Cushion Systems

Thin membranes exhibit a drum-like acoustic response when subjected to rainfall. The resulting rain-induced noise can have a substantial impact on acoustic comfort and may negatively influence communication quality, speech intelligibility, and cognitive task performance. Consequently, a detailed assessment of rain-induced noise characteristics on membrane structures in different architectural configurations is of considerable importance. Moreover, the inherent differences between natural and artificial precipitation -particularly in terms of droplet size distribution and terminal velocity- highlight the need for validation of artificial rain test scenarios against natural rainfall conditions.This study presents a comparative investigation of acoustic responses generated by natural rainfall and by controlled single-droplet impacts performed in a semi-anechoic chamber on Ethylene Tetrafluoroethylene (ETFE) membranes. Although theoretical and numerical models are available to describe the acoustic response of individual droplet impacts, it remains necessary to evaluate the extent to which a superposition of such events- designed to replicate the size and velocity distribution of natural rainfall- can adequately represent real rain conditions. In this framework, in the here presented work, theoretical droplet-impact modelling was employed to support the interpretation and validation of the experimental results.The analysis was further extended to two structural configurations: a single-layer ETFE membrane and a three-layer ETFE cushion system. The outcomes of this study are expected to contribute to improved simulation strategies for natural rain noise and to support more reliable subjective and objective assessments of rain-induced acoustic environments in architectural spaces.

Speaker: Majid Lavasani (Slovak University of Technology)

280 Measurement of creep in the dynamic stiffness of resilient materials

ISO 9052-1, currently under revision, defines the test method for determining the dynamic stiffness of resilient materials used under floating floors, a key parameter for assessing acoustic performance. However, it does not account for the evolution of this property over time. Material ageing leads to a gradual increase in dynamic stiffness, resulting in a decrease in impact sound insulation, which may reach 1 dB after one year and up to 4 dB after ten years.In this context, the present study investigates a method for evaluating the long-term creep of dynamic stiffness, based on an adaptation of EN ISO 16534, originally developed for compressive creep. To this end, a long-term experimental campaign, initiated in March 2021, monitored the evolution of dynamic stiffness for several floating screed underlays over a five-year period. The results are analysed to assess the relevance of this methodology for characterising and extrapolating the creep behaviour of dynamic stiffness.

Speaker: Charlotte Crispin (Buildwise)

Full paper (PDF)

281 Measurement of dynamic stiffness of resilient studs used for multilayer constructions

Sound insulation of mechanically coupled multilayered constructions, such as lightweight floors and walls, depends on the resiliency of line connections (a.k.a. studs or bars) which are mechanically connecting the rigid layers (boards). Typical examples are vertical or horizontal wall studs and floor or ceiling joists. The transmission of mechanical vibration reduces when the dynamic stiffness, K’, of the studs reduces. There is no standardized method of measuring K’ of resilient studs. Our purpose is to present a method for that purpose. The measurement apparatus contains a vibration exciter, two planar load masses hanging from ceiling, force and vibration transducers, and a two-channel real-time analyzer. The stud samples are mounted between the load masses. Thirty geometrically different resilient studs were measured. The profile width of studs varied from 17 to 125 mm. The resulting database includes the K’ and loss factor values of many typical stud profiles available in the market. The measured range of K’ was 0.08–6.91 MN/m2, and range of loss factor was 0.01–0.03. The most significant properties that were different between individual studs were the profile geometry, thickness of the material, and the cut openings in the studs. The method is complete and it could also be considered in standardization committees. Full paper is available in Ref. [1].

Speaker: Jukka Keränen (Turku University of Applied Sciences)

Full paper (PDF)

282 Acoustic Characterisation of Steel Studs in Double-Leafed Walls

This research is focused on improving the acoustic performance of steel-studded double-leafed partition walls. The effects of the different components of the wall and their contribution to the system as a whole are investigated with a specific focus on the steel stud. Steel studs in walls have largely come to replace wooden studs. The design of steel studs minimises raw material cost while maximising structural strength. Steel studs are usually very thin (<1.0 mm) and composed of a flange and web system and have more complex modal behaviour compared to wooden studs. Replacing wooden studs with steel studs has a complex effect on the Rw transmission loss profile low frequencies (<500 Hz) where the steel stud components introduce modal effects resulting in overall decrease of the sound transmission efficiency. In this study the analysis of wall acoustic performance is conducted with help of numerical solutions and experimental verification. Improvements in the stud-wall system are proposed to maintain high sound transmission loss and structural strength.

Speaker: Tilde Jemima S. Resare (University of Sheffield)

A06.02 Electroacoustics and transducers arrays: S038

Conveners: Manuel Melon (Université du Mans - LAUM - UMR 6613 CNRS), Thomas Gmeiner

283 Crosstalk Reduction in Spherical Loudspeaker Arrays via Decoupled Enclosures

Compact spherical loudspeaker arrays enable the synthesis of controllable directivities using loudspeakers distributed over a rigid spherical surface. However, most existing designs rely on a shared rear enclosure, leading to acoustic coupling between loudspeakers. This coupling requires the use of multiple-input multiple-output crosstalk cancellation systems.In this paper, we propose an alternative approach based on the physical decoupling of loudspeakers through a parametric enclosure design using spherical fraction geometries. This design significantly reduces acoustic coupling, allowing the control problem to be simplified to a set of independent single-input single-output equalization filters.A prototype consisting of 18 loudspeakers arranged according to a Popov cubature rule is designed and manufactured. Acoustic crosstalk is experimentally characterized using laser Doppler vibrometry. Results show an average crosstalk reduction of approximately 20~dB compared to a reference spherical array with shared enclosure. Simulated directivity patterns, based on a spherical cap model combined with the measured acoustic coupling, demonstrate that this per-driver equalization is sufficient to synthesize directive beams.

Speaker: Pierre Lecomte (LMFA UMR 5509 CNRS, Ecole Centrale de Lyon)

Full paper (PDF)

284 Random phase arrays

Array design typically aims to direct the beam towards or away from a given direction. However, certain applications — such as surround channels in sound reproduction or acoustical measurements — require a wide polar pattern. This paper describes a method for achieving such a pattern using arrays in which individual transducers are fed with randomly phased signals.Traditional array design relies on precisely optimised phase and amplitude for each transducer. Established approaches to wide polar patterns include frequency-dependent shading applied to edge transducers, adjustment of element spacing, and delay-based processing. The extreme case of delay-based processing is wave field synthesis, in which the sound field of a virtual source is emulated. Bessel arrays offer an alternative approach, defining transducer weights via Bessel functions; in notable special cases, these weights are unity magnitude with alternating sign. The properties of these existing methods are reviewed.The method proposed applies a unity-magnitude, random-phase filter individually to each transducer. The resulting polar pattern varies with frequency; however, its frequency average approaches that of a single transducer, regardless of the number or spacing of transducers. The smoothness of the average polar pattern within a given bandwidth is influenced by array properties and filter length. The random-phase array can further be interpreted as an ideal distributed mode loudspeaker (DML). Numerical examples illustrating array behaviour are presented and discussed.

Speaker: Juha Backman (Bang & Olufsen)

Full paper (PDF)

285 Dual Personal Sound Zones Based on Low-Latency Real-Time Frequency- Domain FxLMS Adaptive Filtering

Personal sound zones (PSZ) aims to create bright and dark acoustic zones within a space by filtering a reference signal to drive multiple loudspeakers. Adapting the control filters is challenging due to variations in acoustic conditions or shifts of the zones, since error microphones are assumed to be fixed in the zones and the estimated acoustic transfer functions cannot be updated.Under these constraints, a real-time adaptive filtering approach was developed for two movable sound PSZ zones. Each zone consists of five loudspeakers in front of six microphones mounted on the same frame to ensure a fixed geometry. The ten filters are implemented in the time domain to minimize latency. The filter coefficient are updated using a frequency-domain block Filtered-x Normalized Least Mean Square (FxNLMS) algorithm. The system was deployed on a real-time controller.Performance, evaluated in terms of contrast and error, was tested under varying PZS positions and changing local acoustic conditions. The results show algorithm convergence and acceptable performance. The discussion highlights fundamental advantages and limitations of the presented approach, including sensitivity to uncertain and time-varying transfer paths.

Speaker: Manuel Melon (Université du Mans - LAUM - UMR 6613 CNRS)

Full paper (PDF)

286 Comparison of spatial microphone probes in room acoustic measurements

Spatial room acoustic measurements have become common practice in room and concert hall acoustics. The authors have been using different spatial microphone-probes in room acoustic measurements for several years. These microphone-probes typically consist of 4 to 8 microphone capsules either omni-directional or cardioid directivity pattern. Slight differences have been noticed in auralisation use as well as when plotting spatiotemporal visualisations based on the measured room impulse responses. This paper presents a study where the performance of three different microphone probes is compared by measuring room impulse responses in controlled test environment. The analysis of the spatial response is done using spatial decomposition method (SDM)[1]. The aim of the study is to find and build an optimal microphone probe for room acoustic measurements for acoustic consulting use. Based on the analysis of spatiotemporal visualisations, an array with 7 microphone capsules (six capsules used for directional analysis) was found to give more accurate reflection tracking compared to 4 capsule arrays. 4 capsule arrays, however, could be sufficiently accurate for visualizing the overall reflection pattern or a certain space to assess the amount of lateral energy against frontal energy and showing the “shape” of the early reflections for instance. The strong reflections from relatively small surfaces, however, could be difficult to pinpoint accurately with 4 capsule arrays. 6-7 capsule array could be accurate enough to pinpoint individual specular reflections for consulting use. Further work is required to confirm these conclusions in real spaces.

Speaker: Perttu Laukkanen (Akukon Ltd.)

Full paper (PDF)

287 Element-wise Equalization of Line Arrays

Curved line arrays are typically optimized through geometric curvature and, when needed, delay loading to realize a prescribed direct-sound level decay over distance. While this effectively controls broadband level distribution, spectral inconsistencies remain due to discretization, element-to-element variability, and mutual coupling. This work proposes an element-wise equalization method that is directly linked to the curvature design via stationary-phase observation points. For each enclosure, distance‑ and frequency‑dependent responses are measured at its associated observation point, third‑octave‑averaged, and converted to a minimum‑phase representation using real‑cepstrum processing with cepstral windowing. Band‑limited spectral inversion then yields a per‑enclosure equalizer, with RMS normalization preserving the intended decay profile. Concert-hall measurements on a six-element RCF HDL26-A array demonstrate reduced spectral spread compared with a single common equalizer, particularly above 1 kHz. The approach improves frequency consistency while maintaining the desired A-weighted direct-sound attenuation, providing a practical complement to geometry-driven array design.

Speaker: Lukas Gölles (University of Music and Performing Arts)

Full paper (PDF)

288 Characterizing Reversible and Irreversible Stiffness Degradation in Loudspeaker Suspensions

Electrodynamic loudspeaker suspension stiffness degrades over time, altering resonance frequency and compromising long term reliability. The aim of this paper is to characterize stiffness evolution during loading and recovery phases to isolate specific degradation mechanisms. An automated measurement system applies a continuous 20 Hz sine wave to induce high excursion, monitoring applied mechanical work and extracting resonance frequency via the circle fit method on impedance data. A two exponential model characterizes stiffness loss as a function of cumulative work, capturing rapid initial break-in and gradual mechanical fatigue. Subsequently, a time dependent two exponential model describes stiffness recovery during rest periods. Equating the loss and recovery models at the load-recovery transition establishes a mathematical relationship that successfully decouples reversible viscoelastic effects (creep and the Payne effect) from irreversible structural damage (break-in and fatigue). Estimating coefficients from the recovery model isolates purely reversible behavior, allowing for the direct calculation of permanent stiffness loss. Separating the mechanisms prevents temporary viscoelastic variations from obscuring actual structural fatigue, providing critical insight into permanent suspension degradation and improving the accuracy of future loudspeaker durability testing.

Speaker: Louis Duparc (International Master's Degree in Electroacoustic and Audio)

Full paper (PDF)

A07.03 Noise reduction technologies for aeroacoustics (active and passive): S043

Conveners: Anita Schulz (HTW Berlin), Francesco Avallone (Politecnico di Torino), Yves Aurégan (Labo Acoutique de l'Université du Mans)

289 Flow-Acoustic Interactions in a Deep Segmented Porous Chambers: An Experimental Study

Deep axisymmetric cavities integrated into a chamber significantly influence acoustic wave propagation, notably by reducing the effective sound velocity. In our experimental configuration, the effective sound velocity is approximately one quarter of the sound velocity in air (c_e≈c_0/4). This substantial reduction suggests the possibility of introducing a mean flow exceeding c_e to investigate whether sound propagation against the flow can be blocked. However, previous attempts to achieve this effect were stopped by intense whistling instabilities occurring in empty cavities [1] and in cavities fitted with porous rings [2]. The present study investigates whether completely filling the cavities with porous material can mitigate or delay these instabilities while preserving the desired acoustic properties of the structure.Acoustic transmission and scattering measurements were performed on configurations comprising 30 cavities with porous treatments, under flow conditions up to M=0.35. Results demonstrate that no whistling occurs when cavities are fully filled with porous media. A pronounced asymmetry in sound propagation is observed between upstream and downstream directions. Nevertheless, at very low frequencies, transmission against the flow remains possible even at the highest Mach numbers tested.[1] Aurégan Y., and Golliard J., “Linear Aeroacoustics of Deep Cavities,” 30th AIAA/CEAS Aeroacoustics Conference (2024)[2] Golliard J. and Aurégan Y., “Transmission and amplification by deep periodic axisymmetric cavities with flow,” 32th AIAA/CEAS Aeroacoustics Conference (2026)

Speaker: Yves Aurégan (Labo Acoutique de l'Université du Mans)

290 Systematic Performance Analysis of KT and HTLS Methods for Wavenumber Extraction in Lined Flow Ducts

The extraction of complex axial wavenumbers is a key intermediate step in modal sound-field decomposition and liner impedance eduction. This paper presents a systematic benchmark comparison of the Kumaresan--Tufts (KT) and Hankel total least squares (HTLS) methods for controlled two-dimensional synthetic test cases in lined ducts. The analysis covers single-mode benchmarks for varying SNR, noise model, and microphone array geometry, and is complemented by a compact multimodal stress test based on realistic no-flow SDOF Helmholtz-resonator liner trajectories. In noise-free cases, both methods recover the damped-exponential signal model essentially exactly. Under noisy conditions, the extraction error increases by roughly one decade per $20\,\mathrm{dB}$ reduction in SNR, turbulent boundary-layer noise is slightly more detrimental than white noise at low SNR, and HTLS is globally more robust, particularly for modes with positive imaginary part. The geometry study shows that the dominant array-design parameters are the dimensionless products $k_x \Delta x$ and $k_x l$.

Speaker: Anita Schulz (HTW Berlin)

Full paper (PDF)

291 Scattering of Cut-off Modes by Impedance Discontinuity in a Lined Duct

Sound scattering induced by a duct wall impedancediscontinuity located in proximity to a source is investi-gated. It is illustrated that, when a liner is positionedin a source’s near field, evanescent modes can have asignificant impact on the transmitted power, especiallyclose to their cut-off frequency. A parametric study isconducted to highlight how evanescent modes locatedclose to the liner modify the in-duct transmitted power.Sensitivity to various parameters such as frequency,source-liner distance, liner length, liner impedance andsource modes is investigated. It is demonstrated that,for some configurations, the optimisation of acoustictreatments must take into consideration the near fieldin order to accurately predict liner performance.

Speaker: Aymane Chaabaoui (LAUM - Université du Mans)

Full paper (PDF)

292 On Vortex Generating Lining for Jet Noise Reduction

The presence of streamwise vorticity in high speed gaseous jets can be considered beneficial in reducing jet noise. The inner walls of nacelles for turbo-fan propulsion systems are acoustically treated, in order to reduce the amplitude of turbomachinery and combustor noise that propagates through the duct. The German Aerospace Center (DLR) proposed a patent for a bias-flow liner device, called VoGeL (for Vortex Generating Liner) see Ref. [1], where the bias flow also aims the generation of streamwise vorticity in the flow duct. The device has been realized and studied on a single prototype, in the framework of static jet noise in isolated condition and at the presence of a flat plate. We propose a review of the work made by the DLR on the Vortex Generating Liner, Ref.[2, 3]. We also realize and describe a small scale experiment to realise a VoGeL prototype in the inner wall of a round nozzle and demonstrate the effects on static jet noise emissions.[1] A. Bassetti, “Vorrichtung und Verfahren zur Beeinflussung einer Fluidhauptströmung und Düse.” Deutsches Patent-und Markenamt DE102021209284A1.[2] H. A. Siller, W. Hage and A. Bassetti; “Preliminary investigations of jet installation noise influenced by a vortex generating liner at the nozzle inner wall.” The DJINN-ENODISE conference: Aeroacoustic Installation Effects in Conventional and New Aircraft Propulsion Systems, 2023-11-22 - 2023-11-24, Berlin, Deutschland.[3] H. A. Siller, W. Hage, C. Jente and Alessandro Bassetti, “Jet-noise investigations in a small scale facility.” 28th AIAA/CEAS Aeroacoustics Conference, June 2022, Southampton, United Kingdom.

Speaker: Alessandro Bassetti (Hamburg University of Technology)

293 Laser Doppler investigation of Source Location Effects on Near-Orifice Flow and Acoustic Velocity Field over a Liner

The interaction between acoustic waves and turbulent grazing flow over liners strongly affects their aeroacoustic performance. This study experimentally investigates the influence of source location on near-orifice flow dynamics and liner response using Laser Doppler Velocimetry (LDV) in ONERA’s Aero-Thermo-Acoustic facility, providing high-resolution velocity measurements. The analysis focuses on the region close to an orifice, where coupling between the turbulent boundary layer and the acoustic field is strongest. Mean and root-mean-square (RMS) velocities are examined together with the acoustic velocity field, extracted from LDV data by cross-correlation with a reference source signal, allowing the determination of both amplitude and phase. The results show a marked sensitivity of the near-wall fluctuating velocity and coherent velocity distribution to source position. These findings provide new insight into liner–flow–acoustic interactions and highlight the critical role of source location under realistic grazing flow conditions.

Speaker: Ludovic Ambrosiani (Politecnico di Torino)

Full paper (PDF)

294 Experimental Study of Noise Reduction in Swirl Diffusers Using Optimized Blade Geometry

The study demonstrated that the acoustic performance of ceiling swirl diffusers with adjustable blades is strongly influenced by both blade geometry and flow configuration. The results showed that some types of commercial profiles significantly increase noise levels, especially under unfavourable conditions of flow. The best performance was achieved by the profile with aerodynamic shape, which produced the lowest sound power levels and even reduced noise below the baseline diffuser. Profile with serrated edges showed good acoustic properties, especially in swirling flow, where it minimised additional noise. The findings highlight the importance of flow direction. Overall, optimised blade geometry and proper flow configuration are essential for designing low-noise swirl diffusers.

Speaker: Joanna Maria Kopania (Lodz Univerity of Technology)

Full paper (PDF)

A11.05/A15.03 Perception of Music and Musical Instruments: S078

Convener: Kai Siedenburg (CvO University of Oldenburg)

295 Dynamics, compression, and perceived groove in snare-drum patterns

Groove, defined as the pleasurable urge to move in response to music, is a central dimension of rhythmic perception. While previous research has examined the contributions of syncopation, microtiming, and tempo to groove, the role of sound dynamics remains poorly understood. This study investigates how variations in sound level across drum patterns influence perceived groove, and whether dynamic range compression can be used to enhance it. Twenty participants rated the groove, pleasure, and complexity of 159 short snare-drum sequences designed to systematically vary dynamic structure. A linear mixed-effects model revealed significant effects of dynamic pattern on groove ratings, with rankings that remained stable across different levels of gain variability. To predict groove from sound dynamics, we introduce a fuzzy representation of gain levels combined with a ridge regression model. A transition-based feature representation, encoding sequential dependencies between successive gain events, outperformed a distribution-based approach, suggesting that the temporal ordering of dynamic levels is a key perceptual cue. Using the trained model in a generative framework, we further show that modifying sound dynamics with dynamic compression can significantly increase predicted groove across a wide range of dynamic configurations. These findings suggest that compression could be used not only for loudness control, but also as a perceptually motivated tool for shaping groove in live music contexts.

Speaker: Ulysse LEFEUVRE (L-Acoustics)

Full paper (PDF)

296 Selective attention in music: effect of distraction by concurrent melodies

When we listen to music, our auditory system performs a complex task known as auditory scene analysis. This process involves organising overlapping acoustic signals into distinct streams or sources, such as an instrument or a singing voice, and perceiving individual sound events, such as the notes played. The current project focused on investigating selective attention in multi-instrument music scenes – in other words, the ability to attend a target instrument while ignoring other simultaneous sources. A selection of scene features was systematically varied to study the effect of (1) scene size (1,2 or 4 playing instruments), (2) scene density (by changing the distance in frequency between target and distractor instruments) and (3) scene timbre (homogeneous vs heterogeneous set of instruments) on performance. The music pieces were taken from chorales of JS Bach and were chosen for their relative homogeneity in terms of rhythm, harmony and individual melodies. The task was to track a target melody - the “soprano” part (highest-pitched melody) or the “bass” part (lowest-pitched melody) - and indicate each time two consecutive notes were the same. Listeners tended to be better at following the soprano compared to the bass. Performance decreased with increasing scene size, and an interaction demonstrated that this effect was stronger when listeners focused on the bass than on the soprano, which may be due to a “high-voice superiority” effect. Performance also decreased with increasing scene density for both soprano and bass, demonstrating an effect of informational masking related to the distance to the target melody. Preliminary data suggest that timbral heterogeneity helps towards source segregation.

Speaker: Mathilde de Kerangal (UCL Ear Institute)

297 Decomposing the Impact of Musical Structure on Auditory Attention in Multi-Voiced Music

Auditory attention in multi-voiced music typically gravitates towards the highest voice (high-voice superiority; Trainor et al., 2014). Conversely, bass lines often lack this kind of salience (bass inferiority; Bürgel et al., 2021). However, our recent study using pseudo-randomized pure-tone melodies revealed a contrasting "edge effect," where melodies in both the highest and lowest frequency bands exhibited perceptual salience (Bürgel et al., 2024). We hypothesize that the lack of musical realism caused this contrary finding. Therefore, the present study decomposes the impact of musical structure in multi-voice music perception and investigates whether the edge effect persists in more realistic musical mixtures. Stimuli consist of 2-second mixtures selected from four loudness-matched frequency bands with equidistant ERB-spacing (65–1230 Hz). Listeners perform a target-detection task with the cue presented either before the mixture or after. To test the role of musical realism, a regular rhythmic beat grid and a probabilistic Markov model generating tonal melodies in major and minor scales are implemented and will be tested against non-structured (and less musically realistic) rhythmic and tonal choices. Furthermore, tones are presented as either pure tones or naturalistic virtual instrument timbres.We hypothesize that musically structured pure-tone stimuli facilitate detection. In contrast, the complex spectro-temporal features of natural timbres are expected to introduce spectral interference, counteracting the edge effect and shifting attention back toward high-voice superiority and bass inferiority. While data collection is ongoing, we look forward to presenting our findings and discussing their implications at the conference.

Speaker: Michel Buergel

298 Beyond modulation: A large-scale analysis of audio features characterising ensemble sound

Compared to individual instruments, audio features that characterise multiple instruments forming ensembles have received less attention. Apart from reports arguing the importance of decorrelated amplitude modulation of partial tones, it largely remains unknown what audio features describe ensemble sound more generally and how features may vary as a function of instruments, performance variables, and room acoustics. Next to performance factors, a comprehensive study could also consider related audio processing or recording techniques such as chorus, doubling or multitracking. We report on a dataset of currently 6,300 sounds generated using additive synthesis and virtual acoustics to investigate a range of variables, which include between one and ten layers of an instrument, 14 different instruments across their pitch ranges, acoustical factors like room size, source-receiver distance and source spatialisation as well as several chorus implementations. The planned expansion of the dataset will also allow intonation, asynchrony, vibrato, and chorus properties to be studied as random variables. Audio-feature analysis considered both conventional audio descriptors and banks of frequency and modulation filters, with principal components analysis (6 components) explaining 68% and 44% variance in the dataset, respectively. Group medians across layers (N=630) exhibited a clear decrease in low-frequency modulation from one to four layers as well as greater spectral extent with more layers. Subsequent analysis on the growing dataset is expected to provide further insights on remaining variables and could clarify why the blended perception of ensemble sound can be achieved with as few as three or four layers.

Speaker: Sven-Amin Lembke (Anglia Ruskin University)

299 Application of a Template-Based Perception Model in the Context of Audio Scores

Audio scores convey the musical information from the composer to the musician in an acoustic way instead of written notation. Typically, the score is being played to the musician via loudspeakers, and the musician imitates what they hear. Here, psychoacoustics play a major role in the composition, because the timing of perceiving the sound changes and consciously imitating it depends on the musician’s auditory system and the acoustic environment. A critical question from the composers perspective is when the musician is able to perceive a change in timbre because this affects the compositional process. Our work approaches this question from a psychoacoustic perspective. To this end, we modelled the potential of the musician to distinguish between two timbres. Specifically, we used a template-based perception model initially designed to detect perceptual similarity between two pianos. We adapted that model to our task and extended the stimuli by sounds of other instruments. With the help of the model, composers can simulate a virtual musician and test their compositions analogous to a music notation programme that simulates a virtual instrument.

Speaker: Katharina Pollack (Acoustics Research Institute, ÖAW)

300 All You Can Analyze: SInES Tools - An Audio/Video Analysis Lab in your Browser

The Space for Interdisciplinary Experiments on Sounds (SInES, University of Vienna) has developed in recent years more than 80 browser-based applications for audio analysis, AI-driven audio feature estimation, (re)synthesis, and video tracking at https://sinestools.univie.ac.at. Following a "upload file – click analyze – download results" philosophy, these free tools leverage modern JavaScript libraries (e.g., TensorFlow.js, Essentia.js) to democratize advanced methods previously requiring high costs or expert programming skills.Recent updates in 2025/2026 introduce powerful new capabilities: | - Audio Analysis & AI: Harmonic Tension Analyser, Masking/Timbral Blending Score Estimator, a Room Acoustics Tool (room acoustic analysis according to ISO3382), Modulation Power Spectrum Analysis etc. New AI models include Speech Emotion Classifiers, Phoneme Recognizer, and musical instrument/ensemble detection. | - Spectral (Re)Synthesis & Editing: A Spectral Editor allows "Photoshop-style" sonogram editing. New tools enable harmonic-percussive separation, Formant shifting, spectrum- and modulation transfer, spectral gating, sound morphing, and artifact-free time-stretching. | - Tracking & Motion: A Face Muscle Tracking Tool allows for the analysis of 52 facial attributes. A new unified interface (based on human.js) now combines pose, face, and hand tracking. Previous SInES Tools for tracking poses, hands, faces, emotions, and points in videos have been completely redesigned (Frame-by-frame resolution, improved AI models) | - SInES Helpers: The collection of statistics tools (ANOVA, PCA, Cluster Analysis) has been extended by browser-based calibrated audio recording and AI-supported sample-slicing. | These and other SInES Tools will be presented at the Forum Acusticum and are available for immediate use online at https://sinestools.univie.ac.at.

Speaker: Christoph Reuter (Universität Wien)

Full paper (PDF)

A11.06 Material choice and musical instruments: S079

Conveners: Henri Boutin (Laboratoire STMS), Alex Hofmann (Department of Music Acoustics – mdw)

301 Architectured materials for lutherie: Application to the acoustic guitar soundboard

The soundboard is a key element in an acoustic guitar. It both makes the instrument radiate as a result of the strings’ motion, and withstands the static stress induced by their tension. Made out of a thin wooden plate with wooden stiffening bars glued on its back (the bracing), it plays a dominant role in the instrument’s sound profile. This way, a large variety of bracing patterns have been created through the ages, giving the guitar a substantial sound palette available to this day.We show here an alternative soundboard structure, in which the bracing is redesigned into a multilayer geometrical network, whose vibrational behavior can be controlled by varying two geometrical parameters, using recent concepts and manufacturing processes such as architectured materials and laser cutting. Determining the vibrational behavior by a direct method for such a structure would be too computationally expensive, so we need to develop reduced equivalent homogeneous models. Two methods are presented here: one exploiting the dispersion relation of a thin homogeneous orthotropic plate (Kirchhoff-Love), the other adjusting the generalized stiffnesses of a thick plate model (Reissner-Mindlin).The results show the relevance of these models, both with simulated or experimental data. This novel structure thus aims to facilitate the soundboard dynamical behavior optimization, hence potentially widen the sound palette of the acoustic guitar thanks to spatial modulation of the network’s geometrical parameters, giving luthiers an additional tool to design their instruments.

Speaker: Nathan Zwahlen (LAUM)

302 A Comparative Study of 3D-Printed Cello Bridges. Which One Comes Closest to the Original?

The differences between 8 3D-printed bridges, which were fabricated using various 3D-printing technologies, materials and internal structures, were explored and compared to the original wooden cello bridge. Input admittance measurements were used to investigate how the different bridges affect the resonance characteristics of the bridge–body system. These measurements were obtained using an automated impact hammer system mounted on a robotic arm. This allowed the fragile impact hammer to be moved to a safe position while the bridge was being interchanged, and ensured highly reproducible excitations. The response was captured using two accelerometers attached to either side of the bridge under test. A laser and camera system assisted with the repositioning of the bridges and accelerometers. During six measurement sessions all 9 bridges were tested, with 20 impact measurements per bridge. The high number of repetitions and exchange sequences was intended to identify the actual influence of the material on the measurements and to exclude differences caused by repositioning on the instrument. Input admittance, coherence and speed of sound within the tested bridges are compared amongst materials. The findings provide a foundation for the further development of a 3D-printed sensor-equipped bridge.

Speaker: Alexander Mayer (mdw – University of Music and Perform)

Full paper (PDF)

303 Casting for Sounding: An Interdisciplinary Case Study of Southeast Asian Copper-Alloy Drum

Copper-based drums, emblematic of Southeast Asia and southern China, can be admired in several French museums. Appearing during the protohistoric period (c. 5th century BC), their production has continued to the present day. Varying in geometry and size – with a tympanum diameter from a few centimetres up to 65 cm – drums have intrigued researchers since the early 20th century. Scholars primarily focused on style and functions, retaining that of a prestige good associated with chiefs. While drum casting is a tour de force, technicalstudies remain scarce, leaving artisanal processes unclear. Acoustic studies are even rarer, despite sound production being the drums' primary function.This paper presents a study at the crossroads of archaeology, materials science, and acoustics, aiming to better understand drum casting techniques and acoustics, focusing on a specimen from the National Museum of Archaeology (Saint-Germain-en-Laye). This approach includes visual analysis and condition reports aimed at identifying drums of interest within the museum. A structured-light 3D scanner, X-ray imaging (radiography, tomography) and materials analysis were also used to characterise the geometry and mechanical properties, in particular their elastic properties. Modal analysis of the tympanum and body aims to determine vibration eigenmodes and simulate dynamic behaviour. Finally, sound recordings are used to characterise the radiated sound and to establish a correlation with the instrument’s vibrational properties.

Speaker: Clémence Le Meur (ÖAI)

304 The Negative Bore Concept in Wind Instruments

Musical instruments of the wind family present bore geometries that are usually approximated by acousticians as analytical functions such as cylinders, cones, exponential or Bessel horns. However, a closer look at actual instruments shows that the makers use adjustments of the bore profile to control the intonation of the instruments with a rather fine precision, typically of the order of 1/20th to 1/50th of the bore diameter. Drilling such bores in the workshop is one of the challenges of wind instrument making, whether it involves drilling in wooden instruments or shaping metal instruments on a mandrel. The negative bore concept was first introduced at ISMA 2019. The main idea was to replace the bore with non-ideal geometry with a pipe of ideal profile into which a needle (called "negative core") is inserted, provided that the diameter of the needle is calculated such that the resulting flow cross-section of the pipe is similar at every point along the pipe's axis to that of the original bore. However, this process increases the wall surface, degrading the acoustic equivalence between the two pipes. By studying this effect, the aim is to determine the negative's geometry giving similar acoustic properties that the original bore.The presentation will include a theoretical analysis of the problem in terms of the influence of the viscous and thermal effects on the acoustic wave propagation, together with experimental data, both showing the dependence on the hydraulic perimeter of the pipe including its negative core.

Speaker: benoit fabre (Sorbonne University)

A14.03 Hidden hearing loss & synaptopathy: Diagnostics, treatment & functional consequences: S292

Conveners: Lukas Rüttiger (University of Tübingen), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

305 Auditory Brainstem Responses in Aging Rats are Specific for the Spectral Content in Human Syllables

With aging progressing inner hair cell (IHC) synaptopathy precedesage-dependent elevation of auditory thresholds. We therefore compared central sound responses in young, middle-aged, and older rats for human syllables with spectral and temporal characteristics requiring temporal fine structure (TFS) or temporal envelope (TENV) coding. As stimuli the syllables /o/, /u/, /di/, /bi/, /du/ and /bu/ were selected since they differ specifically in frequency bands below or above the temporal limit for TFS coding (at ca. 1500 Hz). IHC synaptopathy progresses from middle age onward. While middle-aged animals do centrally compensate for the loss of auditory fiber activity through an increase in the delayed auditory brainstem responses linked with the shortening of central response latencies old animals fail to restore central responsiveness and lose temporal resolution in responding to amplitude changes. Therefore we studied the temporal and spectral precision of the auditory brainstem response of rats at different ages. The responses were analysed in the spectral domain for the specificity of sustained reponses to the particular syllable stimuli. The typical brain signatures for processing of the different syllables will be discussed in the view of progressing synaptopathy and high-frequency hearing deficics in the aging rat. Age related synaptopathy and its presumptive role for speech processing may be an important factor for human speech perception.

Speaker: Lukas Rüttiger (University of Tübingen)

306 Auditory-Nerve Spontaneous-Rate Types and Neural Coding of Complex Sounds

The loss of a larger percentage of low-spontaneous-rate (LSR) than high-spontaneous-rate (HSR) auditory-nerve (AN) fibers has been a focus of much of the work on cochlear synaptopathy. This talk will discuss the implications of the loss of different fiber types on hearing of complex sounds, especially in noisy settings. The average rates of LSR fibers are often assumed to be important for encoding moderate to high sound levels, and more generally, spectrum levels. However, quantitative analysis based on optimal decision theory has shown that the rates of LSR fibers cannot explain level discrimination, which improves as level increases. We have proposed an alternative model for neural coding of complex sounds at moderate sound levels based on neural fluctuations (NFs), the low-frequency fluctuations in the probability of firing of AN fibers (of all SR types), which varies along the tonotopic axis depending on spectral level. To maintain the contrasts between NF depths in AN responses tuned near spectral peaks vs. spectral valleys would require broadband adjustment of cochlear gain for complex sounds at low or high levels. The low- and medium-spontaneous-rate (MSR) AN fibers, with wide dynamic ranges, could play a role in this cochlear gain control. In some species, these fibers are the exclusive inputs to cells in the small cell cap in the anteroventral cochlear nucleus, which projects to medial olivocochlear neurons in the brainstem. Using computational models that include efferent pathways, we are exploring this potential mechanism for maintaining NF contrast over a wide dynamic range.

Speaker: Laurel H. Carney (University of Rochester)

307 Does It Matter Which Fibers You Lose? A Computational Analysis of Cochlear Synaptopathy

Animal studies suggest cochlear synaptopathy predominantly affects high-threshold, low spontaneous rate auditory nerve fibers. This pathology has been proposed as a contributor to speech-in-noise deficits despite normal audiometric thresholds. A candidate explanation is that remaining fibers compensate by shifting encoding to alternative frequency channels, but this compensation depends on channels that noise can saturate.We test this computationally by quantifying how selective fiber loss reshapes phoneme level neural representations, and by evaluating perceptual consequences using a neural vocoder and a speech foundation model for transcription.Over 10,000 phonemes from the TIMIT corpus were passed through a model of the auditory nerve and the inferior colliculus, across five fiber-loss configurations and four noise conditions. In quiet, midbrain processing compensated most peripheral distortion, leaving few phonemes affected. In noise, compensation collapsed: most phonemes were distorted at the inferior colliculus, and the increase in speech errors relative to normal hearing nearly tripled. The most vulnerable category shifted from stops to voiced fricatives, because low spontaneous rate fibers provide precise timing in quiet and dynamic range in noise.Because these fibers also drive the olivocochlear efferent reflex, their loss may compound the deficit through reduced noise protection, revealing deficits that shift with listening conditions and escape the audiogram.

Speaker: Marta Campi (University of Zurich, University Hospital Zurich)

308 Effects of Low- and Medium-Spontaneous Rate Fiber Loss on Speech-in-Noise Discrimination: A cross-species Modelling and Experimental Study

Cochlear synaptopathy (CS) is a form of auditory neural damage characterized by the loss of synapses between inner hair cells and auditory nerve fibers (ANFs), often without changes in audiometric thresholds. It is associated with noise exposure and aging and is thought to underlie difficulties in speech perception in noisy environments despite normal hearing sensitivity. CS is believed to disproportionately affect high-threshold ANFs with low and medium spontaneous rates (LSR/MSR). However, isolating and quantifying the contribution of LSR/MSR fiber loss to speech-in-noise perception remains challenging.Here, we address this problem using a cross-species framework combining 1-D transmission line models of the human and gerbil auditory periphery, invasive gerbil recordings, and human behavioral data. A fixed pair of vowel stimuli (/i/ and /y/) is presented in noise across all measurements and simulations. A recurrent neural network, driven by CAP inter-spike interval features, is trained to discriminate between the vowels.The degree of cochlear deafferentation is calibrated using invasive measurements from normal-hearing and kainic acid (KA)-treated gerbils, a model for CS. This framework links ground-truth ANF recordings to human behavioral performance through a unified modeling approach.Preliminary results show that the loss of LSR and MSR fibers degrades speech-in-noise discrimination differently from the loss of high spontaneous rate (HSR) fibers or elevated auditory thresholds. Notably, complete loss of LSR fibers can be more detrimental than a 50% loss of HSR fibers, despite a smaller total reduction in fiber count.

Speaker: Morgan Thienpont (Ghent University)

A14.05 Neural substrates of complex auditory perception and auditory scene analysis: S102

Conveners: Katrin Krumbholz (University of Nottingham), Joseph Sollini (University of Nottingham)

309 Spatial Auditory Attention Decoding Using Gammatone-Based Representations Under Clean and Binaural Conditions

Auditory attention decoding (AAD) aims to determine from neural recordings which sound source a listener is attending to. Here, we formulate spatial auditory attention decoding (SAAD) as a subject-independent end-to-end task, jointly modeling EEG signals and Gammatone-based audio representations to decode the attended spatial direction. While conventional approaches assume access to clean speech signals, real-world conditions provide only binaural mixtures. Therefore, we investigate the feasibility of decoding attention directly from mixture signals, in addition to ear-separated (clean) representations on the DTU and KUL datasets. Results show that decoding performance increases consistently with decision window length for all conditions. On the DTU dataset, binaural mixtures achieve performance comparable to ear-separated signals, with no statistically significant differences across decision windows. On the KUL dataset, the HRTF-based binaural condition outperforms the clean (dichotic) condition at longer decision windows, with significant improvements at 20 s and 40 s. These findings demonstrate that spatial auditory attention can be decoded directly from binaural mixtures in a subject-independent setting and highlight the potential of Gammatone-based representations for realistic AAD systems.

Speaker: Maryam Bajool (Technische Universität München)

Full paper (PDF)

310 EEG Investigation of Human Responses to Low-Frequency and Infrasound Stimuli Using Late Auditory Evoked Potentials

Exposure to low-frequency sound and infrasound is becoming increasingly common in modern environments, yet their neural processing remains insufficiently understood. This study investigates human auditory processing of low-frequency and infrasound stimuli using 64-channel electroencephalography (EEG), with a focus on late auditory evoked potentials (LAEPs), which reflect higher-order auditory processing in the brain. Acoustic stimuli were presented at an equal-loudness level of 70 phon across frequencies ranging from 500 Hz down to 4 Hz (infrasound) in a group of participants. LAEPs could be elicited by all presented frequencies, and source localization showed that brain responses differed between low-frequency and higher-frequency sounds in both spatial and temporal characteristics. In addition, a method was developed during this study to improve LAEP estimation in multichannel EEG recordings. These findings contribute to a deeper understanding of how the human brain processes sound near or below conscious perception and support the development of more robust EEG-based methodologies for investigating low-frequency auditory phenomena.

Speaker: Mohamad Mahdi Mkanna (German National Metrology Institute)

311 Auditory context regulates neural and autonomic reactivity to sudden sounds

Humans frequently introduce music or ambient sound into everyday environments, often without actively attending to it, suggesting that continuous auditory backgrounds may serve functions beyond pleasure or entertainment. We tested the hypothesis that such backgrounds reduce sensitivity to sudden events, whereas silence leaves the brain and body more vulnerable to acoustic surprise. To probe this, unexpected noise bursts were embedded within periods of silence, music, or environmental sounds while neural and autonomic activity were recorded.Thirty-three participants completed a passive listening task during which EEG and ECG were acquired across 20s epochs of each background condition. Noise bursts elicited heart rate acceleration selectively during silence, indicating greater autonomic reactivity in the absence of ongoing auditory context. Relative to silence, both music and environmental sounds increased EEG ongoing beta and gamma activity, consistent with a more externally engaged neural state. Burst-evoked responses also varied across conditions. Silence was associated with larger early auditory (N1) responses, consistent with increased early sensory reactivity. Music reduced N1 amplitude, suggesting attenuation of the initial sensory impact of surprise in a temporally structured auditory context. Environmental sounds reduced later P2/P3 responses, consistent with diminished higher-order evaluation of the burst. Across conditions, auditory surprisal estimated with the D-REX model predicted P3 amplitude, and higher trait anxiety was associated with larger P3 responses.These findings indicate that silence is not a neutral baseline, but a state of heightened sensory and autonomic responsiveness to unexpected events. Continuous auditory backgrounds, and music in particular, may help regulate this responsiveness by buffering the impact of acoustic surprise.

Speaker: Ulysse LEFEUVRE (L-Acoustics)

312 Experience-dependent Sharpening of Human Cortical Spectral Resolution Revealed by fMRI Adaptation

Adaptation is a ubiquitous feature of neural responses and has been used extensively to probe cortical tuning properties with non‑invasive methods, particularly fMRI. This approach typically assumes that adaptation reflects fixed sensory tuning. However, neuro‑ and electrophysiological evidence suggests that cortical adaptation may itself be dynamically shaped by stimulus history.Using ultra‑high-field (7T) fMRI, we measured responses to a fixed 3.8‑kHz probe sound following adaptor sounds that varied in frequency. Adaptors were presented either as a single sustained sound or as multiple brief sounds in rapid succession, matched for total duration. Adaptor frequencies spanned values at or below the probe.The two adaptor regimes produced strikingly different adaptation profiles. When adaptors were repeated, maximal adaptation occurred at the probe frequency regardless of voxel preferred frequency. When a single adaptor was used, maximal adaptation occurred at frequencies intermediate between the probe and the voxel’s preferred frequency. Both patterns were captured by a simple linear depletion (Tsodyks–Markram) adaptation model by assuming different effective bandwidths of the adaptation kernel: narrow for repeated adaptors and broad for single adaptors. Importantly, voxel‑level frequency tuning differed little between conditions and was much broader than expected from the macroscopic tonotopic gradient alone, implying substantial local scatter in frequency preference.We propose that such tonotopic scatter—potentially reflecting a property of hypercolumnar organization—enables dynamic, context‑dependent spectral resolution in auditory cortex. From a predictive-coding perspective, the observed sharpening of adaptational tuning following repeated stimulation may reflect increased precision of sensory priors, corresponding to sharpened sensory expectations.

Speaker: Katrin Krumbholz (University of Nottingham)

313 Where is What and What is Where? Interaural Coherence Differentially Activates Dorsal and Ventral Auditory Cortical Pathways

Cortical processing of auditory scenes is thought to be organised along a ventral "what" and a dorsal "where" pathway; here we ask whether these pathways are differentially engaged by sounds that differ only in the reliability of their spatial cues. Although we perceive auditory objects with unique sources and locations, everyday acoustic environments are replete with echoes and reverberation, resulting in a complex mixture of signals reaching each ear. This superposition introduces fluctuations in binaural cues, rendering them unreliable — an unreliability quantified by interaural coherence (IAC). We designed amplitude-modulated noise tokens in which IAC was systematically modulated over time, with brief segments of high coherence embedded at different phases of the modulation cycle. All sounds were spectrally identical and shared the same energy envelope, rendering them indistinguishable at either ear alone; differences in spatial quality emerged only binaurally. We measured interaural time difference (ITD) thresholds as an index of binaural performance and used functional near-infrared spectroscopy (fNIRS) to record changes in oxy- and deoxyhaemoglobin concentration from superior temporal gyrus (STG; "what" pathway) and inferior parietal lobule (IPL; "where" pathway). These pathways were dissociable in both the magnitude and temporal characteristics of the responses: STG was activated earlier than IPL, with a systematic temporo-parietal delay gradient. STG responses were comparable across all conditions, consistent with their identical spectral content, whereas peak IPL activity was modulated by listeners' ITD thresholds. Our data suggest auditory objects are first represented ventrally in the cortex as objects per se, and only subsequently dorsally with respect to behaviourally accessible spatial properties.

Speaker: Jörg Encke (Macquarie University)

314 Broadband neural inhibition and it's role on perception

Neural inhibition is observed at every nuclei of the central auditory system and has been implicated in a range of functions, such as: spatial cue derivation, controlling gain, sharpening frequency tuning and sharpening temporal sensitivity. In our own work we have observed a role for central inhibition, i.e. inhibition in the central auditory system, in the processing of broadband sounds. We present neurophysiological results demonstrating a role for central inhibition in improving neural correlates of sound perception. Specifically, wideband inhibition in auditory cortex appears to improve hearing in noise, particularly for comodulated sounds. Central inhibition, in response to broadband noise, has consequences for hearing beyond shaping perception. It is also to believed to be involved in the generation of illusory sound percepts. We also present human data supportive of the idea that central inhibition plays a role in the generation of the Zwicker tone, an illusory tone percept heard after the presentation of notched noise.

Speaker: Joseph Sollini (University of Nottingham)

A15.01 Spatial Hearing: Modeling and Applications: S297

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Ville Pulkki (Aalto University)

315 Mitigating Spatial Disorientation via Spatial Audio: A Psychoacoustics Experiment Design

Spatial disorientation (SD) is a perceptual phenomenon particularly affecting pilots, contributing to approximately one-third of all aviation mishaps. While the vestibular, visual, and auditory systems collectively maintain spatial orientation, the potential of spatially controlled auditory stimuli as SD countermeasures remains under-explored. This paper presents the design, protocol, and preliminary observations of an ongoing experiment investigating which spatial audio configuration and stimulus type most effectively mitigate SD induced by the somatogyral illusion. Participants undergo 1-minute yaw rotations in a Bárány chair at 40°/s, while exposed to three stimulus types — pseudo-speech, low-frequency slow bursts, and windowed white noise pulses — delivered through an 8-loudspeaker horizontal ring controlled via Vector Base Amplitude Panning (VBAP). Two spatial conditions are examined: a static 0° frontal azimuth point source and a chair-synchronized continuously panned condition with congruent and incongruent rotation directions. Multimodal physiological data including EEG, EOG, ECG, fNIRS, and eye-tracking are recorded. Behavioral data includes joystick measurements which reflect the subjective experience of SD. Outcome measures include vestibulo-ocular reflex (VOR) gain, alpha and theta band EEG activity, temporoparietal junction activation via fNIRS, nystagmus, SD duration and subjective severity, and perceived rotation direction. Preliminary data suggest stimulus-dependent differences in SD duration.

Speaker: Gamze Cengiz (Middle East Technical University Graduate School of Informatics)

316 Effects of Automatic Attentional Orienting and Cue-Target Similarity on Auditory Spatial Discrimination

Auditory spatial attention is shaped by both stimulus-driven orienting and perceptual interactions between sequential sounds. This study examined whether auditory cueing effects in spatial discrimination arise from attentional orienting or cue–target similarity. Eleven participants judged the direction of a small horizontal shift between two “buzz” sounds while maintaining central fixation. Spatial configuration (cue and target at one of three locations 0°, ±25°), cue type (similar “buzz” vs. dissimilar “noise” to target), and cue validity (valid vs. invalid, always not informative) were manipulated. Sensitivity (d′) was significantly higher for valid than invalid cues, indicating facilitation when cue and target were spatially aligned, while cue type had no effect on sensitivity. Also, response bias (criterion c) was strongly affected by cueing. Invalid cues produced a pronounced bias away from the cued location, especially for lateral targets. This bias was reduced when cue and target were dissimilar. These results indicate that, while cue–target similarity does not influence the attentional effects on discrimination in terms of d’, it does affect how biased the percepts are. Unexpectedly, response bias away from fixation was also observed in valid lateral spatial configuration. And this bias was increased for the dissimilar cue. These findings show complex interactions between automatic orienting, perceptual organization, and eye-gaze direction in spatial discrimination.[The research is supported by HORIZON-MSCA-2022-SE-01 Grant No. 101129903, APVV-23–0054, UPJS VVGS-2026-3871.]

Speaker: Yeganeh Modaresnia (P. J. Šafárik University)

317 Principles of auditory scene analysis emerge in computational audio separation

Auditory scene analysis describes the ability of the human auditory system to separate a mixture of sounds into the individual sources. The separation is based on bottom-up grouping information following Gestalt principles, such as harmonicity, onset synchrony and common modulation in amplitude and frequency, and learned top-down information. We investigated the separation mechanisms of a convolutional time-domain audio separation network (ConvTasNet). The network was trained for two-source separation on a variety of sounds: speech, environmental sounds, and music. The separation mechanisms were investigated by conducting a range of experiments on auditory scene analysis following classical experiments. The network learns to perform the separation and is able to separate the abstract experimental stimuli, thereby showing similar performance bounds as humans. The separation mechanisms exhibit the same principle organization rules as in the human auditory system: harmonicity, onset synchrony and common amplitude and frequency modulation. This suggests that machine learning of separation based on the statistics of sound stimuli results in similar mechanisms as have developed in humans.

Speaker: Bernhard U. Seeber (Technische Universität München)

A16.02 Reliable Characterization of Acoustic Absorption, Reflection, and Scattering: S114

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Marco Berzborn (Eindhoven University of Technology), Jonathan A. Hargreaves (Acoustics Innovation Institute)

318 A Stochastic Modal Energy Decay Model for the Estimation of Absorption

The measurement of Sabine's random incidence absorption coefficient as obtained from reverberation room measurements is well known to suffer from poor inter-laboratory reproducibility. This issue is often attributed to an insufficiently diffuse and distinctly non-uniformly dampened sound field within the laboratory, which are most prominent when the absorbing test specimen is mounted. Characteristic multi-exponential energy decay curves and corresponding multi-modal damping distributions are extensively reported already in the early literature on the topic. In recent years, multiple methods have been investigated to consider non-uniform damping either through explicit or implicit numeric modeling or when inferring the reverberation time of the empty and occupied laboratory.In this contribution we analyze the uniformity of damping using an energy based stochastic modal damping model originally introduced by Kuttruff. Discrete damping distributions are computed for a rectangular reverberation room with non-uniform damping based on the closed form solution for rectangular rooms. Finally, we investigate the suitability of the damping distributions of the occupied and empty reverberation room for the estimation of Sabine's random incidence absorption coefficient.

Speaker: Marco Berzborn (Eindhoven University of Technology)

Full paper (PDF)

319 Increasing reliability in tube measurements for characterization of porous materials

The three-microphone transfer function method has been a tube measurement method widely accepted for characteristic impedance and propagation coefficient of porous materials. When combined with the multiple-microphone method, the measurement frequency range can be extended to almost twice that of traditional measurements. However, discontinuities often emerge in the wavenumber spectrum at high frequencies. This problem is due to the inherent nature of the inverse cosine function. This work unwraps the complex-valued wavenumber spectrum with sequential Bayesian inference. The results demonstrate the reliability of the proposed method for broadband characterization of porous materials.

Speaker: Ziqi Chen (Rensselaer Polytechnic Institute)

Full paper (PDF)

320 Estimating Material Absorption Coefficients Using the Sound Field Scanning Method

The characterization of sound absorption typically relies on standardized methods such as impedance tube measurements or reverberation room testing, with associated sample size, mounting conditions, or test environment limitations. This paper investigates an alternative approach based on time and space resolved acoustic imaging of reflected sound fields.The proposed method uses a rotating linear microphone array implementing coherence scanning acoustic holography to measure the sound field generated by a pulsed acoustic source placed near the array. A material sample is positioned in front of the array, allowing capture of the sound reflected by the sample. By applying temporal windowing to isolate the reflected component and spatial filtering to focus on the reflection region, the reflected sound field is reconstructed and analyzed in the frequency domain.The spectrum of the reflected signal is compared to a reference spectrum obtained with a highly reflective surface under identical conditions. The spectral difference is used to estimate the frequency dependent absorption coefficient of the sample.The objective of this study is to evaluate the feasibility, and limitations of this approach for absorption estimation, particularly for mid- to high-frequency ranges. Preliminary considerations suggest that it may offer a flexible alternative for in-situ material absorption characterization.

Speaker: Antoine Decloux (Seven Bel GmbH)

321 Impact of Positioning Errors in the Two-microphone Method for Surface Impedance Measurements

The two-microphone method, as defined in the standard ANSI/ASA S1.18-2018, is used to determine the acoustic impedance of ground surfaces. To derive the impedance from the sound pressure ratio of the two microphones, the method relies on a specific geometric relationship between the sound source and the two receiving microphones. Therefore, we explore the effect of positioning errors within the measurement setup, i.e. in the alignment between source and receivers.First, an experiment is conducted to characterize the uncertainties in triaxial geometric positioning that occur during manual setup under practical measurement conditions. The resulting probability distribution is used as input for a Monte Carlo (MC) uncertainty propagation across the template method applying the variable porosity model for a sample material. Consequently, the uncertainty is analyzed for the level difference of the two microphone signals, which is a critical intermediate processing step. The microphone signals are computed by adding the geometrical paths of direct and reflected sound emitted by a point source using the Weyl–Van der Pol approximation for spherical waves. The MC simulation yields the resulting uncertainty estimates for the impedance.First results show that uncertainty arising solely from geometric setup variations is relatively small but strongly frequency-dependent. The uncertainty of the impedance is larger at low frequencies than at high frequencies. Positioning errors in height show to have the strongest influence on the output quantities.

Speaker: Jonas Heck (Institute for Hearing Technology and Acoustics)

Full paper (PDF)

322 In Situ Characterization of Non-Planar Sound Absorbers Using Physics-Informed Neural Operators

Accurate characterization of acoustic boundary properties is essential for reliable acoustic simulations, yet existing in situ methods remain limited by model assumptions, computational cost, and sensitivity to noise. In this work, a physics-informed neural operator framework is proposed for directly estimating the frequency-dependent acoustic surface admittance from near-field measurements of sound pressure and particle velocity. The proposed approach employs a deep operator network architecture to learn the underlying neural operator that maps measurement data, spatial coordinates, and excitation frequency to the corresponding acoustic field quantities, thereby enabling the simultaneous estimation of the complex-valued surface admittance spectrum. Notably, this formulation circumvents the need for an explicitly defined wave-propagation model or a computationally intensive numerical forward model. Physical consistency is ensured by embedding the governing acoustic relations, including the Helmholtz equation, the linearized momentum equation, and the Robin boundary condition, into the training process. The method is validated using synthetically generated data from a simulation model of a cylindrical absorber under semi free-field conditions. Results demonstrate accurate estimation of the complex-valued admittance over a wide frequency range from 250 to 5000 Hz. The proposed framework offers a neural network-based yet physically consistent alternative to conventional inverse methods, with particular advantages for the characterization of finite-sized and non-planar sound absorbers.

Speaker: Jonas M. Schmid (Technical University of Munich)

Full paper (PDF)

A17.05 Sound field estimation and reconstruction: S124

Conveners: Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology), Samuel A. Verburg (Technical University of Denmark)

323 Time-domain Behaviour of Spatial Aliasing Artefacts in Spherical Microphone Arrays Processing

Spatial aliasing in spherical microphone arrays (SMAs) arises from insufficient spatial sampling to capture high-order components, which are folded back into lower orders. This phenomenon is well understood: the frequency-independent spatial aliasing matrix characterises the mapping by which higher-order components are aliased into lower orders, as determined by the array’s sampling geometry. In addition, the mathematical expression of spatial aliasing also involves a frequency-dependent term determined by radial functions of different orders. This can be rigorously analysed by poles and zeros in the Laplace domain. Upon this, a closed-form expression for the corresponding time-domain spatial aliasing artefact is derived, which demonstrates that spatial aliasing introduces temporal spreading of higher-order Ambisonic signals measured with SMAs.

Speaker: Yueheng LI (Institute of Sound and Vibration Research)

Full paper (PDF)

324 Measurement of the Spurious Sound Field Generated by a Parametric Array Loudspeaker Using Exponential Sine Sweeps

Parametric array loudspeakers use arrays of ultrasonic transducers to generate highly directive audible fields through the non-linear properties of air in the ultrasonic range. They require very high ultrasonic pressure levels to generate the audible field, but these levels, in turn, lead to measurement inaccuracies produced by the microphone. Known as spurious sound, this measurement error can be circumvented by incorporating a filter in front of the microphone to reduce the ultrasonic levels. However, its effects have been studied mainly in the far field and at a few fixed measurement points. This paper explores this phenomenon by mapping the sound field using an automated robotic system, composed of a 6-degree-of-freedom robot mounted on a linear track. By recording exponential sine sweeps and deconvolving with the corresponding inverse filters, sound field maps for the audible and ultrasonic fields are obtained. Measurements were performed with and without the microphone filter to isolate the spurious sound. Results confirm a correlation between the ultrasonic field and the spurious sound, and show that spurious sound levels increase with the audible difference frequency.

Speaker: Augusto Fantinelli (Human-Environment Research group, La Salle - URL)

Full paper (PDF)

325 Sound Field Estimation in Continuous Time using Kernel Ridge Regression

Kernel ridge regression (KRR) has emerged as an effective approach to sound field estimation (SFE). By constructing the kernel and its associated reproducing kernel Hilbert space (RKHS) appropriately, the optimal KRR solution can be guaranteed to satisfy the wave equation. Such kernels and RKHSs have so far only been derived in the frequency domain and discrete-time domain. Therefore, in this paper a continuous-time kernel and associated RKHS is proposed, which enforces the wave equation by construction. The kernel function has a simple closed-form expression which is shown to have lower computational cost compared to similar discrete-time methods. A KRR-based method for SFE with the proposed model is then presented. The proposed SFE method is flexible, naturally supporting moving microphones and non-uniform sampling. The proposed method is evaluated using simulated data with both stationary and moving microphones, demonstrating the value of the approach in terms of both estimation performance and computational cost.

Speaker: Jesper Brunnström (Uppsala University)

Full paper (PDF)

326 Directional Mixture Models of Continuous Plane-Wave Distributions for Sound Field Estimation

Plane-wave-based sound field estimation commonly relies on discretising the plane-wave domain into a finite set of propagation directions, which can lead to high-dimensional models and unfavourable conditioning in sparse inverse problems. In this work, we investigate continuous directional latent models by parameterising the plane-wave distribution on the unit sphere with compact mixtures. We consider two component families within the same framework: von Mises–Fisher (vMF‍) components, which provide an intrinsic isotropic model on the sphere, and tangent-space Gaussian (TG‍) components, which provide explicit covariance control and analytically convenient local approximations. Inserting these models into the Herglotz plane-wave representation yields compact sound-field models whose atoms are tied to continuous directional distributions rather than to a fixed propagation grid. The unknown amplitudes and directional parameters are estimated from sparse pressure measurements using a variable-projection style procedure: amplitudes are refit by ridge least squares for fixed directional parameters, while directions and spreads or covariances are optimised iteratively. The comparison is motivated by the trade-off between intrinsic spherical fidelity for vMF mixtures and local anisotropic flexibility for TG. Numerical experiments in simulated room-acoustic fields compare reconstruction accuracy and parameter efficiency against ridge and sparse fixed-grid plane-wave estimation. The results show improved reconstruction at matched intrinsic parameter counts in several settings for both methods, with vMF mixtures giving the most consistent high-frequency gains, at the cost of iterative nonlinear fitting.

Speaker: Matthias Blochberger (KU Leuven)

Full paper (PDF)

A19.05 Sonification and visualization for acoustics: S143

Conveners: Katharina Gross-Vogt (IEM), Sandra Pauletto (KTH Royal Institute of Technology), Wolfgang Aigner (University of Applied Sciences St. Pölten)

327 Audio–Visual Redundancy and Complementarity in Sound and Music Contexts

Audio is frequently paired with visuals, and sometimes vice versa, for didactic, exploratory, and engagement purposes. This paper systematically reviews such practices across their main usage context: (a) the academic domain, including standard visualizations of sound and didactic examples of sonification (including auralization); (b) creative tools that generate sound from visual input; including visuals in music players and dedicated tools for sonifying images that are emerging but only sparsely adopted. Finally, (c) selected artistic works such as installations and live performances. While category (a) typically relies on rule-based mappings between audio and visuals, examples in (b) and (c) mostly employ more unsystematic, creative encodings. Our corpus of examples was assembled through a mixed-method approach combining literature and web searches with expert interviews.Across contexts, one modality (audio, visuals) is either used to be redundant (‘re’) or complementary (‘co’) to the other. To explore this relationship, we apply a new framework that characterizes redundancy and complementarity in audio–visual displays along the perceptualization path: from content; over encoding and physical representation; to user interpretation and interaction. For each aspect, we assess the degree to which visuals and audio are re/co and infer the use cases they best support.Academic contexts and creative tools often aim for redundancy but partly fail to produce perceptually equivalent audio–visual impressions. By contrast, artistic works typically seek complementarity. Our analysis reveals gaps in standardized methods and tools, which would be needed to support more inclusive and comprehensible practices, e.g., in didactics and audience engagement.

Speaker: David Adlberger (KUG)

Full paper (PDF)

328 Natural Language Interaction for Morphology-Aware Audification

Audification maps data directly to sound using uniform time scaling. We previously introduced a rule-based framework that classifies time signals into one of 15 morphology classes and assigns class-specific time-scaling factors, evaluated across 591 real-world signals from 8 scientific domains and 251 synthetic control signals. That framework only ever produces a class label, a confidence score, and a scaling factor, telling a user nothing about what to listen for, offering no way to feed domain knowledge back in, and, because the tool is a conventional visual interface, remaining inaccessible to anyone who cannot see it.This paper adds a natural-language interaction layer built on a large language model (LLM) strictly downstream of the pipeline. The LLM receives only structured metadata, never the raw signal, and uses it to generate summaries calibrated to a user-supplied domain description, support reclassification and parameter adjustment with trade-offs stated up front, and answer open-domain questions about the signal. Because the entire interaction is text-based, it lowers the barrier for domain and sonification non-experts while simultaneously opening the tool to blind and low-vision (BLV) users via standard screen readers, with no plot or slider required. Our own survey of 57 sonification tools found only 9 explicitly built for BLV users, making this screen-reader-native workflow unusually direct for the field. Every LLM claim traces back to a number the pipeline already computed, so accessibility is gained without sacrificing analytical grounding. We evaluate using open-weight EU-based models (Mistral API and Ollama) to keep the system reproducible and free of proprietary dependencies.

Speaker: Sofia Vallejo Budziszewski (IEM)

Full paper (PDF)

329 The Role of Sound and Auditory Displays in Telescope Control Rooms: A Pilot Study

Astronomers are often stereotyped as gazing at the stars, but today they receive digital data from remote observatories. One of the most advanced and complex is the Very Large Telescope (VLT) in Chile, managed by the European Southern Observatory (ESO). Multiple streams of information converge in the VLT control room, where telescopes and instruments are managed, with operators handling numerous tasks on dozens of screens in a challenging environment, marked by harsh geography, intense work rhythms, and isolation. We present the results of a pilot study aimed at assessing the current use of sound in the VLT control room. Based on the analysis of questionnaires and interviews conducted with ESO personnel, we identify the key characteristics of the current VLT soundscape, as well as opportunities for improved design strategies.

Speaker: Sara Lenzi (University of Deusto, Ikerbasque Basque Foundation for Science)

Full paper (PDF)

330 Designing Personalised Sleep Sonification

This paper presents a participatory and personalised approach to the sonification of sleep EEG data. While existing sleep technologies primarily rely on visual and abstract representations, this work explores how sound can be used as a meaningful medium for engaging with sleep. An in-depth, single-participant, iterative design study was conducted combining EEG data collection with three one-hour interviews and progressive design refinements. Through this process, the participant actively contributed to shaping how their sleep was represented through sound, influencing both the selection of sound materials and their mapping to sleep features. This exploratory study shows that sound preferences, in this context, are shaped by both acoustic properties and personal associations, and that iterative engagement is important in understanding nuanced expectations around continuity, narrative flow, duration, and uses of the sonification. The resulting sonification was perceived as engaging, reflective, and personally meaningful, though some limitations remained in clarity of sleep stages. These observations point to the effective role of participatory approaches in designing personal sonifications of data and suggest that future personalisation systems for sound-based sleep representations should prioritise user involvement, contextual meaning, and experiential qualities.

Speaker: Abhishek Choubey (KTH Royal Institute of Technology)

Full paper (PDF)

331 Molecular Sonification as Sound-Driven Design Methodology: An Open, Evaluation-Ready Glycine Stimulus Set with Visual Representation

Sound-driven design requires methods that transform complex source data into auditory stimuli in ways that remain transparent, reproducible, and suitable for evaluation. This contribution presents a workflow for generating listening-ready stimuli from molecular vibrational structure and examining how specific design decisions influence the resulting auditory experience. The primary contribution is methodological rather than molecular: glycine serves as a compact demonstration case because its vibrational features are well documented and readily traceable throughout the workflow.Starting from an IR/Raman peak list, vibrational features are mapped to audible carriers through explicitly documented frequency and amplitude transforms. Temporal structure is then introduced through controlled envelope design. Two envelope families are provided: an evaporation-style reference envelope and a multi-phase kinetic-inspired envelope used as a design hypothesis for exploring how alternative temporal trajectories influence listening impressions. To support transparent comparison, the framework also generates envelope-swapped, time-scrambled, and loudness-matched controls. Together, documented mappings, controlled stimulus variants, and reproducible comparison conditions provide an objective basis for evaluating design choices.The workflow is intended for applications where the relationship between source data, design decisions, and perceptual outcomes must remain visible and auditable. Potential uses include sonification practice, product sound-quality research, where envelope design can be assessed as a controllable sound-design variable, and educational settings in which students can modify mapping parameters and evaluate their audible consequences. A synchronized visual representation supports design iteration and teaching activities. All assets are released with metadata, mapping specifications, and integrity records. The resulting contribution is a reusable methodology for generating, comparing, and evaluating auditory stimuli while maintaining traceability from source data to listening outcome.

Speaker: MM AbdAlhamid (Independent Researcher)

A20.08 Perception, Acoustics, and Transmission of Speaker Impact: S149

Conveners: Oliver Niebuhr (University of Southern Denmark), Sara Pearsell (University of Southern Denmark), Ingo Siegert

332 Parliamentary Recordings as a Resource for Speech Research in Less- Resourced Languages

Parliamentary recordings are a well-known resource for obtaining large-scale data for automatic speech recognition. In this contribution, we argue for a much broader use of parliamentary recordings and their accompanying official transcripts, rich speaker metadata, and generally non-problematic privacy and copyright status, especially in the context of less-resourced languages. Beyond speech recognition, such resources provide unique opportunities for large-scale research on acoustic, linguistic, prosodic, and discourse-level phenomena in semi-spontaneous speech. To illustrate these opportunities, we showcase ParlaSpeech, a multilingual collection of more than 6,000 hours of parliamentary recordings aligned with official transcripts in four Slavic languages: Croatian, Czech, Polish, and Serbian. The resource integrates several layers of linguistic and speech-oriented annotation. In addition to extensive parliamentary and speaker metadata inherited from the ParlaMint project, the corpus includes Universal Dependencies syntactic annotation, sentiment and topic annotations derived from the textual modality, and filled pause and primary stress annotations derived from the speech modality. The resulting infrastructure supports the study of interactions between prosody, acoustics, syntax, discourse structure, and speaker behavior across languages. We present several completed and ongoing research directions enabled by this infrastructure. Previous work has explored relationships between acoustic variables and sentiment expression, as well as cross-linguistic patterns in the frequency and duration of filled pause production. Ongoing analyses investigate patterns of primary stress across speakers with varying sociodemographic background, and the syntactic environments associated with occurrence of filled pauses. Future work will extend the collection with additional languages and annotation layers, opening up additional opportunities for under-researched languages.

Speaker: Nikola Ljubešić (Jožef Stefan Institute)

333 Charisma in a few seconds: Prosodic Prominence and the Perception of Vocal Charisma in Short Conversational Speech

This study examines the relationship between prosodic prominence and perceived charisma in conversational speech. We conducted a perception experiment with 92 speech stimuli drawn from a corpus of spontaneous interactions and obtained ratings of charisma, like-ability, self-confidence, and passion from 30 listeners. First, we analyzed the interrelations among these perceptual dimensions and found that charisma ratings of spontaneous conversation excerpts are most strongly associated with perceived likeability, while also showing high correlations with self-confidence and passion. Second, we compared human ratings with prosodic charisma scores derived from the PICSA model, originally developed for longer segments of public speaking. Despite the substantially shorter duration of our stimuli, PICSA scores showed strong correlations with the listener ratings, demonstrating robust predictive validity under conditions of limited temporal input - and for spontaneous conversation excerpts. Third, we examined the distribution of prosodic prominence within utterances, and found that the ratio of prominent words per utterance significantly shaped both human and PICSA-based charisma ratings. Articulation rate, despite influencing PICSA scores, did not contribute to perceived charisma. These findings open new perspectives on the temporal dynamics of charisma perception and the role of prominence and rhythm within the broader set of acoustic-prosodic cues to charismatic speech.

Speaker: Oliver Niebuhr (University of Southern Denmark)

Full paper (PDF)

334 The company we keep: Prosodic similarity within and across organizations

Do speakers from different organizations sound systematically different? Using the COMPASS corpus of more than 500 L2 English business presentations from 18 companies across 10 industry sectors, we analyzed 28 acoustic-prosodic features of speaking style. Linear discriminant analyses and MANOVA reveal clear non-random clustering of prosodic patterns. Both company membership and industry-sector influence speaking style, but company-level effects are consistently stronger. Tempo, vocal effort, formants, and pausing are the most influential company predictors, particularly in organizations where employees interact frequently in stable teams or social-service contexts, while pitch variation becomes more relevant at the sector level, particularly in communication-intensive professions.

Speaker: Oliver Niebuhr (University of Southern Denmark)

Full paper (PDF)

335 Subjective and objective employability for degraded speech in video-mediated job interviews

The quality of audio transmission in video conferencing systems can influence not only speech perceptionbut also higher-level social judgments in communication contexts such as hiring decisions following jobinterviews. The present study investigated how degradations in speech quality affect perceived employability and whether objective audio quality models capture these effects.The audio streams of AI-generated job interview videos were manipulated to create three levels of audio quality (high, medium, low). Participants evaluated the employability of the speakers, while thecorresponding audio signals were analyzed using two objective audio quality models, ViSQOL and GPSMq.Results showed a clear effect of audio quality on employability ratings, with substantially lower ratings inthe low-quality condition compared to medium and high quality, and no significant difference betweenmedium and high quality. ViSQOL reflected these differences by assigning lower scores to degraded signals, whereas GPSMq showed limited sensitivity, including a floor effect in the medium-quality condition.Correlation analyses revealed only weak associationsbetween model outputs and employability ratings, indicating that current audio quality models capture only a limited portion of the variance in social judgments.Overall, the findings highlight that audio quality contributes to employability judgments in video confer-encing contexts, but does not fully account for them.

Speaker: Julia Schütze (UniDistance Suisse)

Full paper (PDF)

336 The Audience Reactor: An Embodied Sparring Partner for Public- Speaking Training

Public-speaking skills comprise a set of interrelatedabilities, including vocal delivery, linguistic appropri-ateness, and audience-oriented communication. How-ever, this multidimensionality makes it challenging toprovide individualized, high-quality feedback, particu-larly within time-constrained curricula and in contextswhere instructors lack specialized training in oral com-munication pedagogy.To address this problem, the Audience Reactor wasdeveloped. The Audience Reactor is a humanoid-likerobot that gives feedback during the user’s speech,allowing them to actively improve and see their per-formance based on the robot’s reaction. The reactionsinclude facial expressions, body colour changes, mo-tion, and sound cues and icons indicating the speechparameters most important to improving the over-all speech quality. The robot is designed to give theimpression of actively listening to the user, with thefeedback being playful to encourage its use. The algo-rithm from which the feedback is derived is a speechanalysis algorithm provided by the company AllGo-odSpeakers. A web application allows the user to seetheir progress and a more detailed speech analysisthrough the saved results from previous sessions.

Speaker: Idun Elvira Christensen (University of Southern Denmark)

Full paper (PDF)

A21.04 Railway noise and vibration: S322

Conveners: Karoline Alten (AIT Austrian Institute of Technology), Christopher Knuth (Empa), Slimane Ouakka (Empa)

337 RailNoise: Future-Proofing Norwegian noise calculations with new source data for Nord2000 and Cnossos-EU

The primary objective of the RailNoise project is to enable reliable railway noise calculations using the calculation methods Nord2000 and CNOSSOS-EU. This requires gathering source emission data for both methods representing all operational train classes currently in regular use, and addressing limitations within standard CNOSSOS-EU, which lacks the 1/3-octave band resolution and maximum sound level (Lmax) results required by Norwegian noise regulations. The noise measurements are based on ISO 3095/NS 8177 but adapted for measuring in-service trains on operational railway lines. Computational back-propagation techniques are applied to account for locally varying ground impedance and to derive absolute sound power levels. Based on measured data, rails and wheels are classified by surface quality, for use in noise calculations.This paper presents the measurement methodology, the back-propagation process, and the process of converting the measured emission data into source data.

Speaker: Leo Heggem Hauge (SINTEF)

Full paper (PDF)

338 Semi-virtual assessment of cab signals in trains

Measuring audible signals in train driver’s cabs currently lacks a normative measurement procedure. New rolling stock must comply with the TSI LOC&PAS which requires the signal level to exceed the background noise level in the cab by 6 dB at maximum speed. Compliance can be demonstrated by direct measurement of the sound pressure level of each sound in the driver’s cab in accordance with EN ISO 3381. The procedure is repeated until a satisfactory signal level is reached. A statement as to whether this procedure is the most appropriate to assess audibility is beyond the scope of the contribution. The purpose is to introduce a new semi-virtual approach that can be summarized as follows•Measurement of source-receiver impulse responses (IRs) in a driver’s cab;•Time-convolution of the measured IRs with the audible signals time series. The input signals for the procedure are any acoustic signals to be reproduced in the driver’s cab. The outcomes are the calculated signals at the receivers’ positions which can be played back and/or level adjusted in post-processing rather than on-the-fly. Initial findings confirm that the semi-virtual approach, within the scope of audibility assessment, yields consistent results with direct measurement. Practical guidance on concerns such as impulse response measurement, level adjustment and system requirements is provided. To conclude, this work aims to provide an innovative alternative to repeated measurements of many warning signals, in turn facilitating a streamlined audibility assessment.This presentation is based on preliminary work carried out by the CEN/TC 256/SC 4/WG 3 “Acoustics” working group, which is responsible for drafting a European standard for evaluating audibility of warning signals in the driver’s cab.

Speaker: Domenico Tallarico (Stadler Rheintal AG)

Full paper (PDF)

339 Audiovisual investigations of noise abatement measures on their effect on annoyance from rail traffic noise

Rail traffic is considered a sustainable mode of transport for people and goods. Nevertheless, rail traffic contributes significantly to noise pollution in Europe. Previous studies suggest that established A-weighted sound pressure level metrics describe the evoked noise annoyance by railway noise only to a limited extent. In addition to psychoacoustic factors, non-acoustic factors (e.g., visual aspects) can also influence noise annoyance. The increased availability of immersive audiovisual recording and playback techniques enables the investigation of such factors in laboratory experiments with high audiovisual plausibility. To this end, we carried out audiovisual train pass-by measurements along two different railways with noise abatement measures (earth embankment, noise barrier) in Germany and Austria. The employed sensor setup was based on EN ISO 3095 and extended with an artificial head, a 3rd-order Ambisonics microphone and a 360-degree video camera. Selected pass-by recordings were used in audiovisual experiments to rate the noise annoyance according to an ISO/TS 15666 scale along with relative magnitude estimation tasks. Besides an investigation of the influence of the noise abatement measures in an audiovisual context, the influence of the playback mode was also examined. While the noise abatement measures expectedly led to a reduction of the annoyance ratings, no clear effect of the presentation mode (audio vs. audiovisual) on noise annoyance was observed. This work contains parts from the research project „Perception-based evaluation of railway noise abatement measures “, initiated by the German Centre for Rail Traffic Research at the Federal Railway Authority, and contributes to the acceptance of rail traffic.

Speaker: Martin Czuka (AIT Austrian Institute of Technology)

340 Perceptual Impact of Rolling Stock Tonality on Railway Passengers and Nearby Communities: In situ Studies

Tonal noise from rolling stock is addressed in railway through dedicated indicators, but its perceptual relevance in real operation is still difficult to quantify. This paper reports in-situ perception studies performed by Alstom and ADIF in collaboration with the Instituto de Biomecànica de València. The work was carried out within the European project FP4 Rail4Earth (Grant Agreement ID: 101101917).Surveys were conducted with passengers during revenue service and with residents living near railway stations, in situations presenting different tonal characteristics related to traction and auxiliary systems. Preliminary steps, including a one week online passenger diary and pilot surveys, were used to define field protocols suited to realistic environments.Large participant samples and structured questionnaires were used to assess comfort and noise-related annoyance. In these real life conditions, no statistically significant relationship was observed between the tonal character of a noise and its perceived annoyance: while tonal sources were associated with low annoyance, others sources (e.g. train passengers or warning signals such as horns) were perceived as highly disturbing for the interior zones or the exterior locations investigated.Practical considerations are proposed. During acceleration and braking, where traction noise is inherently tonal and time varying, target setting should primarily rely on overall noise level. For other interior conditions, tonality may be used as a secondary refinement via a moderate penalty added to overall level requirements when high comfort is targeted. For exterior noise, no clear tonal influence was observed, the results highlight the importance of identifying dominant noise sources and contextual factors in order to enable targeted mitigation strategies.

Speaker: Joan Sapena (ALSTOM Transport SA)

Full paper (PDF)

341 Frequency‑Modulated and Single‑Tone Excitation to Reveal Vibro‑Acoustic Nonlinearities in Loosened Bolted Joints

Preload loss in bolted joints alters structural stiffness,damping, and nonlinearity. This paper evaluates ashaker-based vibro-acoustic method on a rail-vehicle-inspiredbeam assembly. The central M10 fastener(S2) in a three-bolt lap joint between two steel beams,representing bolted interfaces on carbody and roof-mountedcomponents, was tested at 0%, 20%, 40%,and 80% preload. A triaxial accelerometer measuredthe structural response, while a microphone placed300 mm from S2 recorded the radiated sound as a secondarysensing channel. Sine-sweep and narrow-bandexcitation identified a dominant resonance near 130Hz. Under 130 Hz single-tone excitation, the loosestate produced additional high-frequency spectral componentsin the accelerometer response. Frequency-modulated(FM) excitation from 125 to 135 Hz furtherseparated the preload states through harmonic-bandlevels relative to the carrier: the loose-to-80% differenceswere 17.5 dB in the second-harmonic band and36.5 dB in the sixth-harmonic band.

Speaker: Berkay Kullukcu (FG Maschinendynamik und Akustik, Technische Hochschule Wildau)

Full paper (PDF)

A16.03 Comfortable offices: from design to perception: S115

Conveners: Valtteri Hongisto (Turku University of Applied Sciences), Giulia Fratoni (University of Bologna)

342 Prediction of Listening Effort in Meeting Rooms based on Objective Metrics

While traditional room acoustic metrics successfully evaluate speech intelligibility, recent research indicates that high intelligibility does not guarantee low listening effort. In modern meeting environments, signal-to-noise ratios are typically sufficient for high intelligibility due to short communication distances and adjustable playback levels. Consequently, this study reevaluates acoustic metrics for predicting listening effort under equal-loudness conditions to isolate perceptual effects beyond mere loudness. The mean opinion scores derived from a subjective listening test were regressed against established acoustic parameters using ordinary least squares.Leave-one-out cross-validation revealed that a linear combination of reverberation time and direct-to-reverberant energy ratio outperformed classic metrics such as clarity and the speech transmission index.

Speaker: Leon Kaiser (University of Music and Performing Arts)

Full paper (PDF)

343 Dynamic Acoustic Management in Open-plan Offices: a Case Study on Activity-driven Mobile Partitions and Adaptive Ceiling Panels

Open-plan offices frequently struggle to balance speech privacy and speech intelligibility, which can degrade productivity and occupant well-being. This study investigates how acoustic conditions in an open-plan office can be improved using adaptive acoustic elements, including mobile partitions and height-adjustable island-type ceiling panels, both manually and robotically operated. These elements can be reconfigured to support different activity types such as collaboration, communication and focused work. Room acoustic measurements and simulations were performed to evaluate performance against ISO 22955:2021 - Acoustic quality of open office spaces. Activity types were represented through 21 different layout configurations with multiple acoustic element set-ups and combinations. Acoustic performance was assessed using key acoustic parameters related to speech like DA,S, D2,S, Lp,A,S,4m and reverberation time. The results show that changing the position, height and configuration of acoustic elements has a measurable impact on these parameters. Some configurations achieved higher speech attenuation values and lower reverberation times, while others produced different balances between spatial decay and reverberant conditions. The findings also highlight the influence of ceiling height on the effectiveness of partitions. These results provide different insights for designing activity-based acoustic environments that can dynamically respond to changing office use.

Speaker: Pascal van Dort (Rockfon (ROCKWOOL B.V))

Full paper (PDF)

344 Assessment of the Relative Impact of Noise, Density, and Visibility on Occupant Satisfaction

Background: While the acoustic environment is widely recognized as a major source of annoyance in open-plan offices, satisfaction with the physical workspace is inherently multidimensional. The workplace well-being of an individual is influenced by various factors, including noise levels, lighting conditions, thermal comfort, and number of people in the workspace. The objective of this study is to assess the relative impact of various environmental factors on the overall satisfaction of occupants in open-plan offices. Methodology: A large-scale field study was conducted in thirty-one open-plan office spaces, covering a range of activities such as call centers, individual workspaces, and collaborative workspaces. The research integrated objective measurements (sound levels, reverberation time, speech attenuation, lighting levels, temperature, and humidity, as well as real-time occupancy rate data) with subjective observations. A detailed questionnaire was administered to approximately 450 occupants to assess their perceptions of acoustics (e.g., overall noise, speech noise, equipment noise), lighting, thermal environment, visual comfort, and perceived density. Results: Initial analyses, conducted using regression models based solely on subjective data, indicate that while acoustics remains a determining factor in satisfaction, it does not act in isolation. It appears that perceived density and visibility are both key factors in predicting overall environmental satisfaction. The main results will be presented at the conference.

Speaker: Laurent Brocolini (INRS)

345 Longitudinal Assessment Of Acoustic Conditions In Open-Plan Doctoral Offices: Integration Of Objective And Subjective Data

Open-plan offices are widely adopted for their flexibility and space efficiency, yet they often present critical acoustic issues affecting occupants’ comfort and performance. This study presents a longitudinal investigation of acoustic conditions in shared offices used by doctoral researchers, combining objective measurements and subjective evaluations over an extended monitoring period.The study was conducted across four monitoring phases, for a total duration of 20 weeks, enabling the observation of temporal variations under real working conditions. Each phase corresponds to a different acoustic treatment configuration, including variations in workstation screens and the introduction of sound-absorbing elements, allowing a controlled comparison between alternative design solutions.Objective data were collected through continuous and in situ measurements. These include standard acoustic parameters such as reverberation time (T30), speech transmission index (STI), spatial decay rate of speech (D2,S), and sound pressure levels, as well as additional monitored variables such as occupancy (number of people) and environmental conditions. This approach enables a more comprehensive characterization of the acoustic environment beyond single-time measurements performed in each phase.In parallel, subjective data were gathered using structured questionnaires administered to occupants, addressing perceived acoustic comfort, speech disturbance, and overall satisfaction.The longitudinal design allows the investigation of how different acoustic treatments and varying occupancy and environmental conditions influence both measured performance and user perception over time. By integrating objective and subjective datasets, this research aims to provide a robust framework for evaluating acoustic quality in open-plan offices, supporting evidence-based and user-centered design approaches.

Speaker: Riccardo Caradonna (Politecnico di Torino)

A03.02 Measurements and experimental analysis in building acoustics: P461

Conveners: Martin Schneider (Hochschule für Technik Stuttgart), Selina Vavrik-Kirchsteiger

346 Evaluation of Heavy-weight Floor Impact Sound using Tapping Machine

Floor impact sounds are mainly divided into light-weight and heavy-weight, and different measurement equipment and evaluation methods are required according to each characteristic. Using both types of impact sources in a real experimental environment is time-consuming and resource-consuming, and there are limitations in terms of repeatability and reproducibility. Rubber balls, which are heavy-weight impact sources, may experimental errors increase depending on the impact position or drop angle due to the free-fall method. Accordingly, there is a need for a simpler and more reliable measurement method. To simplify this measurement method, this study aims to examine the possibility of evaluating heavy-weight characteristics including low frequency bands (50, 63, 80 Hz) by excluding existing rubber ball impact sources and using only tapping machines. By analyzing the low frequency band response of the tapping machine, we want to explore its applicability as a heavy-weight impact source and present a more practical and consistent evaluation technique. As a result of analyzing the single numerical evaluation amount of the existing rubber ball impact source and tapping machine focusing on the low frequency band, which is the heavy-weight evaluation frequency, the correlation coefficient between the two impact sources is not high. It can be seen that the correlation coefficient for each frequency is very high. This is thought to be a difference according to the method of evaluation rather than a problem with the impact source. If only an appropriate evaluation method can be proposed, evaluation of the low frequency band using a tapping machine is likely to be valid.

Speaker: JINYUN CHUNG (Korea Conformity Laboratory)

Full paper (PDF)

A04.00 Education, Public Outreach and History in Acoustics: P514

Conveners: Lukas Aspöck (IHTA, RWTH Aachen University), Iván Herrero-Durá (SoundPLAN GmbH), Michael Buba (Technical University of Munich)

347 EPSRC Centre for Doctoral Training (CDT) in Sustainable Sound Futures

The EPSRC Centre for Doctoral Training in Sustainable Sound Futures is delivering PhD training on an unprecedented scale for Acoustic Engineering. Over five cohorts, the CDT will support more than seventy interdisciplinary researchers addressing the societal, environmental and technological challenges of sound and noise. The programme is a collaboration between the Universities of Salford, Bristol, Sheffield and Southampton with over 50 partners from industry, third and public sector. The Centre integrates four research themes: new sound sources, perception of humans and other animals, machine learning for acoustics, and positive uses of sound. Alongside their research, students benefit from a cohort-based curriculum that develops broad, interdisciplinary working. By combining world class facilities with a bespoke taught module, residential workshops, masterclasses, journal clubs and summer schools across partner universities, the CDT develops students’ technical expertise and transferable skills.The CDT programme emphasises measurement, simulation, signal processing, statistics, psychophysics, data management and ethical practice, and is designed to foster entrepreneurship, policy engagement, public communication and inclusive research cultures. Placements and industry partnerships accelerate translation of research into standards, products, and policy across national, regional and international sectors. Governance, supervision and partner integration are structured to ensure PhD project relevance and impact, with substantial partner funded studentships. By aligning fundamental research with applied innovation and industry needs, the CDT aims to produce research leaders capable of advancing metrics, mitigation strategies and design principles that promote quieter, healthier and more sustainable soundscapes across energy, transport, built environment and health sectors.

Speaker: Antonio Torija Martinez (Acoustics Research Centre)

348 The WAVES Erasmus Mundus Joint Master Degree: Experience and Impact of a European Programme in Acoustical Engineering Education

The Erasmus Mundus Joint Master Degree (EMJMD) WAVES – Waves, Acoustics, Vibrations, Engineering and Sound is a two-year (four-semester) international Master’s programme in Acoustical Engineering, approved for funding in 2019. Its main objective is to provide students with advanced scientific knowledge and engineering skills to address acoustic problems in both academic and industrial contexts, while responding to the evolving demands of the global job market. To date, EMJMD WAVES has successfully graduated three cohorts of students, and a fourth cohort is currently engaged in Master’s thesis development.The programme is jointly implemented by four internationally recognised European higher education institutions: the Aix-Marseille Université and École Centrale Marseille (France), the Universidade de Coimbra (Portugal), and the Universitat Politècnica de València (Spain), and is supported by an extended network of associated academic partners and industrial partners. This structure enables a fully integrated, mobility-based curriculum that promotes educational expertise and the development of high-level transversal competences.The EMJMD WAVES curriculum was designed as a common and innovative programme, aligned with contemporary societal and economic challenges faced by acoustical engineers. It combines strong foundations in core acoustics with emerging topics and pedagogical approaches, addressing application areas such as architectural acoustics, environmental noise, ultrasound-based monitoring of structures and processes, and noise and vibration control in transport. The experience gained from the first cohorts demonstrates the capacity of Erasmus Mundus joint degrees to foster excellence, innovation, and harmonisation in European acoustics education, preparing graduates for competitive doctoral studies or immediate industrial employment.

Speaker: Paulo Amado-Mendes (University of Coimbra, Dep. Eng. Civil)

Full paper (PDF)

A06.01 Microphone and MEMS transducers: P430

Conveners: Dominik Mayrhofer (Graz University of Technoloy), Petr Honzík (Czech Technical University in Prague)

349 Reduction of nonlinear distortion in condenser microphones: a unified approach for single and dual backplate designs

Nonlinear distortion is an inherent limitation of condenser microphones, originating primarily from the nonlinear change of capacitance between the diaphragm and the backplate. In single-backplate designs, this effect leads predominantly to second-order harmonic distortion, which can dominate the total harmonic distortion at high sound pressure levels. In contrast, dual backplate configurations inherently suppress even-order nonlinearities due to their differential structure; however, residual distortion remains as a result of geometric and electrical asymmetries between the two sides of the diaphragm.This contribution presents a unified approach for describing and reducing nonlinear distortion in single- and dual-backplate condenser microphones, applicable to classical and MEMS designs. The approach is based on low-order nonlinear models of the microphone output, derived from the capacitance variation, which reveal a common structure of the distortion mechanisms despite the differences in design.Based on this unified description, we propose a simple signal-domain nonlinear distortion correction method. The technique relies on a single parameter that can be obtained either from physical microphone properties or from a straightforward measurement procedure, and can be implemented in analog or digital signal processing with negligible computational cost.The results demonstrate that the proposed method provides substantial reduction of harmonic and intermodulation distortion in single-backplate microphones, while also enabling further improvement in dual backplate designs beyond their intrinsic suppression of even-order nonlinearities. The presented approach offers a practical tool for improving the linearity of both single- and dual-backplate condenser microphones.

Speaker: Petr Honzík (Czech Technical University in Prague)

A10.05/A16.11 Advanced Characterization of Acoustic Materials: P481

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Luc Jaouen

350 Influence of binder in the acoustical behaviour of samples prepared from used cigarette filters

Cigarette filters constitute a significant fraction of persistent urban waste, raising growing environmental concerns due to their widespread disposal and limited recycling pathways. In this context, their potential reuse as porous materials for acoustic applications have attracted increasing attention.This work explores the role of different binding agents in the development of acoustic materials derived from used cigarette filters. The incorporation of binders is considered as a strategy to improve the mechanical integrity and handling of the samples, which are typically affected by the intrinsic heterogeneity of the raw material and the processing methods. At the same time, the addition of binders may alter the internal structure of the material, potentially influencing its sound absorption performance.Several types of binders are evaluated, with particular attention to their impact on key acoustic properties. The study aims to identify trends relating binder type and concentration with the resulting acoustic behaviour, as well as to assess the balance between structural stability and acoustic efficiency.Preliminary observations suggest that the selection and dosage of the binder play a critical role in determining the final performance of the material. Further work will complete the experimental assessment and consolidate the comparative analysis.

Speaker: Guillermo Rey-Gozalo (Universidad de Extremadura)

Full paper (PDF)

351 Measurement Verified Biot–JCAL Modelling of Wool Felt

Wool felt is a fibrous material composed of randomly oriented and mechanically entangled elastic fibres. Despite its widespread use in acoustic engineering, the sound absorption mechanisms of wool felt remain insufficiently understood. In this study, a Biot-Johnson-Champoux-Allard-Lafarge (Biot-JCAL) poroelastic modelling approach is employed, in which the felt is represented as an equivalent fluid with an elastic frame formed by the fibres. The model is calibrated using experimentally determined physical, non-acoustic, and geometrical parameters measured by the authors. Impedance tube measurements of sound absorption and transmission loss are further used to validate the proposed model. Both the measured and simulated absorption curves exhibit a pronounced low-frequency absorption peak that cannot be associated with any characteristic microstructural length scale of the material. This feature cannot be reproduced using rigid-frame equivalent-fluid models such as JCA or JCAL, demonstrating the importance of accounting for frame elasticity in wool felt. The presence of this resonance-like absorption behaviour suggests that wool felt may be regarded as a naturally derived acoustic metamaterial. Good agreement is observed between the experimental measurements and modelling results. The present study contributes to a deeper understanding of the acoustic behaviour of wool felt and related non-woven textiles. The reported measurements and validated model provide a basis for the tailored design and optimisation of sustainable sound-absorbing materials for engineering and noise-control applications with enhanced acoustic performance.

Speaker: Maria Miranda Vuin (Tallinn University of Technology)

352 Obtaining Poroelastic Parameters of Porous Materials from Impedance Tube Measurements

The accurate characterisation of the acoustic and elastic parameters of a porous material based on impedance tube measurements presents challenges, in particular, because the specimens’ diameters can deviate from the nominal inner diameter of the impedance tube. If the specimen’s diameter exceeds the nominal diameter of the tube, it is compared and therefore the material parameters are modified. The elastic solid phase of the diphasic poro-elastic material, is modeled using Biot’s model with the parameters mass density, Young’s modulus, Poisson’s ratio, and loss factor. In the fluid phase, viscous and thermal dissipation is modeled with the JCAL-model with the parameters static airflow resistivity, open porosity, tortuosity, static thermal permeability, and viscous and thermal characteristic lengths. The model is restricted to plane waves at normal incidence, enabling an analytical coupling of the two models. In this contribution, the parameters of the diphasic poro-elastic Biot-JCAL model are fitted to impedance tube measurements of the FOAM 02-database. This database contains 864 measurements of the sound absorbtion coefficient obtained in an impedance tube with (±2% diameter deviation). An inverse parameter characterisation model is implemented in Python. The results highlight the effect of diameter inaccuracies on the visco-thermal and elastic parameters of the Biot-JCAL model.

Speaker: Florian Kraxberger (Graz University of Technology)

Full paper (PDF)

A11.00 Musical Acoustics: P470

Conveners: Vasileios Chatziioannou (Univ. Music and Performing Arts Vienna), Charalampos Saitis, Christoph Reuter (Universität Wien)

353 Does the Finish Matter if No One is Looking? Visual Bias in Trumpet Testing

Surface finish is frequently discussed by brass players as a determinant of sound and playability, yet such judgments may be influenced not only by acoustical differences but also by visual information and prior expectations. This exploratory pilot study examines whether perceived differences between three nominally comparable trumpets are larger when their surface finishes are visible and whether corresponding differences are reflected in the acoustic signal. Ten trumpet players performed identical material on three B-flat trumpets of the same model with different finishes (raw brass, silver-plated, lacquered) in blind and open conditions. After each trial, participants rated 13 perceptual dimensions. Across 11 of the 13 dimensions, descriptive differences between the instruments were larger in the open condition than in the blind condition. In the acoustic feature analysis, variability was dominated by residual and participant-related components, whereas the finish-associated component was small. These pilot findings are consistent with the possibility that visual information contributes to instrument evaluation. However, because of the small sample, the pairing of one instrument with each finish, and the exploratory nature of the analyses, the results should not be interpreted as demonstrating a causal effect of surface finish or visual expectation.

Speaker: Michael Oehler (Osnabrück University)

Full paper (PDF)

354 Limits of geometric pluck calibration for repeatable robotic excitation of plucked string instruments

Robotic excitation systems have become increasingly relied upon for acoustic studies on string instruments, utilizing geometric (open-loop) calibration: moving the exciter to fixed coordinates. While such calibration is highly repeatable within a single session, comparative metrology inherently requires remounting instruments. This study shows that open-loop calibration is insufficient for longitudinal testing of plucked string instruments. Remounting and mechanical creep in the setup introduce sub-millimeter shifts which, given the guitar's high sensitivity to initial plucking conditions, cause substantial reproducibility errors.To quantify this remounting issue, a simplified electric guitar underwent 9 complete remounting iterations on a cartesian robotic platform. Remounting precision was verified using caliper measurements across 8 points. A plectrum swept through 17 fixed Z-axis depths (10 plucks per step). Magnetic single-coil and humbucker pickups were used to capture the string signal. We extracted a set of spectrotemporal metrics (e.g., RMS Energy, Spectral Centroid) to analyze acoustic variability across iterations.The contribution is twofold: a quantitative error map showing how acoustic feature variability depends on nominal pluck depth across remounts, and identification of the static-friction-to-release transition as the dominant mechanism driving the irreproducibility. The dataset reveals high inter-mount variability, driven by unpredictable shifts in the spatial threshold where the plectrum transitions from static friction to full release. Consequently, nominal depths yield highly variable responses, with shallow plucks exhibiting a Coefficient of Variation (CV) ranging from 10 to over 60% for both RMS Energy and Spectral Centroid. These findings indicate geometric calibration fails to provide reliable, repeatable data in comparative studies. Longitudinal repeatability requires closed-loop methods such as acoustic feedback control.

Speaker: Jan Jasiński (AGH University of Krakow)

Full paper (PDF)

355 Reversible Wolf Tone Suppression in Acoustic Guitars Using Locally Resonant Metamaterials

The tuning of guitar soundboards is traditionally anirreversible subtractive process, relying on the removal of bracingmaterial. This study proposes a reversible, additive approach utilizingtop plate mounted Locally Resonant Metamaterials (LRMs) to suppresspathological resonances arising from strong string-body coupling,commonly referred to as wolf tones, which cause uneven note decay. UsingScanning Laser Doppler Vibrometry (SLDV), we mapped the operationaldeflection shapes of the guitar's top plate under controlled excitationto identify its eigenmodes. We then experimentally simulated wolf toneconditions by tuning strings to couple directly with a specificstructural resonance. To counteract this effect, custom metamaterialresonators were designed and manufactured to act as subwavelengthmechanical band-stop filters. To distinguish the local resonance effectfrom simple mass-loading, the system’s response was compared against acontrol using equivalent masses. The efficacy of the treatment wasevaluated through comparative vibrometry and audio signal analysis ofthe decay envelope and timbre of a plucked note. Results demonstratethat LRMs can clearly reduce note decay irregularities and achievesuperior suppression and frequency-specific targeting compared to massaddition alone, providing a non-invasive solution for instrument voicingapplicable both during manufacturing and for post-production adjustment.

Speaker: Jan Jasiński (AGH University of Krakow)

Full paper (PDF)

A11.05/A15.03 Perception of Music and Musical Instruments: P471

Convener: Kai Siedenburg (CvO University of Oldenburg)

356 Bootstrap-Based Psychoacoustic Characterization of Playing Techniques in Non-Western Percussion Instruments: A Case Study on the Colombian Alegre Drum

The underrepresentation of non-western musical traditions in open-source datasets poses a persistent challenge for computational psychoacoustic analysis, particularly due to low-quality recordings or insufficient sample sizes that characterize these collections. This study examines the use of statistical inference methods for the psychoacoustic characterization of non-western traditional percussion instruments under conditions of small sample size. The method is applied to three playing techniques of the Colombian Alegre drum — Abierto (Tone), Bajo (Bass), and Quemado (Slap)— analyzed through psychoacoustic features. Recordings were obtained in an anechoic chamber using an artificial head system. To address the limited dataset size (N = 17), the bootstrap method is applied to generate 95% confidence intervals, and Hedge's g is computed to assess effect sizes, enabling a reliable and robust comparison of timbral profiles across techniques. Preliminary results reveal distinct timbral profiles for each technique, with the Bajo exhibiting the lowest centroid and widest bandwidth, the Quemado the highest roughness and spectral centroid, reflecting weak fundamental energy and broad spectral distribution up to 5kHz, and the Abierto the highest crest factor. Timbral profiles for each technique are visualized through radar charts, allowing a direct comparison of psychoacoustic features across playing techniques. These findings demonstrate that meaningful psychoacoustic characterization is achievable with small datasets, supporting the documentation and digital preservation of non-western musical heritage.

Speaker: Cristhiam Fidel Martínez Orellanos (University of Hamburg)

Full paper (PDF)

357 Chills and Betting Bills - Audio Features, Listener Emotions and the Success of ESC Songs

Each year, the Eurovision Song Contest (ESC) generates speculation about the winning song. Previous research about Eurovision Song Contest Songs has mainly examined the musical features of past ESC participants, while the listeners emotions and chill experiences have not yet been combined with it. This study investigated whether acoustic and emotional differences can be measured between highest- and lowest-rated ESC songs based on pre-final betting odds (April 2026), aiming to capture potential predictors of final outcomes. Twenty-seven participants (17f, 10m; 19-57 years old, mean 28) listened to three lowest- and three highest-rated songs, while skin conductance was measured via Mindfield eSense sensors and valence, arousal (Circumplex Model) and experienced chills were reported. After each song, participants were asked to rate the song liking (“not” to “very much”) and how often they have heard it before (“not once” to “many times”). Further, they filled out a personality questionnaire (SEPPO) as research indicates preference and physiological reaction differences. Audio features were extracted using a.o. MIRToolbox and PADMEA.There were significantly more experienced chills for the three higher-rated songs (t(3,first_vs._3,last)=3.264, p<0.05). Differences (t-Tests) between personality traits, song liking, valence, arousal, and musical audio features between the three lowest- and three highest-rated songs were noticed. Based on the acoustic audio feature analysis the valence also depends on the songs Roughness (Vassilakis, t(3,first_vs._3,last)= -7.528, p<0.01) and Brightness (t(3,first_vs._3,last)= 11.935, p<0.01). Further results will be interactively presented (https://muwiserver.univie.ac.at/esc2026/) at the conference.

Speaker: Helene Lindenbauer (Universität Wien)

Full paper (PDF)

358 Geometric and topological analysis of within-note timbre variability

Timbre's temporal micro-variability contributes substantially to source identity and sound quality. However, existing temporal representations are either computationally driven or perceptually motivated but reduced to scalar descriptors. Building on a momentary approach to timbre developed in prior work, we present a framework that isolates temporal variability and represents it as a continuous trajectory in a low-dimensional space.Perceptual grounding is realized through asymmetric sampling in time: parallel feature streams are extracted at short (~30 ms) and long (~200 ms) integration windows. We compare wavelet scattering features against log-mel spectrum and MFCC baselines. The streams are embedded with a diffusion map into a geometric representation motivated by the non-metric, relational character of perceptual timbre spaces. Over this embedding we apply two complementary topological summaries: Mapper graphs to expose the qualitative shape and connectivity of variability trajectories, and density-aware persistent homology (DTM filtration) for quantitative signatures comparable across sources. We are not aware of prior work combining perceptual constraints with a geometric representation of within-note timbre variability.We present an initial case study on sustained notes from instruments spanning wind, brass, and string families. Early embeddings show family-consistent groupings, though we treat these observations as illustrative at the time of submission; a complete evaluation is expected in time for the conference. The framework is intended as a foundation for computable, perceptually valuable descriptions of timbre variability.

Speaker: Szymon Błaszczyk (Adam Mickiewicz University in Poznan)

359 Modeling Instrument Identification in Chord Mixtures

Identification of musical instruments has primarily been investigated for sounds in isolation. However, as musical scene analysis often involves multi-source scenarios, research must be extended to sound mixtures.In this study, thirty-four listeners identified sounds from eight Western classical instruments presented in isolation and in stationary chord mixtures of two to four instruments. For the chord mixtures, the task required listeners to hear out and identify a target instrument within the mixture, guided by a preceding sinusoidal cue. Behavioral results showed that identification accuracy decreased as the number of instruments in the mixture increased. Furthermore, results revealed instrument-dependent confusions and effects of register and voice positioning within the chord.To further explore these findings, we implemented a computational frontend-backend model. In the frontend, a harmonic mask containing the first ten harmonics of the target F0 was applied to the mixture spectrograms to extract target instrument representations. The width of these harmonic windows served as a parameter for frequency resolution in pitch tracking. Following auditory ERB-filtering, spectral and temporal modulations were captured using the modulation power spectrum (MPS). For the backend, a random forest classifier was trained on isolated instrument sounds and tested on the masked mixtures. Internal noise was introduced as a second parameter to impair pattern matching.Our model provides plausible classification results that reflect human performance in both accuracy and confusion patterns of instruments, establishing a foundation for understanding the auditory perception of instrument timbre within complex musical mixtures.

Speaker: Simon Jacobsen (CvO University of Oldenburg)

A15.01 Spatial Hearing: Modeling and Applications: P529

Conveners: Piotr Majdak (Acoustics Research Institute, ÖAW), Ville Pulkki (Aalto University)

360 Pinna2HRTF: A Pipeline for Automatic Calculation of Head- Related Transfer Functions from Pinna Meshes

Individual head-related transfer functions (HRTFs) can be numerically calculated from 3D ear meshes with the help of the software package Mesh2HRTF. However, the process from the individual ear geometry to a valid project for Mesh2HRTF can be tedious as it involves many steps: head generation, ear stitching, mesh grading, and the setup of the project directories. Here, we introduce Pinna2HRTF, a pipeline, automating that process by taking paired (left and right) pinna meshes stored as STL files as input and outputting a binaural HRTF set stored as a SOFA file. The pipeline, implemented in Python, generates an appropriately sized dummy head, closes the ear canals, stitches the ears onto the head, applies frequency-dependent mesh grading, exports ready-to-run project Mesh2HRTF directories, and runs Mesh2HRTF. Further, an optional stage called Mesh2PPM can be used to register ear geometries to a parametric pinna model (PPM), which then provides ear meshes of sufficient quality for the calculations with properties such as a closed surface, no artifacts, and uniform sampling. Pinna2HRTF is open source and available to the public.

Speaker: Felix Perfler (Austrian Academy of Sciences (ÖAW))

Full paper (PDF)

A17.05 Sound field estimation and reconstruction: P494

Conveners: Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology), Samuel A. Verburg (Technical University of Denmark)

361 Learning Object Geometry from Acoustic Scattering via a Brinkman Penalty in a Universal Physics-Informed Neural Network

Acoustic scattering plays an important role in ultrasound, geophysics and object mapping for navigation. When the object geometry is unknown, reconstruction of the scattered acoustic field is challenging as the boundary conditions cannot be explicitly enforced. We incorporate a learnable Brinkman-style penalty into the governing wave equations. The penalty induces attenuation within the object while outside the object the wave can propagate unhindered, thereby enforcing boundary conditions implicitly across the entire wave-field domain. The Brinkman penalty mask acts as a spatial representation of the object geometry. We estimate both the Brinkman penalty and wave-field using a Universal Physics-Informed Neural Network (UPINN). The UPINN uses a neural network to model the unknown terms of the differential equation. By jointly learning the object geometry and the wave-field the method results in an improved reconstruction of the scattered field. This method is demonstrated in the frequency domain by simulating the scattering of plane waves from multiple incident angles onto a cylinder. The framework highlights the potential of UPINNs for data-driven discovery of governing differential equations in acoustics.

Speaker: Stefan Graham (Technical University of Denmark)

A20.00 Speech: P438

Conveners: Philipp Aichinger (Medical University of Vienna), Barbara Schuppler (Signal Processing and Speech Communication Laboratory), Oliver Niebuhr (University of Southern Denmark), Franz Pernkopf (Signal Processing and Speech Communication Laboratory), Peter Balazs (Acoustics Research Institute, Austrian Academy of Sciences)

362 Classifying overlappee and overlapper based on acoustic prosodic features

Conversational overlaps, i.e., instances where two or more participants speak simultaneously, are a prevalent feature of human communication, influencing conversational dynamics and speaker roles. While previous studies have mainly analysed and categorised them based on competitiveness, this paper focuses on the roles of individual speakers within these overlaps. Our study is based on instances of overlapping speech from a corpus of casual, spontaneous dyadic conversations, from which we extract acoustic features and communicative function labels that were created manually. Using XGBoost for classification, we find that overlaps as coming from an overlappee or overlapper, can be discriminated with a balanced accuracy score of approx. 75%. Our analysis of feature importances shows that the communicative function along with the intensity levels are the key indicators in identifying speaker roles during overlaps. This study contributes to our understanding of conversational dynamics and provides insights applicable to making turn-taking in human-computer interaction more natural.

Speaker: Barbara Schuppler (Signal Processing and Speech Communication Laboratory)

Full paper (PDF)

363 Exploring Intonemes in Brazilian Portuguese

Isačenko and Schädlich describe German F0 contours as communicatively motivated tone switches aligned with accented syllables. Stock and Zacharias link these switches to phonologically distinct intonational elements, or intonemes, with three main classes distinguished by sentence modality: information, contact and non-terminal intoneme. Using the Fujisaki model, tone switch timing is captured relative to the syllabic grid mostly by the alignment and amplitudes of the accent component. In the current study we further explore the tone switch approach by applying it to Brazilian Portuguese. Based on acoustic results we developed synthetic stimuli which covered a continuum of utterances between the three categories. Analysis of stimuli by prosody experts of BP showed that the corner stimuli indeed reflect the three intended categories well. In an ensuing perceptual study, we examined whether native speakers of BP were also able to decode the three different basic intonemes. While Information and Contact intoneme were robustly identified, the Non-terminal intoneme was more elusive in naïve perception, despite consistent expert recognition. This asymmetry reflects functional differences between sentence mode and discourse-level intonational cues and the constraints by the identification task that brought these two dimensions together in one experiment.

Speaker: Hansjörg Mixdorff (Berliner Hochschule für Technik)

Full paper (PDF)

364 How 'ahem' and 'also' affect perceived fluency, clarity and understandability

This study investigates the perception of disfluencies in speech, focusing on the fillers aehm" andalso". Using conversational speech recordings from the Austrian German GRASS corpus, we select 72 stimuli similiarly long in number of tokens that contained either aehm" oralso" and conduct a perception experiment with 30 participants. During the experiment, listeners rated for each stimulus how well they understood the content of the stimulus (i.e., understandability), how clearly pronounced or reduced they perceived the stimulus (i.e., clarity), and also how fluently spoken or disturbingly disfluent they perceived the stimulus. Our statistical analysis shows that speech stimuli containing aehm" tend to receive significantly higher ratings in clarity, understandability, and fluency than stimuli containingalso". Further, the results show correlations among listeners' perceptions of these three dimensions. Overall, the findings contribute to a better understanding of how specific disfluency markers influence speech perception and highlight the importance of considering specific filler types when studying spoken language processing.

Speaker: Barbara Schuppler (Signal Processing and Speech Communication Laboratory)

Full paper (PDF)

365 Speech Reduction and Speech Rate: Disentangling Temporal and Prosodic Effects

This analysis of the link between speech rate and speech reduction identifies the degree to which rate alone drives reduction. Using the German part of the Bonn-Tempo corpus, we analyze vowel, consonant, and pause reduction as a function of speech rate and prosodic accentuation. Phone- and syllable-based rate measures were first computed, followed by Pfitzinger’s perceptual local speech rate (PLSR), allowing a detailed characterization of rate distributions across the five corpus-defined speed levels for individual speakers. Substantial inter-speaker variability is observed, indicating that nominal rate categories only partially reflect actual production. Acoustic analyses of vowel formants show that significant differences in vowel centralization occur primarily between very slow, normal, and very fast speech. After Lobanov normalization, vowel centralization increases approximately linearly with PLSR before reaching a saturation plateau at very high rates. Temporal analyses further reveal that vowel, consonant, and pause compression and elision are not uniform processes, but depend systematically on both accent status and underlying PLSR. These findings demonstrate that speech reduction constitutes a structured, prosodically mediated acoustic dimension that cannot be fully derived from speech rate alone. A companion perceptual study addresses how listeners respond to these reduction patterns.

Speaker: Hansjörg Mixdorff (Berliner Hochschule für Technik)

Full paper (PDF)

A20.08 Perception, Acoustics, and Transmission of Speaker Impact: P441

Conveners: Oliver Niebuhr (University of Southern Denmark), Sara Pearsell (University of Southern Denmark), Ingo Siegert

366 Effect of hearing aid amplification on supra-segmental acoustic features for speech production

Speech production is continuously adjusted by auditory and somatosensory feedback loops, allowing speakers to monitor and adapt their vocal output based on multimodal sensory input. In hearing-impaired speakers, a degraded auditory feedback loop results in altered supra-segmental characteristics, such as increased vocal intensity, slower articulation rates, and elevated fundamental frequency. The importance of auditory feedback is revealed by the effects of hearing aid (HA) amplification, which partially mitigates these speech alterations. However, HA amplification doesn’t form a homogeneous effect as gain is computed for individual hearing thresholds and some elements of amplified own speech remains below these thresholds.This study investigates the quantitative relationship between the amount of gain provided by hearing aid amplification and changes in supra-segmental acoustic features in 20 adult HA users with post-lingual, moderate-to-severe hearing loss. Acoustic features were extracted from speech recorded during various tasks, including reading, picture description, and spontaneous speech. Changes in the acoustic characteristics of the speaker's own voice during HA use were quantified via probe-tube measurements and percentile analysis. The complex effects of amplification on speech production are discussed within the context of multimodal feedback, suggesting that auditory rehabilitation can partially compensate for sensory deficits for speech production by hearing-impaired talkers.

Speaker: Christophe Lesimple (Sonova AG)

A24.00 HRTF processing and modeling: P443

Conveners: Lorenzo Picinali (Imperial College London), Julie Meyer (King's College London)

367 Machine Learning Based Refinement of Synthetic HRTFs Toward Measured Fidelity

Numerically simulated Head-Related Transfer Functions (HRTFs) provide an attractive alternative to acoustic measurement, removing the need for anechoic facilities and dedicated measurement equipment. However, their perceptual fidelity remains limited compared to gold-standard acoustic measurements, with residual errors in interaural time differences (ITDs), interaural level differences (ILDs) and monaural spectral, cues which govern lateral and elevation localisation. This seems to be mainly (but not only) due to the underlying mesh acquisition method or the simulation pipeline, which result in lower performances when assessed through perceptual models. The aim of this study is to provide a machine learning based post-processing tool that refines synthetic HRTFs towards acoustically measured fidelity, independently of the mesh source used. To this end, neural network architectures are trained on the Extended SONICOM HRTF dataset (200 subjects), each providing paired acoustically measured HRTFs alongside two synthetic counterparts obtained from a high-resolution three-dimensional scan and from a photogrammetry-reconstructed (PR) mesh, enabling refinement to be learnt across heterogeneous synthesis inputs. Networks are trained subject-independently using different perceptual and non-perceptual loss functions, and evaluated on unseen subjects through numerical metrics (ITD, ILD, log-spectral distortion), auditory-model predictions and a behavioural VR sound localisation test. Improvements are found in each evaluation method, supporting neural refinement as a simple post-processing stage that compensates for synthesis artefacts and brings simulated HRTFs closer to measurement-grade fidelity.

Speaker: Ludovic Pirard (Imperial College London)

15:40 Coffee break

15:40 Coffee break

A18.08 Soundscape in Natural Parks: S289

Conveners: Tin Oberman (University College London), Claudio Guarnaccia (University of Salerno, Department of Civil Engineering)

368 Listening to Norwegian National Parks

Natural areas play a central role in human well-being, particularly in Norway where national parks are widely used for recreation. Soundscapes are increasingly used to assess both biodiversity and human pressures, yet remain poorly documented in Nordic environments.We investigated soundscape patterns in two Norwegian national parks, Femundsmarka and Rondane, by combining ecoacoustic indices, automated species detection (BirdNET), expert validation, manual annotations, and an ISO 12313 survey of hikers. A 4-day study was first conducted using passive acoustic monitoring across three habitats (forest, water, tundra). Seven acoustic indices (ACI, ADI, AEI, BI, entropy, biophony, anthrophony) were computed to assess their ability to discriminate habitats and temporal dynamics. Monitoring was subsequently extended across summer and winter, enabling detailed annotation of biological and anthropogenic acoustic activity under contrasting diel regimes.Results during the summer showed that acoustic indices consistently discriminated tundra from forest and water soundscapes, while partial overlap between forest and water highlights limits in ecological specificity. Strong diel patterns driven by bird vocal activity were observed in Femundsmarka, whereas weaker temporal structuring occurred in Rondane. Notably, the anthropogenic index did not reflect annotated human activity, showing higher values at night despite daytime human presence. Results of the survey during the summer indicated that all habitats were generally perceived as pleasant, but soundscape quality decreased when anthropogenic sounds were heard.Overall, our study highlights both the potential and limitations of acoustic indices and underscores the importance of integrating automated metrics with ecological validation for robust soundscape monitoring in protected areas.

Speaker: Emilie Rojas (Norwegian University of Science and Technology (NTNU))

369 Mapping Quiet Areas in the Nordic Countries: Results from National-Scale Studies in Denmark, Finland, Norway, and Sweden

The concept of “quiet areas” is gaining traction in environmental planning, with potential to safeguard and improve sound environments in agglomerations and open country. Within the framework of the European Union Environmental Noise Directive (END, 2002/49/EC), member states are required to map environmental noise and develop action plans to manage its effects, including the preservation of quiet areas. The directive does not define how such areas should be identified or implemented, resulting in a variety of strategies, some of which have been documented in subsequent guidance. Yet, the spatial characteristics of these initiatives remain largely unknown at larger scales. This study examines the implementation of quiet areas in the Nordic countries, providing a comparative perspective across different planning contexts. The aim is to map and analyse the spatial distribution and characteristics of officially designated quiet areas in Denmark, Finland, Norway, and Sweden. Data were collected through questionnaires to governing bodies and reviews of policy documents, resulting in a GIS dataset of 6,920 designated quiet area polygons, covering approximately 7.4 million hectares. Results show that officially designated quiet areas occupy only a small proportion of national territories, with uneven spatial distribution, particularly in Norway and Finland. Further analyses examined relationships with land cover, overlap with protected areas, and alignment with the Quietness Suitability Index (QSI), identifying both synergies and potential conflicts. The findings highlight substantial variation in implementation strategies and point to the need for clearer guidance to support consistent and effective planning practices.

Speaker: Gunnar Cerwén (SLU)

370 The Role of Listening in National Parks in fostering Nature Connection

Soundscape tourism is gaining importance in and around national parks and protected areas, in part reflecting a growing demand for quietness and the experience of natural sounds. This study examines changes in nature connectedness before and after a listening task, in relation to measured acoustic properties. Understanding how nature connectedness unfolds through listening may have broader implications for the biodiversity crisis (Richardson et al., 2022). In natural settings, focused listening has been associated with a feeling of connectedness to nature and other species (Jiang, 2022), but less is known about how this process unfold in relation to different compositions of nature soundscapes (including disturbances). Soundscape tourism may facilitate these experiences that are important precursors to adopting pro-environmental behaviors and world-view transformations (Van Riper et al. 2020). Data were collected by intercept survey delivered to hikers nearby passive acoustic monitoring sites in Femundsmarka National Park and linked to daily acoustic characteristics of each site (e.g. ADI, SPL). Preliminary findings show that in general, participants experienced an increased sense of nature connection after the three minute listening experience. Participants who heard anthropogenic sources (i.e. motorized, technological sounds) reported a decreased or non-altered sense of nature connection. Support for conservation policy and the reported intrinsic value of nature was also associated with perceived acoustic quality. Analyses of measured acoustic data aligned with soundscape assessment even though the preciseness of sounds identified differed according to urban and rural dwellers. The findings highlight the role of intact natural soundscapes in facilitating nature connection, and point to the potential benefits of tourism approaches centered on listening.

Speaker: Rose Keller (The Norwegian Institute for Nature Research (NINA))

371 Soundwalk in Paklenica National Park, Croatia

Sounds significantly affect our experience of natural environments and presence of people changes the acoustic environment, creating an interactive system driven largely by human perception. This study explores soundscape in the Paklenica National Park (PNP). A soundwalk was conducted in Velika Paklenica canyon during a peak tourist weekend. The soundwalk route connected two educational trails, cultural heritage sites (two historical mills), viewpoints, historical filming locations, an underground interpretation centre, UNESCO-protected primeval beech forest and a world-famous mountain climbing site. The soundwalk with 20 participants applied ISO 12913 standard for the evaluation of acoustic environment in a natural setting along the trail with seven evaluation spots. The participants were invited from three key groups of stakeholders: 1. planning, design and research experts (architects, urban planners, acousticians, natural scientists); 2. local experts (residents and PNP staff); 3. tourists. The study reveals the perception of cultural heritage in a touristic attractive location (dominated by human sounds) and locations dominated by natural heritage (without human sounds). The results contribute towards enhancing management and planning of national parks as well as protecting natural and cultural heritage.

Speaker: Klara Kranjčec (Faculty of Architecture, University of Zagreb)

Full paper (PDF)

372 Crowdsourcing infrastructure for soundscape observations and monitoring in natural parks

Crowdsourcing has become an established and valuable method for environmental observation and monitoring using the participation and collective effort of individuals in various contexts and geographic areas. More recently, crowdsourcing or citizen science tools have been also used to collect multisensory information, such as noise, sound sources, perceived soundscape quality, and experiences and values in different types of natural environments. However, such data is often collected and used in silos. The INAR RI eLTSER research infrastructure project lead by the Finnish Environment Institute aims to bring different citizen data sources together by developing the Open311 based Citobs platform. The platform aims to provide and manage a single digital infrastructure for varied services e.g. by using adaptable questionnaires for different research and planning uses, and user groups with different levels of professionalism. We present a pilot case study of testing the Citobs Platform in Helsinki Central Park, where we use incentive-based approaches to collect data on perceived soundscape quality of the park. We also discuss the broader applications of the platform in natural areas and the opportunities for scaling it up at the national level.

Speaker: Silviya Korpilo (Finish Environment Institute)

373 Screening of the 'Silenzi in Quota' documentary

When we think of mountains, we imagine boundless spaces where silence is a constant, invisible companion. But what happens when human noise invades these remote reaches? Silenzi in Quota was born to measure the mountain soundscape, using both microphones and people as 'sensors' to explore the fragile balance between nature and human presence. Through curated soundwalks, the collective has guided hundreds of participants from the Dolomites to the Scottish Cairngorms, fostering an active listening of the surrounding landscapes and a tranquility that is at risk of vanishing. This documentary chronicles nearly 20 expeditions conducted since 2021: an authentic collection of spatial audio recordings that reveal the tension between our desire to escape the city and a form of 'unconscious tourism' that replicates urban sounds at high altitudes. It is an invitation to pause, to listen, and to envision a more mindful way of experiencing the peaks. "Silenzi in Quota" is a movie directed by Andrea Paternolli and Mario Pedron (in collaboration with Giacomo Gozzi, Simone Torresin and Tin Oberman), first screened at the Trento Film Festival in 2025.

Speaker: Simone Torresin (University of Trento, Dept. of Civil, Env., Mech. Eng.)

16:00 Coffee break

A13.10 Acousto-Optics and Distributed Acoustic Sensing: S097

Conveners: Emmanuel Dekemper (BIRA-IASB), Samuel Dupont (Université Polytechnique Hauts-de-France), Ireneusz Grulkowski (Nicolaus Copernicus University)

374 Acoustic Field Structure Visualisation in Quasi-Collinear Acousto-Optic Cell

Some applications of quasi-collinear acousto-optic (AO) cells, such as laser pulse generation and shaping, microscopy, spectroscopy, and multicoloured optical tweezers, require the creation of precisely defined distribution of the acoustic field frequency and amplitude along the entire AO interaction length. If the actual acoustic field structure inside the AO crystal deviates from the desired one, the spectral components of the input optical pulse are diffracted in the areas with improper characteristics. It leads to a distortion of the diffracted beam spectral profile at the output. Numerical modeling is sometimes insufficient, since the AO cells geometry may differ due to an inaccurate fabrication, which significantly modifies the acoustic beam structure along the AO interaction length. In this work, an acoustic field visualization technique for the quasi-collinear geometry of AO interaction is demonstrated as an effective tool for analyzing the redistribution of acoustic power caused by acoustic anisotropy, acoustic power absorption and temperature gradients. We experimentally obtained high-contrast 3D maps of the acoustic power distribution in a quasi-collinear AO cell based on a tellurium dioxide crystal. The proposed approach provides a more consistent and reliable way to characterize the acoustic field structure.

Speaker: Ekaterina Kostyleva (M.V. Lomonosov Moscow State University)

Full paper (PDF)

375 Challenges in Using AOTF-Based Spectral Imaging for Atmospheric Remote Sensing

A hyperspectral imaging instrument has been developed for the purpose of mapping air pollution in the urban environment, in particular nitrogen dioxide (NO2), a molecule responsible for respiratory diseases. The instrument is based on a large aperture commercial TeO2 acousto-optic tuneable filter (AOTF). In nominal operations, the instrument acquires spectral images of the light-diffusing atmosphere at a number of wavelengths in the 430-490nm domain, building a hypercube. The sub-nm spectral resolution of the AOTF, combined with good imaging performances, allow to detect the spectral signatures of NO2 in the light spectrum collected in every pixel.The method to retrieve the 2D distribution of NO2 from the hypercube relies on fitting each pixel-recorded spectrum with the absorption cross sections of the light-interfering species (NO2, O4, H2O, O3) convolved by the spectral response function (SRF) of the instrument (largely driven by the AOTF). Due to RF chain constraints, a too large acoustic power is sometimes injected in the crystal, taking the effective SRFs away from nominal shapes (sinc^2), and making them vary across the aperture. With the help of lasers and spectral lamps, the SRFs were characterised. The impact on the remote sensing application is discussed.

Speaker: Emmanuel Dekemper (BIRA-IASB)

Full paper (PDF)

376 Physics-Aware Learnable Filtering for Vehicle Detection in Distributed Acoustic Sensing

Distributed Acoustic Sensing (DAS) enables continuous monitoring of traffic activity over long distances. However, noise contamination and the lack of labeled data limit the scalability of existing approaches. This study proposes a framework that integrates a physics-aware learnable filtering module with a data-driven refinement strategy for vehicle detection in DAS signals. The filtered outputs are used to generate and progressively improve proxy labels, enabling robust segmentation without manual annotations. The extracted vehicle regions are further utilized for speed estimation. Results on real-world data demonstrate stable vehicle detection under noisy conditions and reliable speed estimation performance. The findings highlight the potential of combining physics-informed filtering with adaptive learning for scalable DAS-based traffic monitoring.

Speaker: Sasan Farhadi (IGMS)

377 Optical probing of corneal stiffness through focused- ultrasound tissue stimulation

The quantitative assessment of biomechanical properties of ocular tissues remains a significant challenge, primarily due to the limited availability of non-contact, high-resolution techniques capable of probing localized mechanical responses in vivo. In particular, the cornea exhibits complex viscoelastic behavior governed by its microstructural organization, and its mechanical properties are closely linked to physiological function and pathological conditions such as ectasia. The approaches, coupling acoustic radiation force excitation with optical detection of tissue motion, provide a promising framework for localized elastography with micrometer-scale sensitivity. In this study, we demonstrate a method for stiffness characterization of thin soft tissues, combining focused air-coupled ultrasound excitation with phase-sensitive optical coherence tomography (OCT). Localized mechanical perturbations are induced via acoustic micro-tapping (AμT), in which a focused ultrasonic transducer generates a transient acoustic radiation force at the tissue surface, launching surface (Rayleigh-type) waves. The resulting nanometer-scale displacements are captured using OCT, enabling depth-resolved tracking of wave propagation with high temporal and spatial resolution.The proposed approach was validated using gelatin-based phantoms with controlled elastic properties, as well as ex-vivo corneas. Shear wave velocities were estimated from spatiotemporal displacement fields and subsequently used to reconstruct the effective Young’s modulus of the samples. The results demonstrate differentiation of stiffness across phantoms and confirm the feasibility of the method for thin, layered biological tissues. This work highlights the potential of optical and ultrasound methods for non-invasive, high- resolution assessment of corneal biomechanics. The presented methodology is particularly suited for ophthalmic applications, where precise characterization of mechanical properties may support early diagnosis of corneal pathologies and improve patient-specific treatment planning.

Speaker: Raúl Urriza-Arpal (Nicolaus Copernicus University)

378 Rail vibration measurement using Distributed Acoustic Sensing and rail-mounted monolithic fiber optic sensors

This paper presents an experimental evaluation of Distributed Acoustic Sensing (DAS) combined with rail-mounted monolithic distributed fiber optic sensing (DFOS) elements for measuring the dynamic response of railway infrastructure under high-speed operation. The study was conducted during controlled nighttime test runs on a secured track section using dedicated rolling stock, with train velocities incrementally increased to 200, 220, 240, and 250 km/h.The primary objective is to assess the wide-bandwidth capabilities of DAS in capturing complex vibration phenomena, with particular emphasis on high-frequency events occurring in inter-harmonic regions under high-amplitude excitation. Special attention is given to the influence of sensor installation methods on measurement fidelity and acoustic coupling efficiency. To enable a controlled comparison, the impact of installation techniques was evaluated based on measurements performed on a dedicated experimental test field. Furthermore, a low-invasiveness adhesive-based mounting technique for monolithic DFOS segments is introduced. This approach eliminates the need for mechanical pre-tensioning while significantly improving coupling between the sensor and the rail compared to conventional telecommunication fibers installed in conduits.Experimental results demonstrate that the integration of rail-mounted monolithic DFOS segments within a DAS system enhances sensitivity and extends the usable frequency range. The system successfully resolves complex spectral signatures, including high-frequency components associated with wheel–rail interaction and structural resonances at peak operational speeds.The findings confirm that optimized sensor installation plays a critical role in maximizing DAS performance. The proposed hybrid sensing configuration provides a robust and scalable solution for advanced railway monitoring, supporting the development of autonomous structural health monitoring systems under high-speed conditions.

Speaker: Szymon Dlugosz (SHM System/Nerve-Sensors)

A14.03 Hidden hearing loss & synaptopathy: Diagnostics, treatment & functional consequences: S293

Conveners: Lukas Rüttiger (University of Tübingen), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

379 Effect of Aging on Forward-Masking Estimates of Medial Olivocochlear Gain Reduction: Computational simulations and Behavioral Data

In humans, aging is associated with a loss of synapses between inner hair cells and auditory nerve fibers (cochlear synaptopathy (CS)). While CS is expected to affect afferent sound encoding, it may also influence efferent control of cochlear gain by reducing the input to the medial olivocochlear (MOC) system. To date, this possibility has not been extensively investigated. Here, we combined computational simulations and psychophysics to evaluate age-related CS effects in a forward-masking task designed to assess MOC-mediated gain reduction. We measured detection thresholds for a 2-kHz probe (10 ms) presented alone or preceded by a 400-ms noise precursor (20-ms gap) in young (yNH, n = 8) and older normal-hearing listeners (oNH, n = 7). Under these conditions, the precursor-induced threshold shift (dB) can be used to estimate cochlear gain reduction mediated by activation of the MOC system. Simulations were performed using a computational model of the auditory periphery with MOC feedback driven by inputs from the inferior colliculus and auditory nerve. Precursor bandwidth and temporal-envelope characteristics were systematically varied. Simulations predicted reduced threshold shifts in conditions simulating reduced auditory nerve input to the MOC system, particularly for narrowband precursors with weak temporal-envelope fluctuations. In contrast, behavioral results showed no differences between groups for narrowband precursors and significantly larger threshold shifts in oNH for wideband precursors (with and without temporal envelope modulations). These results suggest that forward-masking estimates of cochlear gain reduction reflect more than MOC-mediated gain reduction alone. Future work should combine behavioral paradigms with objective measures of MOC function to better isolate age-related changes in MOC control of cochlear gain.

Speaker: David López-Ramos (STMS (Ircam-CNRS-SU))

380 Auditory Nerve Fiber Loss And Vowel Discrimination: Evidence From Machine Learning Analysis Of Single-Unit Recordings

Traditionally, vowel encoding in the auditory nerve relies on type I spiral ganglion neurons (SGNs) with characteristic frequency (CF) close to global spectral peaks (formants, F). This study investigates contributions of subpopulations of auditory nerve fibers to coding spectral and temporal features of vowels under normal conditions and following kainate-induced cochlear synaptopathy. Four synthetic vowels with varied spectral contrast were used to stimulate the SGNs: /i/, /y/ (F2>1.5 kHz), /o/, and /u/ (F2<1.5 kHz), presented at 70 and 50 dB SPL. Using machine learning methods on single-fiber recordings of auditory nerve fibers’ responses from Mongolian gerbils, we demonstrate that, phenotype-independent progressive ablation of up to 90% only reduced discrimination performance by 5%. On the contrary, loss of ANFs of low spontaneous firing rate (<18 spikes/s; low-SR) and high-CF (>8 kHz) fibers drastically degrades classifier’s performance (by 35% and 30%, respectively). Furthermore, kainate-induced increase of mean population SR by up to 50 spikes/s led to 25% drop in performance even without depletion of ANFs. These findings highlight the functional significance of low-SR and high-CF fibers in vowel coding in the auditory nerve and demonstrate the detrimental effects of their progressive loss on vowel discrimination.

Speaker: Daniil Kiselev (Institute for Neuroscience Montpellier)

381 Differentiating Threshold-Specific Responses of the Human Auditory Nerve via Refractoriness

Auditory nerve (AN) fibers transmit the output of cochlear hair cells to the brainstem. Fibers vary in their threshold, with some sensitive to lower-level sounds and others requiring higher levels to respond. While threshold-specific fiber responses can be directly recorded in animals, non-invasive measures in humans, i.e., auditory brainstem responses or the compound action potential (CAP), lack threshold-specific resolution. Here we designed a refractory-masking paradigm to differentiate the electrophysiological responses of AN fibers with different thresholds in humans. Stimuli consisted of five rapid click trains with one, two, three, four, or five clicks. The clicks in each train were spaced 0.25 ms apart, and each click was 15 dB higher in level than the one that preceded it. Preceding lower-level clicks within a train were designed to mask part of the response to the next click, since lower-threshold fibers would be in a refractory period when the following click was presented. Fiber-specific CAP responses to the final click of each train were isolated by subtracting the response to the train with one less click, canceling out neural responses to prior clicks. Consequently, we were able to isolate the responses from fibers with different threshold rates. Identifying fiber-specific response patterns not only helps to bridge the gap between human and animal research in auditory nerve function but may also hold potential as a proxy for peripheral neural degeneration.

Speaker: Ross Maddox (University of Michigan)

382 Extended high frequency hearing for cortical processing of speech

Young adults hear sounds up to 20 kHz. Therefore, the loss of extended high-frequency hearing (EHF; above 8 kHz) is a hallmark of age-related hearing loss, often progressing from early lifetime. This deficit frequently goes undetected because routine clinical hearing tests and most hearing aids are currently limited mostly up to 8 kHz. EHF hearing has been linked to deficits in speech perception in noise and to self-reported hearing. However, it remains elusive how EHF hearing influences speech intelligibility. Here we recorded neuromagnetic brain responses using magnetoencephalography (MEG) within a frequency-tagging speech paradigm designed to probe hierarchical levels of attention and memory-dependent speech processing and recognition. Auditory evoked cortical magnetic field (AEF) responses were significantly reduced in both left and right brain hemispheres in individuals with impaired EHF hearing compared to those subjects with rather preserved EHF hearing. A gradual reinforcement of left-hemispheric AEF seen over age was not observed in young adults (19-29 y) with preserved EHF hearing. This was linked to stronger auditory brainstem responses (ABR), reflecting better neural synchronized auditory responses at stimulus onset. The reinforced left hemispheric dominance in young adults with impaired EHF hearing, in contrast, was linked to lower ABRs. Our findings suggest that sound energy above 8 kHz contributes through its impact on stimulus-onset synchrony to phase locking of oscillations in the auditory cortex to intelligible speech. Together, the results highlight the need to reconsider the neglect of EHF hearing in both audiological assessment and hearing aid design.

Speaker: Lukas Rüttiger (University of Tübingen)

383 The NExpo Study: Behavioral, Physiological, and Neuroimaging Measures to Assess the Effects of Noise Exposure on the Auditory Pathway

The NExpo study investigates how noise exposure, aging, and noise-induced hearing loss affect the human auditory nervous system. 200 healthy participants were recruited into four groups (n=50 each) based on age, pure tone audiometry (PTA), and Noise Exposure Structured Interview (NESI) score. Group 1 (G1): 18-19 years, low noise exposure, normal hearing; Group 2 (G2): 30-50 years, low noise exposure, normal hearing; Groups 3 and 4 (G3/G4) same age range as G2, high noise exposure, normal hearing (G3); high noise exposure, elevated 3-8 kHz PTA (G4). Participants underwent audiological assessments and MRI to measure auditory nerve structure (cross-sectional area; CSA) and microstructure (diffusion fractional anisotropy; FA), along with whole-brain structural measures and resting-state functional MRI of the auditory pathways.Tinnitus Functional Index (TFI) scores differed across groups, with significantly higher scores in noise-exposed participants. Audiological differences were driven by G4, with no significant G2-G3 differences. While CSA showed no significant overall group or age effects, post-hoc t-tests revealed differences between G1 and other groups (G1-G2: p=0.007, G1-G3: p=310-6; G1-G4: p=0.0003) and between G2 and G3 (p=0.01). FA differed significantly across groups (H(3)=11.96, p=0.008) with post-hoc Mann-Whitney U tests (Bonferroni-Holm-corrected across comparisons) revealing that G1 had significantly higher FA than G4 (p<0.001). Further MRI analyses of the cortical brain measures are ongoing. Early findings from this study suggest that the effect of noise exposure on the auditory pathway is likely much smaller than the effect of ageing or noise-induced hearing loss.

Speaker: Rebecca Dewey (University of Nottingham)

384 Digit‑Triplet Speech‑in‑Noise Performance Is Minimally Sensitive to Noise Exposure History: Evidence from a Large Population‑Based Sample of Middle- aged and Older Adults

Animal studies suggest that the earliest consequences of loud noise exposure may be audiometrically unapparent, or “hidden,” to conventional threshold‑based tests. In humans, difficulty understanding speech in noise (SiN) has been widely hypothesised as a primary functional consequence of such hidden hearing damage, leading to the view that objective SiN tests—such as the digit‑triplet test (DTT)—may serve as integrated markers of auditory peripheral integrity across both subclinical and overt stages. Here, we critically evaluate this assumption by characterising the inter-relationships among multiple hearing measures (DTTs, tinnitus, hearing-aid use, and self-reported hearing difficulty) and their dependence on key risk factors (age and lifetime noise exposure) in a large (N = 14,917) population-based sample from the UK Biobank (UKBB) imaging study. Descriptive analyses showed that, compared with self‑reported hearing measures and tinnitus, DTT performance was more sensitive to age and less sensitive to lifetime noise exposure. Path analysis suggested that the weak association between noise exposure—particularly leisure exposure to loud music—and DTTs may, at least in part, be due to a moderating effect of cognitive effort. Factor analysis indicated that the UKBB hearing measures are influenced by two primary causal drivers: one loading on DTTs and hearing aid use, consistent with clinically overt hearing damage, and a second loading on tinnitus and self-reported hearing difficulty, suggestive of a distinct subclinical damage dimension. Together, these findings challenge the interpretation of digit triplet SiN performance as a unified marker of hidden and overt auditory damage.

Speaker: Katrin Krumbholz (University of Nottingham)

A17.05 Sound field estimation and reconstruction: S403

Conveners: Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology), Samuel A. Verburg (Technical University of Denmark)

385 Superresolution of spatial room impulse responses with boostlet transforms

Reconstructing room-acoustic wavefields from sparse measurements remains challenging in the presence of diffraction and scattering, where multiple propagation components overlap in space–time. Classical representations, such as plane waves, provide global field expansions but require significant coefficients to capture localized interactions and destructive interference. Multi-scale approaches, such as wavelets and shearlets, improve spatial adaptivity but are not explicitly aligned with the physics of wave propagation. In this work, we propose a boostlet-based framework for space–time superresolution of acoustic fields (i.e., recovering missing microphone responses) that captures propagating wavefronts as sparse, local structures in joint space–time coordinates. The reconstruction is formulated as an ℓ1-regularized inverse problem and evaluated under various undersampling rates (half and 1/3 of the data) and severe noise levels (down to 5 dB signal-to-noise ratios). A room-impulse response dataset is used to assess the method against baselines. Our results highlight two main observations. First, the proposed approach improves the overall reconstruction of propagating structures compared to the baselines, particularly in regimes with multiple overlapping wave components. Second, the performance of the proposed approach is governed by a spectral filter that penalizes boostlet coefficients in a frequency-dependent manner: a higher/lower cut-off frequency in the early/late parts. The results demonstrate the potential of boostlet representations for sparse sound field reconstruction and motivate further investigation and extensions to higher spatial dimensions.

Speaker: Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology)

386 From eigenfunctions to sources: Four extremes of Gaussian boostlet kernels

Parametric representations of wavefields are central to sound field estimation and reconstruction, where capturing both propagation geometry and localization is essential. We introduce a parametric framework, called Gaussian boostlet kernels, a variation of boostlets that uses Gaussian functions to control their frequency bandwidth and phase-speed selectivity (geometric spread). We explore the expressive power of these boostlet kernels through asymptotic limits in which the Gaussian widths become infinitely broad or narrow. This analysis reveals four distinct regimes. When both frequency bandwidth and geometric spread vanish, the kernel reduces to a classical monochromatic plane wave. Vanishing frequency bandwidth and broad geometric spread yield cylindrical wavefields represented by Hankel functions, capturing a boosted superposition of plane waves. Conversely, infinite frequency bandwidth and vanishing geometric spread produce transient plane-wave sources localized along characteristic rays, consistent with trace-wave interpretations in array acoustics. In the fully broadband limit, the kernel converges to the space–time Dirac delta, equivalent to the Green’s function source. These findings demonstrate that Gaussian boostlet kernels smoothly interpolate between plane and cylindrical eigenfunctions, and between impulsive plane and point sources. This provides a geometric connection among wavenumber processing, Green’s function formulations, and wavefield decompositions, and shows potential for parametric sound-field reconstruction, source separation, and room-acoustics analysis.

Speaker: Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology)

387 Extending the frequency range of sound field reconstruction methods using perceptual criteria

Sound field reconstruction methods have increasingly been employed to estimate acoustic field quantities of complex sound fields across space. These approaches typically describe the observed field as a superposition of basis functions that satisfy the wave equation. Beyond established applications in sound source radiation, material characterisation, and active noise control, their use in sound field reproduction remains limited by physical constraints, particularly those imposed by the spatial aliasing frequency introduced by sensor positioning. In this work, we address this limitation by formulating the wave coefficient estimation task to incorporate perceptually-motivated criteria. Specifically, we exploit the accuracy of physical models for magnitude and phase estimation below the aliasing frequency. At higher frequencies, inspired by the duplex theory, the proposed method focuses on energy estimation, maintaining the directional properties of the field and providing plausible sound field reconstructions, even when large apertures are considered.

Speaker: Antonio Figueroa-Duran (Universidad Politécnica de Madrid)

388 Sound Field Reconstruction with a Differentiable Source Model

Sound field reconstruction from sparse microphone measurements is an ill-posed problem highly sensitive to modeling choices. Approaches based on elementary wave expansions, such as plane wave expansions (PWE) or  the equivalent source method (ESM), are popular and widely used, but they require domain discretizations prone to model mismatch. More recent physics-informed neural networks (PINNs) are more flexible but only enforce physical constraints weakly. In this study, we propose a sound field reconstruction approach that retains the physical consistency of the ESM while replacing fixed source parameters with learned parameters. We evaluate the method on both simulated and real measurements and observe improved reconstruction accuracy compared with standard PWE, ESM, and PINN baselines.

Speaker: Samuel A. Verburg (Technical University of Denmark)

389 Data-Driven Low Frequency Concert-Hall Transfer Function Reconstruction using Flow Matching Neural Process

Low-frequency sound reproduction in concert-halls can be significantly improved through the use of spatial sound field control methods. Such techniques may require accurate spatial sampling of the room transfer functions (RTFs) across the venue for every subwoofer loudspeaker. Respecting the Nyquist criteria spatially is impractical in real-world concert scenarios where measurement time is limited. To that matter, RTF reconstruction methods estimate RTFs from a reduced number of measurements. Data-driven approaches, including deep learning models trained on simulated datasets, have recently shown promises in this area. We study Flow Matching Neural Processes (FMNPs) for the RTF reconstruction task. FMNP integrates the flow matching paradigm into the neural process framework: a U-Net transformer encoder aggregates measurements at arbitrary spatial positions and generates RTFs by learning a velocity field transporting samples from an initial Gaussian distribution toward the RTF distribution. To train and compare FMNP alongside existing methods from the literature, we introduce a dataset of room impulse responses simulated with a GPU-accelerated finite difference time domain method on large concert hall geometries, with variability introduced through room deformations and varied absorption conditions. This dataset is used to evaluate the practical feasibility of data-driven RTF reconstruction at scale.

Speaker: Louis Reine (LMSSC, CNAM Paris, HESAM Université)

390 Learning-Based Multi-Listener Sound Field Reproduction Using Boundary Velocity Targets

This paper studies learning-based sound field control for a prototype large-audience scenario with ten loudspeakers and multiple listeners distributed over a listening region. Two identical complex-valued multilayer perceptrons are compared. One is trained on a boundary normal-velocity objective and one is trained directly on binaural ear pressures. The velocity-based network is first validated against the analytical controller based on previous work by Shin et al and shown to closely reproduce its region-based behavior. The two models are then compared using common region-based and listener-based metrics, including pressure error and binaural cue errors based on Interaural Level Difference (ILD) and Interaural Phase Difference (IPD). For ten listeners, direct ear pressure training defines an overdetermined 20-ear/10-loudspeaker problem, whereas the velocity-based model provides a more robust trade-off between regional accuracy and perceptual performance. The results indicate that region-based velocity training is a promising approach for overdetermined multi-listener sound field reproduction.

Speaker: Nara Hahn (Institute of Sound and Vibration Research)

Full paper (PDF)

A04.00 Education, Public Outreach and History in Acoustics: S023

Conveners: Lukas Aspöck (IHTA, RWTH Aachen University), Iván Herrero-Durá (SoundPLAN GmbH), Michael Buba (Technical University of Munich)

391 Discover: an exhibition exploring acoustics through reimagined iconic vinyl album covers

Discover is an exhibition that aims to introduce the general public to acoustics through iconic vinyl album covers. Each of the 25 panels in the exhibition features the original cover, the cover redesigned to illustrate a particular acoustic phenomenon, and scientific explanations. Each panel comes with an interactive module offering an experience, an audio clip, a video, or an object to observe. The exhibition was presented in Le Mans (France) from January 17 to February 7, 2026, as part of the Le Mans Sonore sound biennial, and welcomed 800 visitors. This article presents the process of creating the exhibition by describing the general concept, the various choices made, the different stages of development, and the working methods. Finally, the public's reception will be analyzed through feedback obtained during visits or written testimonials left by visitors.

Speaker: Manuel Melon (Université du Mans - LAUM - UMR 6613 CNRS)

Full paper (PDF)

392 Interdisciplinary approaches in the design of music performance spaces: Insights from the ATHENS programme

The design of spaces for music is an inherently interdisciplinary field that affects all stages of the design and construction process, as well as the different technical disciplines involved. While a basic understanding of acoustics is essential, it is equally important to integrate knowledge of structures, building services, lighting, and venue planning.Within the framework of the ATHENS programme—a network of 16 European higher education institutions, including the Technical University of Madrid—a one-week intensive workshop was held in March 2026. The workshop provided introductory training in acoustics, lighting, venue planning, building services, and structures, complemented by technical visits, including the National Auditorium, with the aim of equipping students with practical tools for the design of such spaces.The purpose of this paper is to present the interdisciplinary approach adopted, as well as the main findings derived from the analysis of five European concert halls selected during the workshop.

Speaker: Teresa Carrascal García (Instituto Eduardo Torroja for Construction Sciences - IETcc-CSIC)

Full paper (PDF)

393 The Community Hub for Open-source Room Acoustics Software (CHORAS) in Education

Teaching (room) acoustics to (under)graduate students is generally considered a challenge for many reasons. Firstly, acoustics theory is quite technical and mathematical by nature, and is therefore often perceived as quite abstract by many students. Secondly, acoustics theory does not naturally lend itself to visual representations when compared to, for instance, architectural design, or other courses in their curriculum. Therefore, in acoustics education, the main challenge is to make abstract concepts tangible for the students, while at the same time keeping the solution affordable.This paper presents the use of the Community Hub for Open-source Room Acoustics Software (CHORAS) in an educational context. CHORAS is a web-based user interface that allows users to run room acoustics simulations using open-source simulation packages as developed in the research community. Users provide room geometries, source/receiver positions, and material properties. CHORAS then provides an impulse response, which can be auralized by the interface, as well as room acoustic parameters, including T30, D50 and C80. CHORAS was used during one bachelor’s and one master’s course at the Eindhoven University of Technology (TU/e). Students created their geometry from scratch based on an existing space in which they carried out room acoustics measurements, and used the outputs provided by CHORAS to iteratively improve upon this.

Speaker: Silvin Willemsen (Eindhoven University of Technology)

394 Using a Web-based Reverberation Time Prediction Tool for Teaching Room Acoustics

Reverberation time is a fundamental room-acoustic parameter and a key target in many acoustic design guidelines. It is influenced by room geometry and the absorption properties of surfaces and furnishings. Understanding sound propagation, absorption, and their impact on reverberation time is essential for teaching the fundamentals of room acoustics. Predicting reverberation time is straightforward when applying the diffuse-field assumptions of Sabine or Eyring. However, despite the simplicity of the underlying physical principles and mathematical models, calculations can become cumbersome due to complex room geometries, multiple surfaces, interior elements, and frequency-dependent absorption coefficients in octave bands.A user-friendly web application allows students to experiment with different furnishing arrangements and acoustic treatments, providing immediate feedback on their effects on reverberation time. Additionally, the tool includes target reverberation times based on national standards from various countries, tailored to different room functions, helping students understand room-acoustic design requirements. Students can investigate the impact of various acoustic treatments on reverberation time, analyze how different room functions necessitate specific acoustic conditions, and explore the variations in national standards and their target criteria. The ability to listen to the predicted acoustic conditions fosters a deeper perceptual understanding of the relationship between reverberation time and room acoustics.This paper presents examples of integrating the tool into classroom teaching and digital learning environments, discussing its potential for interactive and exploratory learning in room acoustics.

Speaker: Stefan Weigand (SoundPLAN GmbH)

A12.01 Numerical methods for acoustics and vibration: S081

Conveners: Dionysios Panagiotopoulos (KU Leuven Campus De Nayer), Stefan Schoder (IGTE, TU Graz), Maarten Hornikx (Eindhoven University of Technology)

395 Acoustic Scattering from a Moving Object Driven by a Moving Emitter Using Time-Dependent Physical Optics

We study the problem of calculating the acoustic field scattered by a moving rigid object in response to a moving emitter in three-dimensional space. The object is assumed to be locally planar relative to the wavelength, enabling the use of the physical optics approximation to model wave scattering. However, due to the motion of both the emitter and the object, conventional frequency-domain physical optics formulations are not applicable. Instead, we develop a time-domain variant of this approach that explicitly accounts for their motion. The analysis focuses on the near-field regime, where the combined motion induces a complex wave field that cannot be described by a simple Doppler effect. The proposed algorithm accommodates arbitrary emitter trajectories, arbitrary translational motion of the object, high velocities relative to the speed of sound, and emission of arbitrarily shaped signals.

Speaker: Asaf Kor (Tel Aviv University)

396 High-Performance 3D Wave-Based Modelling of Outdoor Acoustic Propagation

Most available engineering tools for outdoor acoustic simulation rely on ray-based methods, or sometimes on even simpler energy-based formulas. These simplifications have long been necessary because of the limited computational resources and time frames available to engineers. However, modern numerical methods and the increasing computational density of hardware are bringing fully 3D wave-based simulations within reach.In this presentation, I will summarize the research behind SoundSim360, a new simulation tool for large-scale outdoor noise propagation. I will focus on its computational engine, which is based on a high-performance implementation of summation-by-parts finite differences (SBP-FD). SBP-FD enables provably accurate and stable schemes while supporting complex boundary conditions and source models, and it benefits from the efficiency of finite differences on modern hardware.SoundSim360 approximates solutions to the second-order acoustic wave equation, directly computing pressure fluctuations in 3D space and time. It is implemented in custom CUDA C++ code in a matrix-free formulation, enabling highly efficient memory use. On a mid-range consumer-grade GPU, SoundSim360 can compute a noise map from a point source over a domain spanning 120 wavelengths in each direction in approximately 20 minutes.The presentation will give an overview of the numerical method and CUDA implementation. I will also show several application examples where SoundSim360 has been used, including wind turbine, traffic, and underwater simulations.

Speaker: Gustav Eriksson (Uppsala University)

397 Identifying moving acoustic sources in the wavenumber frequency domain

In the past a method was presented, that allows to separate frequency components of a moving source. In the wavenumber frequency domain the Doppler shift leads to a rotation of the axis, as long as direction and velocity of the moving source are constant. Originally, a microphone array with 8x8 microphones was used. This leads to a Fourier transformation with eight bins in the direction of movement. By far not enough to estimate the spectrum with narrowband singularities. A new approach shall use a one directional array with 1024 bins or even more. The results of the new approach are presented.

Speaker: Holger Waubke (Acoustics Research Institute, ÖAW)

398 CRUNA – An Open-Source Finite-Difference Time-Domain Simulation and Optimization Framework based on Adjoint Gradients

Wave-based simulation methods are well suited for analyzing acoustic problems in which wave phenomena such as diffraction and scattering play a dominant role. Among these, the finite-difference time-domain (FDTD) method is a robust and versatile approach.This work presents CRUNA, an open-source 3D FDTD solver with an integrated optimization framework based on adjoint gradient computation.While the modular codebase addresses a wide variety of flow-related problems, this contribution focuses on the acoustic and aeroacoustic components of the solver and their application to wave propagation and optimization tasks.The solver supports several sets of governing equations, ranging from linearized acoustic formulations to more comprehensive aeroacoustic models. It provides explicit time-stepping schemes, as well as explicit and implicit spatial discretization methods. CRUNA is highly parallelized to enable fast execution in high-performance computing environments. Non-reflective characteristic boundary conditions are implemented, and immersed boundary methods based on effective volume and flow resistivity enable an accurate representation of complex geometries and impedance boundaries.Representative examples demonstrate the framework, including aeroacoustic simulations, sound source localization and synthesis, and pinna-related transfer function computations. By providing CRUNA as open-source software, this work aims to support transparency, reproducibility, and further development within the computational acoustics community.

Speaker: Arne Birk Hölter (Technische Universität Berlin, Audio Communication Group)

Full paper (PDF)

399 Irrotational Acoustic Displacement Formulation for Fully Coupled Elastoacoustics

Coupled elastoacoustic problems, involving the interaction between an acoustic field and a vibrating structure, play a central role in low- and mid-frequency vibroacoustics. Displacement-based formulations are attractive due to their symmetry and their natural description of structural dynamics; however, their extension to the acoustic field often leads to numerical difficulties, such as spurious rotational modes and challenges in enforcing interface continuity. In this work, we present a displacement-based formulation that preserves the irrotational nature of the acoustic field by construction. The approach is developed within a Rayleigh–Ritz framework, using acoustic particle displacement and structural displacement as primary variables, and employs spatial derivatives of Gaussian basis functions to obtain symmetric algebraic systems while avoiding spurious circulation modes.A key feature of the method is the treatment of the fluid–structure interface. Displacement continuity is enforced strongly via a nullspace method, eliminating the need for explicit coupling matrices, while traction continuity is naturally satisfied in a weak sense through the variational formulation. This results in a monolithic and numerically robust solution strategy for coupled interaction. The approach is validated through a set of benchmark problems of increasing complexity showing excellent agreement with finite element solutions in terms of modal frequencies and mode shapes.

Speaker: Jie Deng (La Salle, Universitat Ramon Llull)

A16.02 Reliable Characterization of Acoustic Absorption, Reflection, and Scattering: S410

Conveners: Mélanie Nolan (Universidad Politécnica de Madrid), Marco Berzborn (Eindhoven University of Technology), Jonathan A. Hargreaves (Acoustics Innovation Institute)

400 Experimental and numerical investigation of acoustic scattering using boostlets

Modelling the interaction of sound waves with objects and surfaces is challenging due to the complex space–time behaviour of acoustic fields. In practice, surface roughness and finite object dimensions cause sound reflections to deviate from ideal specular behaviour, producing scattered and diffracted fields. While separating the total sound field into incident, reflected, scattered, and diffracted components is conceptually appealing, achieving it from spatial measurements remains a non-trivial problem. This work applies boostlet decompositions to simultaneously separate and analyse these components using a 1D line microphone array near the scatterer. The core idea is that each wave is the result of thresholding sparse boostlet coefficients: only boostlets that match the wave spectrum pass the threshold. The method is applied to a benchmark problem in which a spherical wave impinges on a circular disk of radius 40 cm. Measurements are performed in an anechoic chamber using a 6-DOF robotic arm mounted on a linear track, sampling the acoustic pressure along lines in front of and behind the disk, both on-axis and off-axis. Numerical predictions are obtained using an edge-diffraction-based method that separates the time-domain solution into geometrical-acoustics and diffraction components. Finally, an attempt to quantify the disk’s scattering coefficient using this framework is presented for both simulated and experimental data.

Speaker: Augusto Fantinelli (Human-Environment Research group, La Salle - URL)

Full paper (PDF)

401 Challenges in scattering properties characterization of 3D- printed hybrid acoustic materials

Additive manufacturing technologies enable the development of innovative and high-performance acoustic materials, allowing the design of complex structures. These fabrication techniques support a high degree of geometric customization that challenges the acoustic standardized characterization methods and their feasible application in the design process. One aspect that has been unexplored is the characterization of hybrid acoustic materials (HAMs) that integrate sound absorption and diffusion within a single element. This contribution focuses on the sound scattering performances of 3D printed HAMs, designed and developed by BioG3D (Greece), and explores the challenges of their characterization in a small-scaled reverberation room (Politecnico di Torino, Italy). The effects of ten 3D quarter-circle Truchet tiles combined in two different configurations have been fabricated using a large format industrial LCD 3D printer and daylight resin. The experimental investigations aimed at providing insights on the sensitivity of the measurement procedure of the random-incidence scattering coefficient, as defined in ISO 17497-1:2004, applied to the two different configurations. The results showed that the two tested scattering surfaces do not show significant differences despite the variation of the two configurations, which allows for a higher flexibility in the space design and application. It is highlighted that the characterization method is particularly demanding, as it relies on two custom 3D printed scaled samples which requires higher accuracy and precision in the printing process, consequently increasing fabrication costs.

Speaker: Louena Shtrepi (Politecnico di Torino)

Full paper (PDF)

402 Simulation of the acoustics in open-air venues using bidirectional scattering

In open spaces such as amphitheaters or open-air concert venues, surprisingly large reverberation times can be found, despite the absence of the ceiling and side walls. Previous research on Greek and Roman theaters showed that reflections between the stage building and the steps in the audience area determine the main components in the impulse responses. So far, such models have been based on random-incidence scattering coefficients, and the results showed that a reverberant field can indeed be simulated in good agreement with measurement results in one-third octave bands. Periodic surface structures such as steps, however, lead to a specific directional scattering pattern with distinct frequency-dependent lobes. With more detailed characterization of this effect, particularly with narrow-band analysis, the reverberant field can be analyzed further. In the presentation, the simulation using the concept of bidirectional scattering will be presented, and the application in the above-mentioned case study.

Speaker: Anne Heimes (IHTA, RWTH Aachen University)

Full paper (PDF)

403 Bayesian Impedance Estimation in Room Acoustics Using Model Order Reduction

For computational room acoustic models, a suitable characterization of boundaries is required. Yet, boundary conditions are often not accurately known.One possibility is to infer boundary impedance models from in-situ measurements using a comparatively simple measurement setup. Since measurement noise and modelling errors cannot be avoided, we formulate the estimation problem in a Bayesian framework. By prescribing a suitable noise model, we infer statistical information on the impedance parameters of interest. This provides not only a single best-fit estimate, but also uncertainty information in the form of posterior expectations and variances.If several impedance parameters are unknown, the sampling procedure may require a large number of finite element solutions, which can make the overall computation expensive. To reduce this cost, we employ Krylov-subspace-based model order reduction techniques that allow for faster evaluations of the forward model. In this talk, we present numerical results for the inferred impedance statistics and compare the computational performance of the full-order and reduced-order models.

Speaker: Matthias Blau (Jade Hochschule, IHA)

404 In-situ Estimation of Sound Absorption Coefficients at Grazing Incidence with Boostlets

In practice, the sound absorption of finite objects at grazing incidence is dominated by strong edge diffraction and non-local reaction. This paper proposes a 1D-microphone array method to separate incident, specularly reflected, and diffracted waves using the recently introduced boostlet transform. The method is inspired by classical ideas of separation via time-windowing and extended to a convenient approach with space-time-windowing. As long as the waves are spatio-temporally resolved (e.g., Rayleigh criterion), the separation of each wave amounts to thresholding a cross-correlation of boostlet coefficients. The sound absorption (with and without edge-diffraction) can then be calculated from the signal-energy ratio between the separated waves. The method is tested with a 1D microphone grid scanned with a movable microphone, a highly directive panel sound source oriented at grazing angles of 8° and 30°, and two sample materials: a bioboard and a rigid surface. The method currently requires manual tuning for each array configuration, and further work is underway to automate as many steps as possible. Ongoing efforts focus on applying these ideas to scattering coefficients and parametric studies across multiple sample sizes.

Speaker: Arif Onur Yurek (Aalto University)

Full paper (PDF)

A19.03 Methodologies for sound-driven design and education: S141

Conveners: Stefano Delle Monache (Ircam STMS Lab), Nicolas Misdariis (Ircam STMS Lab)

405 Sound perception and design in higher education: the case of the 'Acoustic Design' course at the University of Technology of Compiègne

Sound perception and design courses are rarely included in higher education, particularly in engineering programs which lead to industrial activities where product design plays a significant role. We present the "Acoustic Design" course offered in the University of Technology of Compiègne. This course addresses the fundamentals in psychoacoustics, audio recording and analysis, source categorization, verbal description, sound synthesis and design. At least two external professionals, from academia, industry or independent sound designers, are invited each semester. Two short projects, one in sound product perception and one in sound design, are conducted by students working in pairs. A focus will be made on the sound design projects: recent subjects include sound logos as well as soundscape generations, first based on sequencers and secondly using quadriphonic loudspeaker configuration. A special attention is paid to the ideation process: students are required to follow a systematic methodology based on preliminary sketches, inspired by sound representations used for the analysis of electroacoustic music. Selected examples of student works will be presented with schematics and audio samples.

Speaker: Arthur Givois (Université de technologie de Compiègne)

406 TWAF Framework as an Educational Tool for Soundscape Ideation in Complex Socio-Technological Environments: A case study in sound-driven design for Intensive Care Units

The TWAF (Designing THE, WITH, AGAINST, sound FOR) framework, originating from sound-driven design, offers a structured yet flexible lens for addressing sound-related challenges in complex socio-technological environments (Delle Monache, 2022; 2024). Rather than treating sound as an isolated artifact, the framework reframes design as a multi-perspective exploration of listening experiences, integrating sonic, experiential, technical, and cultural dimensions. This makes it particularly relevant as an educational tool in early-stage ideation processes, where problem framing can benefit from multiple-perspective input. As an educational tool, TWAF supports designers in systematically expanding the design space. It encourages divergent thinking by prompting students to reinterpret a single problem through multiple lenses, while also fostering convergence by linking each perspective to specific problem and solution spaces (creative, integrative, mitigative, purposeful). In the context of Intensive Care Units (ICUs), as explored in van der Stelt’s listener-centric MSc thesis (TU Delft Repository), sound-related challenges emerge from the interplay between multiple users, technologies, and functional demands within a shared acoustic biotope (Özcan et al, 2022, 20226). Traditional approaches—focused on noise reduction or single sound sources—fail to capture this complexity. The sound-driven design framework addresses this gap by splitting ideation into four complementary perspectives: generating new sonic expressions (THE), integrating sound within broader listener experiences (WITH), mitigating acoustic harm (AGAINST), and aligning sound with cultural and contextual meaning (FOR). Accordingly, by providing a shared vocabulary across disciplines, TWAF facilitated collaboration and supported the development of holistic, listener-centric design strategies that are fitting with the broader requirements of ICUs. As such, the framework promises to enhance students’ capacity to navigate socio-technological complexity with human factors considered.

Speaker: Elif Özcan (Erasmus MC / TU Delft)

407 Understanding sound experiences at work: A diary study in open-plan offices

Noise is a persistent issue in open-plan offices, and soundscape temporal and spectral variations were found to explain part of perceived annoyance. While acoustic treatments and sound masking systems can help reducing the overall level and speech intelligi- bility, they are not designed to adapt to those varia- tions. Other approaches based on a transformation of the existing soundscape have been investigated, but were never applied to the office context. To explore the relevance of soundscape augmentation approaches in open-plan offices, we seek to gain consistent in- sights about people’s sound experiences at work. This paper presents a study based on a human-centred methodology, where participants reported their lived experiences through a diary and were invited to ver- balize about their observations and feelings during semi-directed interviews. Data collected from the di- aries showed that speech was identified as the most frequently perceived sound source, although it was not always badly perceived. Affective judgment such as perceived annoyance and pleasantness seemed to be partly explained by noise sources’ level, relevance and frequency. Also, earphones with music was the most widely used coping strategy. We hope to access a deeper understanding of the noise issue by further analyzing data from the interviews.

Speaker: Clara Boukhemia (Ircam STMS Lab)

Full paper (PDF)

408 Accessible Acoustics: Cost-Efficient Methods for Sound-Driven Architectural Design

In recent years, architectural design has increasingly emphasized performance-based approaches alongside formal and spatial considerations (Hensel and Menges, 2008), yet acoustics, despite its central role in spatial experience, remains underintegrated due to limited access to iterative, accessible evaluation tools (Siltanen et al., 2010; Badino et al. 2020). Conventional acoustic testing depends on costly physical apparatuses or computational simulations limited by standardized material datasets, underscoring the need to rethink accessible methods and prompting the question: how can iterative, resource-efficient processes enable more precise, adaptable control of aural qualities in architectural design?Addressing this gap, Accessible Acoustics introduces an impedance tube system using off-the-shelf components, enabling rapid material-based prototyping and embedding acoustic evaluation within early-stage design workflows and education. Built for under $600, it consists of PVC piping, a full-range speaker, precision microphones, and a two-channel audio interface, coupled with a custom MATLAB script. The system automates logarithmic sine sweep generation (80–6300 Hz), dual-channel signal acquisition, and absorption coefficient calculation using frequency-domain transfer function methods. To validate accuracy, the setup is calibrated against ASTM E1050-12 through comparative testing with a Brüel & Kjær Type 4206 impedance tube. Statistical analysis and correction curves align the custom system’s output with industry-standard results, demonstrating reliable performance. Beyond technical validation, the project positions acoustic measurement as both a design and pedagogical tool, establishing a foundation for data-driven, performance-based workflows. By enabling rapid iteration and direct testing of non-standard material assemblies, Accessible Acoustics functions as both evaluation and representation, informing comparisons, refinements, and sound-driven design strategies.

Speaker: Zackery Belanger (Umbel)

Full paper (PDF)

409 Sound Design for Electric Vehicle – a step ahead towards a tool-based design methodology

Designing sound for Electric Vehicle is an approach being progressively assimilated in the automotive product development and considered one of the strategic assets for the sustainable future of mobility. Nevertheless, this design process presents several challenges, since it requires multidisciplinary expertise and skills, collaboration and decision-making, among designers, sound experts, engineers, product experts, and other professionals. In this context, the Ircam / Sound Perception & Design group has been collaborating since 2009 with the French car manufacturer Renault. We present our latest project, developed in 2024-25, which allowed us to generate and produce not only solutions but also knowledge on both the manufacturer and the research group sides. The project enabled us to advance the experimentation and implementation of an integrated methodology based on novel and proprietary collaborative sound design protocols, methods and tools, thus making this project a good case of sound-driven design and collaboration. The ins and outs of this fruitful and successful project will then be presented, discussed, and put in perspective for possible future works.

Speaker: Nicolas Misdariis (Ircam STMS Lab)

Full paper (PDF)

A20.04 Previously Impossible: Advances in MRI technology for speech research: S148

Conveners: Steven Lulich (Indiana University), Daniel Aalto (University of Alberta)

410 The Iseult 11.7T human MRI project – achievements and prospects

The Iseult project started in 2001, with one fundamental idea in mind being to design and build a human brain “explorer” consisting of a whole-body magnet of 11.7T field. After nearly 20 years of research and development, integration and tests, first images were acquired in vitro with this unique MRI scanner in 2021. Approval from the regulatory agency was then obtained in 2023 to scan a first series of 20 volunteers, leading to exquisite anatomical in vivo images and proving feasibility. A new approval to continue scanning in vivo on a broader scale was obtained in 2025 with new anatomical and functional MRI results. Although there are some daunting challenges to fully leverage the potential of the machine, some of the results are beyond initial expectations and are very promising for future exploration of the human brain. This talk will summarize the 20-year Iseult odyssey covering the history, important technological milestones and achievements (safety, first in vivo results etc), and the hopes it engenders in brain exploration.

Speaker: Nicolas Boulant (Atomic Energy Commission)

411 From Functional Units to Acoustics: Decoding Tongue Biomechanics in Speech Production from Tagged and Diffusion MRI

The human tongue is a muscular hydrostat whose intricate, three-dimensional muscle activity drives nearly every aspect of articulate speech. In this talk, I present an integrated computational framework for studying tongue biomechanics during speech using tagged and diffusion MRI. First, I introduce a sparse non-negative matrix factorization approach that decomposes voxel-level motion into functional units—data-driven groups of cohesive local muscle regions that compress, expand, and move together during protrusion and simple speech tasks—yielding subject-specific building blocks of lingual behavior. Second, I describe how Granger causality analysis applied to fiber-aligned strain time-series reveals the sequential, predictive interactions among individual tongue muscles and among the functional units themselves, illuminating how coordinated motion unfolds over time during utterances. Third, I show how these internal motion representations can be linked back to the acoustic signal through a plastic transformer that synthesizes speech audio directly from tagged-MRI weighting maps, closing the loop from anatomy to motion to sound. Together, these complementary analyses move from anatomy to motion to acoustics, offering speech scientists a quantitative, subject-specific window into normal articulation and a principled foundation for understanding, predicting, and ultimately rehabilitating disordered speech in clinical populations such as glossectomy, ALS, and motor speech disorders.

Speaker: Jonghye Woo (Mass General Brigham)

412 Advanced MRI for Tongue Muscle Architecture, Swallowing, and 4D Speech Imaging

The tongue's complex interdigitating muscle architecture makes it one of the most challenging organs to image. Understanding its structure and dynamics is critical for assessing functional outcomes in tongue cancer patients, who frequently suffer long-term impairments in speech and swallowing after surgery. We first demonstrated that diffusion-weighted MRI with constrained spherical deconvolution (CSD) at 3T can resolve the tongue's crossing muscle fiber architecture in vivo. This work has since been validated against histology and used to quantify muscle strain during speech. Then, to enable the acquisition speeds required for dynamic imaging, we developed a dedicated 12-channel flexible receiver coil providing approximately twice the signal-to-noise ratio of a conventional head-and-neck coil. Using this coil, we developed volumetric real-time MRI of swallowing at 12 frames per second using compressed sensing, capturing full 3D bolus and laryngeal dynamics and enabling detection of aspiration outside a single midsagittal slice. Most recently, we extended this to 4D speech MRI, achieving sub-2 mm³ resolution at 50 ms temporal resolution combined with automated deep learning segmentation of the vocal tract. These advances collectively enable previously impossible visualization of tongue biomechanics, with direct applications in surgical planning and rehabilitation of tongue cancer patients.

Speaker: Aart Nederveen (Amsterdam UMC)

413 Seeing Inside Stuttering: Articulation in Stuttered Speech Revealed by Real-Time MRI

Real-time MRI has transformed the study of speech production by allowing direct visualization of vocal tract movements during running speech. Beyond typical speech, its application to clinical populations has also been increasing. Stuttering offers a tempting test case: stuttering disfluencies have long been identified from auditory-perceptual records, and yet the underlying articulation has rarely been investigated, largely due to challenges in articulatory instrumentation and stuttering elicitation in laboratory settings. We present a real-time MRI approach to studying stuttered speech, combining dynamic vocal tract imaging with a suite of connected speech tasks designed to elicit stuttering during MRI scanning. Seven adults who stutter completed interleaved passage reading, interview, and time-limited picture description tasks. Midsagittal vocal tract movements were imaged using a 0.55T MRI system with high-performance gradients and a custom upper-airway coil. The acquired images have a spatial solution of 2.3 mm and temporal resolution of 10.06 ms. This protocol yielded a rich corpus of stuttered and fluent speech. The resulting data reveal articulatory dynamics that are not recoverable from acoustics alone. Articulation in stuttered speech is largely organized according to the intended linguistic tasks, instead of random, spasmodic movements. Two basic articulatory patterns underpin perceived stuttering: fixation and oscillation of the proximal articulator for the stuttered segment, typically a syllable-onset consonant. Coarticulation between the syllable onset and nucleus appears to be preserved during stuttering, challenging hypotheses that link stuttering occurrences to speech planning disruptions. Together, these findings yield implications for both the phenomenology and theories of stuttering.

Speaker: Yijing Lu (University of Potsdam)

414 Toward Brainstem fMRI Analysis of Sustained Phonation

Very little work has been done on functional imaging of the cranial nerve nuclei in the brainstem in the context of motor control, including only one such study of phonation (Smith et al., 2025, doi:10.21437/Interspeech.2025-2444). That study presented preliminary results from 4 participants, using two candidate data collection protocols and demonstrating that differential brainstem activation patterns could successfully be elicited from sustained [i:] and [a:] vowels, with a potential to localize activations around relevant cranial nerve nuclei. For this presentation, we will present a larger dataset (N=6, including 4 new participants), all using the selected protocol. Furthermore, we report progress toward Region Of Interest (ROI) analysis based on anatomical localization of the relevant cranial nerve nuclei.

Speaker: Steven Lulich (Indiana University)

A24.00 HRTF processing and modeling: S167

Conveners: Lorenzo Picinali (Imperial College London), Julie Meyer (King's College London)

415 HRTFPyKit: A Python Library for HRTF Research Workflows

Head-related transfer functions (HRTFs) describe how the human head, torso, and pinnae filter sound arriving from different directions. They are unique to each listener and are an important part of binaural rendering, spatial audio, and virtual reality systems. HRTF research involves not only acoustic analysis but also the management of Spatially Oriented Format for Acoustics (SOFA) files, public datasets, and reproducible data pipelines for analysis and deep learning model training. In practice, this work often relies on custom scripts combined with existing software packages, resulting in heterogeneous workflows that make experiments harder to reproduce or extend. These challenges motivated the development of hrtfpykit, a Python library for HRTF research workflows. It includes a SOFA layer for reading, writing, and editing SOFA files. The hrtf layer represents HRTF data as a unified object containing synchronized impulse response and transfer-function representations and supports acoustic transformations and spatial and spectral selection. The plots layer provides visualization tools for HRTF inspection and comparison in the time and frequency domains. Finally, the datasets layer standardizes access to public HRTF datasets through declarative specifications and configurable transformations, enabling reproducible pipelines for training deep learning models, particularly for HRTF individualization.

Speaker: Jose J. Lopez (Universitat Politecnica de Valencia)

Full paper (PDF)

416 Nonlinear Harmonic Prediction and Compensation in HRTF Measurements

The exponential sine sweep (ESS) technique became the common approach for measuring head-related transfer functions (HRTFs) since it allows separating the loudspeaker’s distortion from the HRTF by temporal windowing. This, however, fails with overlapping excitation signals. This work assesses the signal-to-noise ratio (SNR) benefit from nonlinear harmonic prediction and subtraction. The measured response is modeled as a superposition of a linear response, nonlinear harmonic distortions, and additive noise. Each harmonic distortion component is represented using a finite impulse response kernel. The kernel is estimated from a reference measurement. For evaluation, broadband and Bark-band limited SNR is computed. Preliminary results for frontal HRTF measurements show broadband SNR values of 30-45 dB depending on the loudspeaker direction, with the proposed harmonic subtraction improving these values by up to about 3 dB for the most affected (most elevated) channel. In perceptually relevant mid-to-high frequency regions, where spectral notches make the HRTF sensitive to measurement noise, mean SNR gains of 1-2 dB are observed. The results demonstrate that the proposed method has potential to reduce non-linear distortion and provide cleaner impulse responses and more accurate noise estimates in overlapping ESS-based HRTF measurements.

Speaker: Payman Azaripasand (Technical University of Munich)

Full paper (PDF)

417 Applicability of 3D-Scan Methods using Smartphones for HRTF Synthesis

Individualized head-related transfer functions(HRTFs) are an essential tool for accurate spatialaudio reproduction, but traditional measurementprocedures are time-consuming and require specializedlaboratory equipment. Numerical approaches enablethe computation of individualized HRTFs from 3Dmodels of the torso, head, and pinnae. Althoughaccurate 3D models present a challenge in itself,recent advances in smartphone-based 3D scanningtechnologies offer a low-cost and accessible alternativeto professional scanning systems.This paper investigates the applicability of threesmartphone-based 3D scanning methods, namely photogrammetry,LiDAR, and TrueDepth, for HRTF synthesisusing Mesh2HRTF. This contribution focuseson evaluating the methods with respect to ease of useand achievable spatial resolution. Therefore, 3D scansof an artificial head are compared against referencemeasurements based on anthropometric parameters,and the resulting mesh quality is analyzed.The results show that smartphone-based LiDAR isnot suitable due to its limited spatial resolution, whilephotogrammetry and TrueDepth can provide sufficientgeometric detail for HRTF synthesis. Amongthe evaluated methods, TrueDepth offers the bestcompromise between acquisition effort and geometricaccuracy, whereas photogrammetry provides highflexibility and device compatibility.

Speaker: Benedikt Bugl (OTH Regensburg)

Full paper (PDF)

418 Listener Acoustic Personalisation Challenge LAP24: Perceptual Evaluation of HRTF Upsampling

The 2024 Listener Acoustic Personalisation (LAP24) challenge benchmarked the spatial upsampling (interpolation) of head-related transfer functions (HRTFs). Seven teams submitted HRTFs that were upsampled from four sparse grids containing between three and one hundred HRTF positions. The submissions included six deep learning-based approaches and one algorithmic approach. They were evaluated with respect to log-spectral distortion (LSD) and broadband interaural time and level differences (ITD, ILD). The learning-based upsampling methods often showed smaller differences to the reference HRTF than the algorithmic approach, especially for very sparse sampling grids. In the current study, we conducted a complementary perceptual evaluation of the LAP24 challenge algorithms with respect to colouration and source direction for a static sound source. This work confirmed the numerical results of the LAP Challenge at least for the top-ranked method. Beyond first place, however, the perceptual ranking differed from that obtained with physical error measures, due to the different aspects captured and their sensitivities. Moreover, an algorithm-specific diffuse field equalisation filter had little effect, suggesting that direction-dependent upsampling errors remain audible even in this case. The upsampled HRTFs, along with audio examples and ratings from the listening test, are publicly available to facilitate benchmarking of future upsampling algorithms.

Speaker: Fabian Brinkmann (Technische Universitat Berlin)

Full paper (PDF)

419 Evaluating the Impact of Head-Worn Device Configurations on Spatial Hearing for Augmented Reality Applications

With the recent surge in augmented reality (AR) applications, devices such as head-mounted displays (HMDs) and headphones are increasingly used as multimodal interfaces that combine visual and auditory information to create immersive experiences. However, to date, studies have focused on evaluating the impact of head-worn devices (HWDs) on head-related transfer functions considering either different headphones or HMDs without considering configurations where two such devices are used simultaneously, as sometimes can be in AR. This study aims at bridging this gap and presents a perceptually-based numerical evaluation of the impact of HWD configurations on HRTFs. HRTFs were measured on a KEMAR mannequin wearing four configurations of HWDs including two pairs of headphones (MushRoom and AKG K1000) measured with and without an HMD (Meta Quest 3). Results are presented in terms of interaural cues, estimates of perceived localisation-, and colouration-induced error metrics.

Speaker: Julie Meyer (King's College London)

Full paper (PDF)

A05.01 Beyond L_den: Citizen Science, Smart Sensing, and Next‑Generation Acoustic Indicators for Environmental Noise Perception: S030

Conveners: Cedric Vuye (University of Antwerp), Qiuya Xiang (University of Antwerp)

420 Bridging community noise perception and the acoustic environment: insights from a citizen science acoustic monitoring network

Strategic noise maps based on sound emission and propagation modelling help identify overexposed areas using weighted annual average noise level indicators such as Lden and Lnight. While providing large-scale, objective data to plan noise mitigation strategies, acoustic metrics retrieved from strategic noise maps can oversimplify and downplay characteristics of the acoustic environment linked to noise annoyance and sleep disturbance. To address this gap, the citizen science project De Oorzaak first deployed a large-scale Internet-of-Things-based acoustic sensor network, for one year in Antwerp, Ghent, and Leuven, with 1,484 citizen scientists installing a Class-2 smart sound sensor outside their bedroom window for continuous six-week measurement periods. In parallel, a questionnaire was sent each week in order to get a detailed view of the respondents’ noise perception. The resulting data will enable a finer-grained exploration of the acoustic environment, including time-varying patterns, soundscape pattern, and identifiable noise events and their sound sources. By linking these objective measurements with the subjective survey responses this study offers a more nuanced perspective on how environmental noise drives annoyance at the population level.

Speaker: Cedric Vuye (University of Antwerp)

Full paper (PDF)

421 Annoyance differences between single and combined transportation noise sources

In the dense modern cities it is very common that multiple noise sources occur at the same time. The total annoyance of all of them is related to the annoyance of each type of noise itself. However, the annoyance of the one type of noise presented alone can differ from the same noise presented in the richer context. To investigate this phenomenon road, tram and aircraft noise were presented to listeners as single and combined noise scenarios. Results showed that while keeping the same levels, mainly tram and sometimes aircraft noise are rated higher when presented together with road traffic noise while the latter is rated similarly in both contexts.

Speaker: Jan Felcyn (Adam Mickiewicz University in Poznan)

422 Pickleball and Padel Noise Assessment Methodology Review

In 2023, the convenor of ISO/TC43/SC1/WG45 on the assessment of environmental noise (ISO 1996) was approached initially regarding whether pickleball noise had come to the attention of this group. This developed into a request for information on how pickleball and the related padel sport are dealt with around the world, and the need for international guidance on how to assess it. Both pickleball and padel are growing in popularity. There are claims that there is a strong difference in the intrusiveness of pickleball and padel in comparison with other sports, due to their acoustic signatures. It has become clear that there is a need to develop assessment methodology for these emerging sports with their unique and unfamiliar acoustic signatures. Existing standardized assessment methodology and dose-response studies are both limited, creating noise assessment challenges. There is an immediate need for the sharing of information and experience from around the world to help experts better develop assessment methodology.ISO/TC43/SC1/WG45 has initiated work to gather information and have agreed to publish findings at an international conference as soon as possible in order to uncover existing and in-development methodology, and to encourage funded international research in this area. This paper describes the background as well as the status of our work to collect existing practice on the assessment of non-professional padel and pickleball sport in outdoor courts. In addition, it describes the plans that the group is forming to initiate work to identify dose-response relationships and appropriate metrics, including the use of Rating Level as defined in ISO 1996.

Speaker: Douglas Manvell (DMdB)

Full paper (PDF)

423 Acoustic effectiveness of the noise barrier: evidence from a phased construction through a multi-metric assessment

Noise barriers have been widely constructed as a common measure for mitigating road traffic noise. However, their effectiveness has often been evaluated primarily in terms of sound pressure level reduction, using LAeq,15 min. and the estimated reduction using simulated values, with less attention given to other acoustic characteristics or long-term performance. To address this gap, this study investigates the acoustic characteristics along a primary road before, during and after noise barrier construction over a six-month period at 20 measurement locations, including noise levels, noise events, and Artificial Intelligence estimated sound sources. The measurement locations were grouped according to their distance from the main road to assess the spatial effectiveness of the barriers. By comparing phased measurements across locations and periods, with additional analysis of temporal variations, specifically between daytime and nighttime and between workdays and non-workdays, this study identifies how the barrier influenced different dimensions of the acoustic environment beyond simple attenuation. The findings provide empirical support for traffic noise mitigation strategies and suggest that the effects of noise barriers can be more comprehensively understood by considering a multi-metric acoustic assessment in broader soundscape characteristics.

Speaker: Qiuya Xiang (University of Antwerp)

Full paper (PDF)

A11.01 Nonlinear dynamics in musical instruments: S074

Convener: Rolf Bader (University of Hamburg)

424 Inverse-Problem solution in musical acoustics using machine-learning analysis of parameter variations of a Dulcken cembalo from 1765

Modeling of musical instruments is most often performed in a forward manner, arriving at an analysis or a sound from a chosen geometry and material parameters. But most often, the inverse problem is of more interest: how to build a musical instrument with a desired sound. Using machine-learning methods trained by sounds of physical model outputs when varying a geometrical or material parameter space, the overall timbre space of an instrument within reasonable variations is found. Varying the material parameters of a Dulcken cembalo of 1765 from the Museum of Applied Arts in Hamburg, the sound parameter space of this type of instrument is investigated. It is assumed that other Dulcken instruments from this builder have been built with the same geometry, but varying wood parameters, and therefore have sounded within such a timbral space. Comparing these variations of other cembalos in the future could lead to a characterization of different instruments within material parameter variations. This could lead to a reasonable reconstruction of historical instrument sounds.

Speaker: Rolf Bader (University of Hamburg)

425 Dynamic Regime Changes in the Impulse Pattern Formulation: From Mathematical Analysis to Musical Dynamics and Articulation

The Impulse Pattern Formulation (IPF) was originally developed as a nonlinear recursive framework for modeling sound generation in musical instruments and was later extended to further topics, including interactions between musicians. In this contribution, the focus is on the performer’s influence on the instrument through dynamically varying control parameters. The IPF describes coupled dynamical systems via interacting impulse trains and captures characteristic acoustic regimes such as stable oscillations, bifurcations, and noise-like behavior. While previous work has mainly addressed the boundaries of these regimes through quasi-stationary regime transitions, musical performance inherently involves time-dependent parameter variations that shape dynamics and articulation. For example, when modeling the tone production of a saxophone, there is a fundamental difference between changing the pitch via fingering, resulting in sudden parameter changes, and increasing the blowing pressure, which induces continuous parameter variation. By analyzing such dynamically evolving conditions within the IPF framework, this work provides a systematic perspective on how performer-driven parameter changes affect acoustic regimes. The results provide new insights into the stability properties of the IPF and contribute to a better understanding of the interplay between instrument physics and performer control in shaping musical dynamics and articulation.

Speaker: Simon Linke (Hamburg University of Applied Sciences, ligeti center)

426 Experimental Study of the Influence of Mouth Cavity Size and Vocal Tract Constriction on the Recorder Sound

CFD simulations of flue instruments suggest two mechanisms how the recorder sound could be altered by its player. First, a larger volume of the mouth cavity could be associated with a more stable amplitude (Ikoga et al., 2024) and clearer higher harmonics (Nordblom and Bader, 2026). Second, shear turbulence inside the mouth could weaken the higher harmonics (Nordblom and Bader, 2026).This work investigates experimentally the sound results of vocal tract modifications after these theories. An alto recorder is blown with a machine. For the different mouth cavity sizes, tubes of various lengths are inserted upstream of the recorder. Moreover, a slotted cap is sometimes added to produce the constriction for the shear turbulence theory. The sound amplitudes, spectra and phase portraits are taken among several psychoacoustic parameters. In general, the results show the same trends as the simulated predictions. This supports that there are at least two different mechanisms how to alter the recorder sound as a player.

Speaker: Naomi Nordblom (University of Hamburg (student))

427 On the Use of Dynamical Systems Descriptors for the Psychoacoustic Characterization of Playing Techniques in African-Rooted Percussion Instruments

Playing techniques from African-rooted percussion instruments focus on exploring the range of sounds produced by striking different parts of the membrane with different parts of the hitting hand. This creates a sonic palette through which culture-defining rhythms are performed. However, the perceptual descriptions found in the literature are usually limited to language metaphors with no clear acoustic correlation, endangering the preservation of such immaterial cultural heritage and exposing the need for accurate psychoacoustic characterizations. In the present study, I explore the use of phase space embedding, correlation dimension and Lyapunov exponents as psychoacoustic descriptors suited for playing technique description. The analysis is grounded in the relation between the short duration of the instrument's transients, the stationarity conditions required for the analysis of time series as well as the constraints they impose regarding sampling rate and signal length; and the temporal windows for auditory processing of timbre and pitch established in the literature. Samples from three different playing techniques (bass, slap and open) performed in the Ghanaian Djembe, the Brazilian Atabaque and the Colombian Alegre —African-rooted single-headed wooden vessel drums— are investigated, systematically varying the parameter settings employed for their projection in the phase space. For each derived embedding, the resulting descriptor values are presented; the gathered results are then examined to assess the influence of the embedding parameters on their calculation and psychoacoustic interpretation in light of the playing techniques. This study is thus a first step toward a psychoacoustic description capable of grounding the digital preservation of playing technique diversity against processes of cultural homogenization.

Speaker: Cristhiam Fidel Martínez Orellanos (University of Hamburg)

A16.06 Active design and control of sound fields in enclosed spaces: S118

Conveners: Jamilla Balint (Rohde Acoustics), Sebastian Schlecht (Friedrich-Alexander-Universität Erlangen-Nürnberg), Tapio Lokki (Aalto University)

428 Contribution of feedback components to the instability of Reverberation Enhancement Systems: a Case Study

Reverberation enhancement systems (RESs) are designed primarily for large concert halls, in which case the energetic and statistical properties of the feedback paths have been extensively studied. However, RESs are also used in smaller and less reverberant spaces, where room acoustic parameters differ from concert halls'. As the stability threshold, or gain-before-instability (GBI), of a RES is commonly estimated via in-situ measurements, the room acoustic properties and transducer setup may greatly influence such analysis. However, the literature does not properly discuss the extent of that effect for non-concert-hall venues, nor it offers guidance on the choice of a stability-improvement technique based on the feedback analysis. This work explores the individual impact of direct sound, early reflections, and late reverberation on GBI estimation in a medium-sized dry venue in octave bands. The comparison with the usual approach of analyzing the entire response shows how the contributions of the three components vary according to the system setup. This work aids in the systematic analysis of RESs for medium and small venues, challenging assumptions regarding reverberant field and GBI estimation, and tackles the improvement of stability based on RES topology.

Speaker: Gian Marco De Bortoli (Aalto University)

Full paper (PDF)

429 On Regeneration and Colouration in Active Acoustics Systems

In active acoustics, controlled feedback is often used to extend a room’s reverberation time by regeneration. This brings an inherent risk of instability, while at stable gains a poorly conditioned feedback system can exhibit quality-limiting colouration. The gain before instability is usually estimated using frequency‑domain eigenvalue measures computed from the open‑loop response, and information about colouration or equalization is sometimes inferred from this. In this paper, we show that while eigenvalue analysis indicates the onset of instability, it does not reliably predict colouration while the system remains stable. We first demonstrate examples where the closed-loop behaviour for stable gains is not predictable by analysing the eigenvalues. We then demonstrate that singular value decomposition (SVD) of the closed-loop operator can be used as a loop-gain-dependent analysis tool. Analysis of simulations and measurements shows a clear correspondence between the characteristics of the SVD analysis and the emerging resonances in the closed-loop response. Furthermore, a spectral selectivity metric is proposed, which indicates the presence of colouration directly from the SVD analysis.

Speaker: William Blayney (L-Acoustics)

Full paper (PDF)

430 Perceptual optimization of active acoustic presets via iterative musician feedback

Active acoustic systems whose algorithms are adapted to a room’s underlying passive acoustics can effectively modify acoustical conditions without compromising the room’s inherent character. This technology is particularly valuable for those rehearsal spaces that are designed to be highly absorbent in order to control sound strength. It can bridge the gap to the concert halls musicians are accustomed to, providing good cross-communication without raising sound strength.This paper presents a novel study investigating the artistic tuning process of an Amadeus ART Acoustics system in the orchestra rehearsal hall of the Finnish National Opera and Ballet in Helsinki.Rather than relying solely on objective measurements, the study employed an iterative, musician-driven tuning process. Several rehearsals were conducted with the orchestra, each comparing two different active acoustic settings. Following each cycle, participants completed an adapted questionnaire based on Mike Barron’s approach to evaluate room acoustics. By continuously assessing musicians’ perceptions and the impact on performance, informed parameter adjustments were made immediately. To maintain plausibility and focus, a maximum of two settings were compared by musicians and conductors. This gradual process significantly improved both the playing experience and the sonic impression, ultimately increasing consensus among musicians on the final acoustic preset. Statistical analysis of the feedback further provided insights into the evolution of survey criteria and the impact of preset adjustments on instrument sections. These findings offer a valuable and reproducible framework for tuning active acoustic systems in comparable situations.

Speaker: Leon Merkel (Amadeus Acoustics GmbH)

Full paper (PDF)

431 Simulation of Quality Parameters for Microphone Coverage in Sound- Reinforcement Planning

In sound-reinforced concert halls, theaters, and multipurpose venues, the microphones of the reinforcement system shape the perceived sound just as much as the loudspeakers do. Yet while loudspeaker coverage is routinely optimized during system design, the way microphones cover a stage or performance area is rarely quantified. Room-acoustic simulations based on geometrical and statistical methods are already used to evaluate loudspeaker coverage, but they typically neglect the directional characteristics of the microphones and the signal paths between sources, microphones, loudspeakers, and listeners.This contribution presents a simulation framework, which incorporates measured high-resolution three-dimensional microphone directivities into a combined geometrical and statistical room-acoustic simulation. From the simulated transfer functions together with the statistically modelled reverberant field, a set of quality parameters for microphone coverage is derived. These parameters describe how consistent and with what timing and spectral balance sources across a performance area are captured, and how robust the resulting system is against feedback and coloration.A central feature of the approach is that each parameter is visualized as a mapping over source or receiver positions. We show that in practice these mappings give system designers and acoustic planners a spatially explicit picture of where a given microphone configuration performs well and where it does not. As such, these mappings complement existing loudspeaker coverage mappings. We demonstrate how this enables more precise evidence-based decisions about microphone type, placement, and orientation during the planning phase and optimization of the reinforcement system as a whole.

Speaker: Julian Ebert (Audio Communication Group, Technische Universität Berlin)

A22.05 Geophysical Applications and Distributed Acoustic Sensing (DAS) in Subsea Environments: S160

Conveners: Abdelghani Idrissi (INFN - LNS, Unict-DFA), Dídac Diego-Tortosa (Institut de Ciències del Mar (ICM-CSIC))

432 Marine Sediment Characterisation Using Interface Waves Recorded by DAS

Distributed Acoustic Sensing (DAS) is an emerging technology that converts conventional fibre-optic (OF) cables into densely sampled acoustic arrays. By analysing Rayleigh backscattering from laser pulses, DAS detects strain changes along the cable, enabling continuous monitoring of subsea infrastructure and seismic events. Shear-wave velocity profiles are critical for assessing sediment stiffness and useful for offshore geotechnical applications. In shallow water, both active and passive sounds generate Scholte waves, propagating along the water-seafloor interface. These waves can be recorded by acoustic sensors deployed on or close to the interface. Scholte wave velocity is closely linked to the shear-wave velocity, and its variation with depth causes Scholte waves to be dispersive. Therefore, the Scholte-wave dispersion property can be used to estimate the shear-wave velocity profiles. This paper presents examples of characterizing shallow-ocean sediments using both active and passive DAS data recorded across various ocean environments. The first example is a seismic survey offshore Norway. DAS data were acquired using an existing FO cable deployed 1 m below the seafloor. The data from airgun shots show dispersive Scholte waves at near-offsets, which were used to estimate the shear-wave velocity profile. The dispersion data resolved two-layer sediments over a half-space. The second example is from a dataset released as part of “The Global DAS Month of February 2023”. Two-hour DAS data of ocean ambient noise recorded by an 80-km FO cable in shallow water in the Southern North Sea were analysed to extract Scholte-wave dispersion curves for estimating the shear-wave velocity profile. The dispersion data resolved a 50 m sediment layer above a half-space.

Speaker: Hefeng Dong (Norwegian University of Sci. & Tech.)

433 Monitoring shipping noise and marine life using DAS at the Portopalo deep-Sea site

The INFN-LNS operates advanced deep-sea electro-optical infrastructures in the southeast of Sicily at the Portopalo site, supporting multidisciplinary research activities. These infrastructures were mainly installed for powering, control of deep-sea observatories and real-time data transmission. The existing optical fibers can also be exploited for optoacoustic sensing techniques. In this direction, within the framework of the LOWNOISER project, an ASN OptoDAS system has been deployed at the Portopalo site on a 96 km long optical fiber, acquiring data continuously since June 2025. At the end of the cable, an array of hydrophones is also connected to the same infrastructure, providing an opportunity to complement and calibrate the DAS measurements. The aim of this project is to assess the impact of underwater radiated noise from shipping on marine biology, with particular focus on fin whales. INFN-LNS also operates an Automatic Identification System (AIS) in the same area to monitor vessel traffic. This allows us to study possible correlations between shipping activity and fin whale vocalizations. Preliminary analyses of the acquired data will be presented, including the automatic detection of fin whale vocalizations using matched filtering techniques. In addition, some recorded acoustic signals will be presented, together with the strategy adopted for long-term data storage and management

Speaker: Abdelghani Idrissi (INFN - LNS, Unict-DFA)

Full paper (PDF)

434 Automatic workflow for fin whale song analysis and source localization using distributed acoustic sensing

Fiber-optic submarine cables instrumented with distributed acoustic sensing (DAS) provide a cost-effective solution for large-scale, real-time monitoring of fin whales. We present an automatic workflow for the detection, characterization, and localization of fin whale songs, developed and tested using data recorded on a telecom submarine optical cable in the Alboran Sea (westernmost part of the Mediterranean, at the transition to the Atlantic Ocean) from late September 2023 to early February 2024. The workflow begins with a kurtosis-value picker (KVP) optimized for fin whale pulse detection. The resulting picks are then clustered using the DBSCAN algorithm to group coherent arrivals across DAS channels. This clustering step enables the characterization of individual notes by extracting key features related to energy, time, and frequency. Based on these features, signals can be classified into 20-Hz notes and backbeats, enabling a detailed characterization of songs. The identified clusters also provide high-quality candidates for localization by exploiting the multi-channel nature of DAS recordings. Time-of-arrival data are selected through a fitting consistent with the cable geometry, followed by grid-search localization. The proposed workflow supports near-real-time processing, while reducing data volumes for downstream analysis, and offers a scalable framework for fin whale acoustic monitoring applicable to other DAS datasets.

Speaker: Dídac Diego-Tortosa (Institut de Ciències del Mar (ICM-CSIC))

435 Fast Super-Resolution Acoustic Imaging for Underwater Distributed Acoustic Sensing Using Improved Richardson–Lucy Deconvolution

Distributed acoustic sensing (DAS) based on phase-sensitive optical time-domain reflectometry enables optical fibers to operate as dense underwater acoustic arrays with large aperture and continuous spatial sampling. This capability creates new opportunities for wide-area underwater acoustic monitoring and target detection. Nevertheless, applying conventional array signal processing techniques to DAS measurements remains challenging in practical underwater environments. Existing approaches based on covariance fitting or subspace direction-of-arrival estimation rely heavily on accurate array manifold modeling and often require prior knowledge of the source-to-array distance in near-field scenarios. In distributed fiber sensing systems, array geometry mismatch, environmental variability, and uncertainty in range information can significantly degrade localization performance, particularly for coherent acoustic sources. To address these challenges, this paper proposes a fast super-resolution acoustic imaging method for underwater DAS arrays based on improved Richardson–Lucy (RL) deconvolution. DAS signals are first processed using delay-and-sum beamforming to generate conventional acoustic images, after which RL iterative deconvolution reconstructs the spatial distribution of acoustic sources. To handle the large data volume of DAS systems, a spatial-resampling-based fast RL deconvolution method is developed to significantly reduce computational complexity while preserving spatial resolution. Tank experiments conducted at Harbin Engineering University and lake experiments in Danjiangkou using real DAS data demonstrate that the proposed approach can resolve coherent near-field acoustic sources with super-Rayleigh resolution, while reducing the mainlobe width by approximately 60%, suppressing sidelobe levels by more than 10dB, and achieving range estimation errors below 0.2m. These results demonstrate that the proposed fast RL-based imaging approach provides an efficient and robust solution for high-resolution underwater acoustic sensing using DAS arrays.

Speaker: Jie Wu (Guangzhou Marine Laboratory)

Full paper (PDF)

Thursday 10 September

A01.04 Active Sound and Vibration Control: S005

Conveners: Alberto Gonzalez, Felix Langfeldt (University of Southampton)

436 Broadband Active Control of Fluid-borne Sound in a Pipeline Using an Actuated Orifice Plate

Pipeline systems are used for the transportation of fluids in a wide range of engineering applications such as water distribution, oil pipelines and marine vessels. Noise from pumps and compressors can propagate efficiently through such systems, often resulting in sound being radiated far from the original source. Existing passive treatments are widely used in industry to minimise the effect this has on radiated noise, but the effective frequency range is limited by mass and size constraints. In contrast, active control systems, that use a secondary source to cancel the sound or vibration from a primary source, generally exhibit high performance at low frequencies. The inclusion of a secondary actuator can, however, adversely affect the fluid flow in the system or require complex modifications to the pipeline. This paper proposes a potential solution where the secondary actuator is integrated into an orifice plate, which are commonly used to reduce pressure or restrict flow in pipeline systems. The actuated orifice plate is used in an optimal feedforward control system to minimise the downstream broadband sound pressure. This work serves as a proof of concept that such an active orifice plate has the potential for controlling fluid-borne noise in pipeline systems.

Speaker: Archie Keys (University of Southampton)

Full paper (PDF)

437 Hybrid active noise control using parametric and dynamic loudspeakers

This paper proposes a hybrid active noise control (ANC) system using a parametric array loudspeaker (PAL) and a conventional dynamic loudspeaker (DL) as secondary sources. The proposed configuration exploits the high directivity of the PAL and the low-frequency efficiency of the DL to realize low-frequency noise reduction while mitigating spillover outside the control region. The control signal is decomposed into low- and high-frequency bands via a crossover filter, with the PAL handling the high-frequency component and the DL generating the low-frequency component. Real-time experimental evaluations confirm that the proposed system provides a well-defined noise reduction region with improved low-frequency attenuation and reduced spillover compared to ANC systems employing only a DL or a PAL.

Speaker: Ichika Miyamoto (Kansai University)

438 OCEAN: An Open Low-Cost Embedded Platform for Active Noise Control Research and Education

Active Noise Control (ANC) is widely used for reducing unwanted noise in ducted systems and enclosed environments. However, hands-on experimentation is often limited by the cost and complexity of conventional laboratory setups, restricting accessibility in both education and research.This paper presents OCEAN (Open Control Environment for Active Noise Control), an open-source, low-cost platform designed to enable practical ANC experimentation at a total system cost of approximately 150 €. The system consists of a compact, 3D-printable duct setup equipped with two loudspeakers and two digital MEMS microphones, interfaced via I²S to a microcontroller-based development board. This configuration supports the implementation of classical feedforward ANC algorithms, such as filtered-x LMS (FxLMS), and can be extended to more advanced control strategies.OCEAN provides a complete workflow from model-based design (e.g., MATLAB/Simulink) to real-time embedded implementation in C. All hardware designs, software models, and documentation are made publicly available, enabling reproducibility and collaborative development.The paper outlines the system architecture, key implementation aspects, and challenges related to real-time processing, latency, and control stability. Experimental results demonstrate the suitability of the platform for typical ANC applications in duct acoustics.By significantly lowering the entry barrier, OCEAN supports engineering education, promotes open research, and facilitates the development and evaluation of ANC algorithms within a broad community.

Speaker: Alessandro Fortino (Fachhochschule Dortmund)

Full paper (PDF)

A03.03 Prediction, digitalisation (BIM) and simulation in building acoustics: S015

Conveners: Andrea Santoni (University of Ferrara), Edwin P.B. Reynders (KU Leuven, Department of Civil Engineering), Antonino Di Bella (Universyty of Padova - DII)

439 Comparison of Airborne Sound Insulation Measured in Reduced Size and Full Size Test Openings

In this paper, the influence of sample size and mounting conditions on the laboratory measurement of the airborne sound insulation of a monolithic wall of aerated concrete blocks is assessed. The same aerated concrete wall was measured in an accredited laboratory in a full-size test opening of 4.00 m x 3.00 m, according to the ISO 10140 standard series, and also in a reduced-size test opening of 1.62 m x 1.87 m, with deviations from the standard conditions: no ratio of the niche depths of 2:1, and no covering of the cavity with reflecting material. The differences in the resulting sound reduction indices between 200 Hz – 5000 Hz fall between 5 and 10 dB, and 8 dB in the SNQs –all in favour of the smaller sample.To achieve better agreement between the two results, further measurements were carried out on the reduced-size sample after modifying the mounting conditions: the acoustic cavity was covered with sound-reflective material, and a thick wooden frame was installed onto the laboratory wall to restore the 2:1 niche depths ratio.The sound reduction indices were also calculated according to ISO 12354-1 versions 2010 and 2017, with the measured longitudinal velocity [m/s] and the mass [kg/m3] of the blocks. Also the influence of sound absorption in the cavity was analysed by means of numerical simulations. The results of these investigations are summarised.

Speaker: Attila Balazs Nagy (Budapest University of Technology and Economics)

440 Numerical analysis of the niche effect on the diffuse sound transmission loss of partition panels

When predicting the sound transmission loss (STL) of a panel for an indoor situation, it is often assumed that the sound fields at both sides of the panel are diffuse. In this way, the STL does not depend on the detailed room geometries but rather represents the average panel performance across an ensemble of source and receiver rooms. Thanks to the reciprocity relation between direct field radiation and diffuse reverberant loading, the STL of a panel of arbitrary complexity can be computed by modeling it in full detail and coupling it to acoustic halfspace models at both sides. So far, a flat baffled geometry has been considered, such that the halfspace radiation stiffness can be obtained via the Rayleigh integral. However, computing the diffuse STL for panels mounted in a niche has remained an open problem. In this work, this problem is addressed by computing the radiation stiffness using a coupled finite element (FEM) – boundary element (BEM) approach. The air volume inside the niche is modeled using FEM, while the coupling to the acoustic halfspace is treated using BEM. This approach is validated on panels mounted in the standardized transmission opening for glazing of the KU Leuven Laboratory of Acoustics.

Speaker: Michiel Lannoye (KU Leuven, Department of Civil Engineering)

Full paper (PDF)

441 Window Glazing Systems Design for Indoor Drone Noise Mitigation

Ensuring acoustic comfort is a key objective in building envelope design. In this context, common glazing systems are typically the weakest facade element, exhibiting sound insulation dips at the critical frequency (fc) for monolithic glass and at the mass–air–mass resonance (f₀) for double-glazed units. Meanwhile, the dominant spectrum of drones (100–2500 Hz), including tonal peaks such as the blade-passing frequency, overlaps with these resonances and can readily transmit indoors. Therefore, this study evaluates the effectiveness of window glazing systems in mitigating indoor noise from drones. The sound insulation of various glazing configurations were analyzed in relation to the acoustic emissions of a delivery drone (2.5 kg payload). Results indicate that, among single-glazed glass 4–8 mm thick, laminated glazing outperforms annealed and tempered glass, achieving reductions of 4.2 dB(A) at fc and up to 6.0 dB(A) in the 50–100 Hz range. For double-glazed systems, increasing the air gap from 12 mm to 50, 100, or 150 mm, combined with thicker glass, shifts f₀ to lower frequencies (48–97 Hz), resulting in improved low-frequency attenuation compared with conventional windows. These framework provide performance-based guidance for window design aimed at reducing drone noise in new residential buildings.

Speaker: Marco Oliveira (Trinity College Dublin)

Full paper (PDF)

442 Accurate and efficient prediction of sound insulation in multilayered plasterboard walls with flexible metal stud frames

Engineering offices and manufacturers of acoustic building systems or materials often rely on measurement data to estimate whether building solutions will provide sufficient acoustic quality. This typically involves numerous costly and time-consuming laboratory measurements, making it difficult to efficiently explore and optimize various design configurations. In this paper, a method is presented to accurately and efficiently predict the sound insulation of multilayer structures, including flexible metal stud frames. Accuracy ensures reliability, while efficiency is crucial for optimization, especially when numerous simulations are needed, for example, to identify the ideal layering or material properties of a layer. The prediction method accounts for arbitrary layering, finite dimensions, boundary conditions and resulting modal behavior, as well as flexible metal stud frames, including the cross-sectional shape, stud spacing and screw positions. Extensive validation has been conducted using numerous examples with a specific focus on plasterboard walls, demonstrating the model’s robustness and reliability; An accuracy of 2-3 dB in single number ratings is generally achieved, independent of the number of plasterboard layers and frame type, which includes single studs, double studs periodically coupled with a small screwed gypsum board plate, and double studs periodically coupled with felt strips.

Speaker: Jasper Vastiau (KU Leuven)

A07.00 Flow Acoustics: S040

Conveners: Stefan Becker (LSTM, FAU Erlangen-Nürnberg), Roberto Camussi, Francesco Avallone (Politecnico di Torino), Stefan Schoder (IGTE, TU Graz)

443 Development of far-field noise prediction model using a multipole expansion method

The development of an acoustic measurement platform using UAVs enables noise field mapping and sound source characterization of outdoor low-frequency noise sources that are, otherwise, difficult to be measured in laboratory environment due to dimensional constraints. A simplified analytical model using spherical harmonics and a multipole expansion of the Green’s function has been developed for estimating the acoustic source signature from near-field measurements. In addition to the simplification of measurement methodology, the multipole expansion method increases the computation efficiency compared to traditional Boundary Element Method. Furthermore, the model predictions are validated with a self-calibrated low frequency laboratory test source in an outdoor free field environment. The methodology developed here can facilitate the quantification of low-frequency noise emission from renewable energy sources and the determination of the propagation characteristics in the lower atmosphere.

Speaker: Ramesh Raja Subramanyam (German National Metrology Institute)

444 Energy-Based Formulation for Solving Galbrun’s Equation

Galbrun’s equation, a displacement-based Eulerian–Lagrangian formulation for aeroacoustics, offers a promising framework for modeling and simulating fluid-structure interaction in flow. The numerical solution of Galbrun’s equation remains challenging. Standard finite element method approaches are prone to instabilities that give rise to spurious nonphysical modes, an effect that is further exacerbated in the presence of a background mean flow. Alternative approaches, including mixed finite element and discontinuous Galerkin methods, have so far yielded only limited improvements in numerical accuracy.In this work, an initial structure-preserving port-Hamiltonian formulation approach is presented with the objective of establishing a physically stable numerical solution. Under the assumption that energy is exchanged between multiple physical domains and that no energy is lost, this approach helps prevent the formation of spurious modes.In the first step, a two-dimensional fluid flow domain together with an Euler–Bernoulli beam are formulated using the port-Hamiltonian systems approach and discretized using the Partitioned Finite Element Method. In a second step, a first attempt to couple the structural and fluid subsystems is made using power-conjugated interconnection ports, ensuring energy consistency at their interface. This is a key step toward a complete port-Hamiltonian formulation of Galbrun’s equation.

Speaker: Michael Buba (Technical University of Munich)

445 Ray tracing of acoustic waves guided in a jet

Ray tracing is performed from a point source located on the axis of a cylindrical shear layer at a Mach number ranging up to Mj=2. In all cases, some of the rays originally directed upstream are guided by the jet flow. These guided jet rays (GJR) can be entirely or partially confined inside the jet depending on the Mach number and the ray initial angle. They can travel in the upstream or downstream directions, or not propagate axially. For subsonic Mach numbers, upstream-travelling GJR are found in all cases, and downstream-travelling GJR are observed above a threshold Mach number. For supersonic Mach numbers, both upstream- and downstream-travelling GJR exist. These results are consistent with the properties of the guided jet waves described in previous studies.

Speaker: Christophe Bogey (CNRS)

Full paper (PDF)

446 Portable acoustic source for in-situ calibration of sensor networks

In the context of smart cities, large-scale acoustic sensor networks are vital for the evaluation and control of acoustic noise. The logistical challenge of ensuring calibrated acoustic sensor networks remains a hurdle. We present a portable acoustic source, that can be used for efficient in-situ calibration and failure detection of distributed acoustic sensor nodes. The source’s internal sound field is measured with a MEMS microphone, from which the sound power output of the source can be predicted using a hybrid experimental-numerical mode-decomposition method. We present experimental confirmation that the source’s sound power is accurately predicted and that the acoustic radiation exhibits uniform spatial directivity. Due to its compactness, the source mounted on ground or air vehicles may be used in the future for autonomous calibration in large-scale acoustic sensor networks.

Speaker: Johan Christian van Aswegen (German National Metrology Institute)

447 Analytical models for predicting the wall-pressure spectrum in turbulent boundary layers

Two families of analytical models for wall-pressure fluctuation spectra in zero-pressure-gradient turbulent boundary layers are reexamined. On the one hand, the formulations by Lysak (Journal of Fluids Engineering 128, 216–222, 2006) and Grasso et al. (Journal of Fluid Mechanics 877, 1007–1062, 2019), among others, are developed in the wavenumber–frequency domain. On the other hand, the approach of Peltier & Hambric (Journal of Fluids and Structures 23, 920–937, 2007) directly models space–time velocity correlations. First, the equivalence of these two families is formally demonstrated under identical assumptions, namely local isotropy and frozen turbulence. An extension is then proposed to account for variations of the mean velocity shear in the wall-normal direction, which improves agreement between the predicted pressure spectra and available experimental data. Model inputs (mean flow, turbulence intensity, one-dimensional turbulence spectrum, among others) are validated against experimental measurements whenever possible to ensure model robustness. Finally, from these analytical models, the relative contributions of different regions within a turbulent boundary layer to the wall-pressure fluctuation spectrum are discussed as a function of frequency.

Speaker: Ajit Rokade (École Centrale de Lyon)

Full paper (PDF)

A08.00 Industrial, Machinery, Equipment Noise and Vibration: S048

Conveners: Ercan Altinsoy (Professur für Akustik und Haptik), Serkan Atamer (Dresden University of Technology), Mehmet Sait Özer (Dresden University of Technology)

448 A Study on the CAE Process for Improving Abnormal Vehicle Body Impact Buzz Noise Caused by Door Closure

This study proposes a CAE(Computer Aided Engineering)-based process to reduce impact buzz noise that occurs when a vehicle door is closed. First, mode analysis on the rear quarter region identifies structural weak points associated with impact buzz noise. Second, DPI(driving point inertance) analysis identifies target frequencies. Third, ODS(operating deflection shape) analysis analyzes deformation behavior during door closing. Fourthly,transient analysis evaluates vibration decay time. Lastly,the damping value derived from the modal model is proposed as a development indicator. Vehicle testing confirms that the impact noise is eliminated by reinforcing the CAE identified weak part, and an improved structure design is derived based on this result.

Speaker: Je-Hyun Son (HMC (Hyundai Motors Company))

449 Unsupervised Classification and Controlled Laboratory Reproduction of Industrial Impulsive Noise

Occupational noise exposure is a leading cause of work-related hearing loss worldwide, with impulsive noise being particularly hazardous due to its sudden high-amplitude pressure peaks, which can cause irreversible cochlear damage even after brief exposure. While its harmful effects are well documented, the scientific community still lacks robust tools to characterize the wide morphological diversity of industrial impulsive signals. Most existing characterization frameworks were developed for military-type signals, which generally exhibit a single dominant pressure peak and an abrupt rise time on the order of microseconds, whereas industrial impulsive sources present fundamentally different temporal structures, with more progressive pressure build-up over several milliseconds and multiple successive sub-peaks of significant amplitude. This gap calls for the development of controlled laboratory tools capable of reproducing realistic industrial impulsive signals.This study addresses this need through a four-step approach. First, a representative acoustic database was assembled from field recordings collected in multiple industrial environments and complemented by laboratory measurements. Second, eight temporal and energetic indicators were selected and adapted to capture the specific characteristics of industrial impulsive signals. Third, unsupervised clustering was applied to the indicator dataset to identify distinct categories of industrial impulsive signals. Fourth, the clusters were used to define acoustic target signatures for the controlled reproduction of industrial impulsive noise in laboratory conditions.Results demonstrate that the proposed clustering approach provides a robust characterization of the acoustic variability across industrial impulsive sources, and that the selected indicators define reliable target signatures for their controlled reproduction in laboratory conditions.

Speaker: Hend Afli (École de technologie supérieure (ÉTS))

450 Case studies of high-performance industrial scale silencer and flue gas system designs

We present three cases of verified high-performance industrial scale silencer and flue gas system design.First is an exhaust silencer system for reciprocating engine test run, where low frequency noise caused nuisance at nearby dwellings. Measurements identified a low frequency tonal component. Complete exhaust system was redesigned to implement a distributed silencer system, including two reactive silencers. Measurements after installation at dwellings indicated the tonal component attenuation by 25 dB.Second is a recovery boiler with distinct induced draft fan noise. Current silencers had modest noise attenuation and flow performance, probably exacerbating the fan noise generation. Measurements indicated a 400 Hz tonal component. The exhaust system was remodeled to improve the fan performance and a low-loss baffle silencer was designed. Design target was 25 dB at 400 Hz and 20 dB A-weighted noise reduction. Commissioning test verified a 25 dB(A) reduction of stack top noise and pressure loss of under 200 Pa.Third is a large ship where low frequency tonal noise was identified in passenger areas. Engine source measurements revealed the causing excitation. Classical resonators were designed and installed in the exhaust ducts considering system performance. Commissioning measurements indicated attenuation of 15 dB in passenger areas at design frequency.

Speaker: Mikko Matalamäki (A-Insinöörit Suunnittelu Oy)

Full paper (PDF)

451 Interpretable Anomaly Detection in Rotating Machines for experts Using Convolutional Autoencoder

Unplanned breakdowns in industrial rotating machinery can cost millions per incident, making fault detection a critical maintenance strategy. This work presents an anomaly detection approach based on a 2D convolutional autoencoder (CAE) trained exclusively on healthy vibration data. The publicly available Mechanical Faults in Rotating Machinery (MFRM) dataset is used, comprising acceleration signals acquired under normal conditions and three fault types: unbalance, misalignment, and mechanical looseness. Vibration signals are first resampled into the angular domain via Computed Order Tracking (COT), then transformed into order spectrograms (OS) whose axes, harmonic orders and shaft revolutions, are directly meaningful to domain experts. A Log-Z standardisation enhances the sensitivity of the CAE to subtle spectral deviations. Trained to minimise reconstruction error on healthy OS only, the OS-CAE produces elevated mean squared error (MSE) when encountering faulty inputs. A threshold calibrated on the combined training and validation MSE distributions separates healthy from anomalous samples. Beyond binary detection, a per-order MSE profile of reconstructed spectrograms localises reconstruction failures at specific harmonic orders, providing interpretable visual cues that hint at the nature of the detected fault without relying on post-hoc explainability methods.

Speaker: Anouck Bruguiere (LAUM)

Full paper (PDF)

A13.04 Photoacoustic Imaging and Spectroscopy: S093

Conveners: Jan Laufer (Martin Luther University Halle), Ben T Cox (University College London), Günther Paltauf (University of Graz), Robert Nuster (University of Graz), Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe)

452 Numerical evidence of multiple scattering of photoacoustic waves in blood

Overall photoacoustic (PA) wave field in blood is strongly influenced by acoustic scattering arising from spatial distribution and acoustic properties of red blood cells (RBCs). In this study, the effect of multiple scattering of acoustic waves in blood with 30% hematocrit is numerically investigated using the convergent Born series (CBS) method. The simulation result is compared with that of a discrete particle approach (DPA). While the DPA estimates the total pressure field as a linear superposition of fields emitted by individual particles, it neglects the contribution owing to multiple scattering of acoustic waves. The CBS method incorporates wave-cell interactions through an iterative Born series procedure. Simulations were performed for an ensemble of RBCs randomly distributed in a computational domain of size 2048x2048 grid points, and the pressure fields were calculated at 88 MHz for three different acoustic impedance mismatch conditions. The CBS technique can exhibit modification of the pressure field occurring due to multiple scattering of PA waves in blood.

Speaker: Ratan K Saha (IIIT Allahabad)

Full paper (PDF)

453 Hard-Constrained Physics-Embedded Neural Networks for Robust Photoacoustic Image Reconstruction

The acoustic inverse problem in photoacoustic (PA) imaging is ill-posed under typical measurement configurations. We propose a hard-constrained physics-embedded neural network that directly embeds a universal backprojection module without learnable weights, surrounded by a learnable FFT filter and a learnable U-Net projection-combination module. This design enables quantitatively accurate reconstruction of initial pressure distributions even with limited training data.We benchmarked various learning-based PA reconstruction schemes, evaluating data efficiency and generalisability by training on 10, 50, 100, 200, and 300 simulated phantom datasets for 25 epochs with matched parameter counts. Quantitative accuracy of the reconstructed initial pressures was validated on simulated data with increasing difficulty and further tested on experimental data both from phantoms and mouse measurements.Using a mix of soft and embedded physics constraints consistently outperforms fully black-box models, already achieving high quantitative accuracy with as few as 50 training examples and showing robust performance on increasingly out-of-distribution situations. Models lacking physics priors during training failed to generalise within the training budget. Embedding a non-trainable physics module improves data efficiency and quantitative reliability, making the approach viable for targeted photoacoustic applications or uncommon detection geometries, where simulating large training datasets is computationally prohibitive.

Speaker: Janek Gröhl (University Hospital RWTH Aachen)

454 Flexible-Noise Self-Supervised Photoacoustic Image Reconstruction

Noisier2Inverse is a self-supervised framework for solving linear inverse problems without ground-truth data. Given noisy measurements Y = AX + N with forward map A : R^n -> R^m, a reconstruction network of the form B : R^m -> R^n is trained by minimizing the surrogate risk E||AB(Y+M)-(Y-M)||^2, and B(Y+M) or B(Y) is used for inference. The theoretical foundation is a surrogate risk equivalence showing that this loss has the same minimizer as the supervised risk E||AB(Y+M) - AX||^2, provided that the added noise M is identically distributed to the noise N. In practice, however, this requirement may be difficult to satisfy or suboptimal. In this work, we generalize the Noisier2Inverse theorem to flexible noise models. For scaled noise M = cN we derive the correct surrogate target explicitly, and under a Gaussian assumption we show that for general independent M and N the optimal surrogate target takes the form Y - (sigma_N^2 / sigma_M^2) M. These results provide a principled basis for choosing the added noise in self-supervised training. As a concrete application, we consider angular deblurring in photoacoustic tomography with finite-size detectors in circular scanning geometry, where the finite detector aperture induces a convolution in the polar domain that is addressed in a self-supervised manner using the generalized framework.

Speaker: Markus Haltmeier (University of Innsbruck)

455 Self-Supervised Reconstruction of Masked Full-Field Photoacoustic Projection Data

In camera-based full-field detection photoacoustic tomography, snapshots of the acoustic pressure field are recorded with an optical camera, yielding two-dimensional projections of the wave pattern at a fixed time after excitation. For a constant speed of sound, the initial pressure projection can be recovered exactly via a Radon-domain backpropagation method: the Radon transform decomposes the wave pattern into two counter-propagating bands, which are shifted back to time zero using d’Alembert’s formula. A key property of this decomposition is that both bands encode equivalent source information, providing inherent redundancy. In practice, parts of the wave pattern may be obscured by the sample, its holder, or the limited field of view, resulting in spatially masked regions with missing data. We propose a self-supervised neural network approach to inpaint the masked wave pattern directly in the image domain. The network takes the incomplete wave pattern as input and predicts the missing regions. Crucially, no ground-truth training data are required. Instead, training is driven by two complementary loss terms: (i) a data-fidelity loss enforcing consistency with the measured pixels and (ii) a Radon-domain self-consistency loss that exploits the redundancy of the two counter-propagating bands. If the data completion is correct, both bands must yield identical reconstructions of the initial pressure after shifting by ±cT. We evaluate the method on simulated data with various mask geometries and compare the reconstruction quality against iterative approaches and supervised baselines.

Speaker: Markus Haltmeier (University of Innsbruck)

456 Estimation of dielectric parameters from ultrasound waves in quantitative thermoacoustic tomography

Quantitative thermoacoustic tomography (QTAT) is a medical and biomedical imaging technique combining electromagnetic contrast and high resolution of ultrasound imaging. In QTAT, a short micro- or radio wave pulse is directed to the imaged target. As the electromagnetic waves propagate in the target tissue, they are absorbed leading to localized thermal expansion and rise in pressure. This pressure relaxes as broadband ultrasound waves that are measured on the boundary of the imaged target. In the inverse problem of QTAT, the dielectric parameters, such as electrical conductivity and permittivity, of the imaged target are estimated from the measured ultrasound waves.In this work, we propose an approach for simultaneous estimation of electrical conductivity and permittivity from the ultrasound waves in the inverse problem of QTAT. The inverse problem is approached in the Bayesian framework utilizing electromagnetic and acoustic forward models based on Maxwell’s equations and the acoustic wave-equation, respectively. The proposed approach is evaluated using numerical simulations. The results show that the dielectric parameters can be estimated accurately. However, the number of electromagnetic pulses and the ultrasound sensor geometry have a significant effect on the accuracy of the estimated parameters.

Speaker: Teemu Sahlström (University of Eastern Finland)

A14.02 Pitch perception impairment and enhancement with hearing devices: S100

Conveners: Etienne Gaudrain (CRNL, CNRS UMR5292, Inserm U1028, Université Lyon 1), Olivier Macherey (LMA-CNRS)

457 Untangling spectral and temporal voice pitch cues in n-of-m cochlear implant strategies

To date, the role of spectral and temporal cues in pitch perception with cochlear implants (CIs) remains debated. Previous research has indicated that pitch perception of CI users depends mainly on temporal cues. However, these studies used either well-controlled but artificial stimuli or indirect measures to assess cues. Using a novel approach, we have directly quantified the relative contributions of spectral and temporal cues to voice pitch perception by controlling their availability in the electrical stimulation pattern.Psychometric functions for fundamental frequency discrimination were collected using triplets of consonant-vowel syllables in a 3AFC odd-one-out task. The population featured 32 CI listeners varying in age and age-at-implantation, all using n-of-m strategies. Stimuli were presented acoustically or via direct stimulation. For the latter, pulse patterns were recorded in advance using the same processor type and the exact settings of each participant. The recorded patterns were then modified to suppress temporal or spectral cues. Temporal cues were suppressed by filtering out F0-related amplitude modulations from the channel envelopes. Spectral cues were suppressed by amplifying or attenuating current levels to produce similar current distributions for the three intervals in each trial. In these direct stimulation conditions, the altered pulse patterns were streamed to the participant using a research interface.Our results show that CI users relied heavily on temporal cues when available. In speech with less temporal cues, CI users had greater difficulty but some did fall back on spectral cues. While the general scores were higher for the younger participants compared to older participants, their relative cue weightings appeared similar.

Speaker: Floris Rotteveel (University of Groningen, University Medical Center Groningen)

458 Temporal Pitch Perception of Multi-Channel Stimuli by Cochlear-Implant Users

Pitch perception by cochlear-implant (CI) users is generally poor both due to technical limitations as well as biological constraints. Nevertheless, delivering sufficient information to provide the best possible pitch percept is important. Here we explore the feasibility of CI processing strategies that aim to improve pitch perception by presenting information on the stimulus temporal fine structure (TFS) in low-frequency channels to the corresponding apical electrodes. Eight users of the MED-EL CI took part in a pitch-ranking experiment with stimuli presented through direct stimulation and consisting of isochronous pulse trains presented concurrently to the four most apical electrodes. When the same pulse rate was applied to all electrodes, pitch ranks increased until about 300 pulses per second (pps). Presenting different rates (100, 200, 300, and 400 pps) to each electrode produced a pitch percept that was ranked between 100 and 200 pps, irrespective of the rate-to-electrode allocation. For both same- and mixed-rate conditions, maximizing the delay between individual pulses on different electrodes generally produced a higher pitch rank compared to when the delay was kept short. Our results show no evidence that CI users can combine the rates of TFS applied to different apical channels so as to estimate the fundamental frequency but do show that the pitch can be affected by the between-electrode delay, likely due to electrical current spread. We conclude that presenting different temporal patterns to adjacent electrodes is unlikely to produce a robust and clear pitch.

Speaker: Evelien De Groote (ExpORL, Dept. of Neurosciences, Katholieke Universiteit Leuven)

459 Perception of Musical Chord Structure in Cochlear-Implant and Normal-Hearing Listeners

Cochlear implants (CIs) reliably restore speech perception in quiet, but are widely assumed to provide no access to musical harmony due to spectro-temporal interactions at the electrode-nerve interface. While temporal fine structure is crucial for pitch perception in normal-hearing (NH) listeners, its role in CI perception is unclear. We therefore investigated whether spectrally sparse coding of musical chords reveals latent harmonic sensitivity in post-lingually deaf CI users by reduced spectro-temporal blurring under different stimulation strategies.In Experiment 1, 6 CI and 6 NH listeners completed an oddball task with frequent standard triads and infrequent deviant triads. Standards were either physically identical major chords or major chords built on varying root F0s, the latter requiring judgments of voice relations (i.e., chord structure) rather than tracking individual voices. Deviants differed by one semitone in one or two voices. CI listeners were tested with envelope-only, fine-structure, or combined stimulation strategies; NH listeners were trained on vocoded versions transmitting only envelope cues. CI listeners performed above chance in all conditions, with no effect of stimulation strategy. NH listeners performed substantially worse than CI listeners on the vocoded stimuli, suggesting that CI but not NH listeners can effectively extract harmonic information from temporal-envelope cues.In Experiment 2, 10 CI and 10 NH listeners judged harmonically conclusive (with respect to Western harmonic rules) versus inconclusive chord sequences. NH listeners showed high sensitivity, whereas for CI listeners half performed above chance, and half showed no sensitivity.These findings reveal latent harmonic sensitivity in CI users and clarify the roles of temporal-envelope and fine-structure cues in harmony perception.

Speaker: Marie-Luise Augsten (Austrian Academy of Sciences)

460 Impact of salience on local–global temporal processing: A comparison of normal-hearing (NH) and simulated aided hearing impaired (aHI*) listening using General Recognition Theory (GRT)

Local-global temporal auditory processing reflects how complex auditory scenes are organized over time, with global processing typically dominating local processing. This study investigates how aided hearing impaired (aHI) listening shapes the local-global processing and its interaction with increased salience on local elements.In a psychophysical task, on each trial participants (young Normal Hearing NH) were asked to compare the local and global levels (divided attention task) of two melodies composed of three sweeps, each moving upward or downward in pitch (local level) with their relative pitches forming an upward or downward pattern (global level). Timbre was sometimes altered in one sweep to create local salience. Two listening conditions were tested (NH vs. aHI* using hearing loss and hearing aid simulators) to minimize high-level confounding factors associated with hearing loss.Data were analyzed using a two-dimensional signal detection theory framework to assess perceptual interferences between local and global dimensions. Results revealed no effect of listening condition, indicating robustness of temporal processing to acoustic variations. Salience impaired both local and global processing, with a stronger impact than previously reported. Finally, GRT uncovered marked inter-individual differences in perceptual maps, which will be discussed.

Speaker: Armand SCHWARZ (STMS (Ircam-CNRS-SU))

A14.08 Computational and AI approaches in audiology: S105

Conveners: Mark Saddler (Technical University of Denmark), Volker Hohmann (Carl von Ossietzky Universität Oldenburg)

461 Non-intrusive, deep-learning-based prediction of speech intelligibility and listening effort across diverse acoustic scenarios

Models that predict auditory perception are important tools to quantify the effect of sound processing, for instance, speech enhancement algorithms in communication systems or in hearing aids. Non-intrusive models which do not rely on a clean reference signal could potentially be applied in real-world settings if they generalize well in broadly varying acoustic conditions. This contribution compares two non-intrusive deep-learning-based systems for predicting two metrics, subjective listening effort (LE) as well as speech intelligibility (SI). SI and LE predictions are compared in complex binaural scenes, based on speech enhanced by hearing aid algorithms, and synthetic speech. The PHOne-based Binaural Intelligibility model (PHOBI) predicts both metrics by measuring phone prediction uncertainties of an automatic speech recognition system. HASANet+ is a deep-learning framework trained to mimic the output of two intrusive speech perception models without a clean reference signal, utilizing the WavLM foundation model for acoustic feature extraction. Both models show strong correlations (r > 0.88) across all tasks, confirming the validity of non-intrusive approaches for speech perception modeling.

Speaker: Hartmut Schoon (Carl von Ossietzky Universität Oldenburg)

462 A Deep Learning Framework for Understanding Cochlear Implants

Cochlear implants (CIs) aim to enable hearing in people with severe hearing loss by electrically stimulating the auditory nerve. Although highly successful, CIs fail to restore normal perception in many listening situations. We propose a deep learning framework to evaluate factors limiting CI-mediated perception in real-world auditory tasks. We optimized deep artificial neural network decoders to recognize words, localize sounds, and selectively attend to a cued talker in multi-talker situations using simulated auditory nerve representations as input. We trained decoders with either acoustically or electrically stimulated auditory nerve input to estimate upper bounds on hearing task performance given either normal hearing or CI-mediated peripheral representations. Once trained, we compared the models to humans with normal hearing and CI users by testing them on the same tasks. To investigate outcomes for different device manufacturers (Cochlear, MedEl, Advanced Bionics), we separately trained models on simulated nerve activity evoked by their respective sound-coding strategies.Models optimized with CI input exhibited impaired speech recognition, sound localization, and selective attention relative to normal hearing models, revealing limitations in the information conveyed by CI stimulation. Performance was similar across the different sound coding strategies. The best-performing human CI users approached the performance of the CI models when tested on the same task, consistent with the idea that the best-performing humans perform about as well as is possible given their CI. The ability to predict real-world behavioral outcomes for candidate prostheses opens the door to large-scale screening of new device strategies.

Speaker: Annesya Banerjee (Harvard University)

463 Local Interpretable Model-Agnostic Explanation (LIME) For Predicting Loudness Classes From Electroencephalography (EEG) Data

The utilization of artificial intelligence in critical and high-risk applications is poised to become a prevalent practice soon and the development of a methodology for testing and approving machine or deep learning models is imperative. As most experience has been accumulated in the medical field to date, this contribution employs a classification task using electroencephalography data to illustrate the preparation of reliable and effective approval and test methods.A deep learning network was trained to assign auditory brainstem responses to three loudness classes. The generation of artificial data was achieved by utilizing audiological experimental reference data, with a local instance being defined to represent the normal hearing of a hypothetical test subject. The time series (trials) were created by employing a perturbation method. The single-layer perceptron model was selected amenable to an interpretation, with the weights indicating the data to which the network is sensitive. To validate a specified relevance parameter derived from the perceptron weights, the accuracy and a metric distance were calculated as performance measures. It has been demonstrated that the perceptron can characterize the relevance of specific data within a local environment surrounding the instance. Furthermore, conclusions have been drawn regarding the rendering of new test data.

Speaker: Christian Koch (Physikalisch-Technische Bundesanstalt)

Full paper (PDF)

464 Harmonizing and Integrating Audiological Data through Data-Driven and Model-Based Approaches

Clinical decision-support systems (CDSS) can support experts’ decision-making by exploiting big data. For example, a classification integrated into the CDSS can provide a statistical proposition of which hearing device a patient would benefit from; or data-driven, unsupervised approaches can characterize patient groups available in the data, showing different profiles of audiological test outcome combinations and different prevalence. The main challenge is to base such a CDSS on international, really “big data”, since local clinical-audiological databases comprise different audiological tests and test conditions, data structure and formats, expert knowledge, or patient populations.Latent variable approaches help deal with these differences by transforming and summarizing diverse input data into a common, harmonized representation that can be derived by experts, machine learning, or auditory models. This presentation focuses on the comparability of speech test data across languages, speech material complexity (words, sentences), and conditions (in quiet, in noise). Results from a model-based approach applied to different international datasets will be discussed. In the future, the proposed approach can be used to transform between different speech tests, to enable combined analysis of various datasets from research and clinics.

Speaker: Mareike Buhl (Medical Physics, University of Oldenburg)

A16.08 Acoustic needs for old and new worship spaces: S119

Convener: Francesco Martellotta (Politecnico di Bari, DARCOD)

465 Worship acoustics: where are we going?

Research on the acoustics of worship spaces received significant attention in the early 2000s, with important contributions addressing different religious traditions, architectural typologies, and liturgical practices. These researches helped to fill a gap in the scientific literature by combining acoustic measurements in existing buildings, subjective investigations, and the development of design criteria for spaces intended for speech, music, ritual, and communal participation. In recent years, studies on worship acoustics have become more widespread but less systematic. Much recent work has focused on the acoustic characterization of historic buildings, often in relation to documentation, conservation, or restoration. While this has produced valuable knowledge, less attention has been devoted to the changing acoustic needs of contemporary worship spaces, including the use of sound reinforcement, multifunctional layouts, evolving liturgical practices, and the expectations of diverse congregations. This study aims to identify current acoustic design trends in worship buildings, considering both new and existing spaces across different religious creeds and regions. The analysis focuses on how design strategies address speech intelligibility, musical quality (where it is relevant), ritual atmosphere, flexibility of use, and the preservation of architectural and cultural identity. By examining recent research and design approaches, the study seeks to clarify whether contemporary worship acoustics is converging towards "purely functional" design principles or remains primarily shaped by worship-specific requirements.

Speaker: Francesco Martellotta (Politecnico di Bari, DARCOD)

Full paper (PDF)

466 Exploring the Acoustics, Architecture, and Use of Three Arnold Brunner Synagogues

Arnold Brunner (1857–1925), the preeminent American-born Jewish architect is attributed with designing several notable and imitable synagogue buildings. While some of these synagogues, like the Moorish/Byzantine Temple Beth-El, have been demolished, three of his landmark synagogues remain in New York City today, including buildings erected for Congregation Shearith Israel, Temple Sharaay Tefila, and Temple Israel of Harlem. Brunner's adoption of the Classical style had lasting influence in American synagogue design in the late nineteenth and early twentieth centuries. While the Sephardic community Shearith Israel still occupies their Brunner building, the other two congregations, both affiliated with the Reform movement, have relocated to mid-twentieth century buildings. Shaaray Tefila's former building is now occupied by the Ukrainian Orthodox Cathedral of St. Volodymyr, and Temple Israel's building now houses the Mount Olivet Baptist Church. The use of the space and the acoustic needs vary greatly across these faith traditions. Acoustic measurements of these historical Brunner buildings are presented and contextualized by the changing needs of the congregations that inhabit them. The current buildings of Sharaay Tefila and Temple Israel are also surveyed. Acoustic issues related to sanctuary arrangement and furnishing, speech intelligibility, liturgical and musical suitability, and the use/disuse of amplification are discussed.

Speaker: Elliot Canfield-Dafilou (Yale University)

467 A preliminary investigation on the acoustics of contemporary Algerian mosques inspired by the Ottoman style

The historical Ottoman mosques continue to exert a significant influence on mosque architecture in Algeria, even in the late 20th century. Within this framework, a novel interpretation of classic Ottoman mosques has contributed to the architectural landscape of Algerian cities. However, the primary question remains: do these models affect (and replicate) the acoustic conditions in the prayer hall, or is their influence limited to the shape? This study aims to address this question by conducting a comparative analysis of two contemporary mosques in Algeria inspired by the Ottoman style. The first is more conservative and is located in the city of Boussaâda, in the south of Algeria. It was constructed by the initiative of the local population. The second is situated in the north of Algeria, specifically in Oran. It was donated by the government of Turkey and represents a new interpretation of Ottoman heritage. The study method is based on in-situ measurements to assess the primary acoustical parameters complying with current ISO standards. Results show that in both cases very long reverberation time is measured, resulting in poor clarity and speech intelligibility. Comparison with historical ottoman mosques of comparable volume shows that contemporary buildings are more reverberant, suggesting the need for improved acoustics control.

Speaker: Mohamed Ladaoui Benfarhat (University of Biskra)

Full paper (PDF)

468 The contribution of domes to speech intelligibility in mosques

Mosques are multilingual spaces in non-Arabic countries. These mosques require intelligible speech in multiple languages, including the languages spoken by the majority of the population in the country and Arabic. For instance in Türkiye, speech should be intelligible in both Arabic and Turkish. Speech intelligibility in enclosures is mainly influenced by reverberation time (RT) and signal-to-noise ratio (SNR). In the literature it is evident that languages are affected by RT and SNR differently. Additionally, due to the sacred nature of mosques, a divine feeling should be present by strong late reflections at certain frequencies. Therefore, early and reflections in mosques should be adequate for each language spoken in the enclosure, while maintaining divine feeling with strong late reflections at specific frequencies. This hypothesis requires multi-layered analysis of field measurements of room acoustic parameters, acoustic simulations and listening tests to achieve an optimization between multilingual speech intelligibility and divine feeling. This paper presents the first phase of an extensive research project. In this phase, our aim was to reveal the effects of dome geometry on speech intelligibility; therefore, we analyzed the difference of acoustic characteristics between domed and non-domed mosques. We conducted field measurements in Mogan Lake Mosque in Ankara, which is a unique example of non-domed mosques, followed by computer simulations evaluating an alternative multi-dome configuration of the same mosque. The results of the study will reveal the contribution of dome geometry to speech intelligibility in mosques.

Speaker: Kivanc Kitapci (TOBB ETU)

Full paper (PDF)

A18.00 Soundscape, Environmental Quality, Health and Well-being: S127

Conveners: André Fiebig (TU Berlin, Department of Engineering Acoustics), Arezoo Talebzadeh (Ghent University)

469 An expert discussion on the current and future value of AI in the field of urban sound

As part of the International Noise Awareness Day, in January 2026 the German Acoustical Society (DEGA) and the German Environment Agency (The UBA) organized an expert panel to discuss the use of artificial intelligence (AI) across various fields related to sound. The goal was to identify from different disciplinary angles both challenges and useful applications to deal with urban sound. With this presentation it is intended to bring the discussion to the European level of expert groups. The panel focused on sharing experiences with AI in environmental acoustics, noise control, and soundscapes, highlighting applications, opportunities, obstacles and risks. It is evident that AI has already conquered personal and professional contexts, mainly as a support tool still requiring expert oversight. In the field of urban sound, AI offers potential to improve data quality, enable smarter traffic management, enhance urban and noise planning, and support more precise sound analysis and personalized noise assessments. The experts also recognized that AI can improve efficiency, support public communication, and enable new approaches like sensor-based monitoring and citizen science. However, identified key challenges include data privacy, legal constraints, dependency on high data quality and extensive training data, and the limited transparency of AI models. There is also concern about overreliance on AI without sufficient critical evaluation along with unclear implementation pathways. Survey results of the experts show positive attitudes toward AI, though risks-especially data privacy-remain a concern. Stronger regulation and greater integration of AI into education and training are recommended.

Speaker: André Fiebig (TU Berlin, Department of Engineering Acoustics)

Full paper (PDF)

470 Urban Soundscape With UAV

The rapid expansion of unmanned aerial vehicles (UAVs) in urban environments introduces a new and distinctive sound source that transforms our soundscapes. This paper explores the characteristics, propagation, and human impacts of UAV noise, and considers how urban planning and design can respond effectively. Compared with conventional urban sounds, UAV noise exhibits unique spectral and temporal features, including tonal components, broadband energy, and dynamic fluctuations linked to flight behaviour. Through a simulation-based case study in Shenzhen, the research shows that UAV noise distributes unevenly across space. Evidence from prior studies suggests that such noise is often perceived as more intrusive and annoying than traditional transportation sounds. Although emerging findings indicate physiological stress responses, results remain inconsistent and require further validation. To address these challenges, the paper advocates for integrating soundscape principles into UAV-related urban planning, with strategies such as optimizing flight paths, increasing operational altitude, adjusting speed profiles, and leveraging existing urban noise for masking.

Speaker: Tianjing Feng (Institute for Environmental Design and Engineering)

Full paper (PDF)

471 Traffic Noise Perception Assessment for Traffic Management: An Empirical Study Based on Facial Action Units

As traffic noise has become a major source of noise in urban public spaces, key traffic factors such as vehicle speed and honking patterns have been shown to have a significant impact on people’s perception of their environment; consequently, more refined measurement and research methods are required to support relevant traffic management. This paper proposes a method for measuring the perception of traffic noise using Action Units (AUs) based on facial expression recognition (FER), and conducts perception experiments to investigate the effects of vehicle speed and honking patterns. The results show that medium- to high-speed traffic noise significantly reduces valence (by 0.04–0.10), whilst low-speed traffic noise has no significant effect; when foreground horn sounds were introduced, the initial decrease in valence was similar across conditions, whilst the subsequent recovery following the cessation of the horn varied between 0.06 and 0.14 across different patterns. This methodology and its findings help to elucidate the non-verbal emotional responses triggered by traffic noise and provide a basis for refined traffic management measures, such as speed control and horn usage regulation.

Speaker: Xuejun Hu (Harbin Institute of Technology)

Full paper (PDF)

472 Urban Soundscape Generation from Video: Domain-Specific Fine-Tuning of a Multimodal AI Model

Soundscape research has received growing attention in recent years, particularly in relation to machine learning-based prediction and multimodal applications. Recent multimodal audio generation models have shown promising performance on general audiovisual content, but their applicability to urban soundscape generation has received limited attention. This paper presents a pilot study on adapting the pretrained MMAudio model for video-conditioned urban soundscape generation. Urban scene recordings were segmented into 8 s clips using overlapping temporal sampling with a stride of 4 s, resulting in a dataset of 6,946 clips. A lightweight adaptation pipeline was developed on top of the original MMAudio framework, including custom clip loading, streaming TensorDict memmap feature extraction, and full fine-tuning in 16 kHz mode on a single NVIDIA L4 GPU. The study establishes a practical baseline workflow for urban soundscape generation under limited computational resources. Preliminary results suggest that domain-specific fine-tuning is feasible and can move generated outputs closer to the target urban soundscape domain, while also highlighting limitations related to data scale, simple textual conditioning, feature completeness, and the need for fuller evaluation. The work provides an initial step toward video-conditioned urban soundscape generation.

Speaker: Yuqi Liang (UCL Institute for Environmental Design and Engineering)

Full paper (PDF)

473 Expected Reproduction Levels of Soundscapes

Humans appear to possess internal expectations of reproduced sound pressure levels of auditory scenes, but they might not be very accurate. Previous studies have shown a systematic underestimation of reproduction levels for speech and traffic soundscapes. This study investigates the factors underlying such underestimation by examining level expectations across six real-world soundscapes reproduced using a multichannel loudspeaker array. Participants adjusted playback levels to match their expectations of the soundscapes' real-world counterparts, both with and without added recording noise. In addition, they rated perceptual attributes in accordance with ISO/TS 12913-3. Results confirm a strong dependence of level mismatch on the true sound pressure level of a soundscape: levels of louder scenes were consistently underestimated, while quieter scenes were slightly overestimated, indicating a compression effect in level expectations. The true level of the soundscape was a better predictor of estimation error than perceived pleasantness. Added recording noise had only a minor influence, primarily affecting quieter scenes, while familiarity with the soundscape showed no credible effect on accuracy. These findings suggest that level expectations are anchored to specific sound sources and biased toward perceptually comfortable levels.

Speaker: Nils Meyer-Kahlen (Aalto University)

Full paper (PDF)

A01.05/A17.03 Automotive Audio and Active Control: P420

Conveners: Filippo Maria Fazi (Institute of Sound and Vibration Research), Fabio Cagnetti (BdSound S.r.l.)

474 A Practical Use of Multiple Coherence Function on Active Road Noise System Optimization

This study explores the effect of utilizing the Multiple Coherence Function (MCF) analyzation in optimizing reference sensors and installment positions of Active Road Noise Cancellation (ARNC) system. The ARNC system broadly consists of speakers, microphones, sensors which provide reference of noise source and electronic controller which filters the unwanted noise through algorithm. ARNC is influenced diversely in differenct frequency domains of road noise: booming noise, tire-pavement interaction noise, and rumble noise. Since a vehicle contains various systems within a limited installment space and budget, optimization of ARNC system became necessary. MCF is one of the analyzations using coherence to outstand the most referring sensors to a road noise out of others. Throughout the experiment, two different vehicles, the vehicle speeds, and the number of people loaded conditions were varied while the vehicles ran on the same testing road on each trial to observe the consistency of the result. Three-axis accelerometer sensors were attached to each wheel side, and target and error microphones were placed near the ear sites at each seat. In conclusion, the contribution of each noise source sensor in main road noise frequency domain can be evaluated by using the MCF. The MCF also allows a coherence comparison between reference noise source sensors and the effect of combination of those sensors. It suggests that the closer the sensors are to the cabin, the higher the contribution level of providing the reference of the road noise source. Furthermore, it suggests at least two sensors in opposite locations to strengthen the contribution.

Speaker: Nayoung Lee (HYUNDAI MOBIS)

475 Acoustic Propagation Modeling for Sound Source Localization in Urban Scenarios

Acoustic perception can complement vision-centric sensing in autonomous driving by providing safety-critical cues such as emergency-vehicle sirens and horns, which often include strong tonal components and distinctive modulation patterns. Acoustic sensing for autonomous vehicles spans event detection/classification, separation, localization, and tracking. This work focuses on sound source localization in urban scenarios. Recent approaches increasingly combine classical array processing with machine learning to improve robustness, but their effectiveness depends on the availability of large, labeled datasets. Model-based alternatives can account for propagation effects, yet high-fidelity urban acoustic modeling is computationally expensive and typically requires environmental information, including acoustic properties, that is rarely available in uncontrolled outdoor scenes. This paper investigates approximate acoustic modeling strategies for urban-like scenarios where the source may be hidden from direct view, for example around corners. We propose a propagation model based on virtual sources and evaluate the localization performance when combined with sequential measurements acquired along the vehicle trajectory. By exploiting motion, spatial diversity can be achieved over time, allowing pressure measurements to be collected at different locations sequentially and reducing the sensor count. To enable controlled evaluation, we employ a simulation-based framework to generate reference data for a moving receiver, capturing reflections, diffraction, Doppler effects, and absorbing boundary properties. The results indicate promising potential for acoustic propagation modeling in urban scenes to support localization under non-line-of-sight conditions, while keeping sensing requirements, model complexity, and computational complexity low.

Speaker: Giovanni Battista Pirro (Virtual Vehicle Research / KU Leuven)

476 Electroacoustic Characterization of a Multichannel Automotive Demonstrator Utilizing Distributed Piezoelectric Actuators

In the automotive domain, flat-panel loudspeakers (FPLs) offer a compelling alternative to conventional electrodynamic transducers due to their reduced form factor and lower moving mass, which facilitate integration into interior trim and improve overall power consumption. However, the acoustic performance of these systems depends on coupled design parameters. The most relevant include surface density, bending stiffness, number, and position of actuators and their actuation principle.This study provides an analysis of piezoelectric actuation within a vehicle cabin, progressing from numerical modeling to a full vehicle audio demonstrator. A multichannel architecture is developed in which piezoelectric actuators are embedded into the headrests, door panels, and an OLED center console. To overcome their low-frequency limitations, full-bandwidth reproduction is achieved using a hybrid setup: a conventional dynamic subwoofer reinforces low frequencies, while piezoelectric elements handle the low-mid to high-frequency range.Furthermore, frequency-dependent directivity is characterized for both the headrest units, critical for personalized audio content, and the OLED center console. The former controls directivity through geometric design. The latter integrates four piezoelectric actuators on the same surface, driven by custom signal processing to exploit structural–acoustic coupling and produce a desired radiation pattern.We present acoustic results for individual components and the complete audio system by analyzing FEM simulations and experimental recordings following procedures recommended by the AES Automotive Audio Committee and complemented by perceptual metrics.This study validates the use of piezoelectric actuators for broadband automotive sound reproduction within a distributed multichannel architecture, highlighting key design trade-offs and limitations.

Speaker: Fabio Cagnetti (BdSound S.r.l.)

477 Experimental Investigation of Vehicle Occupancy Effects on Secondary Path Characteristics and ANC Performance

In automotive active noise control (ANC), the secondary path—the transfer from the loudspeaker to the error microphone—has a strong effect on control performance. In real driving, the number of occupants changes often, but its impact on the secondary path has not been studied much through experiments. This paper tests how occupant count changes the secondary-path magnitude and phase in a production passenger car, and how those changes affect ANC. We measure secondary paths under different occupancy conditions while keeping the same loudspeaker–microphone setup, and we compare the results frequency by frequency. We then run ANC using the measured paths to evaluate noise reduction and filter convergence. The results show that occupancy does not change the secondary path in a uniform way across frequency. Instead, some bands show phase spread and noticeable shifts in the transfer response. In those bands, noise attenuation drops and convergence becomes less stable, even when the reference signal has high coherence. These findings suggest that occupancy is a major factor affecting the secondary path in passenger car

Speaker: Joowon Park (Hyundai Mobis Co., Ltd.)

Full paper (PDF)

478 Redundancy-Aware Reference Signal Selection for Automotive Active Road Noise Control

This paper proposes a strategy for an FxLMS-based automotive active road noise control (ARNC) that accounts for redundancy among reference signals and is suitable for application in production vehicles. Usually, multiple reference signals of vehicles with ARNC system are measured by accelerometers mounted at each wheel hub. However, due to the structural characteristics of vehicles, these reference signals are often strongly correlated with each other, and the increased number of references does not necessarily yield proportional improvements in noise attenuation performance, while raising computational load as well as requiring additional memory resources. The proposed method evaluates independent contribution of each candidate reference signal across the operating frequency band and progressively selects the subset that maximizes noise attenuation potential among all reference signals, while excluding reference signal channels whose contribution is already explained by the selected subset, which are redundant signals. Real-vehicle experiments demonstrate that comparable noise attenuation performance is achieved with a reduced number of reference signals, even with significantly reducing computational load and memory resource consumed.

Speaker: Dongmin Yang (Hyundai Mobis Co., Ltd.)

Full paper (PDF)

A03.03 Prediction, digitalisation (BIM) and simulation in building acoustics: P427

Conveners: Andrea Santoni (University of Ferrara), Edwin P.B. Reynders (KU Leuven, Department of Civil Engineering), Antonino Di Bella (Universyty of Padova - DII)

479 An empirical study for the investigation of BIM acceptance in building acoustics

Building Information Modelling (BIM) is at the core of digital transformation in the Architecture, Engineering, and Construction (AEC) industry. While its adoption is consolidating in primary disciplines such as structures and HVAC, specialist fields such as building acoustics still face significant obstacles due to data fragmentation and a lack of standardised information exchange protocols. This study empirically investigates BIM acceptance among acoustic consultants using the Unified Theory of Acceptance and Use of Technology (UTAUT). Using a quantitative survey approach combined with Structural Equation Modelling (SEM), the research rigorously tests the statistical significance of four primary constructs: Performance Expectancy, Effort Expectancy, Social Influence, and Facilitating Conditions. Furthermore, the model incorporates key moderating variables, including gender, age, professional experience, and voluntariness of use, to capture a comprehensive demographic profile of adoption. The findings map the hierarchy of factors driving behavioural intentions, providing a scientific framework for optimising digital workflows. The study emphasises the need for robust data standards and agile, web-based solutions or dedicated plugins to reduce perceived effort and ensure the seamless integration of specialist acoustic data within the broader BIM ecosystem.

Speaker: Andrea Gerbotto (Politecnico di Torino)

A05.01 Beyond L_den: Citizen Science, Smart Sensing, and Next‑Generation Acoustic Indicators for Environmental Noise Perception: P451

Conveners: Cedric Vuye (University of Antwerp), Qiuya Xiang (University of Antwerp)

480 Low-Frequency Noise Complaints in Styria: A Project-Based Assessment of Selected Cases Using DIN 45680 and Statistical Hearing-Threshold Criteria

Since 2021, the Styrian provincial authority has received an increasing number of complaints concerning low-frequency noise immissions. By March 2024, 46 complaints had been registered, of which eight cases were investigated in detail through night-time third-octave-band measurements following DIN 45680. A-weighted sound pressure levels inside affected dwellings were typically below 15 dB, corresponding to very quiet indoor conditions. One of the eight cases, an air-source heat pump in a pure residential zone, exceeded the DIN 45680 reference values for significant annoyance, while two further complaints were attributable to immissions in the audible range rather than to actual low-frequency components. To account for individual differences in low-frequency hearing sensitivity, the assessment was complemented by comparison with statistical hearing-threshold curves derived using ISO 389-7 and ISO 28961 (1 % and 10 % percentile thresholds) and by evaluation of the G-weighted infrasound level. In most cases, the measured spectra lie well below the extended assessment criteria, indicating that low-frequency noise is unlikely to be the cause of the complaints. These observations suggest that the unusually quiet conditions in the affected dwellings themselves may contribute to the reported perceptions.

Speaker: Florian Lackner (Office of the Styrian Government)

Full paper (PDF)

A05.03 Urban sound planning / The quiet city for life quality: S032

Conveners: Robert Arcos (Universitat Politècnica de Catalunya), Laura Estévez-Mauriz (Universidad de León), Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology), Trond Maag

481 The National Register for Environmental Acoustic data in The Netherlands

All environmental acoustic data relevant for planning, monitoring, and policy publicly available via a central digital platform: that is what we built in the Netherlands. It is important that planners, acoustic experts and policy makers work with the same environmental acoustic data. Therefore, this data has been made FAIR. The central digital platform makes the data Findable and Accessible; in order to make the data Interoperable and Reusable, a data model for environmental acoustic data was developed. To maximize acceptance and adoption, users were consulted from the start, both in the development of the digital platform and the data model. Data provision by central and local authorities to this National Register is ensured by legally enforced regulations. Involving and engaging stakeholders is crucial for success and requires continuous effort. We present the National Register, the process of stakeholder involvement and showcase its initial success.

Speaker: Dorien Lolkema (RIVM)

482 Measuring the noise footprint of products: a provisional framework

Noise pollution is a critical yet historically underrepresented environmental pressure in global sustainability assessments. To address this gap, this research introduces the concept of the Product Noise Footprint (PNF) and establishes its foundational calculation methods. The PNF is designed to capture complex source–receiver dynamics across entire product supply chains through a structured indicator offering two possible levels of resolution. First, the emission-based noise footprint (PNFE) acts as a core inventory metric, objectively quantifying the total acoustic energy output from sources across all supply chain phases. Second, the impact-based footprint (PNFI) translates this raw emitted energy into localized environmental and human impacts by accounting for propagation and exposure severity. To ensure comparability across diverse noise sources, similarly to the CO2 equivalent, the framework introduces the Road Traffic Noise equivalent (RTNeq) as a baseline unit. The PNF measure offers a robust, modular tool to formally integrate noise pollution into sustainability assessments.

Speaker: Angelos Tsaligopoulos (University of Trento, Dept. of Civil, Env., Mech. Eng.)

Full paper (PDF)

483 Temporal Structure of Percentile Sound Levels across Urban Road Typologies

Urban planning is a key tool for improving the quality of acoustic environments. The characteristics of urban roads, the elements they contain, and the services located along them directly influence their patterns of use and, consequently, the resulting noise levels. Most studies analyzing the relationship between urban variables and sound levels rely on energy-based average indicators, as these are the standard metrics used in noise mapping. However, recent research suggests that percentile sound levels may exhibit a stronger association with noise-related effects. In this context, percentile sound levels, by capturing temporal variability, may provide additional insight into the characterization of urban acoustic environments. Based on this hypothesis, temporal records collected in in two Chilean cities were analyzed to assess the relationship between the temporal structure of percentile levels and the characteristics of the road typologies where monitoring stations were located. The results show that percentile sound levels are able to differentiate road functionality and provide improved insight into the temporal variability of urban noise. These findings suggest that percentile-based approaches may complement traditional energy-based indicators in urban noise assessment

Speaker: Guillermo Rey-Gozalo (Universidad de Extremadura)

Full paper (PDF)

A08.00 Industrial, Machinery, Equipment Noise and Vibration: P464

Conveners: Ercan Altinsoy (Professur für Akustik und Haptik), Serkan Atamer (Dresden University of Technology), Mehmet Sait Özer (Dresden University of Technology)

484 Acoustic Monitoring for Predictive Maintenance in Water Supply Infrastructure

Reliable water supply is becoming increasingly challenging due to climate change and shifting consumption patterns. Failures in water supply facilities, e.g. pumps and valves, compromise supply security, making early fault detection a key priority for water suppliers.This paper presents the development and implementation of a multi-modal monitoring system for installations at regional water supply facilities. The system combines two complementary data streams: (i) acoustic sensors capture both airborne and structure-borne sound, and (ii) operational machine data, e.g., pump flow rates and power consumption. These modalities are incorporated in an autoencoder structure. This system builds a real-time picture of pump health that neither data source alone could provide.Machine learning algorithms continuously analyze the fused sensor data to detect anomalies indicative of imminent component failure, before an actual breakdown occurs. This enables 24/7 monitoring of critical infrastructure and supports predictive maintenance strategies by providing early warnings that allow timely procurement of spare parts and scheduling of maintenance interventions.The technology is currently validated at two regional water suppliers in Styria and Burgenland, covering representative facilities under real-world operational conditions. The findings contribute to the broader goal of building resilient, data-driven water supply management systems capable of meeting future demands in an era of increasing climate uncertainty.

Speaker: Stefan Grebien (Joanneum Research Forschungsgesellschaft mbH)

485 Development of an FBS Hybrid Model for Suspension Input Force Identification

To accurately predict vehicle NVH performance, subsystem-level validation should be conducted prior to full vehicle development. The Load Transfer Ratio (LTR) is widely used to evaluate force transmission characteristics of suspension systems. However, its reliability strongly depends on how the suspension input force is defined. This study focuses on the identification of suspension input forces for LTR evaluation, rather than direct LTR measurement. Conventional methods rely on dynamic force sensors, which inevitably include excitation system dynamics. To overcome this limitation, a Frequency-Based Substructuring (FBS) hybrid model is developed to identify wheel-center interface forces independently of actuator boundary conditions. Blocked forces at the excitation interface are obtained using an in-situ Transfer Path Analysis (TPA) method and combined with experimentally measured transfer functions through FBS synthesis to derive representative suspension input forces. To assess the influence of input force identification methods, the results are compared with forces obtained using a classical TPA approach. All forces are transformed to the wheel-center reference using Virtual Point Transformation (VPT) and represented as six-degree-of-freedom force vectors, enabling reliable LTR-based suspension NVH assessment.

Speaker: Sangwon Park (HYUNDAI MOBIS)

Full paper (PDF)

A12.01 Numerical methods for acoustics and vibration: P493

Conveners: Dionysios Panagiotopoulos (KU Leuven Campus De Nayer), Stefan Schoder (IGTE, TU Graz), Maarten Hornikx (Eindhoven University of Technology)

486 Effects of vibration fatigue on mechanical parameters of representative numerical models of painted wooden heritage objects

Painted wooden heritage objects are complex multilayered systems strongly affected by environmental conditions. In addition to variations of thermohygromechanical properties among layers, which can induce stresses leading to delamination and cracking, observations in museum environments have reported damage likely caused by vibrations such as paint flaking and cracking, leading some institutions to implement preventive measures. While the impact of climatic variations has been widely studied, the consequences of long-term low-level vibrations remain poorly understood, whereas they generate material fatigue that may catalyse other degradation mechanisms. Understanding and quantifying the effect of long-term vibrations on such multilayered systems would greatly improve conservation strategies. For this purpose, this work proposes to assess the vibration fatigue condition of a material by measuring the evolution of its mechanical properties and correlating it to the microcrack formation. The study is carried out on wood/gesso/paint/varnish beams, that have been artificially aged to replicate the actual conditions of artworks displayed in museums. These samples are subjected to vibrations in the 1–100 Hz range, a frequency band in which objects displayed in museums are typically excited. The evolution of vibration properties and (micro)crack formation are studied with finite element models taking into account damaging, viscoelasticity and plasticity.

Speaker: Jason Caubrière (SATIE UMR CNRS 8029, CY Cergy Paris University)

A12.06 Validation and Benchmarks in Computational Acoustics: P517

Conveners: Stefan Schoder (IGTE, TU Graz), Marcus Maeder (Technical University of Munich)

487 Multi-GPU k-Wave for Large-Scale Ultrasound Simulation: A Comparison of Global and Local FFT Approaches

k-Wave is widely used for time-domain ultrasound simulation, but large three-dimensional problems often exceed the memory and performance limits of a single GPU. Multi-GPU implementations are therefore required to enable larger computational domains while reducing simulation time.This work compares two multi-GPU implementations of the Fourier collocation scheme employed in k-Wave. The first preserves the original pseudospectral formulation using distributed Global FFTs over the entire computational domain, while the second replaces them with Local FFTs evaluated on overlapping subdomains connected by halo exchange. Both implementations were evaluated on a single node equipped with eight NVIDIA A100 GPUs interconnected by NVSwitch.The results show that both approaches enable substantially larger simulations than a single GPU while providing good strong scaling. By replacing global all-to-all communication with nearest-neighbor halo exchange, the Local FFT implementation reduces inter-GPU communication by approximately one order of magnitude and achieves lower runtime for most tested domain sizes. The resulting approximation maintains relative errors on the order of 10^-3 for appropriately chosen halo widths, providing a practical trade-off between numerical accuracy and computational performance.

Speaker: Oliver Kunik (Brno University of Technology)

Full paper (PDF)

A12.08/A16.12 Numerical Methods for Room Acoustics: P507

Conveners: Maarten Hornikx (Eindhoven University of Technology), Albert Prinn (International Audio Laboratories Erlangen)

488 Surface segmentation for acoustic radiance transfer modeling: a case study

In the context of acoustic radiance transfer, one piece of advice is frequently seen in the literature: the reflecting surface should be segmented in patches of approximately equal area (uniformity). Another often-repeated suggestion is that individual surface patches should be compact, i.e., one should avoid using patches which are long but narrow (compactness). Very few works in the literature provide concrete reasons for either of these recommendations, and none discuss the possible repercussions of their violation. It is therefore possible that abiding to these guidelines imposes unnecessary constraints, making models less efficient than they otherwise could be. This paper describes a case study which challenges such conventional assumptions, using environments from the benchmark for room acoustic simulation (BRAS) dataset. Each environment is simulated several times using different criteria for surface segmentation, including ones which deliberately infringe the aforementioned recommendations. Comparing the simulation results to ground-truth recordings shows that the performance of acoustic radiance transfer is not meaningfully affected by either of the challenged guidelines (uniformity and compactness). All of the tested surface meshes are available online, alongside the code to generate and simulate them.Data will be made available upon submission of the full paper.

Speaker: Matteo Scerbo (DTU Electro)

Full paper (PDF)

A12.08/A16.12 Numerical Methods for Room Acoustics: S085

Conveners: Maarten Hornikx (Eindhoven University of Technology), Albert Prinn (International Audio Laboratories Erlangen)

489 Wave-based simulation of a benchmark room with detailed diffuser modeling and surface scattering

Room acoustic simulations often require geometry simplification to manage computational complexity, potentially introducing modeling bias. In this study, we demonstrate that a finite-difference (FD) framework can accurately simulate a measured benchmark room with complex diffusing structures while avoiding explicit geometry simplification.A geometry-based surface scattering approach is introduced, in which boundary surfaces are perturbed using a three-dimensional noise function to mitigate unrealistically specular reflections at high frequencies.The method is evaluated quantitatively against a dense measurement dataset, showing good agreement and supporting its suitability for realistic room acoustic simulation of complex environments.

Speaker: Jan Wouter Smits (University of Edinburgh)

Full paper (PDF)

490 Wave-based simulation of a small reverberation room with local and extended reaction modeling

A finite-difference room acoustic simulation method is proposed that can model both locally-reacting boundaries and extended-reaction effects via propagation in porous materials. Its suitability for wideband simulations in automotive cabin scenarios is explored by comparing simulations of a small reverberation room with experimental measurements, including configurations with varying absorber geometries and material specifications.

Speaker: Jan Wouter Smits (University of Edinburgh)

Full paper (PDF)

491 A Surface-Based Geometrical Acoustics Algorithm Including Diffraction

Raytracing is an established method for computing the late-time part of room impulse responses, but it has the drawback that only very crude Monte Carlo models of boundary scattering and diffraction are possible to include without it losing its attractive computational cost scaling. This happens because higher-resolution models of these processes output multiple child rays for every parent ray received, making the number of rays grow with reflection order. An emerging solution is so-called ‘Surface-Based’ Geometrical Acoustics. Here the distribution of rays arriving at a boundary is mapped onto a predefined set of spatial elements and angular interpolation functions, producing a vector of coefficients. Re-radiation of subsequent reflections is then a matrix multiplication, with the steady state solution being solvable via a Neumann series. Since rays only ever propagate one reflection order before being collected, diffraction and scattering process that cause them to multiply can be included without issue. In the paper an approach to efficiently represent edge diffraction in such algorithms is presented. In involves discretising the sound energy arriving and departing an edge via angle and representing the diffraction via a matrix multiplication relating these coefficients. Indicative results are shown demonstrating the approach.

Speaker: Jonathan A. Hargreaves (Acoustics Innovation Institute)

Full paper (PDF)

A13.01 Guided waves for NDT & SHM applications: P501

Conveners: Lynda CHEHAMI (UPHF), Markus Saurer (University of Graz), Robert Nuster (University of Graz), Theodosia Stratoudaki (University of Strathclyde)

492 Residual Stress-Induced Shifts of Modal Frequencies Measured by Laser Ultrasound

Residual stresses strongly influence the mechanical performance of metallic components and require reliable non-destructive characterization techniques. In this work, modal vibrations of cuboidal aluminum and copper samples with and without residual stresses are investigated using Laser Ultrasound (LUS) and compared to finite element simulations in Ansys Mechanical. LUS provides contact-free excitation and detection, making it particularly suitable for the investigation of modal behavior.The experiments were performed using a pulsed, non-focused excitation laser positioned at the center of the underside of the sample, while a continuous-wave detection laser was scanned across the topside to measure the out-of-plane displacement with a stabilized Michelson Interferometer. Modal frequencies were extracted from long time-domain signals (~20 ms) using Fourier analysis.The measured modal frequencies of the aluminum sample are compared with numerical simulations and show good agreement. The copper samples are cut from a cold-rolled block along both the rolling and transverse directions in order to investigate the influence of residual stress orientation on the modal response. The measurements reveal mode-dependent shifts in the modal frequencies that correlate with the direction of the residual stress.The results demonstrate the sensitivity of modal LUS measurements to residual stress and highlight the potential of the method for non-destructive testing. Future work will focus on the inverse problem of extracting elastic material parameters and residual stress states from the measured frequency spectra.

Speaker: Robert Nuster (University of Graz)

A13.04 Photoacoustic Imaging and Spectroscopy: P497

Conveners: Jan Laufer (Martin Luther University Halle), Ben T Cox (University College London), Günther Paltauf (University of Graz), Robert Nuster (University of Graz), Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe)

493 Bioengineered extracellular vesicles as next-generation bioinspired optoacoustic contrast agents

Near-infrared (NIR) optoacoustic (OptA) contrast agents are still dominated by synthetic organic and inorganic materials. Although these agents can generate strong signals, their broader preclinical and clinical use is often limited by concerns over toxicity, poor biocompatibility, and high manufacturing cost. Extracellular vesicles (EVs) offer a promising alternative because they are naturally derived, biocompatible, can be produced in bulk, and exhibit low immunogenicity. Most reported EVs-based contrast agents, however, are developed by loading them with synthetic dyes or nanoparticles, which may compromise vesicle integrity, reduce targeting performance, and introduce batch-to-batch variability. In this study, we present a fully biological strategy for optoacoustic imaging by developing genetically modified EVs (GM-EVs) from bioengineered mammalian cells. These vesicles are simple to produce and purify, cost-effective, and show strong optoacoustic performance from deep tissue. We further incorporated RVG peptides to support blood-brain barrier crossing after intravenous administration. In a murine brain tumour model, GM-EVs enabled clear tumour visualisation following systemic delivery, likely due to enhanced tumour accumulation. Importantly, the vesicles were well tolerated in vivo, supporting their potential as bioinspired, bioengineered, and biocompatible NIR optoacoustic contrast agents for in vivo imaging.

Speaker: Vipul Gujrati (Institute of Biological and Medical Imaging)

Full paper (PDF)

494 Optical-resolution photoacoustic microscopy with ultraviolet excitation for label-free imaging of biological tissue

Photoacoustic microscopy with optical resolution (OR-PAM) is a method that uses optical absorption contrast for generating microscopic images of mainly biological samples. The resolution is given by the focus dimensions of an optical objective, through which a pulsed laser illuminates a sample. Local heating in the focal volume gives rise to ultrasound transients, which are detected with a high-frequency, focused ultrasound receiver. With an excitation wavelength in the ultraviolet (UV) range, this method is able to visualize cell nuclei owing to the strong optical absorption of DNA molecules.In this study, we used excitation wavelengths between 258 and 264 nm to visualize UV-absorbing structures in biological samples. The photoacoustic images were compared to optical microscopy images that were obtained using standard hematoxylin-eosin (H&E) staining, showing the ability of OR-PAM to target cell nuclei. Special attention was drawn to the optimization of the scanning technique, using a fast galvo scanner to steer the excitation beam across the sample, thereby achieving a 512 µm wide scanning range with a static, focused ultrasound receiver. A special reconstruction method compensates variations in acoustic sensitivity and time-of-flight across the sensitive area of the detector. Due to the targeting of optically absorbing structures, the sample does not necessarily need to be optically transparent, such as in transmission optical microscopy. This is demonstrated by imaging not only standard histological sections with a thickness in the µm-range, but also up to 20 µm deep into about 2 mm thick slices of fresh biological tissue.The results show the ability of the method to obtain diagnostic information from fresh tissue without using the time-consuming standard staining techniques.

Speaker: Günther Paltauf (University of Graz)

A14.05 Neural substrates of complex auditory perception and auditory scene analysis: P530

Conveners: Katrin Krumbholz (University of Nottingham), Joseph Sollini (University of Nottingham)

495 Sound Exposure in Cities: From Annoyance to Acoustic Amenity

The soundscape concept characterizes the acoustic environment as perceived by individuals within their context. To classify these sounds, the categories of geophony, biophony, and anthropophony have been proposed for natural, animal, and human-produced sources, respectively. Urban environments present a unique soundscape to which inhabitants are continuously exposed, significantly impacting mental and physiological well-being. Although growing research addresses subjective judgments and general physiological responses to soundscape exposure, a detailed understanding of which specific acoustic features drive distinct physiological changes remains lacking. This gap is particularly pronounced in real-world settings.Our research addresses this methodological gap through a multimodal approach conducted both in the laboratory and in the field. By monitoring cardiac, neural, respiratory, and electrodermal activity with portable devices, we employed statistical methods to correlate sound events with physiological markers of sympathetic and parasympathetic autonomic activity. Specifically, our laboratory experiment supports the hypothesis that urban soundscapes elicit higher sympathetic (arousal/stress) and lower parasympathetic (relaxation) activity compared to natural soundscapes, while offering insights into the role of distinct acoustic elements. Concurrently, preliminary in-situ data collection validated the recording system for mobile applications and confirmed trends consistent with laboratory conditions. These findings pave the way for a deeper understanding of the interaction between acoustic and contextual elements in soundscape perception.

Speaker: Boris Gourevitch (Institut de l'Audition, Institut Pasteur)

Full paper (PDF)

A15.02/A24.05 Hearing Research in Virtual Environments: S108

Conveners: Janina Fels (IHTA, RWTH Aachen University), Carolin Breuer (IHTA, RWTH Aachen University), Nils Peters (Trinity College, The University of Dublin)

496 Localization of real and virtual sound sources using headphone-based virtual acoustics in a living room environment

Virtual acoustic environments are increasingly used in psychoacoustics and spatial audio research, with growing importance for virtual and augmented reality applications. However, achieving perceptual validity comparable to real-world scenarios remains a critical challenge. This study investigates human sound localization performance by directly comparing perception of sounds played from physical loudspeakers in a typical home environment (living room lab) with headphone-based, real-time virtual acoustics using the Low-latency Interactive Virtual Environment and Room Acoustics Simulator (liveRAZR). Dummy (non-individualized) head-related transfer functions (HRTFs) and interactive head tracking were used for the virtual sound source auralization. Normal-hearing listeners were seated on a sofa in the living room and always wore open “floating” headphones and a head tracking device. They were asked to indicate the perceived spatial position of a sequence of pink noise bursts on a computer screen displaying an image of the room from the listener’s perspective. The sounds could originate from either real or virtual loudspeakers located in the same room and in a neighboring room connected by a door. Real and virtual sounds were randomly interleaved. Localization accuracy was quantitatively assessed by evaluating the spatial error for the real and virtual sounds and the results are discussed. This research aims to provide a quantitative comparison between real and simulated room acoustics regarding spatial perception, important for tele presence as well as for extended and augmented reality applications with the goal to render virtual sound sources in a real environment, ideally indistinguishable from real sound sources.

Speaker: Yuqing Li (Medizinische Physik, Universität Oldenburg)

497 A perceptual comparison of modal sound fields in real and virtual environments

Modal sound fields arise from wave interference in enclosed spaces and dominate the low-frequency acoustic response of a room. Capturing this behavior in virtual acoustic environments requires numerical wave methods, which remain computationally expensive for real-time interactive applications. Geometrical methods achieve this necessary efficiency but may not incorporate accurate modal information. This study examines the perceptual impact of omitting a modal sound field in an interactive virtual environment (IVE) through a loudness-matching task. Participants equalized the perceived loudness of a sound field at positions throughout an interactive real environment (IRE) and an IVE using controlled stimuli. Relative level differences of stimuli achieved over related positions in the IRE were analyzed and compared to the same evaluation in the IVE, which does not support a modal field. The matched levels throughout the different position and frequency combinations demonstrate the presence of perceived modal-field in the IRE, while the IVE does not produce comparable trends. Furthermore, listeners were markedly less consistent in the IVE than in the IRE. These results suggest that geometrical method-based auralization does not convey the perceived spatial structure of a modal sound field.

Speaker: Thomas Robotham (International Audio Laboratories Erlangen)

Full paper (PDF)

498 Spatial Auditory Attention Detection Using Dry Electrode Electromyography and Eye Tracking

This study investigates the potential of decoding spatial auditory attention using biosignals, specifically the vestigial postauricular muscle reflex (PAMR) and eye movements. Surface electromyography (EMG) using dry electrodes and eyetracking were recorded concurrently while participants were exposed to acoustic stimulation (broadband noise and narrowband sine bursts) across nine spatial locations and four sensation levels. Results reveal a functional difference between the two systems. The PAMR acts as a rapid, orienting mechanism that strictly requires high-intensity, broadband acoustic energy to trigger a robust spatial gradient, actively suppressing frontal and narrowband targets. The eye movements exhibit a lateralized orienting response that is more robust across all tested spectral conditions.

Speaker: Mira Hesse (TU Ilmenau)

Full paper (PDF)

A16.01 Design, Fabrication, Characterization and Perception of Innovative Acoustic Surfaces: P483

Conveners: Louena Shtrepi (Politecnico di Torino), Zackery Belanger (Umbel)

499 Acoustic Performance of a Bio-based Multilayer Wallpaper Using Hemp and Straw Fibers

The acoustic behaviour of multilayer wallpaper made from hemp stalk and rice straw is investigated in this study. The work was developed from a broader study on bio-based interior materials, but the present paper focuses on acoustic performance. The proposed wallpaper consists of a porous core layer based on hemp–straw mixtures and a surface layer made of Xuan paper, jute fabric, and dried botanical elements. Samples with different hemp–straw ratios and different thicknesses were prepared under controlled conditions to examine the influence of material composition and specimen thickness on sound absorption. The normal-incidence sound absorption coefficient was measured by means of an impedance tube over the frequency relevant to interior environments such as hotel rooms and bedrooms. The measurements show that the wallpaper provides sound absorption in the speech-frequency range, and that the acoustic response changes with both thickness and fiber composition. The results suggest that bio-based wallpaper can be considered not only as a decorative wall finish, but also as a functional interior layer with acoustic value.

Speaker: LULU JIANG (POLITECNICO DI MILANO)

Full paper (PDF)

500 Ninefold: A Distributed Resonant Field

Edgard Varèse described music as “organized sound.” and Iannis Xenakis, in collaboration with Le Corbusier, developed spatial systems in which sound, form, and movement are composed as a single environment. Building on this framework and lineage, Ninefold treats composition as an organizational system where architecture, music, and fabrication converge through relational interactions among frequencies, surfaces, and bodies. The installation constructs a spatial field in which sound and form operate as co-constitutive systems. A harmonic spectrum of nine interdependent overtones ranging 50–1800 Hz generates evolving combinatorial polyrhythmic textures that converge and disperse across space. Sourced from everyday waste, recycled plastic becomes the material substrate for the additive fabrication of nine acoustic mediators designed to hold and redirect sonic energy. Acoustic structure, geometric variation, and material behavior converge through distinct artifacts to produce an immersive harmonic field shaped by sonic gradients. Subtle shifts in geometry produce perceptible differences in acoustic response, establishing gradients (propagation, diffusion, reflection, and concentration) rather than boundaries. Different modules selectively bias the spectrum: elongated and tighter cavities amplify upper bands (800–1800 Hz), producing softer, while flared volumes and smoother interiors reinforce low–mid frequencies (50–400 Hz), yielding fuller, more resonant output. Intermediate ribbed geometries introduce localized reflections, accentuating midrange bands (400–800 Hz), generating an array of difference tones and subtle beats experienced through movement.Ninefold composes an environment where artifacts function as sonic and visual bodies. Listening becomes embodied and situated, unfolding as a temporal sequence of encounters shaped by position and proximity rather than fixed composition.

Speaker: Misri Patel (Georgia Institute of Technology)

Full paper (PDF)

A16.01 Design, Fabrication, Characterization and Perception of Innovative Acoustic Surfaces: S113

Conveners: Louena Shtrepi (Politecnico di Torino), Zackery Belanger (Umbel)

501 Stagger: The design, installation, and acoustic validation of a wall of chairs

Stagger is an experimental acoustic surface that proposes to expand the role of furniture in architecture. Vintage Haworth stackable chairs from the 1990s, which were designed to have minimal footprint when not in use, were repurposed to intentionally clad a wall in an acoustically beneficial and visually engaging arrangement. The chairs retain their original purpose, and are removed from the wall and used for events as needed.Such an arrangement can readily be classified as acoustic diffusion, a function which was studied using FDTD simulation techniques. Although comprised entirely of hard plastic and steel, the installation was also tested for absorption per ISO 354 at Riverbank Acoustical Laboratories as part of their Novelty Testing Program, which supports testing of everyday objects and surfaces. The results show mild broadband acoustic absorption, providing further evidence for the potential benefits of intentional acoustic design in furniture systems.

Speaker: Zackery Belanger (Umbel)

502 Contextualising Acoustics: Prototyping Conceptual Boundaries in Acoustic Design

Within architecture’s inherently multisensory framework, contemporary acoustic design often remains secondary, addressed through applied absorptive treatments rather than embedded spatial strategies. Despite advances in computational tools, most built environments still rely on standardized reflective materials like concrete, glass, plywood, resulting in a limited, binary understanding of acoustic performance. This research challenges that paradigm by positioning acoustics as an integral design driver and exploring material behavior as a continuous spectrum rather than fixed categories. The study investigates how emerging design and fabrication technologies can recalibrate architectural thinking, enabling acoustics to inform form, materiality, and spatial experience from the outset. It focuses on materially consistent yet geometrically varied surfaces, leveraging additive manufacturing to encode acoustic performance directly into form. Through controlled variation, a single base material is designed to exhibit a range of sonic behaviors, including reflection, diffusion, and absorption. A catalogue of 25 prototypes was developed, from which four specimens of equal weight but distinct formal characteristics were selected for testing. Acoustic performance was evaluated using ASTM E1050-12 impedance tube measurements. Results demonstrate significant variation in absorption coefficients across the samples, confirming the presence of an embedded acoustic gradient. The findings situate these prototypes within an intermediate performance range, between highly reflective glass and absorptive fiberglass, an area that remains underexplored in architectural acoustics. By integrating digital fabrication with acoustic performance, the research proposes a shift toward embedded, context-specific sonic environments. It extends the provocations: How might designers and acousticians extend the acoustic continuum between material and materiality architecturally?

Speaker: Misri Patel (Georgia Institute of Technology)

Full paper (PDF)

503 Exploring the sound absorption potential of open-cell ceramic foams fabricated via Viscous Thread Printing

The advancement of 3D printing of ceramic materials has opened several research lines regarding the fabrication of acoustic porous building materials for a range of applications. To this end, this paper presents a first proof-of-concept experimental characterization of sound-absorbing panels produced via Viscous Thread Printing (VTP), a previously unexplored fabrication method for room acoustic applications. VTP is an extrusion 3D printing process in which the induced filament instability spontaneously generates an open-cell structure without requiring an explicit toolpath design. We evaluated the sound absorption of three typologies of ceramic panels and of resulting layered configurations, addressing the impact of design parameters such as thread diameter and associated pore size, panel thickness, and layering order. The acoustic measurements were conducted in the Small scale Reverberation Room (SSRR) of Politecnico di Torino. Measurements on cylindrical samples of the base ceramic material were performed with an Impedance Tube to gain initial insight into the contribution of the three-dimensional foam structure. Experimental results show, at greater thicknesses and in layered configurations, panels with absorption coefficients in the range 0.3 to 0.65 in the mid-high frequency spectrum, framing VTP as a fabrication method with potential for further research in the field.

Speaker: Marco Palma (TU Wien)

Full paper (PDF)

A12.01 Numerical methods for acoustics and vibration: S307

Conveners: Dionysios Panagiotopoulos (KU Leuven Campus De Nayer), Stefan Schoder (IGTE, TU Graz)

504 State-space modelling of head-related transfer functions

Any linear time-invariant system admits a state-space realization. In contrast to classical transfer-function models, state-space formulations remain numerically robust at high orders. They can also be more computationally efficient than block-wise convolution-based rendering, especially for MIMO systems. Another advantage is access to model order reduction methods that produce compact models with a priori error bounds.We argue that state-space models deserve renewed attention in acoustics. While more common in earlier work, they have partly fallen out of focus. They are beneficial in many areas, but head-related transfer functions (HRTFs) are especially well suited to highlight their advantages.Despite this potential, state-space HRTF models are still rare in virtual acoustics. Such models can be constructed from simulations or measured data; here, we focus on head-related impulse response measurements. The few approaches reported in the literature indicate clear advantages, but they have struggled to achieve very high fidelity because of computational bottlenecks in the reduction algorithms rather than in the state-space models themselves.In previous work, we have already demonstrated the feasibility of modern numerical linear algebra methods such as the randomized or tangential Eigensystem Realization Algorithm for acoustical applications. Building on this foundation, we show that these advances remove the relevant bottlenecks for HRTF modelling. High-fidelity state-space HRTF models can thus be constructed reliably and with negligible error. We further provide user-friendly open-source Python implementations of the reduction algorithms and efficient solvers for rendering. Overall, the talk aims to show that state-space models are again a practical framework for acoustical modelling, with HRTFs providing an instructive example.

Speaker: Art J.R. Pelling (Technische Universität Berlin)

Full paper (PDF)

505 Efficient solution of multiple input multiple output vibro- acoustic Finite Element systems through an accelerated deflation strategy

The design of vibroacoustic systems commonly requires repetitive high-fidelity finite element evaluations across wide frequency ranges and for numerous excitation scenarios, at significant computational cost. Conventional model order reduction (MOR) techniques, which are often employed to alleviate this cost, apart from inherently trading off solution accuracy for computational gain, typically face notable limitations in multiple-input multiple-output (MIMO) settings: moment-matching approaches scale poorly with the number of inputs, while modal methods struggle with large scale systems. To overcome these challenges, this work proposes an accelerated deflation-based solution strategy for large-scale vibroacoustic finite element MIMO systems. Rather than using MOR to approximate the system response, the proposed approach leverages it to construct a deflation preconditioner that significantly accelerates GMRES convergence when a frequency sweep needs to be solved for multiple inputs. The reduction basis is generated using an Automatic Krylov Subspace Recycling (AKR) algorithm, which collects relevant subspaces based on a maximum iteration criterion. To handle the high condition numbers typical of vibroacoustic finite element models, two preconditioning strategies are incorporated: scaling of the vibration, acoustic, and interface sub-matrices by their Frobenius norms, and LU-based preconditioning at the mid-frequency of a given input. Together, these measures enable practical deflation with moderate subspace sizes. Both theoretical cost analysis and numerical results on a coupled plate–cavity benchmark demonstrate substantial reductions in iteration counts compared to conventional iterative strategies. As with traditional MOR approaches, the dominant cost is transferred to an offline phase; however, the resulting online MIMO frequency sweep is not only highly efficient but also free of any compromise in solution accuracy.

Speaker: Dionysios Panagiotopoulos (KU Leuven Campus De Nayer)

506 Building web-page boundary elements software for modeling sonic crystals scattering pattern

The rapid development of computational acoustics has enabled the creation of digital twins for a wide range of engineering applications. Among the most versatile numerical approaches are the Finite Element Method (FEM) and the Boundary Element Method (BEM), both of which are now available in advanced commercial packages supporting full-scale acoustic design and verification.However, innovative research concepts often require transparent and customizable tools that expose the underlying numerical formulation capabilities that commercial software rarely provide. This motivates the use of programmable, open modeling environments that allow researchers to inspect, modify, and extend the core algorithms, particularly in BEM-based scattering analysis.Building on our earlier contribution, this paper presents mm-bem, a lightweight web-browser application implementing a classical BEM formulation for acoustic scattering from mesh-defined objects. The tool runs entirely client-side, requires no installation, and provides interactive visualization of both near- and far-field responses. As a demonstration, we analyze the scattering behaviour of a periodic lattice of spheres, illustrating the applicability of the approach to sonic-crystal configurations.

Speaker: Marek Moszyński (Gdansk University of Technology)

Full paper (PDF)

09:40 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:20 Coffee break

10:20 Coffee break

10:20 Coffee break

10:20 Coffee break

A01.04 Active Sound and Vibration Control: S416

Conveners: Alberto Gonzalez, Felix Langfeldt (University of Southampton)

507 Acoustic Emulation of Scattering from an Object for a Moving Emitter via Equivalent Surface Control

We address the problem of emulating, in real time, the acoustic scattering from a stationaryobject under a moving emitter in three-dimensional space. Our approach employscontrol sources distributed over an artificial surface that surrounds the regionwhere the object is to be emulated. The control algorithm is based on the Kirchhoff– Helmholtz formulation, which is extended to the time domain to account for emittermotion and potentially nonstationary probing signals, and is combined with a dedicatedtime-dependent physical optics model. Reduction of the control-source extent is also exploredthrough the introduction of an equivalent internal surface.

Speaker: Lea Beilkin-Sirota (Tel Aviv University)

508 Force Cancellation Motor

This paper presents the Force Cancellation Motor (FCM), a microspeaker concept designed to reduce vibrations in ultra‑thin mobile devices without increasing module height. FCM integrates a compact actuator into the speaker’s center magnet, enabling active vibration suppression based on the known excitation signal while remaining independent of acoustic output. Prototype measurements show strong vibration reduction in perceptually relevant low‑frequency ranges and a 71% power‑efficiency improvement over a height‑optimized dual‑woofer reference. Additional benefits include optional activation of the compensation feature and the actuator’s use as a haptic driver, making FCM a promising solution for space‑constrained audio systems.

Speaker: Friedrich Reining (Sound Solutions Austria GmbH)

Full paper (PDF)

509 Modelling of a Magnetic Actuation of a Piano String

Active control enables the attribution of targeted temporal and spectral characteristics to vibrating structures. Applied to musical instruments, this technique allows for the production and the radiation of novel sounds throughout the interpreter's rendition, in contrast to the use of sound effects on recorded music.This study is part of a general project aiming to actively control a complex system composed of piano string triplets struck by the same hammer. For that purpose, a controller within a feedback loop is fed by the measured strings’ vibration and aims to calculate the force signal to be applied to the string triplets. Transducers composed of electromagnets are used as actuators and sensors.In this work, a model of magneto-mechanical interaction is established between the current intensity in the transducers and the vibration of a simplified model of a piano string. First, by referring to formulations derived from the Maxwell’s equations for magnetostatics, we model the magnetic field intensity created by the transducers placed near the magnetized string. Then, the modelling of the resulting force density distribution along the string is investigated using methods such as the integration of the local application of Maxwell’s stress tensor. These multi-physical models are expressed and connected using the Port-Hamiltonian Systems (PHS) formalism, which ensures the system’s passivity during simulation and control.

Speaker: Orlane Mammar (Laboratoire STMS)

510 Development of an acoustic nonlinear metasurface for noise mitigation

Low frequency noise mitigation can be hardly conducted using fully passive materials at the cost of very cumbersome and expensive set-ups. Many alternatives exist, among them, active noise control enables the use of energy to enhance noise mitigation of electro-mechanical devices. Using a model inversion technique, the first mode’s parameters of a loudspeaker can be tuned for efficient acoustic passive mitigation at low frequencies. However, linear behaviors offer limited performances, while nonlinear behaviors benefit from good performances over a large frequency bandwidth, in addition to targeted energy transfer and fast transients behaviors. The main drawback of nonlinear behaviors is their activation threshold at very high excitation amplitudes that can be circumvented using model-inversion pressure-based current-driven control techniques, enabling to create nonlinear behaviors at low excitation amplitudes. This study presents the development of a nonlinear metasurface composed of individual nonlinear electroacoustic resonators at low excitation amplitudes. Numerical analysis is conducted and an experimental bench composed of a waveguide and seven electroacoustic resonators is introduced and show excellent transmission loss performances.

Speaker: Emanuele De Bono (Università di Napoli Federico II)

511 Programmable Acoustic Reflection Through Spatial Modulation of an Electroacoustic Resonator Array

Active impedance control offers a versatile framework for shaping wave propagation by dynamically adjusting generalized boundary conditions through feedback. Electroacoustic resonators are an attractive platform for this purpose, as their target acoustic impedance can be individually programmed through a pressure-based, current-driven control law. This reconfigurability makes them potential candidates for implementing spatially varying boundary conditions without mechanical modification. This work investigates the effect of spatially modulating the control law parameters of an electroacoustic resonator array on the reflected acoustic field. A periodic modulation is applied to the target stiffness along the metasurface, and the resulting scattered field is analyzed using a Bloch–Floquet expansion that predicts the wavenumbers of the generated spatial harmonics. Finite element simulations confirm that the modulation generates harmonics whose propagation angles are governed by the modulation wavenumber, enabling controlled beam splitting away from the specular direction. The results demonstrate programmable reflection control through active spatial modulation of the surface impedance, providing a path for the analysis and design of reconfigurable acoustic metasurfaces and active metaliners.

Speaker: Leonardo Ferreira (Université Marie et Louis Pasteur)

A03.03 Prediction, digitalisation (BIM) and simulation in building acoustics: S351

Conveners: Andrea Santoni (University of Ferrara), Edwin P.B. Reynders (KU Leuven, Department of Civil Engineering), Antonino Di Bella (Universyty of Padova - DII)

512 Acoustic performance of raw earth construction systems: Measurement and prediction

The building sector is confronted with two key challenges: minimizing the environmental impacts while ensuring occupants well-being. In this context, the use of sustainable materials with high thermal and acoustical performances appears as a necessity. Therefore, raw earth appears as a strategic low-carbon material, as it is naturally abundant, locally available and renewable. It also offers interesting thermal properties to improve summer comfort, thus impacting the energy efficiency and comfort of buildings.The main objective of the CarAc’Terre project is to remove regulatory and normative obstacles to raw earth construction techniques relative to acoustics. A multi-scale methodology for characterizing the acoustic performance of raw earth construction systems is proposed using both a performance-based and a perceptual approach. Measurements are carried out at the material, wall as well as building scale, such that contributions from materials formulation, walls composition, system type, as well as construction implementation can be investigated. This paper describes sound insulation results obtained on different types of raw earth based walls and floors. Comparison between predicted and measured acoustic performance, in terms of sound insulation (and impact sound level for floor systems), is also presented and discussed in detail. Based on this analysis, guidance for a quick evaluation of the acoustic performance to system not tested in laboratory is briefly introduced.

Speaker: Thibaut Blinet (CSTB)

Full paper (PDF)

513 Prediction of flanking sound transmission through periodic cross-laminated timber junctions with Bloch-Floquet analysis

Lightweight building materials such as cross-laminated timber (CLT) are highly susceptible to flanking sound, where vibrational energy is transmitted between two building components across a common junction. Resilient strips or pads can be added at the junction to decouple the components and, as a result, reduce flanking sound transmission. While effective, the performance of these solutions is difficult to predict. Typically, flanking sound transmission is analyzed with models based on plate or shell theory, but these are limited to simple homogeneous junctions with thin panels. While full-scale finite element analysis is possible, its computational cost at high frequencies is prohibitive as very fine element meshes are required. In the current study, an approach is presented which exploits the spatial periodicity often exhibited by CLT junctions. This allows the use of Bloch-Floquet analysis of the junction itself and the connected building components. The junction is treated as an elastic solid with an arbitrary level of complexity and spatial periodicity in the junction direction. Statistical Energy Analysis (SEA) is employed to compute the vibration reduction indices of the junction. The prediction model is applied to varying CLT junctions with rigid connections, continuous strips or discrete pads.

Speaker: Stijn Moons (KU Leuven, Department of Civil Engineering)

514 Digitalization of Acoustic Requirements in Building Design: Developing a bSDD within an Open BIM Framework

The digitalization of the construction sector is a major driver of innovation in Europe, as shown by the increasing adoption of Building Information Modeling (BIM). BIM implementation advances at different speeds: countries such as Denmark and Finland have long established requirements, while others, including Italy, are still consolidating their processes. In Italy, Ministerial Decree 560/2017 and the new Public Contracts Code (Legislative Decree 36/2023) make BIM mandatory from January 1, 2025 for public works exceeding 2 million euros. The objective is to improve transparency, quality, and control over public works while reducing inefficiencies and managerial risks. The private sector increasingly views BIM as a competitive asset enhancing accuracy, coordination, and overall project reliability. Within this framework, the Open BIM approach, based on open and non-proprietary standards, is essential for effective interoperability among stakeholders. The IFC (Industry Foundation Classes) format remains the international reference for structured and cross-disciplinary information exchange, ensuring continuity across the building lifecycle. This article examines the role of the bSDD (buildingSMART Data Dictionary), a tool designed to standardize concepts, properties, and classifications within information models. In building acoustics, a dedicated bSDD could offer significant benefits by harmonizing national and international standards, integrating performance requirements directly into IFC models, and supporting automated verification processes. Standardizing acoustic parameters such as sound insulation, sound pressure levels, and flanking transmission would enhance data consistency, regulatory transparency, and communication among designers, verifiers, and contractors. Integrating BIM methodologies and classification tools like the bSDD supports the development of an interoperable information ecosystem aligned with evolving European regulatory needs.

Speaker: Costantino Carlo Mastino (University of Cagliari - DICAAR)

Full paper (PDF)

515 Reducing Vibration Transmission Through Thin Plate Junctions Using Resonant Metamaterial Strips

Reducing vibration transmission through plate junctions is a challenging problem for limiting flanking sound transmission. Conventional approaches rely on structural decoupling or addition of mass, often involving increased structural complexity, material use, and limited low frequency effectiveness. This work explores an alternative, lightweight solution using resonant metamaterials, implemented as a strip of subwavelength resonators along the junction line. Such resonators can introduce bandgaps that inhibit free traveling waves within targeted frequency ranges. To predict vibration transmission through plate junctions treated with a resonant metamaterial strip, a two-step prediction model is proposed. First, wave propagation through junctions between semi-infinite thin plates connected at arbitrary angles is described using a wave-based analytical model. Within this model, multiple wave types are considered, and the metamaterial treatment is included via homogenized effective properties. This allows the computation of frequency dependent reflection and transmission coefficients, accounting for wave conversion between different wave types. Second, these results are used to compute coupling loss factors and are incorporated in a statistical energy analysis model to predict vibrational energy exchange between the connected plates. The proposed framework is experimentally validated on a junction of two aluminum plates connected at different angles, both with and without resonant metamaterial strips. Different strip widths are considered. Predicted velocity level differences agree well with those obtained from the measurements, showing reduced vibration transmission near the local resonance frequency.

Speaker: Sofie Becuwe (KU Leuven, Department of Civil Engineering)

Full paper (PDF)

A05.03 Urban sound planning / The quiet city for life quality: S368

Conveners: Robert Arcos (Universitat Politècnica de Catalunya), Laura Estévez-Mauriz (Universidad de León), Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology), Trond Maag

516 Urban morphology and greenery as contextual modifiers of traffic noise annoyance: A review and synthesis of the current literature

Traffic noise annoyance is influenced not only by sound exposure levels but also by the spatial characteristics of the urban environment. In urban planning and design, morphological and greenery parameters are commonly reduced to their sound propagation effects, thereby omitting their non-acoustic role in noise annoyance. While numerous studies report effects of the built environment and urban landscape on traffic noise annoyance, a structured synthesis of recent evidence is lacking.This study presents the results of a systematic literature review on the effects of urban morphology and urban greenery on road, rail, and aircraft noise annoyance. Following the PRISMA framework, 59 studies were analyzed, covering morphological characteristics such as façade and window orientation, street width, and access to a quiet side, as well as indicators of urban greenery, including NDVI and view of green.Several morphological characteristics showed consistent associations with reduced noise annoyance, most notably the presence of a quiet side and bedroom or living room orientation. A view of the road was, in some cases, associated with lower annoyance, suggesting a potential role of source visibility. Indicators of urban greenery were generally associated with reduced annoyance, although effect direction and magnitude varied by noise source and level, with less consistent or even reversed effects observed for aircraft noise and at high road traffic noise levels.Overall, the findings demonstrate that urban morphology and greenery act as meaningful contextual modifiers of traffic noise annoyance and calls for their joint, rather than independent, consideration in noise-sensitive urban planning and design.

Speaker: Gustaf Wuite (TU Delft)

517 Towards a Framework for Evaluating and Integrating Quiet Areas into Urban Planning

Existing EU approaches for quiet area identification often rely on noise level assessment, giving less weight to in situ measurements and subjective data. Recent research, alongside EEA and ETC/ATNI reports, recognises quiet areas through acoustic conditions, spatial qualities, visual attributes and cultural context. While emerging studies acknowledge quiet areas as multidimensional constructs, they do not provide a structured methodology that links spatial metrics, acoustic indicators and user experience. This paper proposes a mixed-methods framework that integrates objective and subjective data for quiet area evaluation and their integration into urban planning. Combining soundwalk and space syntax methods with natural and cultural heritage evaluation will enable a multi-layered interpretation that considers accessibility, land-use compatibility and heritage value. Relating soundscape perception to inherent spatial configuration and built structure, will reveal areas where physical form and sensory experience converge to achieve conditions for urban quietness and pleasantness. Particular attention is given to layers of urban forests and cultural heritage whose characteristics shape spatial networks and experiential characteristics of the city. The proposed methodology aims to support a more comprehensive understanding of quiet areas as socio-spatial constructs and contributes to the development of context-sensitive urban planning strategies to enhance overall quality of life.

Speaker: Klara Kranjčec (Faculty of Architecture, University of Zagreb)

Full paper (PDF)

518 Land-Use-Based Modelling as a Complement to Strategic Noise Maps: Opportunities and Limitations

Strategic noise maps (SNMs) represent one of the few available data sources on acoustical information at high-spatial resolution. However, SNMs provide traffic noise predictions, not accounting for other acoustic properties. Recently, studies demonstrated promising results in modelling additional acoustic properties based on land-use (LU) data.Here, we assess the predictive performance of a LU-based model trained and externally validated to the Ruhr region, Germany. The model generates high-spatial predictions for five acoustic properties: Total Noise, Intelligibility, Sharpness, Biophony, and Acoustic Dominance. Acceptable predictive performance on external test data was achieved only for Total Noise and Acoustic Dominance.Building upon these results, we compare LU-based Total Noise predictions to Traffic Noise predictions from SNMs. Predictions were made for 737 participants from a population-based cohort-study. LU-based predictions were on average 6.8 dB higher than SNMs estimates (range: 49.8–69.7 vs. 35.2–72.2 dB resp.), and the two measures correlated only moderately (Pearson's r = 0.64). This suggests that Total Noise is only partially related to Traffic Noise, likely because LU-based models capture a broader composition of sound sources. Overall, these results indicate that LU-based models may represent a complementary approach to conventional SNMs, but currently only for a limited set of acoustic properties.

Speaker: Timo Haselhoff (Insitute for Urban Public Health)

519 BaLSaM – Audio‑visual simulation of mesoscale urban design for sound‑informed planning in the Rhenish mining region

Traffic noise is a major environmental health burden in European cities, yet it is often addressed at a late stage in urban planning processes, mainly through legally required assessments that rely on technical indicators and do not directly incorporate human perception. The BaLSaM project developed and tested an audio visual simulation that links traffic data, auralization and urban design to support sound informed decisions at the mesoscale of neighborhoods. In collaboration with four municipalities in Germany’s Rhenish mining region, we combined parameterised urban design variants (building height and depth, block porosity, street layout) with noise simulations and listening tests in Virtual Reality (VR). This setup can be applied at the earliest stages of urban design, turning acoustic conditions and perception into explicit design parameters and allowing non experts to experience and evaluate possible acoustic futures.The contribution presents: (1) the audio visual VR framework, (2) key findings on how mesoscale urban morphology affects exposure and perceived acoustic quality; and (3) lessons learned from applying the tool in planning processes. We argue that combining noise indicators with perception based assessment provides a robust interface between acoustic science, urban design and municipal governance, and can support a shift from reactive noise control to proactive sound informed urban design.

Speaker: Moritz Lippold (RWTH Aachen University Chair of Urban Design)

Full paper (PDF)

520 Identifying Emerging Environmental Noise Challenges and Future Priorities in Ireland: Insights from the First National Noise Summit

Technological advancements are reshaping environmental noise environments, with emerging sources such as electric vehicles and drones increasingly present in urban areas. In this context, the EPA-funded Noise 2050 project in Ireland explores future noise environments towards 2050 and how to prepare for emerging challenges. To support this objective, Ireland’s first National Noise Summit was organised, bringing together stakeholders from government, research, consultancy, infrastructure and industry. In advance of the event, participants completed a structured questionnaire addressing current noise challenges, key barriers, priority emerging sources, and future challenges. Transport noise remains the dominant concern, while emerging technologies and energy-related sources are expected to increase future noise complexity. Key barriers relate mainly to governance and implementation capacity. Findings suggest that future noise challenges are shaped by both source diversity and policy constraints. The survey informed summit discussions and provides insights into future environmental noise challenges in Ireland and informs the strategic direction of the Noise 2050 project.

Speaker: Priscila Wunderlich Villar (The Galway Sound Lab, School of Engineering,University of Galway)

Full paper (PDF)

A08.00 Industrial, Machinery, Equipment Noise and Vibration: S465

Conveners: Ercan Altinsoy (Professur für Akustik und Haptik), Serkan Atamer (Dresden University of Technology), Mehmet Sait Özer (Dresden University of Technology)

521 Hidden in Plain Sight: Characterization of Airborne Ultrasound Sources

The increasing spread of ultrasound technology is accompanied by a spread of the associated airborne ultrasound exposure. Current risk assessments and exposure limit recommendations are based on outdated measurements and are primarily aimed at occupational sound exposure. However, technological developments have led to many new high-frequency noise and ultrasound sources, especially for use in private and public areas. Unfortunately, only a few studies have been carried out to date on public exposure to airborne ultrasound. Thus, the existing body of knowledge on potential hazards posed by ultrasound noise in private and public spaces is limited. Against this background, the Physikalisch-Technische Bundesanstalt (PTB) has been commissioned by the German Federal Office for Radiation Protection (BfS) to investigate the spread of airborne ultrasound in private and public environments. The aim of this study is to identify and assess possible risks to the general population arising from airborne ultrasound exposure. In addition to an extensive literature research, a series of measurements was conducted both in the field and in laboratory settings to identify and describe airborne ultrasound sources. For qualitative and quantitative characterisation, a range of measurement instruments was employed, such as acoustic cameras, an ultrasonic sound exposure meter and the scanning and calibration facilities of PTB. Selected examples from these investigations are presented.

Speaker: Christoph Kling (Physikalisch-Technische Bundesanstalt)

Full paper (PDF)

522 Assessment and Mitigation of Industrial Environmental Noise in a Mixed Residential–Industrial Urban Area

This paper assesses environmental noise generated by industrial sources at a production facilites located in a mixed residential–industrial urban area. The objectives were to identify the dominant noise sources, evaluate their contribution at the most exposed residential receivers, and propose mitigation measures capable of reducing the noise impact to acceptable levels. The study combined field measurements of total and residual noise, near-source sound pressure level measurements, and technical data from equipment specifications. An acoustic propagation model was developed and calibrated using measured data. The dominant sources were identified as an elevated exhaust fan, a compressor unit near the steam exhaust, a compressor room with ventilation openings, a mill unit, and cooling-related systems. Nighttime measurements showed an assessed noise level of 50 dB(A) and a residual noise level of 40 dB(A), indicating the need for further reduction of industrial noise emissions. Based on the analysis, a phased mitigation strategy was proposed, including acoustic louvers, absorbent treatment of ducts and openings, vibration isolation of motors, and improved sound insulation of façade openings. Model results indicate that the proposed measures can reduce the contribution of the dominant sources so that the residual noise level at critical residential locations would not increase by more than 1 dB(A). The findings confirm the technical feasibility of effective industrial noise control in densely built urban environments.

Speaker: Antonio Petošić (University of Zagreb)

523 Acoustic assessment and improvement of wooden enclosures used in modern kitchens for built-in household appliances

In modern and stylish kitchens, most of the time, a so-called built-in household appliance is enclosed in wooden furniture to enhance aesthetic attribute and elegance. These housings are definitely impacting the sound and vibration of the household appliance itself, to the extent that was not studied before. Like any other enclosure for a sound source the impact could be used for shaping of sound pressure level and perceived sound quality. At the same time, the sound emission of household appliances was improved significantly by manufacturers over last years, reaching noise floor levels, notably for refrigerators and dishwashers. Authors of this paper decided to study the acoustics of wooden enclosures, aiming at understanding their strengths and weaknesses in terms of contribution to overall appliance sound output. The experiments consist of placing omnidirectional speakers inside the enclosures and testing, with selected microphones, the amount of acoustic energy that traveled outside. At the end, some concepts with typical and novel materials were tested attempting an improvement of overall sound insertion loss.

Speaker: Ercan Altinsoy (Professur für Akustik und Haptik)

Full paper (PDF)

524 Sensitivity Study of Measurement Procedure and Parameters in Experimental Modal Analysis of Ceramic Components

This work presents a comprehensive sensitivity study on the influence of measurement procedures and parameters in the experimental modal analysis (EMA) of ceramic blocks, with a focus on optimizing the identification of resonance frequencies and damping ratios for ceramic components used in power products applications. The study systematically investigates how variations in sample mounting, actuator placement, mesh density, and measurement settings affect the accuracy and reliability of modal parameters.A series of preliminary tests were conducted to evaluate the impact of actuator orientation and distance from the sample edge, revealing that while global actuator position has negligible effect on resonance frequencies, proximity to the edge significantly enhances dynamic response. The measurement methodology was further refined by analyzing the effects of mesh density on frequency and mode shape identification.Across a range of sample thicknesses, this study identifies trends in resonance frequency shifts and damping values, providing insights into the relationship between geometric parameters and modal behavior. The findings underscore the critical role of measurement procedure optimization in achieving accurate EMA results and offer guidance for future experimental campaigns and numerical model development.

Speaker: Michal Kozupa (Hitachi Energy Research)

Full paper (PDF)

A12.08/A16.12 Numerical Methods for Room Acoustics: S504

Conveners: Maarten Hornikx (Eindhoven University of Technology), Albert Prinn (International Audio Laboratories Erlangen)

525 Experimental measurements of the diffusion coefficient for the diffusion equation of the Schoenenbourg fort tunnels

The diffusion equation model has been used for room acoustics simulations because of its computational efficiency. Traditionally, the diffusion coefficient of the diffusion equation has been treated as a constant depending on the room's volume and total surface area. However, numerical investigations have demonstrated that in elongated spaces, this assumption does not hold: the diffusion coefficient varies with spatial location, source position, and absorption coefficient of the boundaries. Despite these theoretical insights, experimental validation of spatial dependencies is lacking. This study addresses this gap by presenting experimental measurements of the diffusion coefficient within one of the tunnels of the Schoenenbourg fort, a structure belonging to the historic Maginot line in France. The methodology relies on intensity and energy density measurements taken along the tunnel, allowing the diffusion coefficient to be derived directly from its definition given by Fick's law. The measurement results demonstrate that the diffusion coefficient cannot be regarded as a constant in elongated spaces but instead depends on spatial conditions. In addition, a preliminary validation was performed by incorporating the experimentally obtained diffusion coefficient into the diffusion equation model. The comparison revealed that the model, when used with the measured diffusion coefficient, successfully reproduces the tunnel sound pressure level and reverberation time.

Speaker: Ilaria Fichera (Eindhoven University of Technology)

Full paper (PDF)

526 Characterisation of Sweeping Echo Intensity Patterns in Shoebox-shaped Rooms: A Computational Study

Sweeping echoes are an acoustical phenomenon present in impulse responses of shoebox-shaped rooms generated with either geometrical or wave-based room acoustics modelling software. They are also present in audio recordings convolved with the responses, introducing a distinctive change in pitch that contradicts the usual room reverberance characteristics. Based on an analytical expression derived for the rigid walls image source solution of the wave equation, it has recently been hypothesised, and supported with a modest amount of simulation data, that room setups with a higher degree of regularity lead to stronger sweeping echoes. The main objective of this study was to further test this hypothesis in shoebox-shaped rooms of various levels of regularity in room dimensions, for a number of source-receiver positions. To this end, sweeping echoes present in impulse responses obtained with the image source method were evaluated using the sweeping echo intensity (SEI) measure. The results, based on more than 140 million of impulse responses, the corresponding SEI values, and an informal non-blind listening test, indicate that the sweeping echoes are present and audible at various source-receiver positions not only in the shoebox-shaped rooms with regular parameters but in general, i.e. irregular shoebox geometries, though to a lesser extent.

Speaker: Andrea Andrijasevic (University of Applied Sciences Rijeka)

Full paper (PDF)

527 Room Ratio Optimisation for Listening Rooms with Non-Rigid Boundaries

The optimisation of room dimensions for listening environments has been the subject of several studies in the field of room acoustics. Existing approaches, however, have assumed rigid boundaries and therefore dismissed the impact of non-rigid partitions, such as drywall, on modal behaviour. To investigate the implications of this effect, simulations were performed using the finite element method (FEM), which considers complex wall impedances as boundary conditions, to enable a more realistic analysis. Room ratios for small, medium, and large volumes were examined to determine optimal room dimensions by analysing room configurations, evaluating the uniformity of the resulting frequency response and illustrating the results using two-dimensional maps. These maps allow robust room ratios to be identified in which particularly uniform frequency responses are achieved. Comparing these results with existing models shows that non-rigid boundary conditions alter the structure of these patterns, leading to different optimal room ratios. This suggests that there are no universally optimal spatial configurations, but rather that they depend on the specific assembly of the room boundaries.

Speaker: Paul Simon (Audio Communication Group, Technische Universität Berlin)

528 Physics-Informed Neural Networks for Room-Acoustic Simulations within Bayesian Framework

Accurate prediction of sound energy decay in enclosed spaces is essential for acoustic design in performance halls, classrooms, and recording studios. Statistical room-acoustic theory, such as the Sabine reverberation formula, provides fast but spatially limited estimates, while wave-theoretical simulations are physically accurate but computationally expensive. This work presents a Physics-Informed Neural Network (PINN) trained to solve the acoustic diffusion equation, using finite-difference time-domain (FDTD) simulations for supervised learning. The network takes spatial position and time as inputs and predicts sound energy density without requiring re-simulation. The maximum likelihood enforces the governing partial differential equation, boundary conditions, and initial conditions alongside supervised data. Preliminary results show close agreement with FDTD ground truth across the full decay range, accurately reproducing the temporal decay process at the receiver positions. Full temporal coverage of the decay process provides stronger training than spatially dense but temporally sparse sampling. The current work also explores Bayesian probabilistic evaluations to quantify uncertainties during training with the goal of generalizing across room geometries and absorption coefficients without retraining.

Speaker: Ning Xiang (Rensselaer Polytechnic Institute)

Full paper (PDF)

A14.02 Pitch perception impairment and enhancement with hearing devices: S523

Conveners: Etienne Gaudrain (CRNL, CNRS UMR5292, Inserm U1028, Université Lyon 1), Olivier Macherey (LMA-CNRS)

529 Voice pitch contours in emotional prosody and cochlear implants

Hearing and identifying emotional prosody (the tone and manner of speaking that conveys emotions) is a key element of spoken communication. Voice pitch inflections are a dominant cue for emotional prosody, but sensitivity to voice pitch is reduced in cochlear implant patients relative to normally hearing peers, likely accounting for their impaired performance in emotion identification. In this presentation, I will review the literature on emotional prosody identification by adults and children with cochlear implants, and then focus on our recent findings on how cochlear implant users with different auditory experiences utilize voice pitch contours, duration cues, and intensity cues for emotion perception. I will also describe our findings on the extent to which their use of these cues accounts for individual variability in their identification of emotions in speech.Key findings I will discuss are summarized as follows: 1) the utilization of both voice pitch contour cues and duration cues accounts for emotional prosody perception by CI users; 2) prelingually deaf, young CI users and postlingually deaf, middle-aged and older CI users show some differences in cue-utilization; 3) in young prelingually deaf CI users, duration of device experience is positively associated with voice pitch contour cue utilization, while earlier age of implantation plays a small but significant role. Finally, I will discuss implications of these findings for CI users’ daily social exchanges.

Speaker: Monita Chatterjee (Northwestern University)

530 Perception of Intentional Prosody in Sensorineural Hearing Loss

Social interactions require interpreting intentions and mental states, often conveyed via speech prosody. This study investigated the impact of sensorineural hearing loss and hearing aids on the perception of paralinguistic prosody conveying intentions. The performance and the confidence of adults with typical hearing and of experienced hearing aid users were evaluated in two classification tasks: 1) distinguishing between sincerity, sarcasm, teasing, or telling a white lie in audio and audio-video recordings of short dyadic exchanges, and 2) distinguishing whether audio recordings of grammatical questions convey genuine questions or indirect requests. Hearing aid users were tested unaided and aided. In both tasks, linear mixed-effects modeling revealed that hearing loss decreased confidence, and that hearing aid use increased it. By contrast, hearing loss did not decrease classification performance, at least not when accounting for age. Hearing aids improved performance only when participants distinguished between sincerity, sarcasm, teasing, and white lies. Visual cues benefited both performance and confidence, and the benefit on confidence was strongest for participants with hearing loss unaided. These findings suggest that changes in confidence should be viewed as an important challenge of hearing loss and should be considered in rehabilitation strategies.

Speaker: Frederic Marmel (ORCA Labs)

Full paper (PDF)

531 Vocal expression of emotion in children with hearing loss: Insights from three analysis methods

Emotional prosody plays an role in social communication, yet little is known about how children with hearing loss express emotions through speech. This research combines findings from three studies investigating emotional prosodic expression in children with cochlear implants, children with hearing aids, and children with normal hearing.Across the studies, speech recordings of Dutch-speaking children aged 7–18 years were collected during tasks in which they produced emotional utterances expressing happiness, sadness, and anger. Emotional expression was examined using three approaches: perceptual evaluation by normal-hearing listeners, acoustic analysis of prosodic features, and machine-learning–based speech emotion recognition (SER).Perceptual evaluations showed that both children and adult listeners were generally able to recognize emotions expressed by children with hearing loss nearly as accurately as those expressed by children with normal hearing. However, happiness was identified less accurately in children with cochlear implants and was more often confused with sadness in both cochlear implant and hearing aid users. Acoustic analyses demonstrated that the prosodic patterns of children with hearing loss were comparable to those of normal-hearing peers, although some reduced contrasts between emotions were observed. The machine-learning–based SER system classified emotional speech produced by children with normal hearing more accurately than that of children with hearing loss, particularly for happiness.Together, these findings suggest that children with hearing loss retain the ability to express emotions vocally, with only subtle differences compared to their normal-hearing peers. These results highlight the resilience of children with hearing loss in developing emotional communication skills, while also pointing to areas that may benefit from research and targeted support.

Speaker: Jantien Vroegop (Erasmus MC)

A14.08 Computational and AI approaches in audiology: S332

Conveners: Mark Saddler (Technical University of Denmark), Volker Hohmann (Carl von Ossietzky Universität Oldenburg)

532 Mapping the Hearing Loss Landscape: A Data-Driven Phenotyping Framework Beyond Pure-Tone Audiometry

The aetiology of hearing impairment is multifactorial, involving numerous structures within the auditory system. Although pure-tone audiometry remains the clinical gold standard for hearing device fitting, it lacks sensitivity to certain pathologies, including hidden hearing damage. Previous approaches to phenotype classification either relied solely on audiogram data or incorporated additional tests but were limited by small sample sizes. Here, we analyse a large clinical dataset from the Rigshospitalet University Hospital (Denmark) spanning 1995 to 2022, comprising ~300,000 audiograms from over 85,000 unique adult patients, including audiometric thresholds and speech audiometry results.We investigated Principal Component Analysis (PCA) and Uniform Manifold Approximation and Projection (UMAP) for two-dimensional visualization. While PCA captured major audiogram variance, it proved insufficient for multimodal audiological data. UMAP demonstrated greater utility by exploiting both global and local data structures, and was adopted as the basis for our Hearing Loss Map.Comparison with existing phenotype classifications revealed that up to 40% of the data remains unclassified, underscoring the need for a more comprehensive framework. We applied multiple clustering approaches to a sensorineural hearing loss (SNHL) subset — Gaussian Mixture Models and Leiden community detection — identifying novel phenotypes not captured by conventional schemes. A trajectory analysis further represents patients transitioning between phenotypic states over time, enabling longitudinal characterisation of hearing loss progression.We propose this UMAP-based Hearing Loss Map as a powerful tool for navigating the SNHL patient space, with applications in clinical decision-making, patient counselling, evaluation of novel treatments, and development of next-generation hearing technologies.

Speaker: Gerard Encina-Llamas (University of Vic - Central University of Catalonia)

533 Towards Portable Inclusive Automated Hearing Assessments for All: the ATHENA Project

Two of the main challenges that the hearing care workforce currently face concern the limited resources and the increasing number of older adults who are in need of hearing screening. New technology tools, such as automatic speech recognition (ASR), have the potential to revolutionise the field. However, state-of-the-art ASR systems are predominantly trained on speech from healthy young adult native speakers. The ATHENA (Automating hearing assessments for all) project proposes the use of open-source ASR systems to automatically score four traditional speech audiometry tests: speech in quiet, speech in noise, speech in speech masker, and vocal emotion recognition.As a first step, four open-source ASR systems in Dutch (Kaldi NL, Whisper, Voxtral and NeMo) are evaluated when decoding participants’ spoken responses per speech audiometry test, focusing on the impact that decoding errors have on the test results. Three participant groups are included: native Dutch-speaking children, non-native Dutch-speaking adults, and native Dutch-speaking adults. As a second step, the ASR performance is tested with a wider group of non-native and native Dutch-speaking participants, including older adults of varying cognitive statuses, and hearing-impaired individuals. As a third step, we look into ASR customisation, including limiting the ASR lexicon per test (to words related to or included in the test speech material), alongside fine-tuning and data augmentation techniques. As a final step, the best-performing customised ASR systems coupled to the speech audiometry tests are implemented on engaging interfaces (such as socially assistive robots), and evaluated with respect to system robustness, reliability, and accessibility for clinical and non-clinical use.

Speaker: Gloria Araiza-Illan (University Medical Center Groningen)

534 Remote Cochlear Implant Fitting in Real-Life Listening Situations

Objectives: This study investigated the feasibility, acceptability, and perceived benefits of conducting cochlear implant (CI) remote fitting in real-life listening situations.Methods: Twelve post-lingually deafened CI users participated in four sessions: a baseline in-clinic visit, two remote fitting sessions (at home and in a real-life setting), and a final in-clinic evaluation. Participants completed a Client Oriented Scale of Improvement (COSI) survey and a custom post-study survey.Results: Most participants reported high satisfaction with remote fitting. Participants frequently cited reduced travel time and improved satisfaction in specific listening situations as benefits, while technical issues and reduced non-verbal communication were noted as drawbacks. All participants indicated they would recommend remote fitting to other CI users.Conclusion: Remote fitting in real-world settings is both feasible and well accepted by CI users, particularly those with relatively good aided speech intelligibility. It adds value by enabling fitting and subjective assessment in listening situations that, according to their COSI, are important to CI users.

Speaker: Jan-Willem Wasmann (Radboudumc)

535 Individualized, App-Based Phoneme Training for Adult CI Users at Home

ObjectivesThis study introduces a novel approach aimed at improving phoneme confusion errors among adult CI users through individualized phoneme training via a mobile app.MethodsTwenty-five experienced adult CI users with a post-lingual onset of severe-to-profound hearing loss were invited to participate in a four-week phoneme training program using a mobile app. Participants were instructed to train for 20 minutes daily. The training materials were tailored to each participant's unique phoneme confusion errors, as determined by the outcomes of a phoneme test in a quiet environment.ResultsTwenty-one participants completed the required training. On a group level, participants significantly improved (>10%) on their overall targeted errors. On an individual level, both for trained phonemes and carry-over to word perception, training had variable results. All participants reported having enjoyed the training and would make use of the app in the future. ConclusionsOur innovative approach to individualized bottom-up phoneme training may enhance foundational speech perception at the phoneme level among adults making use of a CI. We anticipate that these improvements may contribute to improved speech understanding and communication outcomes. This app-based, self-paced training approach allows for greater autonomy and empowerment for these individuals.

Speaker: Jan-Willem Wasmann (Radboudumc)

A15.02/A24.05 Hearing Research in Virtual Environments: S453

Conveners: Janina Fels (IHTA, RWTH Aachen University), Carolin Breuer (IHTA, RWTH Aachen University), Nils Peters (Trinity College, The University of Dublin)

536 Exploring the Impact of Audiovisual Noise on Listening Effort and Presence

Traditional auditory attention research often relies on controlled but artificial laboratory settings. Virtual reality offers a more realistic alternative, but added visual information may influence not only performance but also the subjective listening experience. This study examined how visualization of target and distracting sound sources in a virtual classroom affects perceived listening effort and presence.Participants performed an auditory selective attention task in four environments of increasing visual complexity: an empty room, a classroom with animated distractors, a classroom with animated target speakers, and a setting including visible target and distractor speakers. After each condition, participants rated listening effort and presence. These ratings were related to results from the attention task. Preliminary analyses suggest only subtle effects. Increased audiovisual noise showed trends toward higher presence, while visible target speakers may have slightly reduced listening effort. These patterns are consistent with earlier findings that informative visual cues can support auditory attention.The results highlight the importance of combining behavioral and subjective measures when studying listening in immersive audiovisual environments.

Speaker: Carolin Breuer (IHTA, RWTH Aachen University)

537 Listening to Conversations in Close-to-Real-Life Settings: A Virtual Reality Study of Noise and Turn-Taking Effects on Content and Source Memory

Listening attentively and remembering spoken information are essential for many everyday life communication scenarios. We rely on these cognitive abilities when engaging in conversations, attending lectures, or following political debates. Prior research has examined how acoustic and audiovisual factors influence listeners’ ability to retain heard information (content memory), but far less is known about their impact on content-to-talker attribution (source memory). This Virtual Reality study investigated the effects of steady-state background noise and gestural turn-taking cues on listeners’ content and source memory in conversations focusing on audiovisual aspects. Participants were immersed in a simplified virtual laboratory environment in which they listened to dialogues between two embodied conversational agents (ECAs), one male and one female, positioned at ±60° azimuth relative to the listener. After each dialogue, participants answered content- and source-memory questions. In a within-subject design, dialogues were presented either in silence or in white noise at –3 dB SNR, and either with or without gestural turn-taking cues. Regarding turn-taking gestures, we explored two competing possibilities: whether they improve memory by facilitating audiovisual integration or impair it due to additional processing demands. Background noise was expected to impair both content and source memory, with potential mitigation of these effects in the presence of turn-taking cues. Results of this study will be presented and discussed with regards to their practical implications for real-life listening situations as well as their theoretical relevance in auditory cognition research.

Speaker: Isabel S. Schiller (Work and Engineering Psychology, RWTH Aachen University)

538 How Well Do Avatars Speak? Framework for Avatar Facial Animation Control

Hearing research has studied the acoustic basis of communication extensively. However, speech understanding is not purely an acoustic process: visual information from the speaker’s face, especially lip movements, plays a key complementary role. These cues become particularly valuable in challenging or noisy listening conditions. As virtual environments expand in both research and everyday communication, digital avatars are increasingly used not only as a medium for interaction but also as a tool for investigating human communication processes. Recent studies suggest that speech intelligibility in such contexts strongly depends on the animation methods used to drive avatars, yet remains lower than with natural video recordings. One major reason is that the visual speech cues produced by avatars often lack the precision required to reliably discriminate certain phonetic elements.This work presents a pipeline for the animation and control of realistic avatars in virtual environments. The proposed approach relies on video recordings as reference material for the production of facial animations, rather than on direct or fully automated methods, allowing for a more precise treatment of complex visemes. The pipeline is designed in a modular way to integrate within our virtual environment. Within this context, particular attention is given to issues of audio-visual alignment, especially the temporal relationship between speech and lip movements. Finally, we offer initial insights into the potential benefits of such animations for speech intelligibility.

Speaker: Antoine Bourachot (Technische Universität München)

Full paper (PDF)

539 Spatial and Immersive Audio as a modulator of therapeutic mechanisms - a literature analysis

This paper explores the therapeutic potential of Spatial Audio and Immersive Audio in the context of mental health interventions. While a broad evidence base shows that audio-based approaches, including music listening, music therapy, and other sound-based interventions, can reduce pain, anxiety, stress, depressive symptoms, and fatigue, the additional contribution of spatial reproduction remains unclear. Using an integrative narrative approach, this review synthesizes evidence from media psychology, psychoacoustics, and clinical research on the relation between therapeutic mechanisms and the perceptual and physiological effects of Spatial Audio. Across audio-only, audiovisual, and interactive VR settings, spatial reproduction can improve localization and externalization, strengthen presence, immersion, attention, imagery, and emotional engagement, and elicit measurable physiological responses. Most findings are mechanism-oriented, and few clinical studies directly evaluate the effects of Spatial Audio on therapeutic outcomes. The available evidence indicates that Spatial Audio can intensify exposure-related activation, while immersive applications incorporating spatial sound can reduce tension, anxiety, and negative mood. Spatial Audio is therefore understood as a technological contributor to Immersive Audio and a modulator of therapeutic mechanisms rather than a stand-alone treatment. Its value depends on reproduction technology, sound design, content, audiovisual context, and the intervention goal. Future research should compare equivalent interventions and combine psychoacoustic, behavioral, physiological, and clinical outcome measures.

Speaker: Sylvia B. Grabe (Hochschule Mittweida)

Full paper (PDF)

A16.01 Design, Fabrication, Characterization and Perception of Innovative Acoustic Surfaces: S408

Conveners: Louena Shtrepi (Politecnico di Torino), Zackery Belanger (Umbel)

540 Inclined Wiremesh Gratings: a Psychoacoustic Assessment

Locally resonant metamaterials can be designed to achieve higher absorption compared to conventional porous materials. The sound absorption performance of both kinds of structures is strongly reduced when incident waves impinge them at angles different from normal incidence. Among other metamaterials alternatives, inclined wiremesh gratings exhibit broadband and omnidirectional absorption in rigid backing configuration. Meanwhile, in a configuration allowing for sound transmission, the absorption is reduced for incidence angles close to the orientation of the inclined layers. This behavior is in accordance with the impedance matching produced by the reciprocity property of the gratings. In addition to sound absorption performance, we assess sound quality of reflected-transmitted sound by means of psychoacoustic analysis. In particular, we investigate the dependence of loudness and tonality on the metamaterial configuration, and the relative location of the source and the receiver. The structure impulse response of the metamaterial is obtained using numerical computations. The results demonstrate a significant reduction in loudness at normal incidence or at incident angles opposite to the wiremesh orientation. Additionally, the tonality exhibits low variation even when narrow band noises are considered.The European Commission is gratefully acknowledged its support of the Marie Sklodowska Curie program through the Horizon Europe DN METAVISION project (GA 101072415). Views and opinions expressed are however those of the authors only and do not necessarily reflect those of the European Union, the European Union cannot be held responsible for them.

Speaker: Juan Pablo Escudero (LAUM, UMR 6613, IA-GS, CNRS, Le Mans Université)

541 Visual and Performance Impacts of AI-Generated Diffusion Panels

Diffusion is a key, but often underweighted characteristic of acoustic surface design in interior spaces. While absorption and reflection dominate practical acoustic treatment, diffusion governs the uniformity of sound propagation, particularly in larger auditoriums and multipurpose halls. Algorithmically generated diffusive geometries deliver calculable scattering performance but are often visually repetitive and aesthetically constrained, limiting their integration into architecturally expressive interiors. This study investigates whether image-based generative AI can produce visually diverse diffusion panels that remain acoustically viable compared with algorithmic baselines.Two-generation strategies are compared. The first is an algorithmic pipeline producing diffuser geometries tuned to a defined target frequency band. The second is an AI-driven pipeline that combines image generation with language-model-assisted prompting for aesthetic variation, followed by translation into surface geometry suitable for acoustic evaluation. Candidate panels from both pipelines are filtered using geometric heuristics related to depth variation, spatial distribution, and feature scale before acoustic analysis.Selected panels are evaluated through simulation at both panel and room scale, building on prior cross-validation work by the authors. The comparison aims to quantify the acoustic cost of aesthetic freedom and to identify conditions under which AI-generated diffusers remain within an acceptable margin of algorithmically generated optima. The work sits at the intersection of architectural acoustics, parametric design, and generative AI, extending a previously published generative pipeline by specializing it for acoustic diffusion and introducing quantitative acoustic validation as a selection criterion.

Speaker: Zlatka Madzharova (UACEG)

542 Objective and subjective investigation of the application of 3D-printed hybrid acoustic materials in Living Lab conditions

The 3D printed architecture together with the expanding capabilities of contemporary parametric design platforms has opened remarkable opportunities to complex geometric explorations and function integration. In this context, perceptual investigations result important to go beyond single parameter objective values and highlight the benefits of design solutions allowing for a more optimized use of complex surfaces. Therefore, this work investigates the application of 3D-printed hybrid acoustic materials (HAMs), designed and developed by BioG3D (Greece), which combine sound absorption and diffusion properties in a single office setting within the Living Lab of Politecnico di Torino (Italy). Both an objective and subjective investigation has been performed to compare 6 configurations obtained by varying HAMs location at different heights and their geometric distributions, while preserving the same number of elements. The different settings have been characterized by monaural, binaural and spatial impulse responses measurements using two different type of sources, i.e. an omnidirectional dodecahedron and a Talkbox. The objective room acoustic parameters (Reverberation Time, Clarity and STI) have been estimated and the aspects related to the perceived speech clarity improvements have been investigated comparing the different configurations through a subjective ABX test. The results clearly demonstrate the crucial role of design control variables and perceptual investigation in achieving optimized conditions.

Speaker: Louena Shtrepi (Politecnico di Torino)

Full paper (PDF)

A21.02 Road traffic noise and vibration: S152

Convener: Michael Cik

543 Amendment Proposal for Noise Emission Data of Motorcycles and Mopeds in CNOSSOS-EU

A proposal is made for revised sound source levels for motorcycles and mopeds in the CNOSSOS-EU traffic noise calculation model, which are considered too low. For motorcycles they are partly below those for cars. L-vehicles have comparatively high impact in terms of annoyance and health effects on affected residents, while comprising a small percentage of overall traffic.The proposed amendment includes the factors for propulsion noise for constant speed, for acceleration/deceleration conditions and for gradients. Vehicle tampering and driving behaviour are taken into account with an increased sound level and weighting for the percentage of vehicles in the fleet with tampering and/or high rev driving.The analysis is based on several measurement data sources and on vehicle noise modelling, taking driving cycles, engine speed and tampering into account.

Speaker: Michael Dittrich (TNO)

Full paper (PDF)

544 Impact of Noise Mitigation Solutions for L-Vehicles

In the EU project LENS, potential noise mitigation solutions for noise from L-vehicles were proposed. L-vehicles include motorcycles, mopeds, trikes, quads/ATVs and microcars. An impact analysis and cost-benefit analysis was made based on average Lden sound levels and on single events. Single events are the main cause of annoyance and complaints. In this paper, both methods are presented and the cost-effectiveness is ranked for the mitigation scenarios. These include improved type testing methods, reduced tampering, improved driving behaviour, access restrictions and fleet replacement. The results are indicative but suffice to rank the scenarios, showing that reduced vehicle tampering is highly effective, followed by access restrictions and improved driving behaviour.

Speaker: Michael Dittrich (TNO)

Full paper (PDF)

545 Noise reduction potential of electric passenger cars at urban intersections using measured vehicle trajectories

This contribution presents results of the NEẍUS project on noise from battery electric vehicles (BEV) under acceleration in urban settings. The potential of replacing internal combustion engine (ICE) passenger cars with BEVs to reduce noise levels for residents is investigated with scenario calculations using measured traffic data and emission models generated from pass-by measurements. Individual vehicle trajectories and driving states were extracted from aerial drone videos of urban intersections. To create the emission models, a measurement campaign was conducted with various driving profiles, which were enforced by means of real-time feed-back to the driver. Sound exposure levels of nine pairs of comparable BEV and ICE passenger cars were captured and used to derive spectral emission models. The model coefficients were determined with an iterative optimization algorithm, solving the inverse problem considering a variety of speeds and accelerations. Combining these emission models with real-world trajectories enables predicting noise exposure levels for virtual residents near the intersections for either BEV or ICE vehicle fleets. The BEV-only scenario reduces LAeq near investigated intersections by 1.5 dB on average, and up to 3 dB near stop lines at times with high traffic volume. At low frequencies, differences up to 10 dB were observed.

Speaker: Dominik Wöllstein (Empa)

Full paper (PDF)

546 Beyond Decibels: Traffic Noise Severity and Acoustic Complexity in Indian Cities with Heterogeneous Traffic Conditions

Road traffic noise is a major contributor to environmental noise pollution in Indian urban environments, where its severity is driven by highly heterogeneous traffic conditions. These include high traffic volumes, mixed vehicular composition, variable speed and acceleration patterns, frequent lane changes, and extensive honking. Such characteristics create complex, non-stationary acoustic environments that are not adequately represented by conventional traffic noise models.This study aims to quantify and demonstrate the severity and complexity of traffic noise in Indian cities. Field measurements were conducted in New Delhi across multiple street typologies and land-use categories. A total of 720 short-term noise recordings, each of 10-minute duration, were collected. These recordings were analysed using standard environmental noise descriptors along with psychoacoustic indicators.The dataset was further classified into 18 distinct noise environment categories, reflecting a high degree of acoustic variability and complexity. The analysis identifies honking as a dominant impulsive component, significantly elevating overall noise levels and contributing to increased perceptual disturbance.The findings highlight that traffic noise in Indian cities is both quantitatively high and qualitatively complex, underscoring the limitations of existing assessment frameworks. The study emphasises the need for context-specific noise prediction models tailored to heterogeneous traffic systems, along with targeted interventions to regulate honking behaviour. These insights are critical for developing effective noise management strategies in India and other rapidly urbanising regions with similar transportation dynamics.

Speaker: Manish Manohare (Transportation Research and Injury Prevention Centre, IIT Delhi)

A07.00 Flow Acoustics: S341

Conveners: Stefan Becker (LSTM, FAU Erlangen-Nürnberg), Roberto Camussi, Francesco Avallone (Politecnico di Torino), Stefan Schoder (IGTE, TU Graz)

547 On-board Acoustic Measurements on Drones: Self-Noise Characterisation and Detection of External Sound Sources

On-board microphones on drones enable external sound acquisition without the need for ground-based arrays; they can be positioned flexibly and quickly to suit individual requirements and allow for wide-area sound measurements. However, measurement performance is significantly influenced by high self-noise, as well as by flow conditions, microphone arrangement and wind protection. This paper presents a measurement setup featuring a reference acoustic source and several reference microphones on the ground. Various reference signals are used at different measurement distances to validate the measurement capability. Furthermore, the measurement configuration and equipment on the quadcopter are explained, including the mounting of the on-board microphones and the measures taken to protect against flow-induced effects. Finally, the data processing is described, the aim of which is to reduce the influence of self-noise on the acoustic measurement. On this basis, the measurement capability is evaluated across a broad frequency spectrum, taking into account the influence of self-noise. From this analysis, the possibilities and limitations of the applied methodologies are ultimately identified.

Speaker: Stefan Becker (LSTM, FAU Erlangen-Nürnberg)

Full paper (PDF)

548 Experimental Parametric Investigation of Aerodynamic and Aeroacoustic Performance of Small-size Commercial UAV Propellers: Comparison of Conventional and Toroidal Designs

The spreading use of small UAVs makes their noise emissions a compelling issue for their social acceptance and public health impact. In this context, understanding the mechanisms of noise generation and identifying effective strategies for noise mitigation could favour a sustainable integration of UAVs into populated environments. To this aim, a parametric experimental study was conducted on a set of commercial off-the-shelf 5-inch UAV propellers. The study systematically examined the influence of design parameters such as blade pitch, number of blades, and blade geometry on the acoustic emission across different rotational regimes. The propellers were tested in single configuration as well as installed on a quadcopter. A comparison with recently introduced toroidal propellers was made to evaluate whether this unconventional geometry provides tangible benefits in terms of noise reduction without compromising thrust generation. Simultaneous measurements using a hot wire anemometer and microphones located in the nearfield and the farfield with respect to the source enabled the correlation of acoustic emissions with the underlying flow structures and unsteady phenomena responsible for sound generation.The results provide insight into the aeroacoustic performance of these systems, offering guidance for the development of quieter UAVs.

Speaker: Luana Georgiana Stoica (Universita degli Studi Roma Tre)

Full paper (PDF)

549 Effect of Blade Tip Geometry on the Noise Spectrum and Sound Pressure Level of FDM 3D-Printed Drone Propellers

Propellers are the dominant source of noise in drones,with the blade tip region playing a critical role due tohigh local flow velocities and strong aerodynamic interactions.In addition to overall sound pressure level,the spectral content of the emitted noise significantlyinfluences human perception and annoyance.This study investigates the effect of blade tip geometryon A-weighted sound pressure level with FASTtime weighting (LAF ) and the frequency spectrum offused deposition modeling (FDM) 3D-printed dronepropellers in the hover regime. A baseline propellergeometry was modified by varying blade tip sweepangle (−30◦, 0◦, and +30◦) and tip chord length (baselineand 50% of the baseline). Acoustic measurementswere conducted under zero inflow conditions over arotational speed range from 5000RPM to 8000 RPM.The results show that blade tip modifications influenceboth tonal and broadband noise components, leadingto measurable differences in sound pressure level andspectral characteristics across operating conditions.In particular, certain configurations exhibit reducedvariability in both acoustic and performance metrics,indicating more stable aerodynamic behavior.These findings contribute to the understanding ofaeroacoustic mechanisms in small drone propellersand support the development of low-noise propellerdesigns.

Speaker: Ivor Santak (FSB, University of Zagreb)

Full paper (PDF)

550 'Acoustic Fluid–Structure Interaction in Automotive Vibro-Aeroacoustic Applications'

In recent years, simulation methods for two-way coupled acousticfluid–structure interaction (AFSI) have significantly advanced and areincreasingly applied to complex automotive configurations. Theseapproaches enable a detailed analysis of coupling mechanisms betweenflow, acoustics, and structural response, and allow identification ofdominant excitation pathways and their implications for modellingstrategies.In this contribution, three representative automotive applications areinvestigated: the vibration of an aluminium underbody panel in an SAEreference body, the vibration of a side window in a near-productionside-mirror/side-window configuration, and the vibration of a bonnetin a series-representative vehicle configuration. For all cases, goodagreement between simulation and experimental measurements isachieved, providing a reliable basis for further analysis.Across the investigated cases, both hydromechanical and acousticpressure fluctuations contribute to structural excitation; however,their relative importance varies significantly depending on theconfiguration. The role of two-way coupling is analysed for each case,highlighting when feedback effects between structure and fluid areessential for accurate prediction.Furthermore, it is shown that, within a two-way coupled simulationframework, particular attention must be paid to the structural modalmodel. In particular, experimentally validated modal models mayalready include ambient-air effects, which can lead to double-countingwhen used in fully coupled simulations—especially in the presence ofenclosed air volumes. The implications for modelling strategies inautomotive vibro-aeroacoustics are discussed.

Speaker: Florian Schwertfirm (KM Turbulenz GmbH)

A12.01 Numerical methods for acoustics and vibration: S525

Conveners: Dionysios Panagiotopoulos (KU Leuven Campus De Nayer), Stefan Schoder (IGTE, TU Graz)

551 A Robust Adaptive Wavenumber Sampling Approach For Efficient 2.5D Acoustic Simulations

The 2.5D acoustic finite and boundary element methods efficiently solve wave propagation problems for longitudinally invariant geometries by employing a wavenumber decomposition in the invariant direction.However, evaluating the improper Fourier integral across the wavenumber spectrum remains a significant computational challenge, as it relies on the numerical integration of a continuous spectrum resolved via discrete sampling strategies. While adaptive sampling schemes reduce the number of required 2D calculations, existing methods relying on purely local relative errors experience numerical instability when the spectral amplitudes approach zero. Furthermore, standard numerical quadrature struggles with the highly oscillatory inverse Fourier kernel, introducing additional integration errors. This paper proposes a robust adaptive wavenumber sampling approach that approximates the spectrum using strictly piecewise linear segments. This deliberate geometric restriction enables the exact analytical evaluation of the oscillatory integral via Filon-type quadrature, effectively eliminating numerical integration errors. To drive the sampling scheme, a hybrid local error norm is introduced to ensure stability near zero, combined with a global convergence check on the physical spatial pressure. Numerical examples demonstrate that the proposed approach drastically reduces computational effort while maintaining the strict accuracy of dense equidistant sampling.

Speaker: Likun Luo (Institute of Mechanics and Mechatronics)

Full paper (PDF)

552 Towards a localised POD-Galerkin reduced-order modelling approach for time-varying structural dynamics

Parametric reduced-order models (ROMs) for linear structural dynamics commonly focus on time-invariant parametric settings. However, when parameters vary in time, the governing operators become time-dependent and ROM accuracy must be ensured over a family of admissible parameter trajectories rather than a bounded parameter domain. Direct sampling of this trajectory space is computationally prohibitive, and approximating the solution manifold using a single global basis often leads to an impractically high-dimensional reduced subspace. We therefore adopt a regime-based ROM library approach, where each local ROM is constructed by computing a proper orthogonal decomposition (POD) basis from full-order snapshots at selected constant parameter values and applying a Galerkin projection onto the resulting reduced subspace. The parameter values are chosen using a low-cost library refinement indicator based on discrepancies between adjacent local ROMs. In the online phase, the model selects the appropriate local ROM according to the active parameter regime and a mass-orthogonal projection transfers the reduced states across subspaces at regime transitions. The method is demonstrated on a Kirchhoff-Love plate with a time-varying boundary stiffness parameter, which induces a time-dependent stiffness matrix affinely. Results show that this method can achieve sufficient accuracy with substantially smaller reduced dimensions than a single global ROM.

Speaker: Shing-Cheung To (University of Southampton)

Full paper (PDF)

553 Efficient CAD-based broadband acoustic analysis and shape design using reduced-order IGABEM models

Broadband acoustic analysis and design of realistic computer-aided design (CAD) geometries are challenging due to the non-affine frequency dependence of the Boundary element method. When combined with isogeometric analysis (IGABEM), exact CAD geometry representation, high accuracy, and relaxed mesh requirements are offered. However, the use of spline-based basis functions increases the computational cost per degree of freedom, making analyses and iterative design loops computationally demanding. In that context, this work presents an efficient IGABEM-based framework that enables broadband acoustic shape exploration by combining IGABEM with a non-intrusive two-step model order reduction (MOR) strategy. The frequency-dependent boundary element system is approximated and reduced using Krylov-subspace recycling MOR and reduced basis method, allowing fast evaluation of acoustic responses over wide frequency ranges. Shape variations are performed directly at the CAD level by modifying NURBS control points supported by a topological ring-based smoothing strategy to ensure smooth and robust geometry updates. The methodology is demonstrated on an academic example, a multi-patch non-conforming flat plate subjected to broadband acoustic excitation. The result shows that geometry smoothing leads to more natural shape variations, while the reduced-order model makes acoustic evaluations computationally feasible.

Speaker: Philip Le (KU Leuven Campus Diepenbeek)

Full paper (PDF)

A13.04 Photoacoustic Imaging and Spectroscopy: S506

Conveners: Jan Laufer (Martin Luther University Halle), Ben T Cox (University College London), Günther Paltauf (University of Graz), Robert Nuster (University of Graz), Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe)

554 Photothermal tuning of ultrasensitive Fabry-Pérot sensors for parallelized PA measurements

Photoacoustic (PA) imaging combines optical absorption contrast with the high imaging depth of ultrasonic techniques. It has been well established for the investigation of biological tissue at shallow (millimeters) depths. Imaging multi-centimeter depths in heterogeneous tissue is more challenging. For example, strong ultrasonic attenuation in bone tissue hinders applications such as human brain imaging—driving the need for more sensitive ultrasound detectors.We have developed ultrasensitive plano-concave Fabry-Pérot (FP) sensors for deep PA imaging. The sensor geometry allows mode matching to the interrogation laser beams. In combination with soft, thick and impedance matched spacers, this results in high optical Q-factors and a high, broadband acoustic sensitivity.The sensors detect sub-Pascal pressure amplitudes over a bandwidth of more than 1 MHz. While single sensors have been used for PA imaging, parallelised detection using a sensor array would require separate interrogation lasers for each sensor element, which is economically infeasible. Also, if a single laser was used to interrogate multiple sensors, the fabrication tolerances would be in the sub-nanometer range for a millimeter-sized spacer to ensure sufficient spectral overlap of the resonances.In this study, we demonstrate a method for simultaneous photothermal tuning of multiple plano-concave FP sensors using a single tunable interrogation laser for synchronization and interrogation, enabling simultaneous PA measurements. This method makes use of resonance-dependent photothermal heating induced by the interrogation laser, which affects the optical path length and therefore the resonance wavelength. We outline a general approach for designing synchronizable ultrasensitive FP sensors by controlling their thermal properties and the resulting photothermal shift.

Speaker: Thomas Kirchner (Martin Luther University Halle)

555 Fabry-Perot tomograph using a camera-based detection scheme with a continuously adjustable optical magnification for photoacoustic mesoscopy

Fabry–Perot tomographs have produced striking photoacoustic images, as they combine very small detection elements with high acoustic sensitivity and a wideband frequency response. We recently introduced a camera-based tomograph that incorporates a Fabry-Perot sensor with uniform optical thickness, a collimated, expanded green interrogation beam, and a high-speed intensified sCMOS camera. This setup enables high-resolution, three-dimensional backward-mode imaging, achieving lateral and axial resolutions of several tens of micrometers. The acoustic lateral resolution is primarily determined by the dimensions of the active element size. When a collimated interrogation beam is used, the active element size is defined by the camera’s pixel pitch and the magnification of the imaging lens that projects the Fabry-Perot sensor onto the camera. Fixed-focal-length lenses and an objective lens with continuously adjustable magnification (0.5x - 2x) were used to change the detection aperture and, as a consequence, the active element size. This study investigated how acoustic lateral resolution is influenced by the optical magnification. In addition, backward-mode imaging for photoacoustic mesoscopy was demonstrated by recording acoustic fields from phantoms.

Speaker: Jan Sievers (Martin Luther University Halle-Wittenberg, Institute of Physics)

556 Developing Highly Sensitive Fused Silica Fabry Perot Ultrasound Sensors for High Resolution Deep Tissue Photoacoustic Imaging

Photoacoustic imaging (PAI) is a rapidly advancing biomedical imaging technique in which pulsed light is used to excite ultrasound waves inside living tissue. To enable high-fidelity PAI systems, ultrasound sensors should provide high sensitivity, small element size, and broad bandwidth. A suitable choice is a Fabry-Perot ultrasound sensor (FPUS), an optical cavity with an ultrasonically compressible spacer layer. FPUSs can be used to achieve photoacoustic images with 50-100μm spatial resolution and penetration depths of 10-15mm, enabling detailed, non-invasive imaging of small blood vessels and other structures. Current applications, including monitoring skin cancers and assessing superficial vasculature morphology, could be expanded through the use of higher sensitivity FPUSs, enabling deeper imaging. Higher Q-factor cavities can be achieved by increasing mirror reflectivity, however this is limited by losses within the typically polymer spacer layer which occur due to optical absorption and surface roughness. In this work, we investigate replacing the spacer layer with one made from fused silica in order to take advantage of its very low absorption coefficient and fabrication methods which can decrease the surface roughness. Increased thickness uniformity may also enable single wavelength readout systems. To test this design concept, we fabricated and characterised several fused silica FPUSs and used them to acquire photoacoustic images. Preliminary results suggest sensitivity values could far exceed those of polymer FPUSs, indicating fused silica FPUSs are a promising alternative which could enable new PAI applications such as deep tissue imaging.

Speaker: Isobel Henderson (University of Birmingham)

A18.00 Soundscape, Environmental Quality, Health and Well-being: S333

Conveners: André Fiebig (TU Berlin, Department of Engineering Acoustics), Arezoo Talebzadeh (Ghent University)

557 Designable Factors in Urban Courtyard Soundscape: A Scoping Review

As a quintessential urban element, courtyards foster restorative spaces within dense cities. However, their soundscapes remain underexplored compared with open urban spaces and indoor environments. Utilising the soundscape description system, this scoping review synthesised empirical evidence from 90 studies on the acoustic environments of urban courtyards. First, regarding sound characteristics, courtyards are not inherently quiet (SPL IQR = 49.15-59.07 dBA) but provide substantial sound attenuation. Because overall SPLs are moderate, design strategies can shift from noise mitigation to amplification of natural, human, and context-specific sounds. Second, diverse spatial configurations influence reverberation (RT = 1.05s), with boundary conditions, openness, and adjunct building elements (e.g. covered corridors, pavilions), serving as effective physical intervention strategies. Third, reported social and demographic aspects of users are primarily related to residential courtyards, exhibiting a dichotomous design narrative between domestic and public urban courtyard soundscapes. Finally, observed mesothermal climate classification and visually dominated environments fundamentally accentuate auditory perception. Regulating physical environmental conditions can facilitate feasible microclimatic soundscape strategies. To advance this, the review identifies underexplored opportunities for designing courtyard soundscapes through spatiotemporal acoustic properties, spatial geometry, and interactions between environmental domains.

Speaker: Veerapong Eawpanich (UCL Institute for Environmental Design and Engineering)

Full paper (PDF)

558 Soundwalking through a Hong Kong Wet Market

Wet markets are integral components of everyday life and culture in Hong Kong. Documenting and understanding the sensory environment of wet markets can better support their protection and management as intangible cultural heritage. Using a soundwalk method, this study provides a baseline assessment of the soundscape of Shek Kip Mei Market in Hong Kong, examining and comparing its environmental quality and people’s experiences. Two soundwalk sessions were conducted at separate time points across eight stall locations in this market, with acoustic measurements taken and on-site questionnaire ratings (n = 96) of soundscape quality collected from six participants assigned to the sound-related role. The results suggested that (1) human voices and fan noise were the dominant sound source types in Shek Kip Mei Market; (2) human voices were positively associated with pleasantness and eventfulness; and (3) pleasantness and eventfulness ratings were similar across the two walks, while agreement for individual sound-source ratings was generally lower. We discuss Shek Kip Mei Market’s soundscape characteristics and outline future research on multisensory environmental perception in Hong Kong, including soundscape and smellscape. The findings provide baseline insights into the research and practical management of Hong Kong wet markets.

Speaker: Qianxi Jiang (City University of Hong Kong)

Full paper (PDF)

559 Pedestrian Walking Speed in Relation to Traffic Noise Sharpness

Urban streets are shaped not only by their physical form but also by their sensory environment, particularly sound. While previous research has focused on perceptual responses such as annoyance, less is known about how acoustic conditions relate to actual pedestrian behaviour in real-world settings. This study examines the relationship between traffic noise and walking behaviour using a non-intrusive observational approach on a London street. Using simultaneous binaural audio recording and video tracking, the research extracted psychoacoustic indicators—acoustic sharpness, a key metric reflecting the high-frequency spectral balance characteristic of urban traffic noise—to analyse its correlation with pedestrian walking speed. Results show that acoustic sharpness exhibits a significant, non-monotonic relationship with pedestrian movement. Specifically, walking speed increased at moderate levels of sharpness, followed by a decrease as the auditory stimulus reached higher, more "piercing" intensities. The magnitude of these behavioural modulations—ranging from 0.11 to 0.12 m/s—represents a quantifiable shift in the rhythmic use of public space. This suggests that sound acts as a subtle yet persistent regulator of the urban tempo, providing a new perspective for the development of more responsive and pedestrian-oriented urban spaces.

Speaker: Kening Guo (Institute for Environmental Design and Engineering)

Full paper (PDF)

560 Sonic Awareness as Lifelong Literacy: A Soundscape Exploration Study with Primary School Children

Auditory development is a lifelong process, beginning with prenatal exposure to acoustic cues and maturing through late adolescence as a complex intersection of biological, cognitive, and social factors. While biological maturation provides the hardware for hearing, the reflective, active engagement with one’s acoustic environment requires an intentional pedagogical framework. We argue for a continuous soundscape curriculum from pregnancy through high school, using a foundational study with children to demonstrate the necessity of early educational interventions for lifelong sonic awareness. We employed a child-appropriate method with three workshops with 12 students (ages 7–8) at the International School Delft (NL). Utilizing the ISO 12913 frameworks, activities included sound-walks, perceptual mapping, and sound visualization. Findings indicate that while primary-age children are highly sensitive to salient sound events, such as mechanical traffic or human vocalisations, they initially perceive such sounds as isolated occurrences rather than a soundscape. However, repeated listening exercises and creative mapping (using colour and abstract forms) successfully shifted student perception toward understanding the auditory environment as a collective, integrated system. These results suggest that because primary-age listeners are still constructing the ability to interpret auditory scenes, continued education is vital. As students transition into middle and high school, curricula should evolve from basic event identification to the critical evaluation of varying degrees of impact of sound and soundscapes on humans. By fostering soundscape literacy early, we prepare children to move from passive reception to active ownership of their urban sonic environments.

Speaker: Elif Özcan (Erasmus MC / TU Delft)

Full paper (PDF)

561 Perfect absorption of audible sound and optimal design of absorbers

The scattering matrix of a system that perfectly absorbs the acoustic energy has properties that can be used effectively to design complex structures. These properties are first examined and analyzed. Examples of systems that provide perfect absorption when bounded by a rigid backing, by an impedance boundary condition (baffled duct), or unbounded (transmission problem) are then presented. Finally, the conditions for their optimal design are provided and illustrated with examples.

Speaker: Jean-Philippe Groby (LAUM, UMR 6613, IA-GS, CNRS, Le Mans Université)

13:00 Lunch break

A03.07 Acoustic comfort, descriptors and subjective perception in building acoustics: S019

Conveners: Herbert Müllner (Staatliche Versuchsanstalt TGM), Rok Prislan (InnoRenew CoE, University of Primorska)

562 Perception of Vibrations Induced by Impact Sound Sources in Buildings

Impact sound is reported as one of the most disturbing noise sources in lightweight buildings, yet its perceptual mechanisms remain insufficiently understood. Existing evidence indicates that annoyance depends strongly on the type of impact source, while correlations with single-number ratings are often weak. In particular, low-frequency components – typically underrepresented in standardized metrics – appear to play a critical role, suggesting the need for spectral adaptation approaches. However, within this frequency range, it remains unclear to what extent perception is mediated purely by the auditory system or by whole-body vibrations. This study examines the role of structural vibrations in the perception of impact sound. Independent measurements of airborne sound and structure-borne vibrations were taken in buildings for characteristic impact sources and two floor assemblies with different insulation performance. A controlled blind listening experiment was conducted in which participants evaluated the perceived intensity of vibrations without knowing whether the vibrations were physically reproduced. Statistical analysis of the responses reveals that, for sufficiently strong excitations and in cases of poor impact sound insulation, the structural transmission path significantly contributes to perceived vibration intensity. These findings show that impact sound perception cannot be fully explained by airborne sound alone and are important for better understanding impact sound annoyance.

Speaker: Rok Prislan (InnoRenew CoE, University of Primorska)

563 In situ measurements of combined effect of railway noise and vibration on annoyance assessed with physiological responses

In order to define appropriate mitigation measures to limit noise and vibration due to railway transportation in residential areas, annoyance criteria have to be precisely defined. In this context, the aim of the VibAcouPhysio project is to explore the potential of physiological measurements to objectivate the perceived annoyance. In a first step, the VibAcouPhysio project investigated the annoyance due to railway noise and vibration in laboratory conditions. A significant correlation was found between physiological metrics, the annoyance evaluated in the questionnaires and the indoor noise and vibration descriptors derived from the noise and vibration signals. In that respect, noise integrates both aspects: air-borne noise (through the façade) and structure-borne noise (radiated by building structure submitted to vibration). The second phase of the project proposes a similar analysis based on in-situ measurements on a few voluntary participants living along railway lines. These measurements include continuous monitoring of physiological, noise and vibration signals, as well as perception surveys several times a day. The participants and their residences have been monitored for a week, day and night, allowing to also address the issue of sleep disturbance. The paper describes the different measurements performed in-situ and proposes a preliminary analysis of the collected data.

Speaker: Apolline Brisson (CSTB)

Full paper (PDF)

564 Exploring the potential influence of visual conditions on the acoustic perception in an office setting

Indoor environmental quality (IEQ) strongly influences occupants’ health, comfort, and productivity. Therefore, it is essential to study the factors that determine the quality of the indoor environment, which typically pertain to the thermal, visual, acoustic, and air quality domains. In this context, efforts have been made to address possible cross-domain effects among multiple IEQ factors. However, the current state of knowledge in this area is still limited. This paper aims to expand the existing IEQ literature by focusing on the potential influence of the visual environment on acoustic perception. To this end, data collected within the framework of a recent research project is used. This project involved an empirical office-based experiment with 78 participants, and the results showed a small but significant effect of the acoustic environment on participants’ visual discomfort. The present contribution examines the influence in the opposite direction, that is, whether visual conditions influence participants' acoustic perception. The results did not display an impact of visual exposure (as characterized in terms of prevailing degrees of glare) on acoustic perception (as expressed in terms of both sensation and satisfaction) for any of the three acoustic settings considered in the experiments. This suggests that cross-domain interactions may not be bidirectional.

Speaker: Irene Martínez Muñoz (Technische Universität Graz)

Full paper (PDF)

565 Towards a Standardized Method for Assessing the Acoustic Environment in Residential Buildings Through Perception Surveys

The acoustic classification methods for residential dwellings—and particularly the one developed in the ISO/FD 19488 Technical Specification—employ several performance indicators to characterize the acoustic quality of housing. In order to better correlate occupants' perceptions with these indicators and, ultimately, with acoustic classes, a new Documentation Fascicle (FD S30-109) establishes a methodology for conducting perception surveys.This methodology will be presented in its entirety, beginning with the acoustic parameters upon which it is founded. A detailed description of the survey’s structural framework will follow, supplemented by an outline of the various contextual and psycho-social factors to be considered. Key principles and best practices will be discussed, along with recommendations for correlating survey results with acoustic measurements. Finally, selected excerpts from a sample questionnaire will be presented and analyzed.

Speaker: Catherine Guigou-Carter (CSTB)

Full paper (PDF)

A05.04 Outdoor sound propagation: S244

Conveners: Christian Adams (Graz University of Technology), Timothy Van Renterghem (Ghent University), Martin Czuka (AIT Austrian Institute of Technology)

566 Outdoor propagation of impulsive sound

The impulsive sound of large caliber weapons or explosions can travel over long distances. To calculate the sound levels at some distance from the source, a linear propagation model and therefore a linear acoustic source strength is used, in general. Nonlinear propagation effects for the shock wave, relatively close to the source, are not taken into account. By measuring the source strength at a larger distance, typically 100 to 200 meters the linear acoustic source strength can be determined.A new method to determine the linear acoustic source strength has been developed, based on using closer measurements. At shorter distances, typically 50 meters or less from the source, deploying microphones around the weapon is simpler and faster. Also, less space is required for the measurement setup.In a validation setup, sound measurements were carried out at two distances from a large caliber weapon: at 150 and at 50 meters. At 50 meters nonlinear effects can be observed; i.e. a faster than linear decrease of the peak sound pressure over distance and a broadening of the shockwave in time resulting in a frequency shift of the corresponding spectrum. An analytical nonlinear model is used which allows to estimate the shock wave in the time domain at various distances. Special attention is paid to the influence of the ground, as this can require to correct with several dB’s in certain octave bands when determining the source strength.

Speaker: Frits Van der Eerden (TNO)

Full paper (PDF)

567 Ground Attenuation in Environmental Noise Models: Benchmarking MRG/ISO and CNOSSOS

In the Netherlands, two models are used for official environmental sound propagation calculations: the national MRG model (similar to the ISO 9613-2 model) and the European CNOSSOS model. A key difference between these models is the calculation of ground attenuation, which can lead to differences of up to 10-20 dB in certain octave bands. Such differences can have a significant impact on the outcomes of environmental noise assessments. The observed discrepancies therefore raise important questions about the validity of each model.This study provides a systematic comparison and validation of the ground attenuation calculation method of both models by combining numerical simulations with field measurements. A Finite Element Method (FEM) is used to generate numerical benchmark data under various ground types and meteorological conditions. To validate these simulations, a measurement campaign is conducted at an airport, where a controlled artificial sound source is used to measure sound propagation simultaneously at multiple distances (up to 500 m) and heights over both acoustically hard and soft surfaces. Ground impedance and meteorological parameters are also recorded. Results from the MRG/ISO and CNOSSOS models are compared with both FEM simulations and experimental data.By providing a comprehensive benchmark, this research contributes to improved reliability of sound propagation models widely used in environmental acoustics. The presentation will discuss some background, numerical approach, and experimental findings.

Speaker: Lennart Bouma (National Institute for Public Health)

568 Quantification of Modelled Wind Turbine Noise Ground Reflections

Ground reflections play an important role in the propagation of wind turbine noise and significantly affect noise assessment. For simplicity, most best-practice wind turbine noise prediction models treat the entire wind turbine as a point source, typically located at the rotor hub. Nevertheless, the main noise generation mechanism in modern wind turbines is better represented by a line source along each blade. This paper proposes a method for assessing ground reflections that accounts for the distribution of spanwise acoustic intensity through an approximation with multiple point sources. Results for a 5-MW reference wind turbine show that most acoustic interference occurs below 5 kHz. Moreover, loss of phase coherence due to atmospheric turbulence further reduces the acoustic interference. Using too few point sources in the distribution leads to aliasing effects, which will affect the accuracy of blade-element-based wind turbine noise models. Overall, the proposed method leads to lower acoustic interference than common single-point-source models, suggesting that current prediction methodologies are likely overestimating the effect of ground reflections in their calculations, potentially leading to misleading results (e.g. amplitude modulation analyses) that may fail to match experimental data.

Speaker: Josephine Pockelé (Delft University of Technology)

Full paper (PDF)

569 Exploring Incoherent Scattering from a Dynamic Water Surface

Environmental sound propagation over water is usually modelled by treating the surface as acoustically hard and, perhaps more importantly, stationary. This simplification does not account for the possible influence of surface motion on the coherence of the surface-reflected sound field. In these tests, surface motion was generated using a simple ripple-making system in an outdoor shallow-water basin, allowing a range of ripple states to be produced. While maintaining a fixed acoustic source--receiver geometry, transfer-function (TF) measurements above the water surface were used to identify changes in the reflected sound field. For each measurement, water-surface activity was characterised by approximate ripple height and wavelength. The results showed that increasing water-surface activity deepened the incoherence around sensitive frequency bands associated with the source--receiver geometry.

Speaker: Sean Horsman (University of Galway)

Full paper (PDF)

570 An Experimental Study Of High Amplitude Shock Noise From Explosive Depth Hardening

Explosive Depth Hardening (EDH) is an industrial process in which plastic explosives are used to com press and work-harden steel. Depending on the steelthat needs hardening, the energy equivalent of up to 10 kg of TNT is used. Depending on the shipments, between one and 4 rails are hardened simultaneously.The nearest noise sensitive receptor is approximately 2 km from the source, and under specific weather conditions, EDH has the potential to exceed 130 dB(C)peakat this location. This noise source is also known to be audible and identifiable at receptors more than 10 km away.In this study, a comprehensive measurement campaign has been undertaken to investigate the propagation of the pulse in the non-linear and linear acoustic regions. These field trial measurements were conducted in-situ on the DNV Spadeadam site during their usual daily operations. The characteristics of the blast wave, including peak pressure, shock duration, and shock speed, have been measured using piezoelectric pressure sensors.

Speaker: Zachary Simcox (University of Salford)

Full paper (PDF)

A07.00 Flow Acoustics: S344

Conveners: Stefan Becker (LSTM, FAU Erlangen-Nürnberg), Roberto Camussi, Francesco Avallone (Politecnico di Torino), Stefan Schoder (IGTE, TU Graz)

571 Direct surface pressure-based source localization on a NACA0012 airfoil

Aerodynamic source localization is conventionally performed by reconstructing acoustic source distributions from real or virtual microphone-array signals after propagation to the far field. Delfs and Ruck [1,2] proposed a diffraction-filtering methodology that reconstructs the acoustically active surface pressure and an associated surface source quantity directly from simulated surface pressure, thereby localizing sources without beamforming or far-field propagation. Here, the methodology is implemented within a frequency-domain Ffowcs Williams–Hawkings (FW-H) framework and applied to wall-modelled large-eddy simulations (WMLES) of a NACA0012 airfoil at 0° and 6° angle of attack, computed with the MGLET solver [3,4]. A mesh-convergence study identifies a converged surface-pressure solution, from which surface source distributions are reconstructed for both angles of attack. As an independent localization reference, conventional delay-and-sum beamforming is applied to microphone cross-spectral matrices synthesised with the in-house FW-H solver FWH-next and processed in Acoular [5,6], using the CFD surface as a common scanning grid. The reconstructed maps identify the trailing edge as the dominant source, resolve finer spatial detail than beamforming, and preserve the radiated far-field spectrum despite suppressing the surface pressure by up to about 40 dB.

Speaker: Geervan Ramasawmy (Institute for Aerodynamics & Flow Tech)

Full paper (PDF)

572 Isentropic formulation for subsonic computational aeroacoustics with two-way acoustic–flow coupling

In this work, we revisit a unified numerical framework for simulating subsonic aeroacoustic phenomena based on an isentropic formulation of the compressible Navier–Stokes equations. The approach relies on a velocity–pressure setting in which the energy equation is omitted under isentropic assumptions, yielding a system that remains well conditioned in the low-Mach limit while reducing the computational cost compared to fully compressible solvers. Unlike acoustic analogy and hybrid approaches, the proposed method solves the flow and acoustic fields simultaneously within a single model, enabling intrinsic two-way coupling between aerodynamics and acoustics and allowing the simulation of acoustic feedback mechanisms associated with self-sustained oscillations and flow instabilities, such as in subsonic cavity flows.Special attention is devoted to the treatment of boundary conditions, where mean flow and acoustic fluctuations are separated to allow outgoing acoustic waves to leave the computational domain without spurious reflections, even under velocity-prescribed boundaries. The formulation is implemented within a stabilized finite element framework. Numerical examples demonstrate the ability of the method to accurately capture both sound generation and propagation across a wide range of subsonic Mach numbers.

Speaker: Oriol Guasch (La Salle, Universitat Ramon Llull)

573 Localization of Vortex Driven Acoustic Noise Source in a Whistling T-junction

Flow induced noise in piping systems can lead to unacceptable acoustic emissions, vibration, and long term structural damage. Identifying the physical mechanisms and spatial regions responsible for sound generation remains challenging, particularly in turbulent flows where convective pressure fluctuations coexist with radiating acoustic waves. In this work, a rectangular T-junction with a cavity, whistling under turbulent flow excitation is investigated. The whistling aeroacoustic instability results from the constructive interaction of the acoustic waves in the cavity and the turbulent shear layer. In the present study, we revisit a previously published dataset, combining Particle Image Velocimetry (PIV) and acoustic Multi-Microphone Method characterization to identify and localize aeroacoustic sources using vortex-sound theory. Howe’s acoustic analogy is applied to the resolved flow fields to compute the vortex sound source term, enabling the spatial localization of regions that contribute most strongly to acoustic radiation.In addition, an attempt at distinguish pseudo sound—convective, non-radiating pressure fluctuations—from true acoustic waves is made. Correlations between pressure signals, vorticity, and far-field acoustic observables are employed to separate hydrodynamic pressure components from propagating sound using the correlation between pressure signals measured at different locations.

Speaker: Claire Bourquard (Eindhoven University of Technology)

574 Acoustic Mode Decomposition in Atmospheric Ducts

Winds and temperature inversions in the lower atmosphere can create waveguides for acoustic waves. These waveguides are similar to acoustic ducts, such that the sound field can be represented as a superposition of a finite set of propagating atmospheric modes. These modes are obtained by solving a cubic eigenvalue problem derived from the linearized Euler equations. This approach enables the simulation of long-range acoustic propagation in vertically stratified atmospheres, as demonstrated through benchmark cases. We present numerical examples illustrating how this modal decomposition facilitates efficient hybrid simulations of long-range acoustic wave propagation. While the resulting computations are efficient, a limiting factor in terms of computational cost is the solution of the cubic eigenvalue problem. Therefore, we discuss the scalability of the approach using distributed eigenvalue solvers. In the future, more efficient eigenvalue solvers could further extend the applicability of the modal decomposition method toward real-time predictive modeling.

Speaker: Stefan Jacob (German National Metrology Institute)

575 Influence of Protective Grilles on the Pressure-Side Acoustics of a Cooler-Fan-Assembly used in Construction Site Machinery

The fans in the cooling system contribute significantlyto the noise emissions of heavy construction equipmentsuch as rotary drilling rigs. Protective grilles are usedto prevent people and animals from coming into con-tact with the rapidly rotating fan blades. It is knownfrom the relevant literature that such protective grilleson axial fans can increase the fan’s radiated soundpower by several decibels. In this study, we thereforesystematically investigate the acoustical influence ofprotective grilles mounted on the pressure side of thefan of a cooler-fan combination used in constructionmachinery. Four typical protective grilles with sig-nificantly different geometries are used. The grillesare first acoustically characterized in an aeroacousticwind tunnel under low-turbulence flow conditions andthen mounted on the pressure side of the cooler. Thecooler, which is installed in an axial-fan test rig, isacoustically measured at various operating points ofthe axial fan. It is found that the grilles have no significantinfluence on the sound emissions of the overall systemat any of the feasible operating points. A similarityanalysis leads to the hypothesis that pressure-side pro-tective grilles primarily influence the emitted soundspectrum in low-pressure axial fans.

Speaker: Jörg Riedel (LSTM Erlangen (Germany))

Full paper (PDF)

576 An Integrated Reduced Order Model for Aeroacoustic and Sound Quality Prediction of Automotive Fans in Heat Exchanger Installed Conditions

The present work proposes an integrated aero-psychoacoustic framework based on a reduced-order model (ROM) for the rapid prediction of rotor noise and its perception. This study focuses on an automotive axial fan installed downstream of a heat exchanger. While traditional design relies on conventional acoustic metrics (e.g., A-weighted SPL, OASPL), this methodology prioritizes human perception. Specifically, it couples Amiet’s (leading/trailing Edge) and Hanson’s (tonal) theories with a Psychoacoustic Annoyance (PA) model, employing the SQAT toolbox for sound quality calculation. The estimation of sharpness, a key perceptual parameter often dominated by trailing-edge noise, is addressed. Input parameters are derived from viscous panel-code simulations (XFOIL), enabling fast design iterations. The model is validated using directivity measurements of a fan conducted in an anechoic chamber, including configurations with and without heat exchanger to isolate installation effects. Results demonstrate that the ROM effectively captures sound quality metrics, such as loudness and sharpness. The two quantities drive annoyance and by these means the degradation of perceived sound quality can be assessed. This integrated approach facilitates perception-driven engineering, enabling designers to enhance sound quality during the early development stages of installed rotating systems. Consequently, it ensures that the final product is not only functionally efficient but also perceived as having an appropriate sound quality.

Speaker: Jacopo Grassi (University of Ferrara, Department of Engineering)

A10.04/A12.04 Numerical methods for acoustic materials and metamaterial: S063

Conveners: Luís Godinho (University of Coimbra, Dep. Eng. Civil), Olivier Dazel (Le Mans Université)

577 Subwavelength Corrugated Surfaces for Planar Focusing: A Parametric Analysis

Subwavelength corrugated surfaces (SCSs) provide a compact and broadband approach for planar acoustic focusing, enabling efficient wavefront manipulation without relying on bulky curved or refractive elements. While the focusing capabilities of SCSs have been demonstrated experimentally and numerically, a systematic understanding of how geometric parameters influence focal efficiency and operational bandwidth is lacking. This work presents a parametric analysis of planar focusing SCS configurations, introducing the Focal Concentration Efficiency metric to quantify the fraction of acoustic power localized near the focal point. Simulations based on both a full-wave model and a reduced-order phased-array model are employed to explore the effects of groove number, spacing, and overall surface dimensions over a broad frequency range. The results reveal clear geometry-driven trends, highlighting intrinsic efficiency limits and providing practical guidelines for the design of compact, broadband SCS-based acoustic focusing devices.

Speaker: Letizia Chisari (University of Sussex)

Full paper (PDF)

578 Modelling the interaction between a controllable Helmholtz resonator-based acoustic metamaterial and an enclosure using impedance boundary conditions

Helmholtz resonators (HRs) are widely used in the context of room acoustics to reduce the sound pressure peaks induced by room modes. A shortcoming of HRs is that they are very narrowband and can become ineffective as soon as target room eigenfrequencies change, e.g., due to occupancy changes. To address this, a controllable HR-based acoustic metamaterial (HRAM) with resonance frequencies that are reconfigurable over a wide frequency range is proposed here. Due to the large number of degrees of freedom required to model the thermoviscous dissipation in the HRAM design, finite element method (FEM) modelling of its interaction with an enclosure is highly computationally intensive. The use of an adequate boundary layer mesh to accurately capture the thermoviscous dissipation in the boundary layer of the HRAM further increases the complexity of the FE model. Instead of modelling the three-dimensional geometry of the HRAM explicitly, an alternative method is employed. In this contribution, the complexity of the FE model is reduced by employing a surface impedance method in which the HRAM is modelled using the specific acoustic impedance value evaluated at its opening (using a separate simulation model of only the HRAM). Simulation results of the pressure field inside an enclosure are presented, comparing the explicit HRAM modelling approach with the surface impedance method. The results show that the computational effort can be significantly reduced, with the surface impedance method achieving almost the same accuracy as the explicit HRAM modelling approach up to a limiting frequency at which the acoustic wavelength becomes comparable to the HRAM unit cells.

Speaker: Yuan Kit Sim (University of Southampton)

579 Application of the wave finite element method for dispersion prediction in materials with frequency-dependent properties

The study of periodic structures (phononic crystals or locally resonant metamaterials) for noise and vibration reduction has led to a plethora of effective modelling tools. For many purposes, in particular the identification of band gaps, dispersion analysis based on the model of a unit cell is a first step in the assessment of the desired functionality. If the unit cell has a complex shape, this calculation is typically done through the wave finite element model (WFEM) approach. The unit cell is meshed and represented by a mass, stiffness, and damping matrix. Periodicity is invoked by applying the Bloch theorem to the appropriate degrees of freedom. Commercial software typically allows indirect dispersion calculation where the frequency f is calculated as a function of a chosen wave vector k. The direct determination of the dispersion (calculating k for a fixed f) is straightforward but leads to expensive polynomial eigenvalue problems.The dynamic properties of periodic media can be enhanced by including materials that are classically used for noise and vibration treatment, such as acoustic absorbers and viscoelastic media. They are represented in the frequency domain by frequency-dependent homogenized properties. Viscoelastic media are described by a complex modulus, where the real part increases with frequency and the imaginary part represents the damping. Porous absorbers can be represented by two complex and frequency-dependent properties: density and wave speed. This works shows how the WFEM of 1D and 2D periodic structures can be performed in multi-material domains, where at least one material is frequency-dependent.

Speaker: Bart Van Damme (Empa, Materials Science and Technology)

580 Wave-based analysis of sound transmission through finite-sized thick anisotropic laminates

Thick anisotropic laminated panels exhibit complex vibro-acoustic behavior in which not only direction-dependent material properties, but also shear and cross-sectional deformations, play a crucial role. These effects are not adequately captured by (equivalent) plate theories and instead require a full three-dimensional elasticity analysis. Although modeling approaches such as the transfer matrix method (TMM) and the wave finite element method have advanced the field, they can suffer from numerical ill-conditioning and computational cost issues, respectively. In addition, finite-size effects and the associated modal behavior of panels have received limited attention. This work presents a methodology that alleviates those limitations. It rests on the construction of the exact impedance matrix of an anisotropic layer in the frequency-wavenumber domain, which is significantly simplified by avoiding the Helmholtz decomposition of the displacement field. The response of a finite-sized panel itself is obtained by analyzing the corresponding infinite panel under anti-symmetric loading conditions, while the radiated sound power is computed directly in the frequency-wavenumber domain. The proposed method is shown to be computationally efficient, numerically robust, and accurate, as demonstrated through validation examples involving cross-laminated timber panels.

Speaker: Edwin P.B. Reynders (KU Leuven, Department of Civil Engineering)

581 Hybrid FEM–MFS Modeling of Low‑Height Sonic Crystal Noise Barriers with Embedded Resonators

This work presents a frequency‑domain numerical model based on the coupling of the Finite Element Method (FEM) and the Method of Fundamental Solutions (MFS) for the analysis and design of low‑height sonic crystal noise barriers with embedded resonators. The application focus is the mitigation of railway‑induced noise from trains and trams, in configurations where the barriers are placed in close proximity to the track. Detailed modeling of the barrier geometry, resonant inclusions, and near‑field acoustic effects is achieved using the FEM, which also includes a limited region of the surrounding air to accurately describe the local wave–structure interaction. The MFS is subsequently employed to represent the unbounded acoustic domain, allowing the correct treatment of sound radiation and propagation in an infinite half‑space without truncation or artificial boundary conditions. The proposed coupled FEM–MFS strategy enables an efficient and accurate numerical assessment of metamaterial‑based noise barriers in realistic deployment scenarios. Numerical examples are presented to demonstrate the influence of resonator design and barrier placement on acoustic attenuation performance, highlighting the capabilities and potential of the method for railway noise control applications.

Speaker: Luís Godinho (University of Coimbra, Dep. Eng. Civil)

Full paper (PDF)

A10.10 Sustainable Materials for Acoustic Applications: S068

Conveners: Francesco Pompoli (University of Ferrara), Philippe Glé (Cerema, Univ. EIffel, UMRAE)

582 Orientation-dependent acoustic behaviour of foam-formed reed and lupine fibre absorbers: limitations of homogeneous porous modelling

Foam-formed plant fibre materials are promising sustainable sound absorbers, but their manufacture creates dense drainage layers, producing internal densitygradients that violate the homogeneity assumption of standard equivalent-fluid models such as the Johnson–Champoux–Allard–Lafarge (JCAL) model. We investigatefoam-formed reed and lupine fibre materials using impedance tube measurements from both the open surface (Side A) and the drainage-mesh surface (Side B). The measured absorption depends clearly on orientation: with the dense layer near the rigid backing (Side A), JCAL inversion and analytical fibrous models reproduce the absorption well, whereas with it facing the source (Side B), the enhanced low-frequency performance cannot be captured by homogeneous modelling. Model parameters obtained by inverse characterisation are therefore configuration-dependent rather than intrinsic. Standard airflow resistivity measurements are further shown to be dominated by the thin drainage layer. These results establish Side A characterisation as the appropriate basis for deriving intrinsic fibre parameters in inhomogeneous foam-formed materials.

Speaker: Janis Heldmann (Aalto University)

Full paper (PDF)

583 Impact of a bio-based flame retardant treatment on the sound absorption performance of vegetal wools

Bio-based materials, such as vegetal wools, are promising alternatives to conventional building materials due to their carbon sequestration potential and sustainability. However, their high flammability necessitates the application of fire-retardant treatments, which may alter their microstructure and, consequently, their acoustic properties. This study evaluates the impact of a recently developed bio-based fire-retardant treatment, based on phosphorus grafting onto the fibers, on the acoustic performance of loose fiber assemblies and thermobonded vegetal wool panels.An impedance tube with a three-microphone method was used to measure sound absorption and indirectly determine the characteristic parameters of the pore network. Additionally, self-consistent homogenization models (Tarnow and Umnova) were applied to estimate the equivalent fiber radius, providing insights into the treatment’s effect on fiber diameter.Results indicate that the phosphorus-based fire-retardant treatment increases the fiber radius, which correlates with a reduction in airflow resistivity, viscous dissipative effects, and overall acoustic absorption. These findings highlight the trade-offs between fire safety and acoustic performance in biosourced materials. The study opens perspectives for optimizing fire-retardant treatments to minimize their impact on the acoustic properties of vegetal wools, thereby enhancing their applicability in sustainable construction.

Speaker: Clément PIEGAY (UMRAE Strasbourg)

584 Modeling the Sound Absorption of Polydisperse Plant-Based Wools: Evaluation of Self-Consistent Homogenization Methods

Vegetal wools, composed of vegetal fibers and bicomponent polymer fibers, offer highly relevant multifunctional properties for green buildings but exhibit limited low-frequency acoustic absorption when used in thin panels. To optimize their performance, it is possible to rely on self-consistent homogenization methods. However, these methods require key data such as porosity and the effective radius of the fibres. However, vegetal wools present particular challenges due to the non-cylindrical shape of their fibres and the high polydispersity of their fiber diameters, making it necessary to analyze the distribution of their fiber diameters.This study establishes a measurement protocol using scanning electron microscopy (SEM) to observe the cross-sections of several vegetal wools, approximating fiber profiles as ellipses to establish radius probability distributions. Four averaging methods are evaluated to determine the most appropriate effective radius for modeling.Three self-consistent models, Tarnow-Brinkman, Umnova-Johnson-Champoux-Allard-Lafarge, and Piégay, are used to simulate visco-inertial and thermal effects. A direct approach (single effective radius) and a composite approach (two effective radii distinguishing fiber types) are compared for sound absorption calculations. The simulation methods are then evaluated against acoustic measurements to identify the most accurate analytical estimates.This analysis concludes that, to calculate static airflow resistivity, it is preferable to use Tarnow's model in a direct or composite approach, taking as the effective radius an arithmetic or quadratic average. For sound absorption, the best approaches are those of Umnova-Johnson-Champoux-Allard-Lafarge in a direct or composite approach, and Piégay's in a composite approach with an arithmetic average radius.

Speaker: Clément PIEGAY (UMRAE Strasbourg)

585 Development and Characterization of Hemp Fiber Biocomposites with Biopolymer Binders for Acoustic and Thermal Insulation Applications

Given the global transition to a circular economy, there is a growing demand for sustainable, bio-based alternatives to synthetic materials used in construction. This study investigates the possibility of developing sustainable hemp fiber composites with biopolymer binders (chitosan, starch, and nanocellulose) and examines their applicability as acoustic insulation wall panels. Four types of composite panels were synthesized using the husk-free fibers of industrial hemp (Bialobrzeskie variety). The binder material contained 4% chitosan and 2% chitosan, while potato starch and nanocellulose (CNF) were added in two different variants. The 2 cm thick panels were produced by pneumatic spraying followed by hot pressing at 100 °C. Tests included three-point flexural strength measurement (EN 310 standard), thermal conductivity estimation (ISO 8301), and determination of sound absorption efficiency (α) using the Kundt impedance tube method. The panels were shown to have low to medium density, i.e., between 0.264 and 0.282 g/cm³. Acoustic analysis showed good high-frequency performance with the highest absorption coefficient recorded in the 3125–4000 Hz region α ≈ 0.9; this makes it quite effective for sound reduction at speech frequencies. In contrast, low-frequency performance (≤500 Hz) was observed to be poor (α < 0.2), which is expected from a porous fibrous material without air gaps. Thermal conductivity ranged from 0.047 to 0.052 W/(m·K), and heat resistance (R) varied between 2.41 and 2.79 m²·K/W, indicating good insulation properties. CNF-reinforced panels showed increased flexibility and energy dissipation, while chitosan-starch bonded panels reached a maximum flexural strength of 0.46 MPa.

Speaker: Thomas Frater (Wood K plus)

A12.02 Data-driven methods in acoustics and vibration: S082

Conveners: Marcus Maeder (Technical University of Munich), Elias Zea (Marcus Wallenberg Laboratory, KTH Royal Institute of Technology)

586 Generalization of Deep Reinforcement Learning for the Design of Broadband Bandgap Structures

andgap materials play a critical role in controlling wave propagation for applications in vibration isolation and acoustic filtering. Traditional topology optimization methods can effectively maximize bandgaps but are computationally expensive and limited in scalability. Data-driven approaches, including generative models and supervised learning, accelerate design discovery but remain constrained by the distribution of the training data and often lack strong generalization to unseen design environments. In this work, we propose a discrete soft actor–critic (SAC)-based deep reinforcement learning framework to autonomously design elastic structures with broadband bandgaps. The agent is first trained to optimize the bandgap in a 4 cm × 4 cm design space and subsequently evaluated in a smaller 3 cm × 3 cm environment to assess its generalization capability. The results demonstrate that the proposed reinforcement learning framework successfully discovers structures withbroadband bandgaps and generalizes to unseen design domains without additional training. Full-wave harmonic simulations confirm that the bandgap pre-dictions obtained from unit-cell analysis are consistent with the frequency response of the corresponding periodic structures. Compared with conventional generative modeling approaches, the proposed method learns the underlying relationship between geometry and bandgap formation through environment interac-tion, enabling improved versatility and physics-aware design discovery.

Speaker: Semere B. Gebrekidan (Institute of Mechanics and Mechatronic)

Full paper (PDF)

587 Data-Driven Equation Discovery in Weakly Nonlinear Sound Beams

In this study, we build on our previous article, in which we presented the possibilities of data-driven equation discovery in the field of nonlinear acoustics. Now we focus on another benchmarking task: finding equations for weakly nonlinear sound beams. To this end, we rely on simulations of the compressible NavierStokes equations. In this case, we are dealing with wave propagation from a piston-oscillating boundary condition. After the appropriate transformation toretarded time and making use of the weak formulation, we again apply our algorithm and are able to re-create the textbook case to a high degree of accuracy: thewell-known Khokhlov-Zabolotskaya-Kuznetsov (KZK) equation is discovered from data.

Speaker: Tereza Filipská (Czech Technical University in Prague)

Full paper (PDF)

588 Data-driven Dynamical Energy Analysis for modelling high-frequency noise and vibration

Predicting high-frequency noise and vibration in complex structures is a challenging task where standard numerical methods for wave-based models struggle with the rapid oscillations involved, leading to prohibitive computational costs. While methods like Statistical Energy Analysis (SEA) or ray tracing are popular alternatives, Dynamical Energy Analysis (DEA) offers a more robust approach by using a transfer operator to track ray densities in phase-space. The result is a high-frequency modelling tool that is neither subject to SEA’s strong restrictions on the wave behaviour within carefully chosen sub-structures, nor restricted by the reflection order. This research introduces a data-driven version of DEA based on Extended Dynamic Mode Decomposition. Here, the data driven construction of the transfer operator facilitates a division of the phase-space into clusters where the ray flow map is locally smooth. The accuracy may then be further enhanced by using Legendre polynomial basis expansions within these clusters and/or the further sub-division of the phase-space.

Speaker: David Chappell (Nottingham Trent University)

Full paper (PDF)

589 Data-driven modeling of strong acoustic fields in piston-driven resonators with boundary layer losses

The investigation of strong acoustic fields in closed, piston-driven resonators is a classical problem in nonlinear acoustics. One of the possible approaches consists in deriving a modified Burgers’ equation for this case and solving it by projecting onto a set of (coupled) harmonics. This procedure even provides certain asymptotic closed-form solutions. In this paper, a similar problem is addressed using data-driven methods. Specifically, symbolic regression is used to provide interpretable analytical expressions that relate the system’s inputs (material parameters, geometry, and excitation amplitude) to its outputs (the magnitude and phase of individual harmonics). The advantage of incorporating prior expert knowledge lies in its ability to constrain the search space and subsequently obtain extensions beyond the asymptotic regime, while seamlessly incorporating boundary layer losses, otherwise formulated through a cumbersome integro-differential expression. The resulting relations very well reflect the energy distribution along the harmonic cascade, and interesting information can also be drawn from the phase shifts when boundary layer losses are taken into account. The results presented correspond to the asymptotic solution by design, and their comparison with numerical validation is satisfactory.

Speaker: Viktor Hruška (Czech Technical University in Prague)

590 Structural Intensity Estimation from Noisy Plate Vibration Measurements Using Physics-Informed Neural Operators

Structural intensity provides valuable insights into vibrational energy flow and dominant transmission paths within mechanical structures. For thin-walled plates, it can be derived from the transverse displacement field and its spatial derivatives. However, obtaining reliable estimates from experimental data is difficult, since numerical differentiation amplifies measurement noise, especially at higher derivative orders. This contribution introduces a machine learning framework based on a physics-informed deep operator network that predicts structural intensity directly from noisy vibration measurements. Rather than differentiating the raw measurements by finite-difference schemes, the approach constructs a smooth, fully differentiable neural network surrogate representation of the displacement field, from which higher-order spatial gradients are obtained through automatic differentiation. This strategy avoids the noise-induced instabilities inherent in numerical differentiation and yields robust predictions of structural intensity from experimental data. The methodology is evaluated on two test cases: an analytical benchmark problem involving a simply supported plate, and laser Doppler vibrometry measurements of a plate structure. The results demonstrate accurate reconstruction of displacement and structural intensity fields over a broad frequency range, capturing both energy-flow magnitudes and directional patterns with high fidelity. Compared with standard numerical differentiation techniques, the proposed data-driven approach consistently provides higher robustness to noise and increased predictive accuracy, supporting its applicability in experimental structural dynamics and noise control engineering.

Speaker: Johannes D. Schmid (Technical University of Munich)

A12.12/A16.13 Diffuse Sound Fields and Reverberation: S088

Conveners: Franz Zotter (University of Music and Performing Arts), Albert Prinn (International Audio Laboratories Erlangen)

591 Quantification of Diffuseness: Toward Understanding Sound Fields in Reverberation Chambers

The existence of a diffuse sound field is a fundamental assumption underlying sound absorption measurements in reverberation chambers. Despite decades of research, diffuseness remains complex and challenging to characterize quantitatively. This contribution discusses recent advances in sound field analysis, including array-based characterization techniques, which have substantially improved our understanding of the physical mechanisms governing sound field diffuseness in reverberation chambers. These developments shed new light on the origins of the limited reproducibility of reverberation-chamber sound absorption measurements across laboratories. We present perspectives on the measurement of sound absorption, the future role of reverberation chambers, and the definition and use of absorption parameters in acoustic modeling.

Speaker: Mélanie Nolan (Universidad Politécnica de Madrid)

Full paper (PDF)

592 Why 2.5D Sound Field Synthesis Isn't Going to Produce Diffuse Sound Fields Well Enough

Reproducing surround audio via a horizontal ring of 3D point-source loudspeakers is termed 2.5D synthesis. Such layouts were recently proven incapable of producing spatially extended diffuse sound fields when fed with mutually uncorrelated signals. Could feeding the loudspeakers via 2.5D sound field synthesis help?We show that the horizontal ring layout of vertical line sources is the ideal 2D layout and permits full control with convolutional driving functions. Such functions are general enough to also include 2.5D WFS or 2.5D Ambisonics with spectral division. But no convolutional driving function can help a 2.5D layout synthesize extended diffuse fields from uncorrelated input. While active sound intensity must vanish in the ideal diffuse field, convolutional driving functions for 2.5D inevitably produce net inward sound intensity.

Speaker: Franz Zotter (University of Music and Performing Arts)

Full paper (PDF)

593 Statistical wave field theory: main results

The statistical wave field theory mathematically establishes the statistical laws of the solutions to the wave equation in a bounded domain. It provides the closed-form expressions of the power distribution and the correlations of the wave field jointly over time, frequency and space, which hold at high frequency and after many reflections, in terms of the geometry and the specific admittance of the boundary surface. This paper summarizes the main results of the theory.

Speaker: Roland Badeau (Télécom Paris)

Full paper (PDF)

594 Assessing the anisotropic features of late reverberation in a variable-acoustics hall

Late reverberation has long been assumed isotropic, i.e., showing a uniform distribution of energy over all angles of incidence. However, recent works leveraging Spherical Microphone Array (SMA) measurements and the associated spatial analysis tools have shown that the isotropy assumption is never perfectly realized. The extent and spatial characteristics of this phenomenon, and its dependence on room geometry, boundary conditions, and transducer positions, require further analysis. This work presents the results of a series of measurements conducted using the em64 Eigenmike, which supports 6th-order Ambisonic encoding, in the Espace de Projection at Ircam, a shoebox-shaped variable-acoustics hall. Spatio-temporal analysis of these measurements reveals a concentration of energy in horizontal axial and tangential modes, influenced by the room geometry. The absorptive boundary conditions induce a noticeably higher anisotropy than the reflective and diffusive ones. Furthermore, in the case of directive sources, the source orientation has a strong influence on the spatial patterns of anisotropy.

Speaker: Alice Pain (STMS)

Full paper (PDF)

595 Exploring theoretical limits of statistical room acoustics

Applicability of statistical room acoustic equations always assumes a diffuse sound field and a non-modal frequency range. In real-world applications this assumption and limit is rather theoretical, but still, these equations are used and are working satisfactorily in most cases. The paper discusses results of experiments to see practical limits of applicability of statistical approximations in more detail & beyond the well-known modal distance-based Schroeder-frequency.

Speaker: Andor Fürjes (aQrate Acoustics Ltd.)

Full paper (PDF)

596 A 2D numerical study of a diffuse field index

Absorption coefficient measurements of material samples in reverberation chambers depend on diffuse field theory. The diffuseness of a sound field depends on the geometry and sound absorption of a given chamber. In this numerical study, a spatially averaged intensity-based diffuse field index is computed as a function of random-incidence absorption coefficient and frequency in a 2D geometry, specifically the Bunimovich stadium. It is shown that the diffuseness of the sound field is reduced when the surface absorption is non-uniform. The effect of scattering and the number of sources on the diffuse field index are also briefly considered. These results highlight the challenge practitioners face when attempting to reliably measure absorption coefficients in reverberation chambers using methods that depend on diffuse field theory.

Speaker: Albert Prinn (International Audio Laboratories Erlangen)

Full paper (PDF)

A13.01 Guided waves for NDT & SHM applications: S090

Conveners: Lynda CHEHAMI (UPHF), Markus Saurer (University of Graz), Theodosia Stratoudaki (University of Strathclyde)

597 Ultrasonic Evaluation of Agra Marble Inlays (Parchin Kari/Pietra Dura) Using Machine Learning for Acoustic Signal Classification: A Heritage NDT Approach

Agra marble inlay work (also known as Parchin Kari, Pacchikari or Pietra Dura) is a traditional craftsmanship wherein carefully shaped hard stones are inserted into an engraved marble matrix, with the Taj Mahal being the best-known example. In spite of its historical importance and popular application to modern items, inlayed objects present significant challenges for material classification due to their complexity and heterogeneous nature.In this paper, we propose an ultrasonic system driven by machine learning algorithms for automated material classification in Agra-type marble inlay work. The acquired immersion ultrasonics data are subjected to pre-processing steps such as alignment, removal of noise artifacts, and feature extraction involving temporal, spectral and energy-related parameters. Material classification is performed based on ranking of features and training of the corresponding classifier.Our findings illustrate the potential of using non-destructive techniques together with a data-centric analysis workflow in order to classify the different materials of interest in complex inlays. In addition to the presented case study, the developed method can be extended to other cultural heritage objects as well.

Speaker: Esma Tuzovic (George W. Woodruff School of Mechanical Engineering)

Full paper (PDF)

598 Ultrasonic Ellipsometry by 3D Laser Vibrometry for SAW-based Viscoelastic Characterization

Viscoelastic materials are widely used in industrial structures, cultural heritage objects, and biomedical applications. Monitoring their mechanical properties is essential, as these properties may evolve due to mechanical loading, aging, or environmental variations such as humidity and temperature fluctuations. Such changes can indicate potential degradation or loss of structural integrity.For isotropic viscoelastic materials, a complete mechanical characterization requires identifying the frequency dependence of the complex Lamé coefficients. This is commonly achieved by analyzing wave–material interactions. Ultrasonic acoustic waves are particularly suitable for this purpose because they are non-invasive and can be generated and measured remotely. In most existing approaches, the estimation of the complex Lamé coefficients is performed in two steps: shear wave analysis is first used to determine the second Lamé coefficient, followed by longitudinal wave analysis to estimate the first coefficient.This work proposes a method to estimate both coefficients simultaneously by studying the multi-component interaction of surface acoustic waves (SAWs) with viscoelastic materials. SAWs exhibit elliptical polarization characterized by the H/V ratio and the orientation angle of the particle motion ellipse. These polarization parameters are estimated using the Quaternion Fourier Transform (QFT), while the complex wavenumber is extracted using the Prony algorithm. The inverse problem is solved using theoretical models of SAW propagation in viscoelastic media. The method is validated using both numerical simulations and experimental measurements.

Speaker: Aziz BOUZZIT (SATIE Lab - CYU)

599 On the role of shear-horizontal plate waves for anisotropic stiffness characterization

Elastic waves in plates are commonly classified by their polarization. Shear-horizontal (SH) waves exhibit in-plane displacement orthogonal to the wave vector, whereas Lamb waves involve the remaining two displacement components. In isotropic plates, these polarizations are fully decoupled, and Lamb waves alone suffice to characterize the material completely. This contribution examines how far this picture holds for anisotropic media. The analysis is complicated by the fact that SH and Lamb polarizations generally couple in anisotropic plates, forming a single family of guided waves. While this suggests that measuring any such wave should in principle allow full material characterization, the conclusion is not straightforward once measurement selectivity is taken into account. Our laser-ultrasonic system has a polarization bias: the interferometric detection is sensitive only to out-of-plane displacement, while the laser excitation on a metallic sample is dominated by in-plane forcing. As a result, only Lamb-like waves are readily observed. To quantify this bias, we introduce the polarization angle — defined as the angle between a wave's polarization vector and the SH direction — so that 0° denotes a pure SH wave and 90° a pure Lamb wave. Using this metric, we perform a sensitivity analysis of laser-ultrasonic measurements with respect to the stiffness parameters of an orthotropic medium. The analysis demonstrates that SH-like waves carry information that is not accessible from Lamb-like waves alone and that their acquisition is essential for unique identification of the full orthotropic stiffness tensor. These findings are illustrated through a two-dimensional wavefield scan on a cold-rolled steel plate.

Speaker: Daniel A. Kiefer (Institut Langevin, ESPCI Paris, Université PSL, CNRS)

600 The Use of a Spatial Light Modulator for Increased Flexibility and Efficiency in Laser Ultrasound Array Imaging

This paper explores how a spatial light modulator (SLM) can improve 3D imaging using laser induced phased arrays (LIPA), by reducing data acquisition time, increasing efficiency and adding flexibility to the existing LIPA platforms.2D LIPAs have enabled remote, couplant-free volumetric imaging for non-destructive evaluation (NDE) [1]. Current LIPA platforms, that scan a single point-focused laser beam, easily exceed the damage threshold of the irradiated component. NDE imaging necessitates laser beam attenuation, which makes inefficient use of laser energy resources. Extensive signal averaging is used to improve low signal-to-noise ratio (SNR), at a significant cost to data acquisition time. Full Matrix Capture, where all generation–detection combinations are captured, enables a wide range of imaging angles and the detection and characterisation of diverse defect types but is time-consuming. Reducing single measurement time significantly impacts overall acquisition time. The SLM facilitates arbitrary laser generation patterns, distributing the incident laser energy to an array of focused points, each below the damage threshold, thereby making more efficient use of a high-power laser. This multi-point excitation enables spatial encoding techniques such as Hadamard multiplexing for parallel averaging of multiple signals with a single detector, reducing data acquisition time. The results presented will be compared with those from scanning-based LIPA platform.Additionally, this SLM-based setup with multi-channel laser ultrasonic generation introduces flexibility into generating arbitrary array designs using all optical means. 2D spiral and random LIPA configurations will be presented to compare array performance and flexibility of array optimisation between platforms. [1] P. Lukacs, et al. IEEE Trans. Ultra. Ferro. Freq. Cont., pp.1, 2025.

Speaker: Hilde Metzger (University of Strathclyde)

601 Ultrasonic imaging of internal defect incurred by corrosion in structural steel used in lattice tower

Electricity transmission towers situated in open environments are continuously exposed to atmospheric and industrial impacts. Corrosion occurs inevitably under the impacts of these external factors. This study discusses the methodology of detecting and quantifying the internal defect created by corrosion using ultrasonic wave technique. Experimental sample was part of an onsite tower and made of painted steel, it has a slide-shaped corrosion defect. Experimental results reveal that wave energy measured in a particular corroded area is at least 10 times higher than that measured in pristine area. Also, a shift in ultrasonic wave vibration to low frequency range is also recorded in that area. Then the corrosion geometry is profiled and imported to a finite element model for further investigation. However, there is no such effect observed from the simulation. This suggests the energy amplifying effect is not caused by the corrosion geometry, rather by the internal feature of the corrosion defect. Further finite element analysis suggests that by analyzing the frequency content of the signals, the size of the defect and its approximate internal features, for example its depth to the surface and its approximate shape can be revealed.

Speaker: Kim Ming NG (SATIE Lab - CYU)

A15.06 Psychoacoustics of everyday sounds: S111

Conveners: Daniel Oberfeld-Twistel (Johannes Gutenberg University Mainz), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

602 Annoyance penalty of impulsive noise

Many kinds of sources are impulsive, e.g., shooting ranges, logistic centres, playing fields, and construction sites. Many guidelines require an adjustment (penalty, sanction) to be applied, if noise is impulsive. Penalty is usually a constant value, k [dB], that is added to the measured sound level, LAeq [dB]. Adjusted value, LAeq+k, should describe noise annoyance better than LAeq alone. ISO/PAS 1996-3 (Model A) provides model where the penalty k depends on measured impulse properties: onset rate (Ron) and level difference (DL). However, experimental evidence showed that Model A overestimates subjective annoyance for many types of impulsive sounds. Our purpose was to test alternative models AE and to determine their accuracy using experimental psychoacoustic data. Model B was a modification of Model A. Model C was based on European directive. Model D was based on alternative level difference. Model E was based on statistical sound level variability. All new models BE predicted both subjective annoyance, and annoyance penalty, with higher precision than Model A. Model E is recommended for practical measurements because of its simplicity. The model is applicable for the impulsive noise within 4565 dB LAeq, which is the most usual range occurring in residential yards.

Speaker: Valtteri Hongisto (Turku University of Applied Sciences)

Full paper (PDF)

603 Annoyance penalty of broad-band steady-state noise due to the spectrum shape

Most guidelines involve a penalty procedure, if noise contains features which increase noise annoyance, such as impulsive or tonal components. In most guidelines, the penalty is a constant, k [dB], that is added to the measured sound level, LAeq [dB]. The adjusted sound level, LAeq+k, is expected to describe noise annoyance better than LAeq alone. However, soundscapes, both indoors and outdoors, have very often a broadband steady-state (BBSS) nature, which lack temporal variations and tonal features. Evidence suggests that the annoyance of BBSS sound depends on spectrum shape. Our purpose is to summarize the psychoacoustic research, which developed a model to calculate the penalty of a BBSS sound, when the spectrum is measured in 1/3-octave bands. Kuusinen et al. (2023) reported the noise annoyance of 23 spectrally different BBSS sounds at levels 32, 40, and 48 dB LAeq. The annoyance penalty values were within 0 and +12 dB, which suggests that penalty is justified also for BBSS sounds. The penalty obtained for a specific BBSS sound was independent on sound level within 3248 dB LAeq which supports that their penalty values might be applied also beyond that range. In the follow-up study, Kuusinen & Hongisto (2026) developed a calculation model, which reproduced the experimentally determined annoyance penalty values with high reliability (R2=0.79). The input variable of the model was spectral centroid, which is calculated from the 1/3-octave band spectrum of BBSS sound. The model can be applied in the assessment of noise annoyance for measurements conducted outdoors and indoors.

Speaker: Valtteri Hongisto (Turku University of Applied Sciences)

Full paper (PDF)

604 Perception of Dissonance in Time-Varying Technical Sounds

While dissonance is traditionally studied in musicalcontexts using rather stationary tonal structures, technicalsounds require a perceptual perspective in whichdissonance may also be assessed for time-varying spectralcontent.This study investigates the perception of dissonanceand unpleasantness in synthetic acceleration and decelerationsounds typical for electric vehicle interiornoise. A laboratory listening experiment was conductedwith 12 participants across two appointments.Long (8 s) instationary sounds and short (1 s) stationaryand time-varying excerpts were rated. The stimuliconsisted of broadband background noise combinedwith tonal orders, presented at two nominal tone levels.Participants rated unpleasantness and perceiveddissonance on a categorical scale.Results show that perceived dissonance is primarilydetermined by interval structure, as long as tonesare sufficiently audible. Sounds are perceived moreunpleasant, when higher frequencies are present andfor more dissonant interval combinations, particularlyfor the higher tone level. Deceleration sounds arerated slightly less unpleasant than acceleration sounds,while dissonance ratings show no systematic differencebetween the two conditions. Comparisons betweenstationary and time-varying short stimuli reveal onlyminor differences.

Speaker: Benedikt Bugl (OTH Regensburg)

Full paper (PDF)

605 Cue weights in auditory time-to-collision estimation for approaching vehicles

For a pedestrian standing at the curb and intending to cross the road, the dynamic spatial sound field generated by an approaching vehicle provides various cues to its arrival time (time-to-collision, TTC). The sound intensity at the listener position increases with decreasing vehicle distance. The azimuthal position of the vehicle and the angular separation of the vehicle sound sources (e.g., left and right front tires) change with distance. Here, we investigated which cues are most relevant for auditory TTC judgments. Using acoustic simulations of approaching vehicles, the TTC and speed signaled by the dynamic intensity cues were shifted against the TTC and speed signaled by the change in the simulated spatial position of the simulated sound sources (i.e., the angular cues). This is only possible in virtual environments. In the real world, these cues are inseparable. The vehicle sound power and the vehicle size were also varied. The results showed a significant association of both intensity cues and angular cues with the estimated TTC. However, general dominance weights indicated a higher importance of intensity cues compared to angular cues. Also, the sound level at the time of TTC estimation, a “heuristic” cue, showed a higher cue importance than the TTC signaled by the dynamic intensity cues. At a given presented TTC, participants estimated shorter TTCs for vehicles with higher sound power, compatible with previous results. Taken together, the data provide novel insights into which acoustic cues pedestrians use for auditorily estimating the TTC of an approaching vehicle. The observed cue weights are discussed in relation to the reliability of the different acoustic cues to TTC.

Speaker: Daniel Oberfeld-Twistel (Johannes Gutenberg University Mainz)

606 Effects of carrier characteristics, stimulus duration, and presentation mode on spectro-temporal modulation detection

Many natural sounds contain modulation across both the temporal and spectral dimensions. The ability of human listeners to detect such spectro-temporal modulations (STM) has often been assessed using stimuli consisting of a STM pattern (moving ripples) imposed on a carrier signal, measuring the listeners’ detection threshold in terms of the modulation depth. The present study systematically investigated how different carrier signals, different stimulus durations, and different presentation modes affect STM detection in a group of young listeners with normal hearing. Using the most common STM configuration with a spectral modulation frequency of 2 cycles/octave and a temporal modulation rate of 4 Hz, stimuli were constructed using either noise or complex-tone carriers. The thresholds for these two broadband carrier types were measured in a reference condition using 1-second stimuli and binaural headphone-based presentation in a 3-alternative forced-choice paradigm. In addition, the effects of reduced bandwidth, shortened stimulus duration, and monaural presentation were measured for the two types of carriers. The results indicate that STM detection was substantially easier with the complex-tone carrier than with the noise carrier in the reference condition. Furthermore, a large variation in performance was found across listeners for the complex-tone carrier but not for the noise carrier. Interactions between the type of carrier and the effects of reduced bandwidth, shortened stimulus duration, and monaural presentation were observed. The potential mechanisms behind these findings and connections to other studies will be discussed.

Speaker: Johannes Zaar (Eriksholm Research Centre)

607 Magnitude of tonal content as function of level

There is an increasing number of sound sources in our environment emitting sounds containing tonal components. Their perception is commonly assessed bymeasuring the magnitude of tonal content, also known as the tonality of the sound. While several studies investigated the impact of the level of tonal content relative to the level of the noise background, little is known about the effect of the overall level on the magnitude of tonal content. This study presents data collected in three German laboratories (Berlin, Magdeburg,Dessau) on the effect of a change in overall level of the sound (keeping the tone-to-noise ratio constant) on the magnitude of tonal content. Technical andartificial sounds were used. The latter sounds were partly basic stimuli (pure tone in noise) and partly stimuli, which were motivated by real sounds of electric vehicles. In total, 77 listeners participated in the experiment; none of them reported hearing impairment. For most stimuli, the measured magnitude oftonal content increases with level. This is in contrast to tonal methodologies such as tone-to-noise ratio and prominence ratio that were defined in national (DIN 45681) and international standards (ISO/TS 20065).

Speaker: André Fiebig (TU Berlin, Department of Engineering Acoustics)

Full paper (PDF)

A18.00 Soundscape, Environmental Quality, Health and Well-being: S339

Conveners: André Fiebig (TU Berlin, Department of Engineering Acoustics), Arezoo Talebzadeh (Ghent University)

608 UAV Noise In Human Resources Research Response Research: A Framework Based On Scoping Review

As multirotor UAVs become increasingly used in urban logistics, emergency response, and future low-altitude transport, noise remains a major obstacle to public acceptance. While existing UAV-noise reviews have focused on propagation, source characteristics, psychoacoustic interpretation, and regulatory assessment, less attention has been paid to the methodological design and documentation of human-response experiments themselves. To address this gap, this paper conducts a scoping review of 35 studies. It first examines the acoustic characteristics of UAV noise, and then reviews how existing studies design and implement human-response experiments. Based on this synthesis, it proposes a structured framework covering both signal-related and procedural aspects, including source acquisition, metadata control, signal preservation, propagation normalisation, participant screening, playback calibration, trial sequencing, and reporting. By integrating acoustic realism with protocol traceability, the framework aims to improve the rigour, reproducibility, and ecological validity of future psychoacoustic and psychophysiological research on UAV noise.

Speaker: Tianjing Feng (Institute for Environmental Design and Engineering)

Full paper (PDF)

609 Impact of Acoustic Characteristics on Authenticity in Historic Districts: Mechanisms and Parameter Optimization

Authenticity serves as a core criterion for evaluating cultural heritage, reflecting the capacity of historic districts to convey tangible and intangible narratives accurately. However, research on authenticity has predominantly focused on visual morphology, leaving the quantitative impact of acoustic characteristics relatively underdeveloped. This study employs audio-visual simulations and interactive experiments conducted via a custom-developed platform to investigate the commercial spaces of the Zhonghua Baroque Historic District in Harbin, China. The research identifies the distributional features of acoustic characteristics, reveals the correlation patterns between these features and perceived authenticity, and determines the optimal parameter ranges for authenticity enhancement. It was found that fidelity of dominant sound and equivalent sound pressure level (LAeq) are the primary predictors of authenticity perception. The results indicate a dynamic compensation mechanism: as background sound levels increase to 55–72 dB, the optimal dominant sound LAeq shifts to 60–65 dB and the required fidelity threshold increases to 35 dB to maintain historical character. In conclusion, this study establishes a scientific basis for the preservation of authenticity in historic urban environments.

Speaker: Xuejun Hu (Harbin Institute of Technology)

Full paper (PDF)

610 A Comparison of the Soundscape of Three Animal Shelters

Animal Shelters are unique facility types with a range of human and animal user groups. These facilities temporarily house animals until they are typically returned to their owners or adopted. One of the most common species in an animal shelter is the domestic dog, whose bark can be loud and dominates the soundscape. The animal shelter building form, particularly the dog housing configuration, can have a large impact on both the noise within dog housing spaces as well as throughout the entire facility. By looking at acoustic metrics such as Reverberation Time, alpha bar, along with kennel occupancy, bark rate at various times of the day, average and maximum bark level and number of dog barks during the day, one can begin to understand how different shelter designs can impact acoustic metrics. By following the sonic flows, acoustic itineraries, measuring acoustic properties, and understanding the acoustic calendar, we can begin to understand the impact of kennel room configurations on the overall soundscape. Comparisons of three very different facilities can inform interventions in existing facilities across the world in conjunction with animal care best practices.Incorporating knowledge of a range of shelter’s soundscapes allows architects and designers to holistically address the components of the soundscape that can be altered by the physical building.

Speaker: Siebein Keely (Siebein Acoustic)

Full paper (PDF)

611 Individualized Tipping Points: How Noise Sensitivity Modulates the Vibrancy-Chaos Crossover in Restaurant Soundscapes

Traditional indoor soundscape research often seeks a universal optimal level of acoustic activity, balancing positive high-energy attributes (vibrancy) with negative ones (chaos). However, this approach assumes a homogenous population. This study presents a secondary analysis of two field datasets (N=806 patrons in 36 restaurants) to investigate how individual noise sensitivity moderates the perception of restaurant soundscapes. Specifically, the "Vibrancy-Chaos Crossover" is examined — the point at which increasing ambient loudness ceases to be perceived as pleasantly "vibrant" and begins to be perceived as stressfully "chaotic." The analysis reveals that while increased loudness generally correlates with higher vibrancy scores, the perception of chaos escalates at significantly different rates depending on the subject's noise sensitivity. For highly sensitive individuals, the tipping point where chaos overrides vibrancy occurs at much lower loudness levels than for less sensitive patrons. Consequently, soundscape pleasantness declines much more steeply for this group as loudness increases, leading to a direct drop in venue recommendation rates. These results suggest that a single "sweet spot" for restaurant acoustics does not exist. To achieve "acoustic fairness" and inclusiveness in gastronomy, room acoustical design should move away from static targets and instead offer diverse, zoned acoustic affordances that accommodate varying sensory thresholds.

Speaker: Jochen Steffens (Hochschule Düsseldorf)

A04.05 Yesterday, today and tomorrow of research in acoustics: S028

Conveners: Stefan Weigand (SoundPLAN GmbH), Iván Herrero-Durá (SoundPLAN GmbH), Kristian Jambrošić (University of Zagreb Faculty of EE and Computing)

612 Approaches and technologies: what changed in environmental acoustics research?

Acoustics' research has dramatically changed over the last 50 years and will further change in the next few years. From a physical-based approach which drove research in the last decades of the past century, where measurement and simulations have allowed characterising, or predicting, existing and future acoustic environments, research started to focus more and more on the effects that noise has on individuals. From the dose-response curves of noise annoyance up to the birth of the soundscape approach in environmental acoustics, perceptual and emotional aspects have gained even more importance, aiming to create environments where the human experience can be controlled or manipulated. A good maturity level reached by virtual reality technologies, as well as the development of even more miniaturised devices, has furthermore extended research in the field of cross-modal interaction of noise, or sound, with the objectives (i.e. vision, light, smell) and subjective (e.g. perceptual, emotional, healthy, cultural and social valence) factors, measured by subjective self-report questionnaires or physio-/neuro-logical responses to the acoustics external stimuli. The collection of data flow from both physical and human's domains will feed the artificial intelligence models to support urban planning and decision-making.

Speaker: Massimiliano Masullo (Università degli Studi della Campania “Luigi Vanvitelli”)

613 30 Years Working with Strategic Noise Mapping: A Retrospective in Europe

Since the early 1990s, environmental noise has become increasingly established as a central element of European public health policy. The formal adoption of the Environmental Noise Directive (END, 2002/49/EC) marked a turning point, establishing a common approach to assessing exposure to environmental noise across the EU and requiring Member States to produce Strategic Noise Maps (SNM) every five years for all major roads, railways, airports, and urban areas. Three decades of experience have generated a growing body of knowledge and fresh perspectives, but enduring challenges remain.From a public health perspective, the urgency of SNM is undeniable. More than 112 million people in Europe are exposed to harmful levels of environmental noise pollution, and road traffic noise contributes significantly to this public health problem in many urban areas. These figures, together with the European Green Deal’s objectives of reducing the number of people disturbed by noise, underscore the importance of having harmonized mapping tools.Methodologically, the harmonization process has been arduous, and it was not until the fourth round of SNM that such process took place, with the CNOSSOS-EU methodology serving as a coherent approach for assessing noise levels from major sources across Europe. This overcame the fragmented national methods that hindered cross-border comparability in the early rounds of the SNM.The presented retrospective examines the evolution of noise management in Europe across four rounds of SNM: its policy achievements, methodological advances, as well as the gaps that still persist in harmonization.

Speaker: Iván Herrero-Durá (SoundPLAN GmbH)

Full paper (PDF)

614 Urban mobility and noise exposure: insights from cohort data analysis and multi-agent modelling approaches

Increasing urbanization intensifies mobility demand, leading to increased noise and air pollution exposure and reinforcing socio-spatial inequalities. Addressing this challenge requires a better understanding of interactions between transport policies and exposure distributions, as well as decision-support tools to evaluate emerging mobility patterns, including multimodality and active transport. The SYMEXPO project developed a systemic framework to assess mobility-related exposures using a multi-agent approach, in which urban dwellers are represented as agents moving through dynamic pollution fields. This presentation focuses on the project’s contributions in environmental acoustics. Based on the MobiliSense cohort, analyses highlight the key role of activity patterns and transport modes in shaping high noise exposure levels. In parallel, integrated modelling chains were developed to simulate spatiotemporal variations in noise at neighbourhood and metropolitan scales in the Lyon Metropolis. These rely on two traffic models, SYMUVIA and MATSim, coupled with the NoiseModelling software. The presentation will highlight key methodological advances, discuss current limitations, and emphasize the contribution of dynamic modelling to multidimensional exposure assessment. Particular attention will be paid to acoustic exposure, with case studies focusing on the identification of critical noise exposure areas and multicriteria assessments of traffic restriction strategies around schools.

Speaker: Arnaud CAN (UMRAE)

615 Seven Decades of Acoustic Measurement Technologies at the Department of Electroacoustics, FER, University of Zagreb

This paper presents the historical development of research and teaching infrastructure at the Department of Electroacoustics of the Faculty of Electrical Engineering and Computing, University of Zagreb, the only university-level department in Croatia fully dedicated to acoustics. Established in 1954, the department’s early research relied on analog instrumentation. The paper outlines several key pieces of equipment that defined this period, including early sound level meters, measurement microphones, audiometers, ultrasound measuring systems, but also analog sound mixers, audio microphones and loudspeakers. These instruments formed the technological foundation for research in room acoustics, electroacoustics, hearing studies, and environmental noise.Particular emphasis is placed on the gradual transition from analog measurement and signal processing toward digital technologies during the late 20th and early 21st centuries. Because research funding was often limited, the department adopted a pragmatic approach to modernization, combining legacy analog equipment with emerging digital systems. The paper discusses how the introduction of digital measurement platforms, computer-based analysis and advanced measurement software expanded experimental possibilities and improved the precision and efficiency of acoustic research. Through selected examples, the paper illustrates how equipment acquisition decisions influenced research directions, laboratory capabilities, and teaching outcomes, highlighting the broader technological transition that shaped acoustic research at the turn of the 21st century.

Speaker: Kristian Jambrošić (University of Zagreb Faculty of EE and Computing)

A13.04 Photoacoustic Imaging and Spectroscopy: S349

Conveners: Jan Laufer (Martin Luther University Halle), Ben T Cox (University College London), Günther Paltauf (University of Graz), Robert Nuster (University of Graz), Nico F. Declercq (GeorgiaTech-CNRS IRL2958, Georgia Tech-Europe)

616 From Graft Quality to Organ Repair: Ultrasound and Photoacoustic Imaging in Kidney and Liver Transplantation

Donor-derived graft fibrosis and ischemia reperfusion injury (IRI) remain a major determinant of transplant outcomes, yet perioperative tools provide limited real-time information on graft fibrosis, oxygenation, and response to emerging organ-repair therapies. One such therapy is mitochondrial transplantation (MTx) which delivers autologous mitochondria to injured grafts to restore bioenergetics, reduce oxidative stress, and support repair. Here, we develop ultrasound (US) and photoacoustic (PA) biomarkers for transplant quality fibrosis assessment and for monitoring mitochondrial transplantation to mitigate IRI.We have pioneered the use of quantitative ultrasound (QUS) analysis of radiofrequency backscatter with multiwavelength PA imaging to interrogate structure and function across clinical and large-animal transplant models. In a trial of 61 patients, pre-perfusion QUS H-scan analysis assessed donor-derived kidney fibrosis, while post-reperfusion PA spectra quantified hemoglobin oxygenation and acoustic-frequency signatures of IRI. In porcine kidney and liver IRI models, US/PA was used during normothermic perfusion or intraoperative reperfusion to evaluate autologous MTx, using oxygen saturation and backscatter/perfusion metrics as readouts of therapeutic response.In human kidney grafts, QUS-derived fibrosis strongly correlated with histology and predicted renal function at one year, while PA oxygenation distinguished reoxygenation differences in living and deceased donors. In liver IRI, MTx improved oxygen recovery and re-normalized the US backscatter patterns relative to controls. In kidney perfusion models, MTx improved renal artery flow, reduced intrarenal resistance, and increased cortical oxygenation.Together, these studies show the potential of US/PA for examining graft quality, monitoring emerging repair therapies in real time, and predicting organ recovery before irreversible dysfunction occurs.

Speaker: Eno Hysi (University of Toronto)

617 High-Resolution, High-Speed OR-PAM Using Fabry-Pérot Sensors and Raman Laser Excitation

Optical resolution photoacoustic microscopy (OR-PAM) is a powerful technology for high-resolution imaging of biological tissues. The integration of optical ultrasound detectors enables highly sensitive and broadband detection of photoacoustic waves up to the high-frequency range (>100 MHz), while maintaining a small detector size. This facilitates high axial and lateral resolution. This work utilizes planar Fabry-Pérot sensors in a coaxial excitation and detection geometry to minimize the distance between source and detector, thereby maximizing the signal-to-noise ratio (SNR). Additionally, a fiber-based Raman excitation laser with a high repetition rate (>100 kHz) was employed to achieve high imaging speeds (>1 Hz, volume = 1 mm³). Time-resolved photoacoustic signals were acquired using both raster scanning and, for higher speeds, continuous scanning modes. For functional imaging of phantoms, two wavelengths were used. A potential application of this highly sophisticated technology in clinical studies is envisaged, for instance in functional imaging of oxygen transport in post-COVID patients.

Speaker: Lea Steigemann (Martin-Luther-Universität Halle-Witten)

618 Photoacoustic Microscopy with Camera based Ultrasound Detection

Photoacoustic microscopy is an imaging technique mainly applied to biological tissue, using short laser pulses to excite acoustic signals in optically absorbing structures. The standard approach in photoacoustic microscopy is the time-resolved detection of the generated sound waves using a single-element ultrasound detector. This has the advantage that, ideally, these sensors are acoustically focused and aligned confocally with the optical focus, thereby achieving high sensitivities. The 3D dataset for image generation is obtained from the recorded ultrasonic depth profiles at various lateral positions on the sample surface. However, to maintain the advantage of high sensitivity in imaging with a large field of view, it is necessary to scan the sample relative to the fixed arrangement of the excitation laser spot and the acoustic focus, which results in reduced imaging speed. This can be addressed, for example, through parallelization using multiple ultrasound sensors simultaneously.This work demonstrates ways in which this parallelization of ultrasound detection using an optical camera can be implemented for application in photoacoustic microscopy. Depending on the specific experimental setup, cross-sectional images or projections of the generated ultrasound field at a specific time can thus be efficiently recorded. In combination with structured optical excitation, such as focused multipoint or multiline excitation as well as coded pattern excitation, a variety of possibilities arise, which are presented in this work using simulations and discussed in terms of their advantages and disadvantages. In addition, results from proof-of-principle experiments are shown that were obtained using the camera-based ultrasonic detection projection method in combination with line-pattern excitation for photoacoustic microscopy.

Speaker: Robert Nuster (University of Graz)

619 Characterising Skin Tone Effects in Clinical Photoacoustic Imaging

Photoacoustic imaging (PAI) is a promising non-invasive technique for clinical diagnostics, combining optical contrast with ultrasound resolution to visualise functional tissue properties. However, melanin absorption in the epidermis introduces significant effects in photoacoustic measurements, leading to spectral distortion and image artefacts. These effects impair the accuracy of key biomarkers, such as blood oxygen saturation, and may compromise diagnostic reliability across diverse skin tones. To quantify the impact of skin pigmentation, we conducted a human imaging study with 42 healthy volunteers across all six Fitzpatrick skin types [Else, 2025]. Multispectral PAI were acquired at multiple anatomical sites alongside colorimetry-based skin-tone quantification. Previous analysis shows a strong dependency of photoacoustic intensity on melanin concentration, demonstrating melanin as a major confounding factor in PAI. To correct these effects, we developed a two-stage pipeline combining orthogonal projection-based acoustic clutter removal with data-driven optical spectral decoloring. The projection method identifies and removes melanin-like superficial signal components that spatially overlap with vascular signals. For spectral decoloring, large-scale optical and acoustic simulations inspired by the clinical study were used to train a digital-twin-based regression model to estimate skin mask, melanin concentration, and wavelength-dependent fluence correction factors.Clinical validation demonstrated improved consistency of arterial sO₂ estimates across participants after correction. In representative arterial measurements, mean sO₂ increased from approximately 0.57 in raw data to 0.80 after acoustic clutter removal and spectral decoloring, approaching physiologically expected arterial oxygenation. These findings show that skin pigmentation effects in clinical PAI arise from coupled optical and acoustic mechanisms, and support clinically validated correction strategies for more reliable imaging across diverse skin tones.

Speaker: Amy Yijie Zheng (University of Cambridge)

A04.05 Yesterday, today and tomorrow of research in acoustics: P428

Conveners: Stefan Weigand (SoundPLAN GmbH), Iván Herrero-Durá (SoundPLAN GmbH), Kristian Jambrošić (University of Zagreb Faculty of EE and Computing)

620 Building Acoustics in Spain: Past Developments, Current Practice, and Future Challenges of Eduardo Torroja Institute

The Eduardo Torroja Institute for Construction Sciences (IETcc), part of the Spanish National Research Council (CSIC), conducts scientific research and technological development in construction and building materials, including the acoustic performance of materials and building systems. Its activity contributes to pre-standardisation research aimed at improving the Technical Building Code (CTE).In recent years, building acoustics in Spain has evolved from prescriptive regulations based on standardised construction solutions towards performance-based design supported by advanced measurement and modelling methods. This paper provides a critical review of this transition, linking past developments, current practice, and future challenges in the field.The “yesterday” stage is characterised by the development of reference solutions and early acoustic regulations. The current national framework, the Basic Document for Noise Protection (DB HR) of the CTE, was published 20 years ago and developed by the Construction Quality Unit of IETcc-CSIC.At present, IETcc-CSIC conducts pre-standardisation research and actively promotes performance-based approaches through harmonised standards, predictive tools, and their integration into regulatory systems. This work is supported by its acoustic testing laboratory and participation in national and international standardisation committees. Looking ahead, ongoing efforts focus on updating the regulatory framework.

Speaker: Teresa Carrascal García (Instituto Eduardo Torroja for Construction Sciences - IETcc-CSIC)

621 Decibels and its use for public outreach on safe listening

The decibel scale is a cornerstone of acoustic measurement, and professionals routinely navigate between equivalent levels, peak values, exposure levels, and a range of weightings and penalties depending on the application. Its legitimacy is solid from an instrumental perspective and deeply rooted in human sound perception, as it conveniently compresses a 140 dB dynamic range into steps close to the minimal audible level difference (Weber’s law). This is also its weakness. In the context of hearing-damage risk, equal‑energy measures are required to quantify the administered dose, which combines average sound level and exposure duration. This dose must be compared to another average equivalent level that also accounts for exposure on surrounding days, since human hearing accumulates effects over periods of weeks, and possibly longer. The personal area networks formed by our many digital devices, together with their access to clouds and user accounts, may in future provide solutions for managing such data and supporting a safe listening lifestyle. Exposure could, for example, be quantified in pascal‑squared‑hours and accumulated as shared information across platforms, allowing services to combine exposure data from both work and leisure. This measure was proposed more than 40 years ago, partly because it decompresses the logarithmic scale and reveals the exponential growth of dose over time. In other words, the numbers conveniently explode as levels call for concern. This paper discusses this measure as the new black in public outreach and explores how noise points can be counted much like footsteps, calories, alcohol units, or fruit intake to support healthy lifestyles.

Speaker: Max Væhrens (Department of Electronic Systems, Aalborg University)

622 Three (More) Years of the EAA's Young Acousticians Network Outreach and Dissemination Activities

This paper outlines the recent advances in communication and community-building initiatives led by the European Acoustics Association’s (EAA) Young Acousticians Network (YAN) between 2023 and 2026. By leveraging a cross-platform digital strategy primarily focused on LinkedIn, Instagram, and Discord, the YAN has achieved significant organic growth, effectively bridging the gap among early-career researchers, industry professionals, and senior academics. Key dissemination series, such as ‘The Science of Acoustics’ and ‘Acoustics Distilled’, are analysed alongside performance metrics that show increased followers and engagement on Instagram and LinkedIn. Furthermore, the paper evaluates the impact of the YAN’s Mentoring Program, an international initiative fostering one-to-one professional development, and the role of specialised online webinars in promoting acoustics knowledge. These efforts collectively strengthen the European research ecosystem by promoting inclusivity, visibility, and knowledge transfer for the next generation of acousticians.

Speaker: Marica Belmonte (Universitat Oberta de Catalunya)

Full paper (PDF)

A10.01 Acoustic, vibroacoustic and elastic metamaterials: P474

Conveners: Steffen Marburg (Technical University of Munich), Bart Van Damme (Empa, Materials Science and Technology)

623 Shaping Acoustic Back Radiation of Loudspeakers via Subwavelength Corrugated Surfaces

The interaction between the enclosure and the acoustic back radiation generated by a loudspeaker driverplays a crucial role in shaping the overall emission characteristics of the loudspeaker system. In this study,Subwavelength Corrugated Surfaces (SCSs), a classof acoustic metasurfaces, are proposed to passivelycontrol the internal propagation of acoustic back radiation and its angular redistribution. The proposedapproach enables the refocusing of the back radiationin lateral directions, while preserving an on-axis soundpressure level comparable to that of a conventionalclosed-box loudspeaker within the frequency rangewhere the metasurface is effective. The behavior ofthe system is assessed through a numerical investigation based on finite-element modeling in COMSOLMultiphysics. Three configurations are analyzed: aconventional closed-box system, a loudspeaker systemwith an open side wall, and a system with open sidewall incorporating the SCS. The simulations allowa direct comparison between the different solutionsand highlight the potential of SCSs to shape loudspeaker acoustic back radiation and extend directivitycontrol through compact, purely passive structuralmodifications.

Speaker: Letizia Chisari (University of Sussex)

Full paper (PDF)

A10.10 Sustainable Materials for Acoustic Applications: P499

Conveners: Francesco Pompoli (University of Ferrara), Philippe Glé (Cerema, Univ. EIffel, UMRAE)

624 Design And Evaluation of New Sound Absorbing Panels in Perforated Honeycomb Cardboard and Cellulose Fibers

This article presents the preliminary results of a study conducted by the Department of Architecture at the University of Florence in collaboration with the Italian National Consortium for the Recovery and Recycling of Cellulose-Based Packaging (COMIECO), focusing on the use of recycled cellulose-based materials for acoustic comfort in indoor environments. The study investigates various configurations of honeycomb cardboard panels (smooth or laser perforated with 1, 5, and 10 mm holes) coupled with layers of cellulose flakes or polyester fibers. The main objective is to overcome the limitations of traditional paper-based materials, which are typically effective only at medium and high frequencies, in order to achieve significant performance across a broad range of the sound spectrum. Experimental tests, conducted in a simplified reverberant environment and validated using a pressure-velocity (p-u) probe, demonstrate the effectiveness of acoustic and membrane resonance mechanisms, the latter being particularly active around 200 Hz in unglued panels. The results confirm that these eco-design solutions can offer performance comparable to traditional synthetic materials, supporting the principles of the circular economy in the construction sector.

Speaker: Lucia Magnini (Department of Architecture, University of Florence)

Full paper (PDF)

625 Evaluation of Grape Stem-Based Porous Materials for Advanced Sound Absorption Applications

In this study, sound-absorbing materials were developed from grape stems, an abundant agricultural byproduct, as porous acoustic absorbers. Four types of porous samples were prepared: (i) loose grape stem particles without binder, (ii) particles bonded with sodium silicate (water glass), (iii) particles bonded with a flour-based, and (iv) grape stems foams with chemical processing. The resulting materials exhibited varying ultrastructural and morphological characteristics and were evaluated using impedance tube measurements supported by JCAL modeling. The acoustic performance of the developed materials demonstrated promising sound absorption behavior, with absorption levels approaching those of conventional porous absorbers over selected frequency ranges. These findings indicate the potential of grape stem-based materials as sustainable alternatives for sound absorbing, with potential for further optimization through microstructural control.

Speaker: Hajizah Azmi Siregar (Slovak University of Technology)

626 Experimental Analysis of Sound Absorption And Flow Resistivity of Pulper Waste

The aim of this research is to analyze potential uses for paper mill pulper waste, a by-product generated by paper mills that recycle paper-based composite materials, which is currently disposed of almost entirely in landfills or incinerators. The paper presents some preliminary results from sound absorption tests in an impedance tube and airflow resistance bench test. The results show promising performance values and encourage the continuation of the study.

Speaker: Lucia Magnini (Department of Architecture, University of Florence)

Full paper (PDF)

A12.12/A16.13 Diffuse Sound Fields and Reverberation: P466

Conveners: Franz Zotter (University of Music and Performing Arts), Albert Prinn (International Audio Laboratories Erlangen)

627 Evaluating Adaptive Reverberation Time Estimation For Non-Standard Conditions

The practices of measuring the impulse response (IR) of a room and obtaining the reverberation time (RT) are standardized in ISO 3382-2. It specifies the use of an approximately omnidirectional sound source and multiple microphone positions (and, where appropriate, multiple source positions) to obtain spatially averaged room acoustic parameters. The standard also specifies minimum distances between the sound source, microphones, and reflecting surfaces to avoid near-field effects and measurement bias. This ensures that the direct sound and the non-diffuse decay of early reflections do not affect the RT measurement.However, these criteria may be difficult to fulfill in practical scenarios, e.g., when measuring with directional loudspeakers or in smaller rooms. To enable RT measurement under such non-standard circumstances, this study investigates approaches that modify the evaluation range of the linear regression used to calculate the RT from the energy decay curve. We inspect modifications based on the degree of non-linearity of the regression as well as adapting the range based on a diffuse field assumption utilizing the loudspeaker-microphone distance, room volume and an estimate of the RT.An evaluation across six acoustic scenarios indicates that these approaches can reduce the error under certain configurations. Specifically, the distance based approach improves the measurement more often than not.

Speaker: Aron Karsai (Institute of Electronic Music and Acoustics)

Full paper (PDF)

628 Taylor-SWFT: Fast Discrete Statistical Wave Field Theory Using Taylor Expansion for Late Reverberation

Dynamic room acoustic simulation aims to render the acoustic effects of an environment in real time while accounting for potentially moving sources and receivers. In this context, the efficient synthesis of the long-term room response, also known as late reverberation, remains challenging because of the intricate relationship between room geometry and acoustic behavior. This paper introduces Taylor-SWFT, an efficient implementation of key results from Statistical Wave Field Theory (SWFT) for the geometry-aware dynamic synthesis of late reverberation. The method is evaluated on the Benchmark for Room Acoustical Simulation (BRAS) and achieves competitive performance compared with classical approaches, while substantially reducing computational cost.

Speaker: Marius Rodrigues (Télécom-Paris)

Full paper (PDF)

A14.07 Computational and model-based approaches to hearing science: P488

Conveners: Helia Relaño-Iborra (Eriksholm Research Centre), Vaclav Vencovsky (Czech Technical University in Prague)

629 A hearing-impaired version of the short-time objective intelligibility (STOI) metric

The short-time objective intelligibility metric (STOI) [Taal et al., 2011, IEEE Trans. Audio, Speech and Lang. Proc., Vol. 19, No. 7] is a widely used predictor of speech intelligibility and is often applied to evaluate the effects of noise-reduction algorithms and other speech-enhancement strategies. However, STOI is not level sensitive and does not account for the effects of hearing loss, which limits its ability to assess the ultimate benefit of such signal manipulations when implemented in hearing-assistive devices.Here we propose a short-time objective audible intelligibility metric (STOAI), derived from STOI, that introduces a level-sensitive processing stage in the form of internal noise in the front end of the model. This stage is adjusted using the audiometric thresholds of individual listeners to account for their loss of sensitivity. Inspired by Plomp’s model of speech intelligibility, we evaluated STOAI on speech reception thresholds (SRTs) in both quiet and noise, such that the impact of the audibility and distortion components of hearing loss could be assessed independently. Overall, individual STOAI correlated well with SRTs in quiet demonstrating that it accounts for the effects of audibility. However, STOAI only partially accounted for elevated SRTs in noise, consistent with the presence of suprathreshold deficits (distortion factor) in the studied population. STOAI extends the STOI framework to account for audibility-related effects and can therefore provide a useful tool for the future development and evaluation of speech-enhancement algorithms targeting hearing-impaired listeners.

Speaker: Helia Relaño-Iborra (Eriksholm Research Centre)

630 From Transmission Line Models to Stuart–Landau Oscillators: A Hopf-Based Description of Cochlear Dynamics

Over the past decades, numerous models of the cochlear partition have been proposed to explain key features of cochlear mechanics, including traveling waves, nonlinear amplification, and compressive responses. Among them, the transmission-line model introduced by Geoffrey Zweig and extended by others represents the cochlea as a distributed system of coupled resonant elements interacting through the cochlear fluid and successfully reproduces the spatiotemporal dynamics of the basilar membrane. Another widely used framework describes active auditory processes as nonlinear oscillators operating near a Hopf bifurcation. The Stuart–Landau oscillator (SLO), the normal form of a Hopf bifurcation, captures essential properties such as self-sustained oscillations, nonlinear amplification, and compressive responses, and has therefore been widely used to model active hair-cell dynamics and cochlear amplification.In this study, we analytically demonstrate that the local dynamics of the transmission-line model undergo a Hopf bifurcation and can be reduced to the Hopf normal form. We further show that this reduced oscillator representation near the bifurcation can describe the local behavior. To illustrate this correspondence, a numerical framework is used to compare basilar membrane displacement with an ensemble of coupled SLOs, revealing close agreement between both representations.

Speaker: Sebastian Handel (Signal Processing and Speech Communication Laboratory)

Full paper (PDF)

631 Towards a Biophysical Dipole Model for Auditory Brainstem Response Wave I Generators

The auditory brainstem response (ABR) provides a non-invasive measure of synchronous neural activity along the ascending auditory pathway and is widely used in auditory neuroscience. Alterations in the ABR waveform serve as biomarkers of altered neural processing and can provide insight into the mechanisms of auditory disorders, including tinnitus. In this work, we present the first steps towards a biophysically grounded dipole model of ABR generators, linking neuronal population activity to scalp-recorded responses. Acoustic inputs are transformed into inner hair cell receptor potentials via the biophysical auditory periphery model of Verhulst et al. (2018). The IHC – auditory nerve fibre ribbon synapse is modelled as a calcium-dependent release process. Fibre-type heterogeneity (high-, medium-, and low-spontaneous rate fibres) is implemented through differences in calcium conductance and release threshold, allowing the model to capture the diversity of auditory nerve responses. The resulting postsynaptic currents drive a population of spiral ganglion cells (SGCs) implemented in a NetPyNE/NEURON framework. The SGC model incorporates Hodgkin–Huxley-type ion channel kinetics and multicompartmental morphologies, capturing key features of auditory nerve encoding, including high temporal precision and broad dynamic range. In response to a click stimulus, the model produces highly synchronous firing across the SGC population, giving rise to a macroscopic population dipole. A head volume conductor model subsequently maps this dipole to scalp-recorded electroencephalography (EEG) signals, corresponding to ABR wave I. To validate the physiological accuracy of the model, simulations of both SGC rate-intensity curves and ABR wave I latencies will be compared to experimental data.

Speaker: Sarah Vandepitte (Ghent University)

A15.02/A24.05 Hearing Research in Virtual Environments: P450

Conveners: Janina Fels (IHTA, RWTH Aachen University), Carolin Breuer (IHTA, RWTH Aachen University), Nils Peters (Trinity College, The University of Dublin)

632 On the impact of target localisation awareness on spatial release from masking

Speech intelligibility decreases in noisy environments, especially for individuals with hearing impairment. Spatial separation between speech and competing sounds improves understanding through spatial release from masking (SRM). This can be measured using loudspeakers or headphones, in the latter case through convolution with Head-Related Transfer Functions (HRTFs). In several use-cases, non-individual HRTFs are employed, despite being known to increase front-back confusion in localisation tests. However, recent studies highlighted that even if participants are not able to “consciously” determine whether the source appears on the front or on the back, the analysis of neurophysiological data allows to detect a difference in brain responses.The present study aims to explore spatial release from masking in the median plane using the Coordinate Response Measure (CRM) corpus in a speech-on-speech task. Participants experience free-field loudspeaker rendering and binaural playback (non-individual HRTFs) rendered via acoustically transparent DIY 3D-printable circumaural headphones. Performance is measured via identification accuracy and front-back discrimination of target and maskers.The free-field presentation is expected to outperform non-individual HRTFs, with the latter performing worse in front-back discrimination due spectral cue mismatches. Consequently reduced SRM benefits are expected.Though, we anticipate that being conscious of the target location will not impact content accuracy. Understanding SRM and target location awareness in different virtual and real world settings can enhance auditory training and can help uncovering generalisation effect from task (front-back discrimination) to task (speech discrimination) in complex environments.

Speaker: Nicola La Magna (Imperial College London)

633 Perceived Room-Related Attributes of Binaural, Stereo, and Monaural Concert Recordings in Audiovisual VR

This study investigates the perceptual differences among various recording techniques used to capture a singer-songwriter performance at the Staatstheater Oldenburg. The performance was recorded utilising binaural (Eigenmike em64, Vikk64, Kemar), stereo (ORTF, AB), and monaural (omnidirectional) configurations. Notably, the venue employs an artificial acoustic system yielding a reverberation time of 1.4 seconds. To evaluate the listener experience, a multi-stimulus rating method was employed to assess specific room-related attributes. These attributes included reverberation, tone colour, source distance, localisation, externalisation, presence, and plausibility. During the evaluation phase, playback was conducted within a virtual reality scenario using a head-mounted display. This setup fully immersed participants in a 180° visual representation of the theatre. The findings clarify the perceptual distinctions among these spatial audio formats and their broader implications for designing realistic concert-hall experiences in virtual environments.

Speaker: Felix Stärz (Jade Hochschule, IHA)

A15.06 Psychoacoustics of everyday sounds: P489

Conveners: Daniel Oberfeld-Twistel (Johannes Gutenberg University Mainz), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

634 Hearing the river: Perceiving hydrodynamics in natural environments

Watercourses are dynamic systems whose acoustic manifestations vary as a function of discharge, flow velocity, and hydrogeomorphological structure. Characterizing how such acoustic manifestations are perceived is essential for understanding the capacity of human beings to monitor crucial geophysical information in their close environment with their auditory system. Building on prior work evaluating the human ability to detect water in natural soundscapes, the present study investigated human sensitivity to hydrodynamic changes in riparian environments.A corpus of acoustic recordings was acquired along a 400-m reach of the Golo River (Corsica, France) using an array of 8 microphones. The site, located downstream of a hydroelectric facility, provided controlled variations in discharge, yielding a range of hydraulic conditions. Estimated discharge measurements were collected concurrently, enabling precise temporal alignment between hydro-geo-morphological parameters and acoustic signals.Stimuli derived from this corpus were presented diotically over headphones in a three-interval, forced choice oddity paradigm. On each trial, human listeners were tasked to identify the interval that differed from the other ones, allowing estimation of auditory discrimination performance as a function of hydrodynamic contrast. In parallel, a computational model of the human auditory system was employed to extract sound-texture statistics at the output of cochlear and modulation filters in response to the same sounds. Simulations evaluated whether sound-texture statistics predict human auditory sensitivity to flow variations.By integrating hyrogeomorphological and psychophysical data with model-based analyses, this work aims to identify diagnostic acoustic features of hydrodynamic change in humans, and to advance the use of sound as a proxy for ecological dynamics.

Speaker: Matthieu Fraticelli (Laboratoire des Systèmes Perceptifs, ENS-PSL)

635 Let’s crack that joint: Psychoacoustics of the noise that gets under your skin

• Background: Joint cracking is caused by tribonucleation, the formation and implosion of gas bubbles in synovial fluid. Many individuals find the associated sound pleasant and report a stress-reducing effect. | - Research Questions: Which joints produce the most pleasant sensations and sounds? And which audio features account for this impression? | - Method: The cracking sounds of 224 finger joints, 54 cervical vertebrae, and 316 other joints of 40 participants (20 female, 18 male, 2 diverse, age 18-45; mean 23.2 years) were recorded with an Austrian Audio OC818 microphone in rooms with comparable acoustics (RT60<200ms) at the University of Vienna and TU Graz. Participants rated each of their cracking sounds in terms of associated physical feeling (e.g. “relaxing to tightening”) and sound impression (six items on property scales). The cracking sounds were then analyzed for over 150 audio features alongside the subjective evaluations and personality traits assessed via the Short Eysenck Personality Profiler (SEPPO). | - Results: Finger and cervical spine joints were the most frequently cracked and preferred, in particular cracking distal joints were perceived as the most satisfactory. Different joint types are distinguishable within a three-dimensional timbre space defined by Brightness, Attack Time, and Acoustic Annoyance (see interactive graph: https://muwiserver.univie.ac.at/body_crackles/brightness_annoyance_attacktime.htm). Finger cracking sounds were found to be brighter and less annoying than spinal cracks. Perceived pain correlated with increased timbral roughness and 500 Hz energy, whereas "liberating" sensations were linked to lower acoustic annoyance. Personality traits also influenced perception: extraverts perceived sounds as "fuller," while higher neuroticism correlated with higher perceived loudness.

Speaker: Christoph Reuter (Universität Wien)

Full paper (PDF)

A16.00 Room Acoustics: S256

Conveners: Jamilla Balint (Rohde Acoustics), Francesco Martellotta (Politecnico di Bari, DARCOD), Mélanie Nolan (Universidad Politécnica de Madrid)

636 Acoustic evaluation of a historic apartment at La Pedrera in Barcelona

La Casa Milà, also known as La Pedrera, is one of the most emblematic works of Antoni Gaudí and a landmark of Barcelona’s architectural heritage. Although its architectural and historical significance has been widely studied, its acoustic behaviour has received little specific attention. This article presents an acoustic experimental study of a historical apartment located within the building. The apartment has been preserved in its original state since 1911, retaining period furnishings and architectural features designed by Gaudí, including mouldings, floors, doors, and windows. The space is characterised by irregular geometries such as undulating ceilings, non-parallel walls, and complex volumes, enhancing sound diffusion. An acoustic measurement campaign was conducted to evaluate the acoustic comfort of typical domestic spaces within the flat. Measurements were carried out in the kitchen, children’s bedroom, living room, and master bedroom, including its adjoining bathroom. Impulse responses were recorded in accordance with ISO 3382-2, using an omnidirectional sound source and an exponential sine sweep excitation signal. Source and receiver positions were adapted to the specific characteristics of each room. Key acoustic parameters were derived, including reverberation time (T30), early decay time (EDT), clarity (C80), and speech intelligibility (D50). The results indicate highly favourable acoustic conditions across all measured spaces, highlighting the intrinsic acoustic quality of Gaudí’s domestic architecture and the relevance of acoustic assessment in architectural heritage studies.

Speaker: Marc Arnela (La Salle, Universitat Ramón Llull)

637 Acoustic Design of Broadcasting Studios in Yammat FM Radio Station in Zagreb, Croatia

The Yammat FM radio station, native to Zagreb, Croatia, was recently relocated to a 100-year-old building in the city centre. In 2023, the new premises were fully restored and redesigned by Croatian architects and reputable contractors. The design project aimed to create a distinctive visual and acoustical accessibility of the designed spaces, with the emphasis on two broadcasting studios facing the street. This paper presents the acoustic design of these studios. The proposed solution strives to solve a complex design problem by addressing the proximity of these studios to a busy road and tram traffic route, evoking the need for adequate sound insulation. Simultaneously, it integrates the studios with the outdoor public space through large front windows, with glass panels that allow a clear view inside from the sidewalk. As a prerequisite for high-quality production of radio programme, room acoustic conditions in the studios have been given special attention. The target design values were defined primarily in accordance with BBC WHP 021 whitepaper and HRN DIN 18041:2012 standard. Room acoustic conditions were designed and evaluated through simulations performed in ODEON 17.13. The proposed acoustic treatment combined acoustic plaster, glass-fibre absorbers, perforated wooden and gypsum-board systems, carpeting, and carefully selected wall and ceiling assemblies. A mid-frequency reverberation time of 0.36 s was achieved in Studio 1 as the larger of the two, while 0.22 s was achieved in the smaller Studio 3. Low-frequency reverberation remined somewhat emphasized in both studios due to visual design constraints. The study demonstrates the integration of acoustically demanding spaces into a transparent, street-visible architectural concept.

Speaker: Mateja Nosil Mešić (UNIZG Faculty of Architecture)

Full paper (PDF)

638 Measurements of Acoustic Quality of Underground Stations of the Porto Subway Network

The Porto Metro is a light rail network in Porto, Portugal, in operation since 2002 and currently in expansion. One of the factors that was relevant during its construction was the acoustic quality of the subway stations, particularly reverberation time control. The Portuguese acoustics legislation requires this parameter to be under a maximum value, for all types of passenger transportation stations, to improve speech intelligibility, particularly important for safety and travel information.To guide the acoustical design and construction works supervision of several new stations of this network, a measurement campaign was planned and executed during 2022. This paper describes the methodology and results of speech transmission index, reverberation time and background noise from 13 existing subway stations. The results obtained are discussed in relation to the efficiency of acoustic materials applied and recent maintenance works, as well as recommendations for future planning of new stations.

Speaker: Tatiana Teixeira (InAcoustics)

A16.09 Acoustic Comfort in Hospitals: P484

Conveners: Simone Secchi (Department of Architecture, University of Florence), Andrea Giglio (European Acoustics Association Young Acousticians Network (YAN)), Juan Negreira, Maria Quinn

639 Acoustic Quality in Healthcare Facilities: an Integrated Approach for Patient Well-Being

This paper presents the findings from a Doctoral thesis in Architecture, investigating the impact of the sound environment on hospital patients’ well-being. Representative inpatient wards were selected within four Italian hospitals to analyze acoustic criticalities, noise sources, and the subjective experience of users, using a mixed-methods approach. The results highlight that WHO limits are systematically exceeded. The complexity of the hospital environment needs an integrated approach that encompasses the various factors influencing the acoustic quality of the spaces. Finally, the research proposes a practical framework, consisting of a replicable investigation protocol and a checklist of design strategies to ensure acoustic quality control throughout every project phase.

Speaker: Veronica Amodeo (Department of Architecture - University of Florence)

Full paper (PDF)

640 Temporal trends in acoustic climate in hospital wards

In recent years, various studies have been conducted to measure noise levels in hospital settings; in most of these studies, noise levels were found to exceed the limits recommended by the World Health Organization (WHO). The aim of this work was to determine, through a systematic review of scientific articles, typical noise levels in different types of hospital environments and to identify the noisiest wards. This work is part of the research activities of the INAIL BRIC Project 2024 (ID07), which aims to investigate the possible extrasensory effects of noise on healthcare staff.

Speaker: Elisa Nannipieri (Department of Architecture, University of Florence)

Full paper (PDF)

A21.02 Road traffic noise and vibration: P433

Convener: Michael Cik

641 Road Traffic Noise Source Inferencing in a Noisy LMIC Megacity

Road traffic noise arises from multiple overlapping sources such as tire friction, engines, exhaust, and honking. Separating these sources is essential for assessing health impacts and developing effective mitigation strategies. Most studies report only aggregate noise levels which are insufficient for environmental impact assessment. Existing source apportionment approaches rely heavily on labelled datasets from developed countries, limiting their applicability to complex urban environments like Low-and middle-income countries (LMICs). The raw audio data used in this study encompasses non-honking vehicular noise, horn noise, and ambient environmental sounds. This study therefore focuses on separating honking and non-honking vehicular noise in heterogeneous Indian urban traffic conditions where drivers exhibit persistent, unregulated honking behaviour varying systematically with road type and congestion level. Class 1 Sound level meters (SLMs) are prohibitively expensive for LMICs, and existing approaches require both an SLM and audio recorder in the field, the SLM for dBFS to dBSPL conversion and the recorder for frequency-resolved source differentiation, imposing a compounded logistical burden, particularly consequential in LMICs. This study addresses both challenges: dBFS to dBSPL conversion using white and pink noise collocation in anechoic chamber yielding octave band wise offsets (125Hz to 10kHz). Using GMM based classification and HMM based temporal sequencing, that requires no pre-labelled data, we found that honking occupies fewer traffic frames yet contributes disproportionately higher noise energy levels with a dominant spectral peak at 4kHz. Applied across 8 routes in Delhi under varying road and congestion conditions, this framework supports traffic noise regulation and environmental assessment in resource-constrained LMIC settings.

Speaker: Kajal Kumari (Indian Institute of Technology Delhi)

A22.04 Propagation, Modeling and Simulation of Underwater Sound Fields: P434

Conveners: María Campo-Valera (Universidad Internacional de la Rioja), Dídac Diego-Tortosa (Institut de Ciències del Mar (ICM-CSIC))

642 Acoustic Leaky Waves Antenna for Underwater Applications: Analysis and Performance

Acoustic Leaky Wave Antennas (ALWAs) offer an effective approach for directional sound radiation in underwater environments. This work analyzes their behavior in a fluid medium, focusing on fluid–structure interaction mechanisms that control acoustic leakage and beam steering. Key performance metrics, such as far-field radiation patterns, are assessed, demonstrating the potential of ALWAs for underwater communication, sensing, and sonar applications.

Speaker: María Campo-Valera (Universidad Internacional de la Rioja)

Full paper (PDF)

643 Broadband Underwater Acoustic Absorption using a Tilted Cylindrical Inclusion in a Simplified Coating Design

A novel underwater acoustic coating design is proposed based on a simplified periodic unit cell containing a single tilted cylindrical inclusion. Unlike conventional coatings employing vertically aligned cylindrical inclusions in compact unit cells, the present design introduces geometric asymmetry by orienting the inclusion at an angle within an enlarged unit cell, thereby modifying the wave-structure interaction mechanisms. Finite element simulations demonstrate that the proposed coating achieves near-unity absorption around 1 kHz and maintains a high absorption coefficient over a broad frequency range extending to several kilohertz. Compared with benchmark coating configurations that exhibit predominantly narrowband absorption, the present design provides a substantial enhancement in absorption bandwidth and overall acoustic performance. The improved response is attributed to symmetry-breaking induced by the tilted geometry, which promotes enhanced scattering, an increased effective propagation path, and resonance-assisted dissipation. In addition to improved acoustic performance, a single inclusion per unit cell significantly reduces structural complexity, offering practical advantages for fabrication and implementation.

Speaker: Vineeth P R (DYSL-SM (DRDO))

644 Dynamic acoustic field modeling of internal solitary waves in the South China Sea and their impact on hydroacoustic detection

The South China Sea (SCS) is one of the most active regions for the internal solitary wave (ISW). During the ISW propagation, strong sound-speed perturbations can be induced, leading to anomalous underwater acoustic propagation phenomena, which have a significant impact on hydroacoustic detection. To quantitatively asses these impacts, this study considers three representative ISW on the northern SCS continental slope. Time-varying acoustic field models are constructed, and numerical simulations are conducted to evaluate the source localization errors and the horizontal deflection of acoustic propagation in ISW-affected environments. The results indicate that, across the three ISW cases, the mean localization error rates in source range and depth reach 10% and 15%, respectively, while the maximum horizontal deflection angle reaches 8°. Furthermore, when the ISWs passes through the source, the mode-1 single ISW causes acoustic waves propagating along the wave-crest direction to be deflected outward, away from the crest. In contrast, the mode-2 single ISW produces the weakest horizontal deflection for its smaller amplitude and the presence of a near-surface low-sound-speed layer. For the mode-1 ISW packet, multiple constituent waves induce repeated bending of acoustic propagation paths, resulting in a more persistent influence over time.

Speaker: Lintai Rong (Northwestern Polytechnical University)

Full paper (PDF)

A23.00 Vibro-Acoustics: P425

Convener: Christian Adams (Graz University of Technology)

645 Symmetric and Asymmetric Embroidered Mass Patterns in Membrane Vibration Modes

This study investigates the transverse vibration of tensioned membranes with spatially periodic surface-mass variations that represent embroidered fabrics. A two-dimensional membrane model is discretized using finite differences with fixed boundary conditions. The resulting generalized eigenvalue problem accounts for heterogeneous surface mass density and is evaluated for four representative designs inspired by the common weave bindings plain, satin and twill, and an asymmetric layout. Computed natural frequencies and mode shapes are compared with the analytical solution for a uniform square membrane. Symmetric patterns largely preserve classical modal symmetry and near-degenerate mode pairs, whereas the asymmetric layout breaks symmetry, lifts degeneracy, and produces larger frequency shifts accompanied by distorted mode shapes.

Speaker: Ahmed Mehrem (German Institutes of Textile and Fiber)

Full paper (PDF)

15:40 Coffee break

16:00 Coffee break

16:00 Coffee break

16:00 Coffee break

A18.07 Sound evaluation in a wider context: S134

Conveners: Timothy Van Renterghem (Ghent University), Manish Manohare

646 Indoor Environmental Quality and Occupants’ Comfort Monitoring in Two Open-plan Offices

Office users’ health, comfort and work productivity are influenced by the Indoor Environmental Quality (IEQ) of the environment in which they work. This study presents a three-months monitoring of the thermal, acoustic, visual and air quality conditions of two corporate open-plan offices, located at the ground and first floor of the building, through a commercial multi-sensor and the PROMET&O (PROactive Monitoring for indoor EnvironmenTal quality & cOmfort) system. It consists of a low-cost calibrated multi-sensor device that monitors the main parameters of IEQ domains, and a questionnaire used to collect feedback on occupants’ comfort. A designed user-friendly interface informs building managers and users about the objective and subjective data monitored. In each office ten PROMET&O multi-sensors, ten commercial sensors and ten tablets were positioned over the occupants’ desks. Results show that a good agreement between the objective conditions and the subjective perception is more evident in the thermal and acoustic domains. Carbon dioxide concentrations fall within the recommended values, but a low air quality perception in the ground floor office is evident. Horizontal illuminance values rarely are higher than 500 lx, with occupants more satisfied in the first floor office than in the ground floor one.

Speaker: Arianna Astolfi (Politecnico di Torino)

Full paper (PDF)

647 Development of an Audio-Visual Interaction Questionnaire

Previous research has shown that audio-visual interactions in environmental perception—and in environmental noise perception in particular—can be effectively leveraged in urban sound planning. However, the extent to which audio-visual cues are integrated varies substantially across individuals. A distinction can be made between individuals who rely predominantly on visual information and those who integrate auditory and visual inputs more evenly. Accounting for such individual differences may be essential for more accurately explaining phenomena such as the benefits of green window views in epidemiological studies. In this study, results from two computer-based tests—a congruent object recognition task (CORT) and the sound-induced flash illusion (SIFI)—were used as objective measures to guide the development of a questionnaire assessing individual audio-visual integration capacity. Subsequently, a broad set of everyday, recognizable experiences was evaluated. From this pool, a subset of questions was identified that demonstrated predictive power for performance on the CORT and SIFI tasks, as well as good test–retest reliability. This refined questionnaire is intended for use in future large-scale surveys conducted outside laboratory settings.

Speaker: Timothy Van Renterghem (Ghent University)

Full paper (PDF)

648 A Systematic Review of Datasets Linking Acoustic Indices to Subjective Assessments and Physiological Responses in Environmental Soundscapes

The correlation between acoustic indices (Bradfer-Lawrence et al. 2025; Dröge et al. 2024) and human responses—both subjective assessments and physiological responses—remains an understudied area in environmental soundscape research. While some recent studies (Durbridge and Murphy 2023; Qin et al. 2025; Uebel et al. 2025) shows subjective-physiological correlations, no study has integrated acoustic indices with both to predict responses across contexts. Preliminary work by (Rey Gozalo et al. 2015) correlated acoustic indices (e.g., LAeq) with subjective assessments, while (Versümer et al. 2025) modeled soundscape subjective assessments using machine learning on datasets like HSDD, ARAUS, and ISD. However, neither study included physiological responses or diverse environments. Currently, no standardized methodology exists to model acoustic indices alongside subjective and physiological data, limiting cross-study comparisons and model validation.The current research work presents a systematic review of available datasets that include subjective ratings and/or physiological measurements. The goal is to identify gaps in annotation standards and data compatibility to outline opportunities for future modeling efforts.Key aspects include:Data collection/validation (e.g., ISO 12913-2 compliance, sensor protocols).Annotation structure (e.g., aggregation level, inter-rater reliability).Challenges in aligning heterogeneous data (e.g., recording conditions, metadata).This review underscores the need for standardized data collection protocols and integrated datasets to enable robust predictive modeling of human responses to soundscapes. Future work aim to ultimately developing a model that predicts both subjective and physiological responses from acoustic indices.

Speaker: Marcello Lussana (Otto-Friedrich-Universität Bamberg)

649 From Noise to Nature: Effects of Prior Exposure on Soundscape Evaluation

Soundscape research has broadened the understanding of environmental sound beyond a primary focus on negative effects, emphasising a wider range of perceptual qualities. These perceptual qualities are commonly operationalised through dimensions such as pleasantness and eventfulness, in line with ISO 12913.However, soundscape evaluations are often treated as momentary entities, whereas real-world experiences are dynamic and affected by prior exposures. In urban landscape design, it has been suggested that contrast between environments may influence perception, such that exposure to loud or demanding sound environments could enhance the perceived quality of subsequent tranquil settings. Yet, only limited empirical work has examined sequential relationships between different soundscapes. The present study investigates the role of contrast in soundscape perception using a between-subject experimental design (n = 66) in an ambisonics laboratory. Participants were exposed to an urban soundscape followed by a tranquil natural soundscape, with initial urban sound levels differing between groups (50 dBA vs 70 dBA). Perceptual responses were assessed using SSID-based self-reports aligned with ISO constructs, alongside physiological measures (heart rate, respiration, skin conductance, and heart rate variability). We hypothesise that higher initial noise exposure will enhance subjective evaluations of the subsequent natural environment and be associated with greater parasympathetic activation during recovery. Preliminary findings will be discussed in relation to environmental planning and soundscape design.

Speaker: Gunnar Cerwén (SLU)

650 Investigating the effects of birdsong on aircraft noise annoyance and noise-induced stress: a controlled within-subject listening experiment

Noise annoyance drives the sustained activation of the sympathetic nervous system and the hypothalamic-pituitary-adrenal axis, resulting in the release of catecholamines and cortisol. Chronic neuroendocrine activation can lead to a pathophysiological cascade characterized by systemic and neural inflammation, oxidative stress, and structural brain changes, favoring the development of endothelial dysfunction, maladaptive coping mechanisms, and mental health outcomes. Identifying interventions capable of mitigating noise annoyance and stress is, therefore, critical to safeguard the health of noise-exposed populations. Environments with abundant natural sound sources (biophony and geophony) offer a potential psycho-physiological buffer against these impacts. Birdsong, in particular, improves perceived soundscape quality, elicits positive affective responses, and promotes perceived restoration and relaxation. When presented alongside road traffic noise, birdsong reduces participants' perceived loudness and noise annoyance, possibly through a combination of informational masking and stress reduction. The effects of birdsong on aircraft noise annoyance and noise-induced stress, however, remain unexplored. In a within-subject listening experiment, participants will be exposed to two binaural audio recordings containing eight aircraft flyovers at 65 dB(A), with and without birdsong, at 48dB(A). Participant physiological arousal levels will be continuously measured using electrodermal activity (event-related SCR), pupillometry (event-related dilation), and respiration rate (RR). Following each condition, participants will report their soundscape appraisal (ISO-12913-2), aircraft noise annoyance (ISO-TS-15666), affective states (valence & arousal), perceived restoration (PRSS), and hypothetical behavior. Combining perceptual and physiological measurements, this study will empirically test the effectiveness of increasing birdsong as a mitigation strategy for aircraft noise annoyance and noise-induced stress in airport-adjacent communities.

Speaker: Rodrigo Vassallo (TU Delft)

651 Influence of Realistic Hearing-Protector Use on Damage Risk Criteria for Impulse Noise

High-level impulse noise from firearms and explosives poses a major risk of hearing injury, particularly for soldiers and police officers during routine training. When source reduction is not feasible, appropriate Hearing Protection Devices (HPDs) are required to limit individual exposure. However, HPDs must provide sufficient attenuation without unnecessarily compromising comfort, communication, and situational awareness. The required attenuation is usually derived from Damage Risk Criteria (DRC), but the DRC implemented in different national standards are not equivalent. Their predictions may differ substantially because they emphasize different signal properties, such as peak level, effective duration, weighted or unweighted sound exposure, waveform statistics, or auditory-model responses.This study compares several DRC under realistic conditions of HPD use. For this purpose, a shooting-noise dataset covering different weapons and indoor/outdoor environments is combined with impulse-noise insertion-loss measurements of active earmuff protectors, including configurations with and without protective glasses and additional headgear. The resulting estimated under-protector exposures and corresponding reference signals are analysed using criteria such as Pfander, A-weighted sound exposure, C-weighted peak level, AHAAH, and other exposure metrics. For each condition, hazard levels and Maximum Permissible Exposures are compared across criteria.The results show that realistic fitting conditions affect not only the transmitted level but also its interpretation by different DRC. In particular, seal leakage introduced by glasses increases under-protector exposure and may change the ranking of HPD configurations as well as the estimated admissible number of shots. Realistic fitting conditions and environmental acoustics should therefore be explicitly considered when deriving operational exposure limits.

Speaker: Marian Weger (French-German Research Institute of Saint-Louis (ISL))

652 Moving Targets: How Hearing Protection Settings Shift Distance Perception

In noisy workplaces, wearing a hearing protection device (HPD) is mandatory but can impair situational awareness and the audibility of environmental sounds. Level-dependent HPDs amplify soft sounds to ensure audibility while still attenuating high-level sounds. However, they can also impair situational awareness, e.g., regarding localization and detection of warning signals. Distance perception is an important part of localization, e.g., in factories or at construction sites, but is rarely investigated with HPDs. In this study, the distance of a reverse warning signal was simulated in eight steps between 1 and 25 m with virtual acoustics for playback in an 86-channel loudspeaker array. Target signals alone and with diffuse pink noise were recorded with an acoustic test fixture with open ears and wearing a level-dependent HPD in three settings. In a listening study, 32 normal-hearing participants had to identify the actual distance of the recorded signals. The results showed impaired performance in all HPD conditions compared to open ears. Moreover, over- and underestimation of target distances varied largely between conditions, indicating risks of misperceptions when switching HPD settings during work.

Speaker: Sina Buchholz (Fraunhofer IDMT-HSA, Oldenburg)

Full paper (PDF)

16:20 Coffee break

16:20 Coffee break

A01.05/A17.03 Automotive Audio and Active Control: S006

Conveners: Filippo Maria Fazi (Institute of Sound and Vibration Research), Fabio Cagnetti (BdSound S.r.l.)

653 Development of Broadband Active Noise Control for Engine Noise Reduction in a Series Hybrid Vehicle

In hybrid vehicles, the contrast between engine‑off and engine‑on states makes engine noise noticeable, and in series hybrid vehicles the engine is used only for power generation and does not need to match the driving state, so engine sound can become annoying. To improve engine sound quality, factors contributing to engine‑noise perception must be quantitatively identified.Using on‑road data from a series hybrid vehicle, a machine‑learning model is trained to predict engine noise perception from vehicle parameters and interior sound levels. Analysis of parameter importance and prediction errors clarifies empirical knowledge: perception depends not only on interior sound level but also on vehicle speed, acceleration state, and driver acceleration intention. The balance between engine noise and background noise (road and wind), together with acceleration intention, is identified as a key factor governing annoyance.Based on these findings, we propose a perceptually motivated broadband active noise control (ANC) targeting engine‑noise components below 300 Hz. Under low‑speed, low‑load conditions with little acceleration demand, where engine noise is noticed, the series‑hybrid architecture allows the engine speed to be held at a constant value, enabling broadband ANC to effectively reduce the target components. During acceleration phases the operating condition varies rapidly and the broadband ANC effect becomes smaller; however, the perception model indicates that large additional attenuation is not required in these conditions.Consistent with the identified perceptual characteristics, the proposed ANC and engine‑speed control strategy adapts across driving conditions. On‑road evaluations demonstrate reduced engine‑noise level in the target scenes and improved interior quietness in an electrified vehicle.

Speaker: Shinichi Suganuma (Nissan Motor Co., Ltd.)

654 Achieving Rapid Convergence in Multichannel RNC Using Short Prediction‑Error Filters

Multi-channel road noise cancellation (RNC) systems in electric vehicles increasingly rely on large arrays of reference signals that generate coloured and highly correlated reference signals. These characteristics result in poorly conditioned co-variance matrices, leading to slow convergence when using adaptive feed-forward algorithms such as FX-LMS. This paper investigates the use of a simple, computationally efficient method of preconditioning the reference signals to reduce the eigenvalue spread of the co-variance matrix, thereby improving the convergence properties. The preconditioning is achieved via a set of short multi-channel prediction error filter (PEFs), which whitens the spectral density matrix of the reference signals.Simulation studies using measured vehicle data demonstrate that the preconditioning methods substantially improves convergence, reducing adaptation times from several minutes to tens of seconds when used in conjunction with secondary plant compensation. Importantly, these gains are achieved with only short, inexpensive PEFs, indicating that the approach offers a practical alternative to more computationally demanding whitening techniques such as spectral factorisation.

Speaker: Laurence Wilmshurst (ISVR, University of Southampton)

655 Sound Source Localization: a Study comparing Angular and Distance Perception Accuracy in a Vehicle Cabin and in a Reference Room

Accurate spatial audio reproduction is increasingly important in automotive applications such as navigation guidance, advanced driver-assistance systems, and immersive entertainment. This study compares sound source localization performance in a vehicle cabin and an ITU-R BS.775-4 compliant multichannel reference listening room to assess whether the latter can serve as a suitable environment for preliminary spatial-audio tuning and verification. Two listening tasks were conducted with 32 participants (16 per environment): angular localization and distance-movement identification. Test stimuli consisted of a navigation speech prompt and a Park Distance Control (PDC) warning chirp. Angular localization performance was comparable in both environments, with 77% and 67% of responses falling within ±20° of the target direction in the reference room and vehicle cabin, respectively, while median localization errors remained nearly identical (11° and 12°). The vehicle exhibited an additional front-left localization bias associated with the asymmetric listening position and loudspeaker arrangement. Distance-movement identification proved substantially less robust than angular localization in both environments. Near-field sources moving away from the listener were frequently perceived as stationary, whereas far-field away movements were identified with high accuracy. Similar response patterns observed in both environments suggest that these limitations arise primarily from perceptual characteristics of auditory distance perception rather than from the vehicle cabin itself. The results indicate that a reference listening room can provide useful preliminary indications of localization performance and perceptual trends, while final validation remains necessary in the target vehicle environment to assess absolute performance and placement-dependent effects.

Speaker: Alessandro Travaglini (HARMAN BECKER Automotive Systems GmbH)

Full paper (PDF)

656 A comparative study of car interior auralization using higher- order ambisonics and the spatial decomposition method

Higher-Order Ambisonics (HOA) is a widely used technique for sound field auralization; however, it presents inherent limitations, particularly in the trade-off between the number of transducers (microphones and loudspeakers) and the achievable timbral and spatial accuracy of the reproduced sound field.More recently, alternative approaches for the acquisition and reproduction of spatial room impulse responses have been developed. These include the Spatial Decomposition Method (SDM) is introduced, whose underlying theory is briefly reviewed in this work.This paper investigates differences in the quality of car interior auralization using impulse responses captured and reproduced with SDM and with HOA. The evaluation combines objective performance metrics with subjective listening tests to provide a comprehensive comparison of the two methods.

Speaker: Dongin Kim (University of Southampton)

657 Listener Position Tracking for Robust Dynamic Crosstalk Cancellation with Headrest-Integrated Loudspeakers

This study examines the impact of real-time listener position tracking on binaural audio reproduction using loudspeakers in automotive settings. The system employs headrest-integrated speakers in combination with crosstalk cancellation (CTC), a technique known for its high sensitivity to listener position. Even minor head movements can lead to substantial performance degradation, particularly at higher frequencies, where small displacements induce significant variations in the acoustic transfer function matrix between the loudspeakers and the listener’s ears. In this work, a non-adaptive CTC system optimized for a reference listening position is experimentally compared with an adaptive approach based on position-dependent filtering. Measurements conducted in a vehicle cabin indicate that the non-adaptive system is highly susceptible to small head movements, whereas the adaptive system maintains consistent performance. The findings highlight the critical role of listener tracking in achieving reliable binaural reproduction for automotive applications employing crosstalk cancellation.

Speaker: Filippo Maria Fazi (Institute of Sound and Vibration Research)

A05.04 Outdoor sound propagation: S367

Conveners: Christian Adams (Graz University of Technology), Timothy Van Renterghem (Ghent University), Martin Czuka (AIT Austrian Institute of Technology)

658 Comparison of the Three Solvers Implemented in Code_TYMPAN, Including ISO 9613-2:2024: A Case Study Based on EDF Power Plants

EDF must ensure the regulatory compliance of most of its electricity generation facilities, including nuclear power plants, wind farms, combined-cycle gas turbines, and diesel engines. With regard to environmental acoustics, this requires verifying sound pressure levels at site boundaries and at the nearest residential receptors. These verifications are based on field measurements. However, for preliminary design studies of new facilities, extensions of existing sites, or the implementation of noise mitigation solutions, numerical predictions are required.The software used to perform these numerical studies is Code_TYMPAN, developed by EDF DTG and distributed under an open-source license. This engineering software follows the recommendations of the international standard ISO 9613; its “historical” solver also incorporates R&D contributions, particularly concerning ground effects. In recent years, numerous developments have improved the stability and usability of the software, and introduced a solver that strictly complies with ISO 9613 (1996 and 2024 versions).This paper focuses on two typical examples: a fictitious model representative of a nuclear power plant and a hydropower site located in steep, confined terrain. The comparison of calculated results provides insights into the validity of the numerical models. These models also make it possible to examine the relevance of acoustic modeling in borderline cases.

Speaker: Thibaud Thénint (EDF DTG)

659 Influence of inclined photovoltaic panels on sound propagation and reflections from distributed infrastructure noise sources

Large‑scale photovoltaic (PV) installations introduce extensive inclined planar surfaces that can act as screening elements but also as reflective structures. This can influence sound propagation from the source to the receiver. While PV plants are often assessed primarily as noise sources themselves; however, the role of PV panels as barriers and/or reflectors affecting sound from other noise sources is less commonly addressed.This contribution investigates the influence of PV panels on sound propagation and reflections for different types of noise sources commonly present. This can include road traffic, battery energy storage systems (BESS), distributed string inverters, central inverters and transformers, and wind turbines. The analysis focuses on geometric effects associated with panel inclination, relative source, panel and receiver positioning, and the interaction between screening and reflection mechanisms.By considering PV panels as part of the sound propagation environment rather than solely the PV plant, as noise‑emitting equipment, the study highlights situations in which inclined panel layouts may alter sound paths and local sound level distributions in hybrid energy and transport infrastructure.

Speaker: Mikel Amatriain (RWE Renewables Iberia)

Full paper (PDF)

660 Validation of the Image Edge Model for Sound Propagation in Urban Environments

The simulation of urban environments poses challenges for geometric propagation models.In these scenarios, diffraction is a significant contribution to sound propagation.The image edge model was proposed as an efficient method to determine diffraction paths.In combination with the Unified Theory of Diffraction (UTD), it can be used to calculate the diffraction contribution to sound propagation.However, the accuracy of this approach has not been thoroughly validated.In this work, a more comprehensive validation of this simulation approach is performed.For this, a larger database of scale-model measurements specifically designed for urban environments with various building configurations and source-receiver placements is used, which focuses on diffraction.Initial results show good agreement between the simulation and measurements, indicating that the image edge model combined with the UTD can provide accurate predictions of sound propagation in urban environments.

Speaker: Pascal Palenda (IHTA, RWTH Aachen University)

661 Outdoor far-field predictions via inverse-modelling

Winds and spatial variations in temperature can strongly affect the outdoor propagation of low-frequency sound. Under downward-refracting conditions, atmospheric waveguides can form, allowing sound to reach greater detection ranges. In our research, we use an inverse method to predict the propagation of sound from outdoor sound sources into the far-field. The method hinges on a two-dimensional normal-mode propagation model, computed under the assumptions of flat ground and a stratified atmosphere, that gets matched to discrete acoustic pressure samples. The modal basis can be computed for a wide range of atmospheric conditions with the Semi-Analytic Finite Element (SAFE) method, and the model is able of predicting atmospheric waveguides. Our approach circumvents reconstruction of complex outdoor sources and captures important physical behavior, while being computationally efficient. In this work, we reach a high-resolution reconstruction of the acoustic field radiated by a point-source by probing discrete points in its vicinity and applying the procedure. In the future, the method could be used in time-efficient acoustic predictions around outdoor sources, such as industrial complexes. wind parks or open-air concerts

Speaker: Rafael Castro Mota (Physikalisch-Technische Bundesanstalt)

662 Long-range Outdoor Acoustic Simulation with 3D FDTD using the Quasi-Wavelet Approach of Turbulence

In the field of outdoor acoustics, the sound propagation is impacted by meteorological conditions, including atmospheric turbulence which causes significant fluctuations in the sound pressure. To numerically model the propagation of sound in an inhomogeneous atmosphere, the quasi-wavelet approach is a useful representation of the random turbulent field created by temperature and velocity-induced turbulence.In this paper, the quasi-wavelet approach is implemented in a GPU-accelerated solver of the 3D acoustic wave equation, using the Finite-Difference Time-Domain method; this solver includes a moving window feature to enable computationally efficient long-range simulations. We predict the sound propagation from a single source point to a distant receiver, both located near a flat terrain and surrounded by an adequate PML formulation. The temperature and the velocity-induced turbulence are first introduced separately and then their effect is combined.For a set of representative cases of meteorological conditions, the coherence loss in ground effect is calculated between the non-turbulent and the turbulent case to assess the scale of each type of turbulence.

Speaker: Sophia Feriani (Empa)

Full paper (PDF)

A09.01/A17.01 Machine learning for array processing: S054

Conveners: Maximo Cobos (Universitat de València), Jose J. Lopez (Universitat Politecnica de Valencia)

663 Towards Microphone Array Geometry Agnostic Sound Source Characterization: A Message Passing Neural Network Approach

Sound Source Characterization (SSC) with microphone arrays has seen significant advances through deep learning. However, most data-driven methods are trained on data from a single, fixed microphone array geometry, making the resulting models geometry-specific and difficult to transfer to other configurations. Moreover, many architectures do not explicitly use sensor position information. To overcome this limitation, a Transformer architecture that uses a Message Passing Neural Network for array-agnostic SSC is proposed. The proposed model operates on a graph constructed from spatial and spectral information of the microphones, from which the location and strength of the dominant sound source in the region of interest is estimated. The model is evaluated across multiple array geometry classes, such as spiral, grid and randomized layouts, and different numbers of channels outside the training distribution, demonstrating robust generalization where a fixed-geometry baseline fails entirely. Furthermore, training with a variable number of microphones is shown to improve performance. To the authors’ knowledge, this is the first work on grid-free SSC generalizing across large-scale array geometries of up to 128 microphones.

Speaker: Siavash Zaid (Technische Universität Berlin)

Full paper (PDF)

664 Direction-Preserving MIMO Speech Enhancement Using a Neural Covariance Estimator

Multichannel speech enhancement is widely used as a front-end in microphone array processing systems. While most existing approaches produce a single enhanced signal, direction-preserving multiple-input multiple-output (MIMO) methods instead aim to provide enhanced multichannel signals that retain directional properties, enabling downstream applications such as beamforming, binaural rendering, and direction-of-arrival estimation. In this work, we propose a fully blind, direction-preserving MIMO speech enhancement method based on neural estimation of the spatial noise covariance matrix. A lightweight OnlineSpatialNet estimates a scale-normalized Cholesky factor of the frequency-domain noise covariance, which is combined with a direction-preserving MIMO Wiener filter to enhance speech while preserving the spatial characteristics of both target and residual noise. In contrast to prior approaches relying on oracle information or mask-based covariance estimation for single-output systems, the proposed method directly targets accurate multichannel covariance estimation with low computational complexity. Experimental results show improved speech enhancement, covariance estimation capability, and performance in downstream tasks over a mask-based baseline, approaching oracle performance with significantly fewer parameters and computational cost.

Speaker: Thomas Deppisch (Chalmers University of Technology)

Full paper (PDF)

665 Adaptive Learned FISTA for Robust Acoustic Source Mapping

Covariance Matrix Fitting estimates acoustic source strengths on a predefined focus grid by fitting a modeled microphone Cross-Spectral Matrix (CSM) to the sample CSM and is often formulated as a nonnegative LASSO problem. Classical algorithms for the LASSO need model selection and many iterations for accurate reconstructions. Algorithm unfolding reduces this effort by learning selected algorithmic parameters from ground-truth source maps.Unfolded algorithms for CMF use a dictionary from an analytic propagation model in each unfolded iteration. In measurements, the unknown true propagation model may differ from this analytic model. Supervised training creates a second mismatch because ground-truth source maps are unavailable for measured array data, so training must rely on synthetic data generated with an analytic propagation model. This work therefore compares supervised training using the Mean-Squared Error (MSE) objective with unsupervised training, focusing on different objectives including LASSO, Cross-Validation (CV), and a smooth approximation of the Bayesian Information Criterion (BIC).In the matched setting, CV and BIC can reach validation NMSEs similar to supervised training using the MSE objective. Under propagation model mismatch, direct unsupervised training on sample CSMs generated from measured transfer functions from MIRACLE-A1 yields sparser source maps than supervised training on perturbed synthetic data.

Speaker: Adam Kujawski (Technische Universität Berlin)

Full paper (PDF)

666 Differentiable, Matrix-Free Time-Domain Wave Expansions

Elementary wave models are at the core of microphone array processing and estimation problems in acoustics. Wave models are often formulated in the frequency domain, and the problems solved one frequency at a time. On the other hand, time-domain formulations are often more suitable, as they preserve the signals' spatio-temporal structure (e.g., temporal sparsity of wavefronts), and can utilize information common to all frequencies (e.g., source location). However, their computation is far more demanding. In this study we propose a computational framework to perform plane wave and point source expansions in time domain, which (a) overcomes memory limitations by a matrix-free implementation of the forward and adjoint operators, and (b) is differentiable with respect to wave coefficients, wave directions and source positions. We present results for direction of arrival estimation and source localization, yet our approach can be used in any acoustic application requiring elementary wave expansions. Our implementation supports execution on both CPU and GPU, and it is easily integrated with standard iterative solvers (to solve e.g., least-squares, ridge regression, and lasso problems) as well as in machine learning pipelines.

Speaker: Samuel A. Verburg (Technical University of Denmark)

667 Multi-Channel Anomaly Sound Detection Based on Spatial-Residual Learning and Dual-Encoder Reconstruction

Anomalous sound detection (ASD) is a key task in industrial applications, where reliable detection of machine faults can prevent costly downtime and ensure operational safety. Although single-channel recordings simplify data acquisition and model design, they often do not capture spatial information that can be critical for distinguishing between normal and anomalous acoustic events in complex industrial environments. To address this, we propose a multichannel approach leveraging inter-channel correlations. Two autoencoder-based models are presented: in both, one channel serves as a reference, and differences with the remaining channels are computed. Two encoders process the reference and the remaining channels, respectively, while decoder designs vary—one model uses separate decoders for reference and remaining channels, and the other employs a single decoder to reconstruct all channels jointly. Our approach captures richer machine behavior representations, improving robustness and detection performance in real-world industrial settings.

Speaker: Clara Luzón Álvarez (Universitat de València)

Full paper (PDF)

A10.01 Acoustic, vibroacoustic and elastic metamaterials: S060

Conveners: Steffen Marburg (Technical University of Munich), Bart Van Damme (Empa, Materials Science and Technology)

668 Automated Design Method for Ventilated Cube-Shaped Acoustic Metamaterials in Sound Transmission Loss Applications

The growing use of design methods based on machine learning and optimization has increased the need for methods capable of generating large, high-quality acoustics datasets. To address this, we present a stochastic 3D geometry generation algorithm for irregular, ventilated acoustic metamaterials that support resonant behavior in sound transmission loss applications. The method operates on a voxel grid, using a flood-growth rule with lateral expansion and inter-layer propagation to synthesize fully connected air duct networks. Finite element simulation model used for the analysis of the generated geometries is validated experimentally sing ten cube-structured geometries fabricated as laser-cut birch plywood specimens. Validation employs four-microphone transmission loss measurements in an impedance tube over 500–6400 Hz. Across all designs, the numerical model reproduces experiments with an average absolute discrepancy of 2.3 dB, standard deviation of 1.4 dB, and maximum of 11.8 dB. Although no specific optimization was targeted, some samples already show a promising broadband transmission loss, reaching up to 58 dB at 2650 Hz. Overall, the proposed method provides a validated foundation for data-driven design of ventilated metamaterials.

Speaker: Emilia Żak-Stefanowska (AGH University of Krakow)

Full paper (PDF)

669 Enhancing Sound Absorption in Labyrinthine Metamaterials: From Stochastic Nonwoven Fibrous Fillers to 3D-Printed Infill

This paper explores the evolution of fibrous enhancements in labyrinthine acoustic metamaterials, transitioning from stochastic fibrous fillers to programmable 3D-printed infill. While previous studies demonstrated that manual packing of natural fibres (e.g., dried dill) into labyrinthine slits significantly enhances low-frequency absorption, it suffers from manufacturing difficulties and uncertain reproducibility. Moreover, due to the uneven and random fibre distribution, analytical modelling may be inaccurate and require complex modifications. To address these limitations, we propose the use of Fused Deposition Modelling (FDM) to integrate controlled fibre networks directly into the channels. A comparative study is presented between a bio-filler reference and an engineered Cross-ply (CP) infill pattern. Experimental results from impedance tube measurements reveal a significant advantage of the engineered 30% CP infill over stochastic natural fibres. While the bio-filler achieves absorption at lower frequencies, the 3D-printed infill generates a significantly broader and higher absorption peak at 1050 Hz, reaching nearly perfect absorption. Consequently, the absorption curve of the fully 3D-printed sample exhibits a broader bandwidth and a higher maximum absorption coefficient than the curve obtained for the bio-filler composite. This demonstrates enhanced broadband sound absorption combined with high manufacturing predictability. This work establishes a digital framework for the design of sub-wavelength absorbers with tunable, reproducible dissipation mechanisms.

Speaker: Michał A. Niedzielczyk (Institute of Fundamental Technological Research)

Full paper (PDF)

670 Numerical analysis of vibration in weakly disordered periodic media

Modeling the vibration of large-scale disordered systems presents a significant computational challenge due to the breakdown of translational symmetry. This contribution introduces an efficient numerical framework for characterizing wave propagation in weakly disordered periodic media. Utilizing a Floquet-based approach, we investigate the impact of stochastic structural perturbations, including node displacements and cell wall length variations, within a honeycomb lattice network. The energy band structure is analyzed from two complementary perspectives: a two-scale model featuring spatially localized perturbations, and a globally perturbed periodic model based on small deviations from the ideal geometry. Together, these two approaches elucidate the transition from coherent wave transport to Anderson-type localization as the degree of disorder increases. The proposed framework provides a predictive tool for quantifying how structural irregularities govern band gap degradation and energy redistribution, with direct implications for the acoustic and vibrational design of physical honeycomb sandwich structures.

Speaker: Stepan Avetisov (Aix-Marseille Univ., CNRS, Centrale Marseille, LMA)

671 A Metamaterial Unit Cell Design Combining Mechanical and Acoustical Resonators for Enhanced Transmission Loss

Locally resonant vibro-acoustic metamaterial panels have been shown to offer high sound transmission loss due to their stopband behaviour. Such treatments can significantly reduce the reradiation of sound energy in the stopband by reducing the local vibration of the panel. However, the increased transmission loss occurs over a narrow frequency band and is usually followed by a pronounced reduction in the transmission loss, hindering the broadband performance. This work proposes a metamaterial unit cell concept that combines mechanical and acoustical resonators to mitigate the dip in the transmission loss and thus enhance the overall transmission loss over a targeted frequency range. A numerical model of the coupled configuration is developed to gain insights into the structural and acoustical dynamics and their mutual interaction. A parametric study is performed to analyse the combined effects of the mechanical and acoustical resonator geometry on the transmission loss. Then, an optimisation procedure is implemented to obtain an optimal unit cell design for maximum attenuation. The numerical results indicate that the hybrid design can significantly reduce the minimum transmission loss while maintaining the stopband maximum, thus enhancing the broadband attenuation. The findings demonstrate that integrating mechanical and acoustical resonators within a unit cell offers a promising route toward compact, effective sound-insulating metamaterials.

Speaker: Adam Cavanagh (University of Southampton)

Full paper (PDF)

672 Noise Mitigation Using Sonic Crystal Scatterers Composed of Porous Concrete with Rubber and Cork Inclusions

Sonic Crystals (SCs) offer a promising alternative to traditional noise barriers through the creation of spectral band gaps. However, the performance of rigid scatterers is often limited by narrow attenuation zones. This study investigates the enhancement of SC barriers by utilizing a novel porous concrete matrix modified with recycled rubber shreds or cork powder.Initially, the acoustic properties of the mixtures were characterized via impedance tube measurements, revealing that rubber additives shift the absorption peak toward higher frequencies, while cork powder significantly broadens the absorption bandwidth. These findings were then applied to a case study of a modular SC barrier. Numerical simulations performed after laboratory measurements of the new mixtures, from which equivalent fluid parameters were obtained, demonstrate that scatterers composed of modified porous concrete provide a synergistic effect: the periodic arrangement maintains Bragg interference, while the intrinsic material absorption complements the insertion loss spectrum. The results indicate that cork-modified scatterers are particularly effective for broadband urban noise, while rubber-modified composites allow for targeted attenuation. This research highlights the potential of functionalized cement-based composites as high-performance, sustainable noise control structures.

Speaker: Nicolas Herrera-Leon (University of Coimbra, Dep. Eng. Civil)

Full paper (PDF)

673 Porous Material with Embedded Wiremesh Grating

The design process of a wiremesh grating can be simplified by replacing the air-host medium by a host-porous material. Hence, the new setup allows frameless wiremeshes while preserving broadband resistance of the layers. In this work, we study a porous material with an embedded grating made of inclined wiremeshes. The proposed structure is analyzed using the effective medium approximation and performing a parametric study. The results reveal that the new setup is able to produce wider omnidirectional subwavelength absorption peak than that when hosted by air medium. The omnidirectional subwavelenth absorption is achieved for low flow resistivity porous material. Furthermore, we show that the absorption coefficient is larger when hosted by a porous material at higher frequencies.

Speaker: Juan Pablo Escudero (LAUM, UMR 6613, IA-GS, CNRS, Le Mans Université)

A10.10 Sustainable Materials for Acoustic Applications: S360

Conveners: Francesco Pompoli (University of Ferrara), Philippe Glé (Cerema, Univ. EIffel, UMRAE)

674 Acoustic Characterization of Bonded Oak-Apple Panels by Normal-Incidence Measurements and Inverse Equivalent-Fluid Modeling

The search for sustainable alternatives to conventional porous absorbers has increased interest in bio-based materials for acoustic applications. This paper investigates the normal-incidence sound absorption performance of bonded oak-apple panels manufactured using two different processing strategies and compares their behavior with that of a commercial polyurethane foam of equal thickness. Measurements were carried out using an in-situ acoustic impedance method based on combined sound pressure and particle velocity sensing. To support the interpretation of the experimental results, an equivalent-fluid modeling approach based on the Johnson-Champoux-Allard formulation was adopted. An inverse parameter identification procedure was implemented to estimate effective macroscopic properties from the measured absorption spectra, including airflow resistivity, porosity, tortuosity, and equivalent viscous and thermalcharacteristic-length factors. The identified parameters are discussed as effective descriptors of a heterogeneous bio-based porous composite rather than intrinsic material constants. The results show that oak-apple panels provide competitive mid and high frequency absorption and confirm the potential of this naturally derived material for sustainable acoustic treatment solutions.

Speaker: Rui Ribeiro (Amplitude Acoustics)

Full paper (PDF)

675 A Comparative evaluation of Sound Absorption Coefficient (SAC) in textile based Acoustic Panels

In the context of circular economy and the growing need to mitigate noise pollution and reduce textile waste, recycled textile materials have emerged as viable alternative to conventional acoustic panels. Despite increasing research interest, studies on textile-based acoustic panels remain at an early stage, and a systematic comparision of sound absorption coefficient (SAC) values across different textile waste types is still lacking. This study aims to provide a comparative analysis of SAC values in textile-based acoustic panels based on existing research data. The analysis identified wool, cotton, polyester, flax, viscose and other fibers as the most frequently utilized materials. The findings indicate that SAC performance is strongly influenced by fiber porosity, tortuosity, material composition, as well as by panel thickness and density. Withing the frequency range of 500-5000 Hz, wool and viscosed-based panels demonstrated superior acoustic performance, achieving SAC values of up to 0.9. In contrast, recycled denim, flax and polyester exhibited moderate sound absorption, with SAC values of 0.53, 0.42 and 0.38, respectively, while 100% cotton panels generally showed lower efficiency. Although existing literature highlights the potential of various textile wastes, further reseaerch is required to enhance SAC performance, particularly in the low-frequency ranges.

Speaker: Hanna Bibi (Kaunas University of Technology)

676 Investigating the effect of soil properties on the acoustic behaviour of earthen constructions

Earthen constructions are promising solutions from an environmental perspective, as they rely on locally available resources, require low transformation processes, and can contribute to CO₂ storage when combined with vegetal particles. However, their acoustic performance remains poorly documented, and further evidence is needed to support and promote the use of various building systems (e.g., light earth, wattle and daub, rammed earth, adobe, bricks, and coatings).The CarAc'Terre project aims to investigate acoustic properties across multiple scales, from material to building. At the material scale, a specific focus is placed on understanding the contribution of soil characteristics (particle size distribution, clay type and content, etc.) to acoustic behavior, in order to account for their effects at the wall and building scales.An experimental design was developed with the following input parameters: four different soil origins, three vegetal particle contents, three levels of water content (affecting binder consistency), and two modes of compaction. The resulting samples were characterized both mechanically (Young’s modulus and damping factor) and acoustically (impedance tube measurements at normal incidence, porosity, airflow resistance, etc.). Finally, modeling approaches are proposed under different assumptions (rigid frame, elastic, and poroelastic), and the influence of constituents and implementation methods is analyzed.

Speaker: Philippe Glé (Cerema, Univ. EIffel, UMRAE)

677 Analytical Modelling of the Thermoacoustic Performance of Sustainable Fibrous Materials: An Integrated Characterization of Posidonia Oceanica Fibers

This work investigates analytical models for the thermoacoustic performance of sustainable fibrous materials. Among the analytical formulations considered, porosity is identified as the key parameter linking the semi-phenomenological acoustic formulations of the Johnson-Champoux-Allard (JCA) model to the effective thermal conductivity k_eff modelling. The primary objective is to propose an integrated theoretical framework to balance thermal and acoustic insulation performance in sustainable insulating systems. The methodology consists of correlating the transport parameters of the JCA model with analytical formulations of the effective thermal conductivity, followed by validation through experimental measurements conducted on Posidonia oceanica fibres. The results indicate that increasing porosity enhances thermal insulation, reducing the effective thermal conductivity to values as low as 0.03 W·m⁻¹·K⁻¹. However, this increase leads to a pronounced decrease in the average sound absorption coefficient, which drops from 0.6 to 0.1. Based on this initial analysis, a first attempt at establishing a multifunctional correlation is proposed, identifying porosity as a governing parameter in the maximization of thermal efficiency, which is, in turn, intrinsically linked to acoustic energy dissipation through viscous mechanisms. Overall, this study represents an initial step toward integrating a thermoacoustic perspective into the development of sustainable bio-based materials.

Speaker: Denilson Ramos (University of Ferrara)

678 Acoustic metamaterial window sustainable design: a preliminary approach to implement LCA of metamaterials

The decarbonisation of building envelopes requires multifunctional façade components that integrate indoor comfort with low embodied environmental impacts. Acoustic Metamaterial Windows (AMWs) are an emerging façade solution capable of providing passive noise attenuation while preserving natural ventilation, yet their material sustainability has rarely been quantified. This paper presents a cradle-to-gate Life Cycle Assessment (LCA) of an AMW prototype produced through hybrid digital manufacturing, combining additive manufacturing of polylactic acid (PLA) metamaterial units with laser-cut polymethyl methacrylate (PMMA) panels. The assessment follows ISO 14040/44 standards and applies the ReCiPe Midpoint method to evaluate climate change, resource depletion, and toxicity indicators. Results demonstrate that PLA 3D printing dominates Global Warming Potential due to filament production and electricity demand, while PMMA panels significantly affect climate-related and ecotoxicity indicators. Circular material strategies yield measurable benefits: replacing virgin PLA with 20% recycled PLA reduces GWP by 6.6%, and substituting PMMA with flat glazing further lowers impacts across most categories. Comparative benchmarking against Environmental Product Declarations of conventional window systems shows that optimised bio-based AMWs can achieve embodied emissions comparable to PVC and recycled aluminium frames. Beyond quantitative results, the study proposes a transferable LCA framework for early-stage metamaterial components supporting evidence-based sustainable material selection decisions.

Speaker: Gioia Fusaro (University of Bologna)

A12.06 Validation and Benchmarks in Computational Acoustics: S084

Conveners: Stefan Schoder (IGTE, TU Graz), Marcus Maeder (Technical University of Munich)

679 Benchmark Finite Element Model of a Building Acoustics Test Facility Based on DIN EN ISO 10140

Within an acoustics-oriented design process, virtual prototypes play a crucial role in early design stages by enabling, for instance, the prediction of building acoustics. In this context, the sound transmission loss is an important quantity which can be measured in building acoustics test facilities according to DIN EN ISO 10140. This contribution presents a simple benchmark finite element model of such a test facility to measure the sound transmission loss of single- and double-leaf walls with and without insulation. In doing so, a simulation chain was used combining SALOME 9.14 for the mesh creation as well as the institute’s in-house research code elPaSo for the matrix assembly and solving procedure. As a preliminary step, the simulation chain was verified using the commercial software COMOSL 6.3 considering a small-scale test facility model. Subsequently, a finite-element model of the large-scale test facility was developed using a frequency- and domain-specific discretisation approach. Finally, the sound transmission loss of the three test specimens was determined in one-third-octave bands from 8 Hz to 630 Hz; the insulated double-leaf wall exhibited close agreement with the theoretical transmission-loss curve reported in the literature.

Speaker: Sebastian Schmidt (Institute for Acoustics and Dynamics, TU Braunschweig)

680 Time or frequency domain? Choosing modelling approaches for sound propagation from simple moving sources

When modelling sound radiation from moving sources such as passing trains or aircraft fly-overs, a key methodological choice is whether to work in the time or frequency domain. Although both approaches can be used to compute the received signal from monopole sources along arbitrary trajectories, their respective assumptions, computational costs, and suitability for different applications are not always straightforward. In this work, we systematically compare time-domain and frequency-domain formulations for moving monopole sources. We analyse their underlying assumptions, numerical efficiency, and accuracy across representative scenarios. The results provide practical criteria for selecting an appropriate approach depending on source motion and modelling objectives. In addition, we explore the integration of frequency-domain formulations into inverse methods for source reconstruction.

Speaker: Jan Simon (Berliner Hochschule für Technik)

Full paper (PDF)

681 A new starting point for benchmarks in room acoustic simulations with CHORAS

There is a standing approach of benchmark studies and round robins in room acoustics modelling methods. These studies are typically carried out by participants who are also the developers of the software packages being evaluated. Although this approach is valuable, it leads to user bias and limits the reproducibility of the results. The rise of open-source software for room acoustics has increased access to the underlying methods for a wider audience. Building on this momentum, the open-source CHORAS platform – the Community Hub for Open-source Room Acoustics Software – was recently launched. It offers a joint framework, that allows multiple open-source room acoustics packages to be integrated in. All methods can be operated and applied to room acoustics problems of choice through a web-based graphical user interface. This platform offers new possibilities for repeatable, transparent and efficient benchmark and round robin studies. This work describes how open-source room acoustics software can be easily evaluated against benchmark rooms within CHORAS, and presents an outlook for future frameworks for round-robins and benchmarks in the field of room acoustics.

Speaker: Maarten Hornikx (Eindhoven University of Technology)

682 An extensible Benchmarking Framework for Computational Acoustics

Computational acoustics deals with a variety of data, mathematical models and numerical algorithms along with a disparate set of workflows to achieve insights into the physics of sound.The diversity of its approaches, while valuable, presents a significant challenge: a lack of systematic means to develop, compare, validate, and evaluate their performance against another under consistent reproducible conditions.The concept of benchmarking emerged precisely to address this very need - yet its definition remains inconsistent across the broader community.Existing benchmarking efforts in computational science remain largely ad hoc, domain-specific, and rarely reusable across studies.The absence of standardized benchmarks obstructs transparent, reproducible evaluation of algorithms or feasibility of modelling approaches.To address this, we present the MaRDIMark benchmarking specification that aims to unify these different perspectives - developed within the MaRDI (Mathematical Research Data Initiative) consortium in Germany.The proposed framework enables a transparent means in which data, algorithms, metrics and workflows of varying complexity and formulations be systematically compared against one another for user-defined metrics.We describe its extensible design and demonstrate its application to representative test cases in computational acoustics.With this, MaRDIMark aims for a community-driven FAIR (Findable, Accessible, Interoperable, Reusable) benchmark ecosystem promoting broader accessibility and reusability.

Speaker: Ashwin Sadanand Nayak (Max Planck Institute Magdeburg)

A13.01 Guided waves for NDT & SHM applications: S355

Conveners: Lynda CHEHAMI (UPHF), Markus Saurer (University of Graz), Robert Nuster (University of Graz), Theodosia Stratoudaki (University of Strathclyde)

683 Subwavelength Damage Imaging Using Nonlinear Guided Wave Propagation

This study presents a combined numerical and experimental investigation of nonlinear guided waves generation and propagation in weakly nonlinear isotropic elastic plates, with particular emphasis on their interaction with a crack-type defect. The analysis focuses on the second harmonic guided wave (SHGW) generated from the fundamental symmetric mode in an aluminium plate. Numerical simulations were conducted to evaluate the influence of excitation conditions and intrinsic material nonlinearity on SHGW generation while eliminating spurious nonlinear effects introduced by measurement instrumentation. Experimentally, guided waves were generated using a calibrated piezoelectric wafer transducer bonded to the surface. The resulting wave fields were measured using a three-dimensional scanning laser vibrometer, enabling full-field acquisition of both normal and tangential surface displacements. This methodology allows reliable isolation and characterization of the second harmonic component. A wavenumber analysis imaging technique is then implemented for damage localization. The results from both approaches confirm that the secondary mode generation is feasible using a narrowband excitation signal. However, the amplitude remains considerably weak in the pristine structure. In the presence of damage, the second and third harmonics are clearly identified. Their amplitudes increase significantly with the severity of the damage. This constitutes a very sensitive indicator for early detection and characterization of defects. The corresponding reconstructions successfully allowed the localisation of the damage.

Speaker: Aziz BOUZZIT (SATIE Lab - CYU)

684 Detection and Scattering Response Estimation of a Heterogeneity in a Thin Elastic Plate using Lamb Wave Beamforming and Reverberation Analysis

Acoustic wave propagation in low-attenuation bounded media produces long-lasting reverberation signals. Traditional non-destructive testing and imaging methods typically rely only on the first-arrivingwave packets, overlooking possible information contained in the later reverberating part (the codas).In this work, these codas are processed in order to compute the scattering cross section of a heterogeneity in a thin elastic plate using a statistical model for reverberation developed in earlier works. More precisely, within the frame of this model, a mathematical expression linking the scattering cross section to the mean of the squared envelope of the reverberated signals. Separately, Chehami etal. developed a beamforming-based imaging method (taking into account the dispersion) enabelingprecise localisation of the defect. Building on these contributions, the present work demonstratesthat defects can be characterized using only two parameters: their scattering cross-section and thecontrast in the corresponding localisation image. Numerical simulations, in which ultrasonic signalsin the 10–30 kHz range (A0 Lamb mode) were recorded at multiple known positions on the platesurface, were conducted on a small cylindrical defect. The scattering cross-section was first estimatedusing nonlinear fitting applied to the averaged envelopes of reverberated signals and compared withboth a theoretical model developed in Norris and Vemula’s works and a conventional approachbased only on direct wave arrivals. Image contrast was then computed from localization results. Finally, a characterization method is proposed adressing the inverse problem by combining thedefect’s scattering response with imaging contrast, expressed as a ratio relative to an isotropic defect.Promising preliminary results were obtained.

Speaker: Lynda CHEHAMI (UPHF)

685 Numerical Study on the Effect of Energy Transfer from Primary to Secondary Wave Field with Non-Phase Matched Conditions

The detection of localized material nonlinearities using guided waves is a valuable approach for nondestructive characterization and damage detection due to their ability to propagate over long distances and interrogate inaccessible domains. Of particular interest is the so-called wave mixing, in which the mutual interaction of two primary waves generates a secondary wave field. Phased-array scanning enables control of the location of the mixing-zone of the primary waves while the transmitters remain in a fixed position.Aggravating damage detection, the amplitude of the resulting secondary wave is not invariably cumulative, especially when the group velocities differ and the mixing zones are finite. In addition to the secondary wave, changes in the primary waves, which may be attenuated or modified because of energy transfer to the secondary field, can be considered for the evaluation of measurements and the detection of nonlinearities.This study numerically investigates how the energy transfer affects the amplitudes of propagating primary waves for the detection and assessment of a finite nonlinear material region in an otherwise linearly elastic plate. In addition a parametric study on the mixing-zone size and its influence is presented.

Speaker: Christoph del Castillo Albildo (University of Siegen)

Full paper (PDF)

686 Influence of the Spot Size in Transient Grating Spectroscopy

The influence of spot size on the quality of the time- and frequency-domain results of transient grating spectroscopy (TGS) was investigated. The experiment was performed on a single-crystalline sample of Ni with a highly symmetrical cut (110). Sharpening of the frequency peaks of the SAW was observed for larger spot sizes. However, the best obtainable sharpness of the frequency peaks strongly depends on the crystallographic direction of the propagating SAW. Ultra-transient effects observable in UTGS measurements depended strongly on the power density of the pump laser.

Speaker: Jakub Kušnír (Institute of Thermomechanics of the Czech Academy of Sciences)

Full paper (PDF)

687 Data-based defect localization with laser ultrasound for samples with dominant higher-order scattering

The advent of novel technologies capable of producing small samples with complex shapes, such as additive manufacturing and 3D printing, has given rise to new challenges for defect detection and visualization techniques, including laser ultrasound imaging. When the ultrasound wavelength is comparable to the sample dimensions, multiple reflections can dominate the wave field, resulting in severe artifacts in conventional physics-based reconstructions. In this study, a simplified numerical test case is used to compare a physics-based reconstruction approach with a data-driven convolutional neural network for defect localization in samples exhibiting dominant higher-order scattering. The results indicate that, within the investigated setup, the data-driven approach can identify defect locations in cases where strong multiple reflections limit the applicability of physics-based reconstructions. For the considered synthetic dataset, accuracies of up to 95% were obtained on unseen test samples, and preliminary tests suggest that the trained model can generalize to selected cases outside the exact training grid. Overall, the study provides a first proof of concept that data-driven methods are a promising option for defect localization in laser-ultrasound scenarios dominated by higher-order scattering.

Speaker: Markus Saurer (University of Graz)

Full paper (PDF)

688 Non-contact low-frequency acoustic methods for crack/corrosion pit detection in pipes

Underground buried pipeline networks play a vital role in sustaining modern urban life by enabling the transportation of essential fluids such as water, oil, and gas. In the UK alone, there are over 300,000 km of clean water mains. However, water leakage in buried pipeline networks is a major challenge due to limited data available on the structural condition and ageing. Much of this loss comes from hidden/background leaks which could be due to small defects such as micro-cracks, corrosion, or wall roughness that remain undetected until they grow into major failures. Non-contact low frequency acoustic methods offer robust and easy to deploy technology capable of detecting in gases or liquids hidden changes in the pipe conditions. In the present work, a perturbation method under low-frequency assumption has been adapted to deal with a waveguide (pipe) problem with localized pit, which resembles the corrosion pits/cracks in walls. The developed methodology is integrated into non-contact pipe condition monitoring tools used in the following related research projects: (i) AI:LINER, an EU multi-institutional project that combines novel acoustic solution, CCTV data, AI-based failure detection, and in-situ monitoring techniques to enhance the asset management life cycle of sewer networks; and (ii) an EPSRC-funded project, to develop bio-inspired micromachine sensors for measuring acoustic quantities.

Speaker: Mriganka Shekhar Chaki (University of Sheffield)

A16.00 Room Acoustics: S112

Conveners: Jamilla Balint (Rohde Acoustics), Francesco Martellotta (Politecnico di Bari, DARCOD), Mélanie Nolan (Universidad Politécnica de Madrid)

689 Level histograms of Schroeder decay-curves: another view on room impulse responses

Schroeder-curves or energy decay curves (EDC) of room impulse responses have a non-positive slope. As a consequence, when quantizing the decay range into level bins, the histogram of probability distribution of EDC levels shows basically the local decay slope or instantaneous decay time for each bin, because heights of bins denote times required to decay in each bin.This simple yet undiscussed representation has interesting and useful features. For example, signal to noise ratio of the response can be estimated without iterative steps. In addition, one can easily identify truncation error or qualify multi-exponential decays.The paper discusses how important features (single or multi decay rates, distinct echoes, truncation, noise) of theoretical decays appear in EDC histogram plots and how these plots relate to basic room acoustic parameters. The paper demonstrates also the application of this representation on measured enery decay curves.

Speaker: Andor Fürjes (aQrate Acoustics Ltd.)

690 Influence of sound source type and position and room acoustic conditions on room acoustics and speech intelligibility measurements

The paper examines the influence a chosen position and type of the sound source have on measurements of room acoustic parameters and speech intelligibility, as well as the influence of the acoustics of the room itself in terms of reverberance. The measurements were performed in two rooms similar in size but acoustically different, one of them being an acoustically treated multimedia studio used for recording, and the other being an acoustically untreated classroom in a higher education institution. Three positions of the sound source were defined: the first one located centrally in the room, the second one in front of the front wall, and third one located in the corner of the room. Three loudspeaker sound sources were examined: a do-it-yourself omnidirectional dodecahedral source, a commercially available single-driver omnidirectional source, and a bookshelf-sized studio monitor as a directional source. The orientation of the dodecahedral source and the studio monitor with respect to the microphone positions were also taken into consideration. A grid of six fixed, centrally located microphone positions was defined in the room. Four additional microphone positions per source position were placed on a line stretching from the source into the room with a two-metre increment in the distance from the source. The measurements were performed according to EN ISO 3382-1 and IEC 60268-16 standards using the integrated impulse response method to obtain reverberation and clarity parameters, as well as the speech transmission index from the measured impulse responses. The results of the measurements were analysed and discussed in terms of the examined potential factors of influence.

Speaker: Marko Horvat (University of Zagreb Faculty of EE and Computing)

691 Quantity over Quality? Comparing measurement error in online and laboratory evaluations of room acoustics

Comprehensive assessments of room acoustic quality rely on perceptual evaluations. Conducting such evaluations online avoids time-consuming laboratory experiments at the expense of experimental control. This seems hardly feasible, given the technological demands of the accurate reproduction of room auralizations. However, since playback system deficiencies differ between subjects, their individual errors would just add noise to the inevitable measurement error, which might be counterbalanced by including more participants. This study therefore compares perceptual ratings of six rooms using the Room Acoustical Quality Inventory (RAQI), obtained from two laboratory studies with 40 and 31 participants, respectively, and an online study with 121 participants. The two laboratory studies employed either dynamic binaural renderings, or a static auralization with fixed head orientation, which was also used in the online study. Bayesian hierarchical regression models based on Multivariate Generalizability Theory were used to quantify measurement error attributable to the participants and to compare true score estimates of the measured RAQI factors. Preliminary results indicate that differences in the estimated true score, participant error variance, and score reliability depend on the specific RAQI factor and are largest within the Quality and Reverberance factors. The results of the laboratory experiment with fixed head orientation showed differences to the dynamic condition that were comparable to those of the online study. While a more diverse set of evaluated rooms would be needed for more conclusive results, the proposed method allows for a sophisticated assessment of measurement error in online experiments, instead of ruling them out from the outset.

Speaker: Markus von Berg (Hochschule Düsseldorf)

692 Ambi vs. Omni: Evaluating the Potential of 3D Microphone Arrays in Room Acoustic Measurements

Room acoustic properties are typically assessed using omnidirectional microphones, from which room impulse responses are derived and used to calculate scalar acoustic parameters. However, spatial measurement techniques are essential for assessing sound fields in spaces equipped with active acoustics systems, since they generate direction-dependent and dynamically changing acoustic responses that cannot be fully characterized by single-point, omnidirectional measurements. In recent years, three-dimensional microphone arrays have become increasingly accessible and are widely applied in other areas of audio processing. However, their use in standardized room acoustic measurements remains limited, as existing standards (e.g., ISO 3382) are primarily based on conventional measurement approaches.In this study, a third-order Ambisonics microphone array is employed to capture spatial room impulse responses (SRIRs) under varying acoustic conditions in a space equipped with an active acoustics system.First, the omnidirectional component of the SRIRs is compared with conventional omnidirectional measurements under identical conditions. Initial results suggest that the resulting acoustic parameters are largely consistent, with only minor deviations observed under specific conditions.Subsequently, the potential of the spatial information contained in the SRIRs is examined for extended analysis. This includes the direction-dependent evaluation of established ISO 3382 parameters, such as reverberation time and clarity, enabling their representation as function of direction. The study discusses the potential role of 3D microphone arrays in future room acoustic measurement practices, and highlights the necessity of establishing three-dimensional measurement techniques to adequately assess the properties of active acoustics systems and spatially complex sound fields.

Speaker: Elias Braun (Institute of Electronic Music and Acoustics)

Full paper (PDF)

A12.12/A16.13 Diffuse Sound Fields and Reverberation: S347

Conveners: Franz Zotter (University of Music and Performing Arts), Albert Prinn (International Audio Laboratories Erlangen)

693 Diffuse Sound Field Modelling and Metrics and their application in the Space Launch Environmental Test Sector

The concept of a diffuse sound field is long established and oft assumed in room acoustics. Metrics to quantify the ‘diffuseness’ of a given sound field also exist, and include cross-correlations between spaced sensors, the ratio of intensity to potential energy, and isotropy. Nonetheless, recent research has shown that there remains knowledge gaps, notably around the relationships between these metrics, their confidence intervals, the source layouts they inform, and even on some details of the underlying mathematical theory. In room acoustics, interest is often in quantifying the acoustic qualities of a space, for example in a reverberation chamber for absorption measurement. In the space launch environmental test sector, where spacecraft and payloads need to be qualified against high intensity sound fields representative of the launch environment, similar usage occurs. But here another more active use is also emerging, as the means to specify the sound field synthesis target in Direct Field Acoustic Noise (DFAN) testing systems, in which large arrays of high-powered loudspeakers positioned close to the test article produce the require noise spectrum. This paper will summarize approaches from that sector and some new developments in metrics and modelling.

Speaker: Jonathan A. Hargreaves (Acoustics Innovation Institute)

694 Towards Diffuse Sound Fields in Immersive Listening Rooms: an Experimental and Numerical Approach

High-quality immersive listening experiences require an adequately diffuse sound field to ensure spatial uniformity and listener envelopment. However, in small and medium-sized rooms, acoustic conditions are often compromised by dominant modal behaviour, leading to pronounced spatial unevenness of the sound field. This study outlines an integrated design approach for optimising an existing immersive listening room (280 m3) by combining in situ measurements with advanced numerical simulations. Initial acoustic characterisation of the study site revealed significant spatial non-homogeneity and a strong modal response, with reverberation times exceeding 1.1 seconds in all the octave bands assessed. A finite-element approach enabled enhanced spatial distribution analysis and the preliminary evaluation of corrective strategies. The proposed intervention features a system of perforated gypsum panels with variable air gaps on the rear wall behind the listener, designed to provide broadband absorption and enhance macro-diffusion through edge diffraction. This is complemented by high-density curtains for mid-high frequency control. Experimental and numerical results show that the implemented treatments successfully bring the reverberation time within recommended target ranges while also contributing to a more uniform sound pressure distribution, particularly at low frequencies.

Speaker: Giulia Fratoni (University of Bologna)

695 Covariance of the Frequency Response Function of (Vibro-)Acoustic Built-Up Systems in the Mid-Frequency Range

In the mid-frequency range, the high modal density of flexible (vibro-)acoustic components makes deterministic methods computationally expensive, while small spatial variations in geometry, material properties, or boundary conditions may have a (random) wave scattering effect. An efficient statistical analysis approach is therefore required, and since spatial and frequency averaging are commonly used to reduce response variability in reverberant systems, obtaining both the mean and the covariance of frequency response functions is essential. This paper therefore investigates the covariance of frequency response functions of built-up systems comprising flexible components, treated as random, connected to stiff components, modelled deterministically. Such configurations are common in practice, such as aircraft fuselages, where stiff frames support thin skin panels, or buildings, where flexible airspaces (rooms) are bounded by stiff walls and floors. The covariance is considered for different response locations, excitation locations, and excitation frequencies.To this end, a cross-frequency extension of the diffuse field reciprocity relationship is derived and applied to the analysis of such systems. This relationship links the cross-spectral reverberant forces of the random components to the system response, and is expressed in terms of the mean and covariance of the frequency response functions of the individual components. Substituting this relationship into the system-level dynamic equations yields expressions for obtaining the mean and covariance of the full system response.The theoretical framework is validated against Monte Carlo simulations of random built-up systems, demonstrating good agreement and confirming the practical potential of the proposed method for efficient uncertainty quantification in mid-frequency (vibro-)acoustic analysis.

Speaker: Cédric Van hoorickx (Eindhoven University of Technology)

696 Influence of the Angular Resolution on the Directional Energy Decay Curve

The sound field in rooms with non-uniform distribution of absorption or coupled volumes is well known to be subject to distinctly anisotropic reverberation. This is often evident as a multi-exponential energy decay with distinct directional properties. In recent years, the directional energy decay curve has been proposed and applied to investigate and quantify such phenomena. However, it was found that the angular filter response of the receiver array has a strong influence on the angular-temporal evolution of the captured directional energy decay curve.In this contribution we introduce a stochastic energy decay model capable of incorporating the angular filter function response. Based on the stochastic model, we investigate how the limited angular resolution of the receiver array influences the angular-temporal separation of directionally dependent decay processes. Finally, different angular filter functions are compared and with respect to their separation performance.

Speaker: Marco Berzborn (Eindhoven University of Technology)

697 Sports halls, Austrian Standards and the (non) diffuse Sound field

Recently, complaints about room acoustics in Austrian sports halls have increased noticeably. This development can be traced back to a rising number of acceptance measurements and the simplified comparability with the specified reverberation time of the latest versions of the local standard (ÖNORM B 8115-3) and guideline (ÖISS RL10). Our measurements and those collected from partners show that (almost) no sports hall meets the new requirements, or the old one for that matter. To determine whether this reflects an actual problem for users of these sports halls or is a symptom of the kind of sound field and how it is calculated and measured, room acoustic measurements were carried out in newly constructed schools in Vienna and compared with the planned acoustic design. It was also investigated whether the sound strength in the respective rooms correlates with the measured reverberation time. The preliminary findings suggest that some but not all deviations between the measured reverberation times and the planned values can be considered problematic, and that evaluating the acoustic quality of sports halls solely based on reverberation time may be misleading.The findings are restricted by the small number of measurements conducted and the limited scope of the measurements from external sources. Further investigations should therefore examine additional sports halls regarding the relationship between sound strength and reverberation time. The insights gained from the present research contribute to present and future standards like the new ÖNORM B 2608 “Sports halls - Guidelines for planning and execution”. Initial outcomes from those committees will be highlighted during the presentation.

Speaker: Sebastian Kraync (Saint-Gobain Austria GmbH)

A15.06 Psychoacoustics of everyday sounds: S359

Conveners: Daniel Oberfeld-Twistel (Johannes Gutenberg University Mainz), Emmanuel Ponsot (STMS (Ircam-CNRS-SU))

698 A rationale for elaborating tests assessing environmental- sound perception

Shafiro and Gygi (2004) – who pioneered psychoacoustic research on environmental sounds – promoted a rigorous methodology for selecting appropriate stimuli to study the auditory perception of this category of sounds by human listeners. Here we propose a theoretical rationale to complement this methodology, based on an ecological approach grounded in recent findings concerning the perception of biotic and abiotic sounds and complex scenes. This rationale emphasizes the importance of choosing acoustic stimuli and interpreting test outcomes based on: (i) the sounds’ spectro-temporal modulation characteristics and causal properties; (ii) the role of low- vs. high-level auditory mechanisms in the perception of environmental sounds, with a fundamental distinction between biotic and abiotic sounds, animal vocalisations and action sounds, and textural and non-textural sounds; (iii) context (i.e., the type of environment: domestic indoor, urban outdoor, and natural outdoor); (iv) the type of monitoring tasks to be performed by participants and their hearing-related functional character and emotional valence in daily life. This rationale should help in designing test batteries that are better suited to assessing the perception of environmental sounds in healthy listeners, patients with hearing impairment and brain damage, as well as in designing rehabilitation devices and audio processing algorithms.

Speaker: Christian Lorenzi (Ecole normale supérieure, CNRS)

Full paper (PDF)

699 Reorganization of Natural Soundscape Perception Under Cochlear Hearing Loss

Preliminary findings indicate that cochlear hearing loss impairs discrimination of natural soundscapes according to habitat, time of day, and season.This psychophysical study investigates how cochlear damage affects natural soundscape perception and the influence of ecological factors (habitat, time of day, precipitation level). Normal-hearing (NH) and hearing-impaired (HI) listeners identified and estimated the number of biological and geophysical sound sources in soundscapes from nine pristine habitats, four times of day, and two precipitation periods. HI listeners were tested with unprocessed soundscapes, and NH listeners with unprocessed or processed soundscapes simulating absolute threshold elevation, loudness recruitment and reduced frequency selectivity. All sounds were presented diotically at 55 dB(A).Results show that cochlear hearing loss reduces the perceived scene richness, attenuates the influence of ecological factors on perception and reorganizes scene composition: biological sound sources (e.g., birds) tend to be reassigned to geophysical sound categories (e.g., wind). These effects mainly result from absolute-threshold elevation and loudness recruitment, with reduced frequency selectivity further degrading auditory segregation between biological sound sources and reducing the number of perceived birds.Altogether, these findings highlight the dramatic consequences of cochlear damage on natural-scene perception.

Speaker: Robin Wendling (LSP, Ecole normale supérieure, ENS PSL)

700 Psychoacoustics in realistic everyday conditions using real- time 6-DoF environment and room acoustics simulation

Perception of everyday sounds typically involves motion of the source and receiver, sound reflections, diffraction, and reverberation. Contrasting classical psychoacoustics with static headphone rendering of simplified stimuli, hearing research in such realistic conditions requires virtual acoustics methods. liveRAZR is a low-latency interactive virtual environment and room acoustics simulator designed for auralization in dynamic environments with full six degrees of freedom (6-DoF) motion for both sources and listeners. It employs geometric acoustics (GA) techniques, including image-source modeling for early reflections, combined with computationally efficient late-reverberation modelling to achieve perceptually plausible and physics-driven results under real-time constraints. To address limitations of purely specular GA propagation models, liveRAZR incorporates edge diffraction modeling. This allows the simulation of sound propagation into shadowed regions and around obstacles, improving realism in complex scenes with partial occlusion. Diffraction paths are evaluated efficiently to remain compatible with interactive performance requirements. The system supports binaural rendering using measured high-resolution head-related impulse responses and integrates source directivity datasets via SOFA files. Directional filtering is applied through partitioned convolution, enabling accurate reproduction of spatial cues during listener and source movement. Implemented as a multi-threaded C++ application, liveRAZR can operate standalone or be controlled externally via OSC messages from a controller application such as, e.g., Unreal Engine. This architecture supports continuous updates of scene geometry and listener state. We discuss the psychoacoustically relevant key features of the suggested system and present use cases involving auditory distance perception, exemplifying how psychoacoustics can be performed in realistic, yet lab-controlled and reproducible conditions, approaching natural behavior and perception of everyday sounds.

Speaker: Kevin de Haas (Medizinische Physik, Carl von Ossietzky Universität Oldenburg)

701 Do you know who you’re talking to? - Priming the identity of a familiar talker improves speech-in-noise intelligibility

Recognition of familiar individuals relies on information from multiple sensory-modalities, including faces and voices. Existing person-recognition frameworks propose that face and voice processing initially occur through separate modality-specific pathways before converging at a modality-independent identity stage. While cross-modal identity processing is known to facilitate person recognition, its broader perceptual consequences remain less well understood. One important example is the familiar-talker benefit, whereby speech from familiar individuals is more intelligible in noisy environments than speech from unfamiliar individuals. In the present study, we investigated whether this familiarity benefit extends to celebrity voices and whether cross-modal identity priming enhances speech perception in noise. Participants listened to speech produced by familiar celebrity talkers and unfamiliar voices under noisy listening conditions. We additionally examined whether presenting identity cues, including a person’s face and name, before speech onset modulated intelligibility. Our findings demonstrate that celebrity familiarity produces a measurable speech-in-noise advantage, indicating that relatively sparse but repeated exposure to famous voices is sufficient to generate perceptual benefits. Importantly, the familiarity advantage was strongest early in the sentence, suggesting that listeners rapidly exploit stored knowledge about familiar vocal characteristics during speech processing. We also found that identity priming, through faces and names, enhanced this benefit, supporting the idea that cross-modal person representations can influence auditory perception prior to speech onset. These findings suggest that seeing a familiar individual may prime neural representations of their voice, preparing the auditory system to process expected vocal statistics. More broadly, the results extend models of person recognition by demonstrating that shared identity representations can influence perceptual processing beyond recognition alone.

Speaker: Joseph Sollini (University of Nottingham)

702 LOUNGE: a living room as a laboratory environment for the research of perceived noise

Research on perceived noise annoyance seeks to identify conditions into circumstances where certain sound events elicit adverse psychological or physiological effects to those exposed. Within this research field two primary methodologies are often applied: laboratory-based studies, utilising sound laboratories or noise simulators to control and manipulate sound exposure, and field studies, which involve assessing the impact of sound events on individuals in real-world environments where they are already exposed to such noises. The advantage of laboratory experiments is that the sound events are fully controlled, while an empirical study measures people in their natural environment. To reconcile these approaches, the LOUNGE laboratory environment has been developed. It offers a controlled environment, combining the benefits of laboratory experiments and field studies, by replicating a natural living room setting, thereby enabling the exposure of participants to various sound events in a comfortable and realistic manner. The LOUNGE is equipped with an audio reproduction system to generate directional sound events and cover a frequency range of 19Hz-20kHz . A simulation system generates the sound events by combining positional data on sound sources with the expected sound signature, for instance of an aircraft flyover sound. A one-way window connects the room with a observation room to examine people’s behaviour. Initial studies that have been undertaken already include the comparison of a VR-based noise simulator with the LOUNGE environment and the evaluation of aircraft noise events while conducting two different cognitive tasks.

Speaker: Roalt Aalmoes (Royal Netherlands Aerospace Centre NLR)

Full paper (PDF)

A18.01 Restorative soundscapes in cities: S128

Conveners: André Fiebig (TU Berlin, Department of Engineering Acoustics), Margret Sibylle Engel (Dresden University of Technology)

703 Field and laboratory soundscape assessments of contrasting urban sites: a case study in turin

Soundscape studies allowed a broader understanding of how sound contributes positively to urban quality of life. Investigating soundscape effects in-situ entails logistical challenges that can be simplified by conducting laboratory-based assessments, which emerged as a viable alternative with sufficient ecological validity. This study aims to further investigate said validity by analyzing the soundscape of two distinct urban sites in Turin, both in-situ and in the laboratory, using third-order Ambisonics reproduction paired with a 360° video headset. To represent contrasting acoustic environments, a heavily trafficked intersection and a small public park were chosen. Participants (N=15) rated each site using the ISO 12913 perceptual attributes under both field and laboratory conditions. In accordance with the findings in the literature, the public park soundscape was perceived as more pleasant in both conditions. Statistical comparison via Wilcoxon signed-rank tests, with Benjamini-Hochberg correction for multiple comparisons, revealed no significant differences between field and laboratory assessments. Results contribute to confirm the ecological validity of laboratory-based soundscape evaluation methods.

Speaker: Lorenzo Dell'Anna (Politecnico di Torino)

Full paper (PDF)

704 Learning from the Application of Criteria for Recognising Potential Health Restoration Soundscapes (HeReS) in Urban Parks: Towards Soundscape-Informed Design

Urban green spaces are increasingly recognised for their potential contribution to psychological restoration and public health. However, conditions for potential Health Restoration Soundscapes (HeReS) are not uniformly distributed within urban parks and may vary across different areas and use conditions. Understanding this variability is essential for supporting soundscape-informed urban planning and design.This study revisits the application of HeReS-Criteria in Quinta Normal Park, a large urban park located in Santiago, Chile. Originally developed as a practical approach for identifying the potential of urban green spaces to support health restoration, the HeReS-C integrate five complementary conditions related to naturalness, sound levels, perceived sound sources, soundscape assessment, and sensescape coherence.The study combines multidimensional and synchronic soundscape data acquisition, integrating objective and subjective assessment techniques simultaneously applied in situ. Data collection included 263 questionnaires conducted at multiple locations throughout the park during both weekdays and weekends.Spatial variability was analysed through the distribution and overlap of the HeReS-C across different park sectors. Results show substantial variability in the fulfilment of restorative soundscape conditions within the same urban park, suggesting that restorative potential should be understood as spatially dynamic rather than as a homogeneous characteristic of green spaces.The findings also suggest the relevance of temporal dynamics in restorative soundscape assessment. Rather than presenting restoration as a fully demonstrated outcome, this work discusses how the application of the HeReS-Criteria may contribute to soundscape interventions, urban green space planning, and soundscape-informed design approaches.

Speaker: Pablo Kogan (ChucaoLab, Dept. of Sound, Faculty of Arts, University of Chile)

Full paper (PDF)

705 Memory retrieval in urban and indoor soundscapes: Affective and comfort associations from soundwalks in Brazil and Germany

Memory retrieval is closely tied to emotional and affective content. Prior research has shown that emotionally charged stimuli enhances recall and shape the subjective quality of remembered experiences. In the context of soundscapes, everyday acoustic environments can act as powerful cues for evoking autobiographical memories that are often intertwined with perceptions of comfort and affective appraisal. This study investigates memory retrieval in urban and indoor soundscapes, focusing on associations with affective and comfort attributes derived from soundwalk methodologies conducted in Brazil and Germany. In Brazil, data were collected at Inter-Noise 2025 with 35 participants who evaluated two contrasting outdoor environments: a busy avenue and a quiet square. In Germany, data were gathered with 20 participants during DAGA 2026, covering two indoor areas within a conference venue and one outdoor area characterised by traffic noise and social interaction. Participants assessed affective and comfort-related attributes, as well as providing classified responses regarding emotions, preferences (like/dislike) , appropriateness, and overall feeling. Reported memories were categorised and analysed in relation to these perceptual dimensions. Using correspondence analysis, associations were identified between specific types of retrieved memories, comfort attributes (noisy and comfortable), and affective attributes (pleasant, not annoying, exhausting and relaxing). The results highlight systematic relationships between soundscape characteristics, affective-comfort evaluation, and memory content, suggesting that both environmental context and perceptual appraisal influence the nature of memory retrieval. These findings contribute to a deeper understanding of how urban and indoor acoustic environments support emotionally grounded memory processes, informing sound design strategies that promote comfort and positive experiences.

Speaker: Marcos Holtz (University of São Paulo (USP))

Full paper (PDF)

706 Multisensory perceptions of tranquillity: Asymmetries in a representative spatial dataset

Tranquillity is recognised as a key component of restorative environments. However, it remains insufficiently conceptualised as a fundamentally multisensory experience and one that spans from the ideal to the pragmatic. This study examines how tranquillity is constituted through interactions between sensory modalities and differs between an attributed ideal landscape and as experienced in the latest nature visit.Empirically, the study draws on a national survey from Denmark (n = 9,125), including georeferenced mappings of perceived tranquillity and open-text responses (n = 1,222). A qualitative coding analysis identified sensory dimensions and environmental features, complemented by spatial analyses of distribution across landscape types and urban–rural gradients.Results show that positive contributions to tranquillity are articulated as immersive experiences, often without explicit sensory reference. When specified, visual and auditory cues dominate e.g. natural landscapes, water, and natural sounds. In contrast, negative perceptions are more explicitly shaped by auditory disruptions, particularly human activity and traffic noise. Whilst the ideal relies on natural landscapes away from the urban core, the urban core holds much of the experienced tranquillity. This reveals a consistent asymmetry: while positive experiences rely on integrated sensory conditions, disruption is mainly associated with sound. Although tranquillity is associated with peripheral landscapes, people find pragmatic tranquillity in the urban ones.Conceptually, the study extends soundscape frameworks by integrating multisensory environmental perception of tranquillity. Methodologically, it demonstrates the potential of triangulating large-scale participatory mapping with qualitative and spatial analysis to capture environmental perception across contexts.

Speaker: Ina Hildeman (Copenhagen University)

19:00 Social Dinner

20:00 Jam Session

Friday 11 September

A03.04 Acoustics of wooden buildings: S016

Conveners: Jean-Luc Kouyoumji (FCBA Institute of Technology), Heinz Ferk

707 Heavy fillings on CLT floors I – Comparative Analysis of Impact Sound Pressure Levels in Small-Scale Tests

Elastically bound granular chippings are widely used in cross-laminated timber (CLT) floor constructions to improve the impact sound insulation of the construction. Although laboratory measurements often demonstrate very low normalized impact sound pressure levels for such systems, significant variability has been observed in practice. In the present study the impact sound pressure measurements of different chipping types in small-scale tests are shown. Various parameters – including impact sound insulation, type of chippings, filling height, and binding agents – are investigated. The results are presented in terms of impact sound reduction in order to more clearly identify frequency ranges of improvement and deterioration, enabling robust conclusions to be drawn. The material properties of the different heavy fillings and binding mechanisms were also investigated and are shown in the second part of this paper.

Speaker: Christian Lux (Holzforschung Austria)

Full paper (PDF)

708 Heavy fillings on CLT floors II – correlation of the material properties and its acoustic performance

Elastically bound granular chippings are widely used in cross-laminated timber (CLT) floor constructions to improve the impact sound insulation of the construction. Although laboratory measurements often demonstrate very low normalized impact sound pressure levels for such systems, significant variability has been observed in practice. Previous research [1, 2] also supports the assumption that the type of the chippings and their binding agent strongly influence the acoustic performance. In addition to the impact sound pressure measurements of different chipping types (loose, elastically bound, mechanically bound, etc.; shown in [3]), the material properties of the different heavy fillings and binding mechanisms were also investigated in the present study. Specifically, the longitudinal wave velocity cL in m/s and the loss factor η were determined experimentally and correlated with their acoustic performance in CLT floors. In this paper a short overview of the measurement method and its development is provided. The impact sound pressure measurements already showed that the type of binding agent has a greater effect on the impact sound pressure level than the thickness (i.e., mass) of the filling itself (see [3]). Furthermore, the findings indicate that the internal damping behavior (loss factor η) of bound fillings is a key parameter for predicting the acoustic performance of timber floor systems.

Speaker: Alexander Stenitzer (Holzforschung Austria)

Full paper (PDF)

709 Design Implications for Impact Sound Insulation in Timber Floors: Experimental Results on CLT Assemblies

Sound insulation in timber construction remains a major challenge to the wider adoption of lightweight building systems, particularly Cross-Laminated Timber (CLT) floors. This paper presents the results of a recent laboratory measurement cam-paign focused on the acoustic performance of floating floor systems incorporating resilient materials installed under screed on CLT slabs. Measurements were carried out under controlled laboratory conditions, in the absence of flanking transmis-sions, in order to assess the intrinsic performance of the tested assemblies. The results are discussed in comparison with pre-vious experimental campaigns on similar configurations, as well as with more recent data available in the scientific literature, highlighting both consistencies and discrepancies. A further objective of the study is to compare the measured impact sound insulation performance with the requirements set by major European acoustic regulations, and to evaluate the compliance of the investigated solutions across different national frameworks. The work will be the basis to extend the knowledge on mate-rials and isolation systems for future comparison on heavyweight slabs, enabling a direct comparison of performance as a function of the structural floor, considering mass and stiffness of the supporting structure, as well as the role of the physical and mechanical properties of resilient materials. The outcomes provide useful insights for the acoustic design and optimiza-tion of high-performance timber buildings.

Speaker: Leonardo Luison (ISOLGOMMA SRL)

Full paper (PDF)

710 Accurate and Efficient Prediction of Sound Insulation using Soprano: a Robust Method validated with Floating Floors

Engineering offices and manufacturers of acoustic building systems or materials often rely on measurement data to determine the adequacy of solutions for specific situations. This typically involves numerous costly and time-consuming laboratory measurements, making it difficult to efficiently explore and optimize various design configurations.In this paper, a method for predicting the sound insulation of multilayer structures is validated against experimental data from real-life floor systems. The method is based on analytical three-dimensional elastodynamic analysis of the system components, so it combines accuracy with computational efficiency. Accuracy ensures reliability, while efficiency is crucial for optimization, where numerous simulations are needed, for example to identify the ideal layering or material properties of a(n) (inter)layer. The prediction method accounts for arbitrary layering, finite dimensions, boundary conditions and resulting modal behavior, as well as frequency-, load- and temperature-dependent material properties.We take a closer look on the prediction of impact sound insulation of floating floors, demonstrating the model’s capabilities. Several different systems, consisting of concrete base floors, cross-laminated timber (CLT) panels, full-surface elements and wet/dry screeds, were tested to show the model’s robustness and reliability.

Speaker: Bart Van de Velde (SonIQ)

A05.04 Outdoor sound propagation: S528

Conveners: Christian Adams (Graz University of Technology), Timothy Van Renterghem (Ghent University), Martin Czuka (AIT Austrian Institute of Technology)

711 Optimizing Civil Defense Siren Placement for Audibility Coverage

Ensuring that warning sirens are audible to the population during natural disasters, technological accidents, or security crises remains a key challenge in emergency planning, while their spatial deployment is still not systematically optimized on their audibility. This study proposes a two-step approach. In France, public warning sirens are tested every first Wednesday of the month at noon. The first step relies on questionnaires distributed to residents after these monthly tests in order to derive an empirical audibility curve based on reported perception of the signal. The second step uses open-source tools for environmental noise modelling and multi-objective optimization: NoiseModelling, based on the CNOSSOS-EU propagation model, and OpenMole, implementing the NSGA-II evolutionary algorithm. Their coupling enables exploration of alternative siren configurations and identification of Pareto-optimal solutions according to two objectives: (1) the number of buildings exposed above 80 dB, and (2) the total area exposed above this threshold. A case study on Saint Barthelemy Island shows that the optimized Pareto front ranges from 7836 to 7858 dwellings and from 15.29 to 15.31 km². These results provide a set of acoustically optimal configurations and illustrate how such modelling tools can support decision-making, while complementing expert-based approaches that also consider non-acoustic constraints.

Speaker: Pierre Aumond (UMRAE)

712 Uncertainty analysis of aircraft noise using a heuristic ray-tracing model

Accurate prediction of aircraft noise requires numerical models capable of considering sound propagation from a source in motion. In Kayser et al. [Acta Acustica, 8, 62], we proposed a ray-tracing model based on a heuristic formulation that explicitly accounts for moving source phenomena, including Doppler shift and convective amplification, together with sound propagation in inhomogeneous atmospheric conditions, while remaining computationally efficient for engineering applications.This contribution presents an uncertainty analysis of the ray-based model predictions using a global sensitivity approach. A large number of simulations is performed, relying on Sobol-based sampling to quantify the impact of uncertain input parameters on sound pressure levels. The analysis is conducted under certification conditions relevant to aircraft noise assessment.Two receiver configurations are investigated for the same test case: a microphone positioned at 1.2 m above the ground, as currently required for certification measurements, and a ground-mounted microphone that suppresses ground interference effects. The results show that the latter configuration leads to reduced uncertainty, highlighting the strong influence of ground-induced interference patterns on sound level variability. These findings illustrate how ray-tracing model can support more robust aircraft noise prediction and inform future certification measurement strategies.

Speaker: Bill Kayser (Cerema)

713 Application and Comparison of NF S 31‑120 Meteorological Data Analysis Methods for Outdoor Sound Propagation: A Case Study Based on In‑Situ Measurements

The French standard NF S 31-120 defines methods for characterizing meteorological conditions and ground properties relevant to outdoor sound propagation. For meteorological characterization, the standard proposes three approaches:(i) a classification method based on a meteorological analysis grid (UiTi),(ii) a method based on measurements of the acoustic velocity gradient, and(iii) a statistical method based on wind speed and wind direction data.The objective of this presentation is to compare the application of these three methods using a common dataset of acoustic and meteorological measurements, in order to identify acoustic indicators representative of unfavorable, homogeneous, and favorable sound propagation conditions. The ability of each method to synthesize information contained in the temporal evolution of the measurements is evaluated. Based on this comparison, recommendations are proposed concerning the appropriate use and limitations of each method.

Speaker: Fabrice JUNKER (EDF-DTG)

714 Analysis of the Sound Propagation Calculation Method according to ISO 9613-2

The author served as chairman of ISO TC43 SC1 WG61, the working group responsible for the revision of ISO 9613-2, and presents key findings from the analysis of the sound propagation model defined in this standard.Although the formalism is simple and clearly structured, particular attention is required for the complex interaction between ground attenuation and barrier attenuation. The impact of proposed revisions on calculated sound levels at distant receiver locations is evaluated using software that implements both the current version of the standard and the planned modifications.Comparative calculations were performed for large-scale industrial sites covering several square kilometres, including models with up to 6,000 individual sound sources. The results demonstrate the expected net effects of the implemented or proposed changes, both for the overall area and for individual sound propagation paths.

Speaker: Wolfgang Probst (DataKustik GmbH)

Full paper (PDF)

A05.06 Advanced air mobility noise: S035

Conveners: Martin Blass (Joanneum Research Forschungsgesellschaft mbH), Franz Graf (Joanneum Research Forschungsgesellschaft mbH), Antonio Torija Martinez (Acoustics Research Centre)

715 Experimental investigation of the noise radiation of an electric motor for urban air mobility vehicles

The electrification of the aviation sector requires new propulsion technologies for future aircraft concepts. A key component of these systems is the electric motor, which converts electrical energy into mechanical power for the propulsor (e.g., open rotors or ducted fans). Within the DLR project VIRLWINT, the psychoacoustic effects of two urban air mobility (UAM)vehicles equipped with electric propulsion systems are investigated.To improve understanding of the electric propulsion system’s contribution to overall aircraft noise, acoustic measurements were conducted on a representative electric motor. These measurements were performed under poor acoustic conditions, with significant background noise and sound-reflecting boundaries. The analysis focuses on the noise characteristics of the electric machine under varying operating conditions, such as rotational speed and load. A local regression algorithm was used to obtain the broadband noise from the dominant tonal features of the spectra. The results indicate that rotational speed influences the amplitude of both tonal and broadband noise components, whereas changes in torque affect only the tonal components, leaving the broadband noise largely unchanged.

Speaker: Philipp Schulze (DLR Institute of Electr. Aero-Engines)

Full paper (PDF)

716 Drones' noise emission under outdoor hovering: on-board measurements

The increasing use of small unmanned aerial vehicles (UAVs) requires measurement methodologies capable of describing their acoustic emissions under realistic operating conditions. Traditional drone noise characterization mainly relies on anechoic chamber or ground-based measurements. However, while the first offers an accurate measurement of the drones’ noise emission and their directivity in stable air, the second allows performing measurements during real flying conditions but under the influence of the existing background noise in the recording situ.This study presents a comparison of noise emissions of drones during stationary flight by means of an on-board single-microphone measurement approach. An experimental campaign was conducted in the campus of the Universidad Católica San Antonio de Murcia on three different drones: Mavic 3 Classic, Matrice 4T and DJI Mavic 3M, with the latter drone being under two different setups with two propeller types. Measurements were performed with the microphone positioned onboard the hovering drone at three different heights above ground under monitored atmospheric conditions to investigate potential influences of ground reflections and air stability. Results compare the overall noise emission of the drones while highlighting the variability of the acoustics and psychoacoustics metrics.

Speaker: Nicola Gravina (Università degli Studi della Campania “Luigi Vanvitelli”)

717 UAS-NoiseCheck – Results and Recommendations for EASA Drone Noise Measurements

The increasing deployment of unmanned aerial systems (UAS) raises important questions regarding environmental noise impact and public acceptance. To support regulatory development, the European Union Aviation Safety Agency (EASA) has proposed guidelines for standardized drone noise measurements; however, practical experience with their implementation remains limited.Within the UAS-NoiseCheck project, comprehensive measurement campaigns were conducted on eight multicopter drones. The sensor setup combines acoustic, visual, GNSS, and meteorological components to enable multisensory data acquisition. Reference microphones in both ground-based and inverted configurations, together with a hemispherical 32-microphone array, enable the determination of standardized noise metrics as well as spatially resolved acoustic analysis. By assessing array localization performance and steered sound pressure levels practical aspects of drone noise certification and regulation are investigated. Improved RTK-based GNSS and electro-optical camera systems provide a precise reference for positioning and target detection.The results confirm the general applicability of the EASA methodology while revealing practical challenges related to environmental sensitivity, flight path repeatability, positioning accuracy, and complex directivity patterns. Based on these findings, recommendations are derived to improve measurement procedures, array integration, and data evaluation. The presented work supports the refinement of standardized drone noise assessment and provides experimentally grounded guidance for future regulatory frameworks.

Speaker: Martin Blass (Joanneum Research Forschungsgesellschaft mbH)

Full paper (PDF)

718 Noise Reduction in Drone Propellers: A Comparative Study of Design Optimization and Active Phase Control

The rapid expansion of civilian drone applications has intensified concerns regarding environmental noise and its impact on human well-being. Due to their pronounced tonal and high-frequency components, drone noise is often perceived as particularly intrusive in urban environments, making noise reduction a key factor for public acceptance.This study investigates the acoustic characteristics of drone propellers and evaluates mitigation strategies combining passive design optimization and active noise control. Several propeller configurations, including toroidal designs and alternative blade geometries, are systematically compared to conventional two-blade propellers.A dedicated experimental test bench was developed to enable time-synchronous acquisition of acoustic and aerodynamic performance parameters, including sound pressure, electrical power consumption and thrust. This approach allows for a comprehensive, multi-criteria assessment of propeller performance.The results are analyzed with respect to trade-offs between noise emission, aerodynamic efficiency, and thrust generation. In addition to passive design measures, active control strategies based on destructive interference are examined. In particular, phase shifts between multiple propellers and adaptive techniques targeting dominant tonal components are investigated.The findings demonstrate significant potential for reducing tonal noise while maintaining aerodynamic performance. The study provides insight into the optimization of drone propulsion systems and contributes to the development of quieter unmanned aerial vehicles for improved environmental compatibility.

Speaker: Lukas Lübbert (Fachhochschule Dortmund)

Full paper (PDF)

719 Integrated Modelling Framework for Noise Exposure and Annoyance in Advanced Air Mobility Operations

Advanced Air Mobility (AAM) platforms exhibit complex aeroacoustic noise that is highly modified by the rotor-blade configuration and operative strategies of the vehicle. The emitted sound has been reported to have complex multi-tonal and broadband noise components with strong directivity. Moreover, those acoustic features are modulated by easy transitional flight manoeuvres and airflow conditions during flight. As a result, the existing noise modelling and assessment tools are not well-suited to these unconventional noise emissions. This scenario limits the accurate assessment and mitigation of AAM noise impacts. This paper outlines ongoing research at the Salford Acoustics Innovation Institute toward the development of an integrated modelling and assessment framework for AAM noise. The framework comprises three core components: (1) Source definition, which integrates experimental measurements and multi-fidelity aeroacoustic predictions into a parametric acoustic source model for representative AAM platforms; (2) Sound propagation, accounting for diffraction, reflection, and atmospheric effects within complex urban geometries using hybrid numerical–analytical methods; and (3) Impact assessment, combining physical noise metrics with perceptual models to estimate human response to AAM noise.Key innovations include the development of parametric noise source definition and the development of simplified sound quality metrics that can be directly coupled with aeroacoustic predictions. These metrics provide a perceptually relevant description of AAM noise, enabling integration of aeroacoustic engineering prediction, the psychoacoustic-based annoyance models and community response. Representative urban AAM operational case studies demonstrate the applicability of the framework.

Speaker: Antonio Torija Martinez (Acoustics Research Centre)

A09.00 Machine learning and artificial intelligence in acoustics: S053

Conveners: Mirco Pezzoli, Alois Sontacchi (University of Music and Performing Arts), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Franz Pernkopf (Signal Processing and Speech Communication Laboratory)

720 A Deep Neural Network for Predicting the Sound Field Radiated by Orthogonally Stiffened Plates

In recent years, the adoption of surrogate modeling in vibroacoustics applications is being explored. Artificial Neural Networks (ANNs) are proving their effectiveness in predicting structural vibration, and more recently their capabilities in sound radiation prediction is being investigating as well, paving the way for new research in the field. This paper explores the use of a Deep Neural Network (DNN) approach for the prediction of the sound field of plates withone orthogonal stiffener. In particular, the surrogate model can estimate the Sound Pressure Level (SPL) over a hemispherical surface enclosing a baffled plateof fixed dimensions, where the position of the stiffener and the forcing point is changed. The dataset for the network training is generated by numericalsimulations combining FEM, Rayleigh integral and modal superposition. The results prove this approach to be promising and effective for tackling complex vibroacoustic problems.

Speaker: Manuele Montrasio (Politecnico di Milano)

Full paper (PDF)

721 HERMES: A Physics-informed Hierarchical Mesh Transformer for Real-time Room Eigenfrequency Prediction

The accurate prediction of low-frequency room acoustics relies on modal frequencies, traditionally computed using expensive numerical methods such as the Finite Element Method (FEM). This paper presents HERMES (Hierarchical Estimator for Room Modal Eigenvalue Synthesis), a physics-informed hierarchical mesh transformer that predicts the first 200 modal frequencies directly from 3D surface meshes of rooms with diverse geometries. The architecture employs a two-phase transformer backbone that leverages object-level structure for efficient processing of complex meshes, combined with an LSTM-based prediction head grounded in Weyl's law to incorporate physical priors. Trained and evaluated on a synthetic dataset of 7,500 rooms with various shapes and sizes, HERMES achieves a mean absolute error of 0.59 Hz across all room types, consistently outperforming an analytical scaled box baseline, while reducing computational cost by approximately three orders of magnitude compared to conventional eigenvalue solvers. Additionally, we present a sensitivity analysis which quantifies how modal frequency prediction uncertainties propagate into derived acoustic pressure fields. By demonstrating that deviations remain within a stable, linear regime for the error range achieved by HERMES, we confirm our neural approach provides both the efficiency and precision required for reliable acoustic engineering.

Speaker: Bence Bakos (Eötvös Loránd University)

Full paper (PDF)

722 Two-stage sound field learning through holomorphic neural networks and Vekua operators

Holomorphic neural networks have recently been proposed as two-dimensional universal function approximators constrained to produce holomorphic outputs. In this work, we combine holomorphic neural networks with Vekua operators toconstruct a machine learning architecture that inherently generates solutions tothe Helmholtz equation. This approach is particularly advantageous for two-dimensional boundary value problems in room acoustics at low frequencies, astraining can be restricted to enforcing only boundary conditions, resulting in substantial speed-ups and improved accuracy compared to traditional physics-informed neural networks. Possible applications include real-time sound field prediction, inverse design and surrogate modelling.

Speaker: Matteo Calafà (DTU Electro)

Full paper (PDF)

723 A neural eigenvalue solver using plane wave representations

Estimating acoustic eigenfrequencies is essential in the design of rooms and buildings, particularly at low frequencies and/or small rooms. While analytical solutions exist for simple geometries, more general shapes and materials require numerical algorithms. In recent years, machine learning approaches for eigenvalue problems have been proposed, but their accuracy and robustness remain limited and unexplored, leaving finite element methods as the preferred standard. Although, a neural network architecture called HergNet has been recently introduced which leverages the plane wave decomposition to achieve superior performance over previous physics-based machine learning approaches. In this work, we present a framework that employs HergNet as an eigenvalue solver, obtaining substantially reduced training times while maintaining high accuracy and robustness. The learning mechanism further allows the search to be restricted to a specified eigenfrequency range. The method is tested against FEM across different geometries and boundary conditions, demonstrating its applicability to a wide range of scenarios.

Speaker: Matteo Calafà (DTU Electro)

Full paper (PDF)

A14.01 New approaches for improving and assessing outcomes with hearing aids: S099

Conveners: Hendrik Husstedt (Deutsches Hörgeräte Institut GmbH), Inga Holube, Florian Denk (German Institute of Hearing Aids)

724 The Effect of Hearing Loss on the Characteristics of Daily Life Conversations

Background. Advances in acoustic feature extraction and machine learning enable conversation detection in real-world settings. This provides unique insights into the acoustic characteristics associated with communication. This study assessed the influence of hearing acuity on conversation frequency, own speech levels, and speaking time.Methods. Acoustic features and Ecological Momentary Assessment (EMA) data were used to detect daily conversations, without accessing raw audio, in 90 participants (aged 40-73 years, mean 59 years) with varying levels of hearing acuity (Pure Tone Average thresholds [PTA] of the better ear, which ranged between -5 and 94 dB HL, mean 28 dB HL). Using the binaural olMEGA microphone system and a random forest classifier trained on EMA-prompted self-reported conversation periods, the likelihood of conversation was predicted throughout the day. The participants’ voice was separated from ambient sound and own speech level was assessed. The relationships between PTA and hearing loss (yes/no), and conversation characteristics (frequency, speaking time, speech level) were then assessed. Results. Using leave-one-subject-out cross-validation, the conversation classifier achieved a mean weighted F1 score of 0.73. Conversation likelihood was associated with noisier environments and hearing acuity. During conversations, speech levels increased with ambient noise. Speaking time was also associated with ambient sound levels, and this effect depended on PTA. Conclusions. This study shows the potential of combining acoustic and EMA data to assess daily communication. The inability to separate background sound from conversation partners’ speech and the short monitoring periods remain a key limitation. Future research should address these and explore individual communication strategies to better understand how hearing acuity affects social interactions.

Speaker: Carlota Sabaté Cao (Amsterdam University Medical Center)

725 Conversing while Doing: From Everyday Experiences of Dual-task Conversation to Controlled Simulation

Conversations often take place simultaneously with other activities. These dual-task situations place high demands on cognitive and sensory resources. Most previous research has overlooked how naturalistic conversations are affected by dual-tasking and how individuals with hearing loss (HL) adapt their communication behaviours under dual-tasking demands. To address this gap, we conducted two studies.In Study 1 we explored what activities people undertake while conversing. Fourteen adults with HL (aged 40–75) participated in semi-structured interviews. Participants described real-life experiences of engaging in conversations while simultaneously carrying out other tasks. Participants reported a range of activities involving visual-motor and cognitive demands, such as driving, cooking, and walking.In Study 2, we will examine how engaging in dual-task conversations influence both conversational behaviour and performance on the other task. Pairs of frequent conversational partners (one with HL) will engage in conversations, while one partner simultaneously performs one of two screen-based tasks designed to increase demands on attentional and visuomotor capacities analogous to the demands introduced by the most frequently reported activities from Study 1. Results from Study 2 will include the impact of dual-tasking on conversational behaviours (e.g., speaking time, speech balance), on subjective experiences (e.g., perceived conversation success), and on error rates for the screen-based tasks.Study 1 contributes to the development of a structured catalogue of everyday dual-task conversation scenarios that could serve as inspiration for ecologically valid assessments of hearing technologies. Study 2 further provides a pilot test of the feasibility of simulating real-life dual-task scenarios in controlled settings.

Speaker: Karolina Smeds (ORCA Labs, WS Audiology)

726 Effects of Stimulus Type and Presentation Mode on Uncomfortable Loudness Perception

Clinical hearing assessment and hearing aid fitting rely mainly on the pure tone audiogram, which emphasizes audibility at low levels. However, many real-world sounds occur at moderate to high levels, where speech understanding can decline and listening discomfort may arise. Current clinical measures capture these effects only to a limited extent. The LoudSounds project aims to develop clinically feasible methods for assessing hearing at moderate to high levels and to evaluate their relation to existing procedures. Data were collected at the University of Southern Denmark and the German Institute of Hearing Aids from older adults with mild to moderate sensorineural hearing loss and young normal hearing listeners. This study focuses on loudness perception and discomfort. Pure tone audiometry included uncomfortable loudness levels as an index of high-level tolerance. Loudness growth was assessed using adaptive categorical loudness scaling with speech and speech shaped noise, yielding estimates of most comfortable levels and upper loudness categories. Speech-based measures of most comfortable and uncomfortable levels were obtained using procedures aligned with pure tone testing. Sensitivity to impulse sounds was evaluated using transient stimuli with increasing levels until discomfort or maximum output was reached. Results confirm that hearing impaired listeners show elevated thresholds while uncomfortable loudness levels remain similar to those of normal hearing listeners. Estimates of most comfortable and uncomfortable levels were consistent across methods. Hearing impaired listeners showed stronger binaural loudness summation, increasing at higher levels. Loudness recruitment was observed for sustained stimuli, while sensitivity to impulse sounds was reduced, suggesting a partly independent perceptual dimension of transient sounds.

Speaker: Hendrik Husstedt (Deutsches Hörgeräte Institut GmbH)

727 Use of an acoustic indicator of the open earcanal wall vibration distribution to mitigate the occlusion effect induced by intra-aural devices

The occlusion effect (OE) induced by intra-aural devices is a major source of acoustic discomfort, often limiting the proper use of earplugs and hearing aids. This phenomenon originates from the amplification of body‑conducted sounds, particularly at low frequencies, caused by the increased acoustic pressure radiated by the vibrating earcanal wall and intra-aural device when the ear is occluded. Understanding earcanal wall vibration is therefore essential to mitigate the OE at its source. Yet, direct measurements on the earcanal wall remain challenging due to its small size and complex anatomy. Analytical models represent the acoustic effect of earcanal wall vibration using a flow source positioned at the centroid of the wall’s normal velocity distribution. An indirect acoustic method employing resonators coupled to the ear has been previously proposed by the authors to estimate this centroid position experimentally. The present study investigates the relevance of this centroid as an acoustic indicator of a point of predominant vibration contributing to the OE. A vibroacoustic finite element model of a simplified outer ear is used to analyze the relationship between wall vibration distribution and centroid position. The acoustic estimation method is then applied to the virtual ear to assess its feasibility and optimize its precision prior to human testing. Finally, series of length‑controlled custom occlusion devices are evaluated numerically to examine the sensitivity of the OE to device-skin coupling with respect to this indicator. Results demonstrate the potential of this approach to inform design strategies aimed at reducing tissue‑conducted sound and mitigating the OE induced by intra-aural devices.

Speaker: Robin Richert (École de technologie supérieure)

A18.04 Indoor soundscape: S131

Conveners: Simone Torresin (University of Trento, Dept. of Civil, Env., Mech. Eng.), Papataya Nur Dokmeci Yorukoglu

728 Multisensory Augmentation of University Study Rooms: The Cross-Modal Role of Scents on Soundscape Perception

Indoor soundscape and smellscape research is transforming the conception of built environments, moving beyond sensory neutrality toward the intentional design of spaces that are pleasant and supportive of occupants’ activities, while explicitly recognizing the multisensory nature of environmental experience. This laboratory study investigated the effects of auditory and olfactory augmentation on soundscape perception in university study rooms. Thirty-six students participated in a full-factorial within-subject experiment, experiencing nine combinations of three binaural auditory conditions (baseline study-room background, background plus forest sounds, and background plus music) delivered via closed headphones, and three olfactory conditions (odourless, Talcum, and Swiss pine) administered through an eight-channel olfactometer. Participants evaluated soundscape and smellscape perception. Preliminary results indicate significant effects of sound type on soundscape pleasantness and eventfulness: music enhanced perceived pleasantness, whereas forest sounds increased eventfulness relative to baseline. Importantly, cross-modal effects of olfaction were also observed. Valence-related dimensions were aligned, with higher smellscape pleasantness associated with greater soundscape pleasantness, while activation-related dimensions were linked, as stronger smell presence corresponded to higher soundscape eventfulness. These findings highlight the potential of combined audio–olfactory strategies to enhance indoor study environments, warranting further validation in real-world settings.

Speaker: Simone Torresin (University of Trento, Dept. of Civil, Env., Mech. Eng.)

Full paper (PDF)

729 Effect of sounds on cognition: a comparison between classroom and laboratory

Research on the impact of environmental quality on cognitive performance is frequently conducted in laboratory settings. However, findings can differ significantly from those obtained in real-world environments. This study compares outcomes from two experiments, involving students aged 11 to 14 years. The former experiment was conducted inside a students’ classroom, while the latter in a laboratory environment. Both the experiments investigated the effect of sound stimuli on calculation abilities. Two main effects emerged. First children performed better in the laboratory probably thanks to a reduction in distractions compared to the regular classroom setting. This was further facilitated by working in smaller groups and consistently performing the test after the laboratory phase. Secondly, the effects of the sound stimuli depended on the setting likely due to an interaction with undetected personal characteristics. Overall, the findings shows discrepancies between the two settings highlighting limited ecological validity of the laboratory and illustrating the importance of minimising lab-field differences when investigating this particular age group.

Speaker: Matteo Pellegatti (University of Ferrara - Department of Engineering)

Full paper (PDF)

730 A Comparative Study on the Role of Soundscape Perception in Place Attachment of Graduate Student Housing

Place attachment reflects the emotional and functional bonds individuals develop with their living environments, playing a critical role in well-being and residential satisfaction. While previous research has extensively examined place attachment through the dimensions of place dependence and place identity, the role of soundscape in shaping these bonds remains underexplored, particularly within residential academic settings. Graduate student housing represents a distinctive living environment where academic routines, social interaction, and environmental conditions intersect, making it a valuable context for soundscape-related inquiry. This study examines how soundscape perception varies across different spatial configurations in graduate dormitories and how these perceptions relate to residents’ place attachment. The research is conducted at Bilkent University, where graduate housing units are located in three contrasting spatial conditions: a dormitory adjacent to a sports field, an elevated and acoustically exposed building, and a newly constructed courtyard-type complex with reflective spatial characteristics. These configurations generate diverse acoustic environments shaped by varying external and internal sound sources. Based on the ISO/TS 12913-2 soundscape assessment framework, the study employed a questionnaire incorporating Method A and a Place Attachment scale, together with in-situ sound level measurements (LAeq). A total of 57 graduate students participated in the study. The findings revealed statistically significant associations between soundscape perception and place attachment, demonstrating that positive residential soundscapes emerge from the interaction between acoustic conditions and spatial context, ultimately strengthening students' emotional attachment to campus housing.

Speaker: Ceren Sahmaran (Independent Researcher)

Full paper (PDF)

731 Developing Descriptors for Indoor Soundscape and Lightscape in Public Self-learning Spaces

Public self-learning spaces are high-cognitive-demand indoor environments in which experience is shaped by multiple perceptual factors rather than by sound or light alone. However, questionnaire-oriented perceptual descriptors for indoor soundscape and lightscape in such spaces remain limited. This paper reports a preliminary part of Phase 1 of a larger doctoral study and addresses one central question: which perceptual descriptors can be preliminarily identified and organised into a questionnaire-oriented wording list for indoor soundscape and lightscape in public self-learning spaces, based on the analysed Chinese focus groups. Across the three Chinese focus groups analysed so far, participants showed a relatively stable shared preference for environments characterised by low disturbance, appropriate visual conditions, comfort, safety, and concentration support. The findings provide a preliminary descriptor basis for later pilot testing and questionnaire attribute selection in public self-learning spaces.

Speaker: Wenxi Zhang (Institute for Environmental Design and Engineering)

Full paper (PDF)

732 SoundSafe: Towards an Integrated Assessment of Soundscape Quality in a Neonatal Intensive Care Unit}

Neonatal ICUs exceed WHO noise limits >90% of time, disrupting infant development, family bonding, and staff performance. SoundSafe develops the first integrated soundscape assessment platform for NICUs, combining privacy-preserving Machine Listening, perceptual modeling, and human-centered design. Given the sensitivity of the environment, we embed privacy by design by automatically filtering speech in audio recordings. We deploy a multichannel audio array in Hospital Universitario de Donostia NICU, filtering speech via Universal Speech Codec while classifying sound types using YAMnet and extracting psychoacoustic features. Families and medical staff rate affective qualities per ISO 12913-2. We develop a SoundSafe Quality Index (SSQI) that integrates acoustic, event, and perceptual layers to predict soundscape quality. Co-designed with healthcare professionals via participatory workshops, the SSQI dashboard provides real-time, actionable insights on hospital tablets. Preliminary measurements show 51-56 dBA daytime levels vs. WHO 35 dBA target. Hospitals need a sustainable solution to assess soundscape quality. SoundSafe advances data-centric soundscape assessment for healthcare, supporting Spain's Digital Health Strategy 2021-26.

Speaker: Sara Lenzi (University of Deusto, Ikerbasque Basque Foundation for Science)

Full paper (PDF)

A18.06 Hospital Soundscape, Hospital Acoustics, Soundscape design: S133

Conveners: Elif Özcan (Erasmus MC / TU Delft), Ela Fasllija (Erasmus MC / TU Delft), Thomas Hampton

733 Absence of Evidence or Evidence of Absence? ICU Acoustics ReSearch for a Needle in a Haystack

The role of hospital acoustics and soundscapes has gained increasing attention, as creating restorative environments for patients and less stressful conditions for healthcare providers are closely linked to the quality of care. Despite this growing interest, the field remains underdeveloped; strong design guidelines and recommendations are still scarce. Intensive Care Units (ICUs), characterized by complex and dynamic sound sources, are among the most extensively studied healthcare environments. However, clinical studies investigating the relationship between the ICU acoustic environment and patient outcomes have yielded mixed, inconclusive findings. Conducting rigorous randomized controlled trials in these settings is inherently challenging due to patient vulnerability, ethical constraints, and the presence of numerous confounding variables. This raises some critical questions: To what extent do current acoustic indicators influence patient clinical outcomes? Are we measuring and analyzing the right variables and outcomes? This paper reviews the clinical literature, examines both acoustic metrics and patient outcomes, identifies key methodological limitations, and outlines future directions, including soundscape-based interventions, multisensory approaches, and more context-sensitive acoustic characterization.

Speaker: Elif Özcan (Erasmus MC / TU Delft)

Full paper (PDF)

734 A human-centred multilayer framework for continuous acoustic characterisation of healthcare acoustic environments

Healthcare acoustic environments (HAEs) are commonly characterised using sound level descriptors (e.g. LAeq, L10, Lmax, etc). These quantify acoustic magnitude and temporal variation but provide limited information about modelled loudness or acoustic context. Complementary approaches are often analysed separately, making their temporal relationships difficult to examine. This conceptual paper proposes a human-centred multilayer framework for continuous HAE characterisation. It organises signal-derived information into three layers: (1) a Physical Layer describing measured acoustic characteristics; (2) a Perceptual Layer representing the temporal evolution of loudness using the Moore-Glasberg-Schlittenlacher method; and (3) a Context Layer assigning each interval to the broad acoustic category it most closely resembles. Timestamps align the outputs while preserving their definitions, temporal resolutions and uncertainties. Local processing retains descriptors and assignments without storing raw audio or extracting linguistic content. Intended users, clinical activities and care objectives guide monitoring design and interpretation, while associations with human responses or clinical outcomes require independent evidence. The framework does not combine the layers into a single score or treat them as direct measures of experience or clinical effect. It provides a traceable basis for examining complementary acoustic evidence over time. Technical and ecological validation are required to establish its accuracy, generalisability and practical value

Speaker: Alejandro Solorio (The Galway Sound Lab, School of Engineering,University of Galway)

Full paper (PDF)

735 A rapid, occupant-facing mobile tool for perceptual soundscape assessment in hospitals

Hospital soundscapes shape patient recovery, staff performance, and the communicative value of functional sounds such as alarms and speech, yet they remain difficult indoor environments to characterise perceptually. ISO 12913 instruments deliver rich individual assessments but require time, explanation, or researcher presence; automated sound monitoring captures event-level acoustic data but not subjective experience. This contribution explores the space between. Extending a lineage of mobile soundscape tools such as Hush City, it presents a mobile web application through which patients, staff, and visitors self-report their momentary perception of the sound environment in roughly sixty seconds — not an automated measurement, but a first-person perceptual judgment requiring no training to complete. The app compresses the ISO/TS 12913-2 Method A questionnaire, rendering Axelsson et al.'s (2010) circumplex through two continuous sliders coupled to a dynamic descriptor cloud drawn from a validated 116-word pool, intentionally trading depth for speed. Each response captures one impression rather than an overall experience; repeated responses accumulate, in an International Soundscape Database–compatible schema, into a temporally and demographically distributed picture — complementing automated monitoring rather than replacing it. A working prototype is demonstrated, and three design questions — input modality, descriptor scope, and data-collection model — are opened to community feedback ahead of a planned hospital pilot.

Speaker: Cees van Wezel (Saint-Gobain Ecophon)

736 Spatial variability of noise exposure in an intensive care unit: Fourteen-day comparison of bedside and nurse station LAeq and LAFmax levels

Environmental noise in intensive care units remains a persistent challenge, with implications for patient recovery, staff performance, and communication. While ICU noise is often characterised using aggregated measurements, less attention has been given to spatial variability within functional care zones. This study presents a fourteen -day continuous acoustic measurement comparing sound exposure at two representative ICU locations; a patient bedside and a nurse station. Equivalent continuous sound pressure levels (LAeq) and maximum fast time-weighted sound pressure levels (LAFmax) were analysed to characterise background acoustic conditions and transient peak noise events. Spatial and temporal comparisons were performed to investigate differences in average exposure, variability, and the occurrence of high-level events across day and night periods. Preliminary findings indicate distinct acoustic profiles between the two monitoring locations, with differences observed not only in average sound levels but also in the frequency and magnitude of transient events. These results demonstrate the importance e of spatially resolved monitoring for capturing acoustic variability in critical care units and suggest that localised noise dynamics should be considered in acoustic assessment and targeted noise mitigation strategies.

Speaker: Donya Dalir (Yeditepe University)

A21.03 Tyre/road noise: S153

Convener: Manfred Haider (AIT Austrian Institute of Technology GmbH)

737 CPX simulator: development of a hybrid simulation model of the CPX method

The close-proximity (CPX) method according to ISO 11819-2 is increasingly applied to monitor and character-ize road surfaces. To better understand the physical mechanisms underlying CPX measurements and to enable predictions of CPX spectra, a hybrid simulation model called CPX simulator is developed. This contribution presents experimental and numerical investigations conducted during model development. First, an experi-mental approach to quantify the influence of the tire tread pattern on CPX spectra is introduced. The results re-veal that the tread pattern of the used standard reference test tire (SRTT) plays a decisive role, particularly for smooth pavements at low driving speeds. The horn effect of a loaded tire is simulated using 3D FDTD. The sim-ulations highlight the impact of the circumferential tire grooves and the acoustic impedance of the pavement on CPX spectra. The influence of the CPX trailer enclosure was examined through narrowband measurements using an artificial sound source. The observed distinct trailer influence at low frequencies can be well repro-duced with a modal sound field description. Finally, predictions with the CPX simulator are compared to CPX measurements performed in Switzerland on various pavement types at different driving speeds. The compari-sons confirm that the hybrid model accurately captures the combined influences of speed, pavement surface texture, and pavement impedance on spectral CPX levels.

Speaker: Reto Pieren (Empa)

Full paper (PDF)

738 A new tyre for the CPX method

The CPX method is a standard method for assessing the acoustic properties of road surfaces. The reference tyres to be used with this method are specified in ISO/TS 11819-3. The H1 tyre (AVON AV4) mentioned therein, which is used to characterize the tyre/road noise of trucks, has been discontinued since the end of 2014. This makes it necessary to find a suitable replacement for AVON AV4.With project funding from the German Federal Ministry of Transport (BMV) and with the support of the German Federal Highway and Transport Research Institute (BASt), a consortium of all providers of CPX measurements in Germany has worked on project FE 04.0340/2021/AGB "Methodology for the selection of a new H tyre for the CPX process" and proposed a new H tyre (hereinafter: H2 candidate). For this H2 candidate, the correction coefficients for air temperature and shore hardness were determined, as well as a formula for converting results with the reference tyre H1 and the H2 candidate. In addition, the procedure used in this project has been described in a manual as a guideline for the future choice of new reference tyres.

Speaker: Beate Altreuther (Müller-BBM Industry Solutions GmbH)

Full paper (PDF)

739 Advancing rubber hardness measurement procedures for CPX reference tyres: Insights from ISO/TS 11819-3 revision

Within the framework of working group ISO/TC43/SC1/WG33, which is dealing with measuring methods for comparing traffic noise on different road surfaces, the revision of ISO/TS 11819-3 is ongoing since 2025. This technical specification concerns reference tyres used in CPX (Close-Proximity) noise measurements. One of the remarks received during the systematic review of this technical specification was to improve the procedure to measure rubber hardness of the reference tyres, which is used as correction factor for the measured sound pressure levels No distinction is currently made between digital and analogue durometers, although some users report different results depending on the device used. A measurement plan for a round robin test was set up where various institutes were requested to repeat measurements on their own P1 tyre by varying only one of the following parameters: durometer type, time setting after contact (for digital durometers), operator (including their level of expertise), measurement day. Participants were asked to measure more points around the tyre and to report all their results, including values that would be discarded according to the current procedure. A substantial dataset from seven institutes was obtained and analysed. Based on this analysis, a proposal for improving the measurement procedure has been put forward. It was determined that the operator’s level of expertise and the device settings are significant parameters. The results will be used as input for the revision of the Technical Specification with the aim of reducing variability and increasing reproducibility of the CPX reference tyre hardness measurements.

Speaker: Anneleen Bergiers (Belgian Road Research Centre)

Full paper (PDF)

740 CPX Method – Crucial Microphone for Low-noise Surfaces Assessment

The CPX method uses two mandatory microphones and up to four additional (optional) microphones to assess road surface noise. The microphones are positioned around a reference tyre and record noise generated at the tyre–road contact. However, the overall noise levels may vary depending on the number and placement of the microphones. This raises the question of whether the currently mandatory microphones are sufficient for the proper assessment of low-noise surfaces.This paper presents noise levels measured by all six microphones on conventional and low-noise surfaces of different ages. Microphones located along the tyre (Nos. 1, 2, and 3) recorded nearly identical values for both surface types. In contrast, differences between microphones 4 and 5 were already evident for low-noise surfaces. The measurement from microphone 5, located at the trailing edge, is strongly influenced by void content. The largest difference, up to 2.5 dBA, was observed on the newest surface.

Speaker: Blanka Hablovicova (Transport Research Centre)

Full paper (PDF)

A23.02 Modelling & Simulation Techniques: S163

Conveners: Giuseppe Petrone (University of Naples Federico II), Marcus Maeder (Technical University of Munich)

741 FreeFem++ for Pipette Aspiration

Pipette aspiration is a widely used biomechanical technique for measuring the mechanical properties of cells and soft tissues. In this study, we present a computational model of pipette aspiration based on the finite elementmethod, as implemented in FreeFEM++. This model simulates tissue deformation under suction pressure, enabling us to analyze the system’s frequency response. Our results demonstrate FreeFEM++’s effectiveness in modelling complex biomechanical systems, providing insights into the viscoelastic properties of tissues under dynamic loading conditions.

Speaker: Raphael Lamprecht (UMIT TIROL)

Full paper (PDF)

742 Unified High-Order 3D Finite-Element Framework for Vibroacoustic Frequency-Response Analysis

A unified three-dimensional high-order finite-element framework is proposed for the frequency-response analysis of finite layered acoustic systems under plane-wave and diffuse acoustic field (DAF) excitation. The formulation relies on factorized high-order expansions of the Carrera Unified Formulation (CUF) type. At the same time, the present benchmark focuses on acousticized structures, namely solid layers recast as equivalent acoustic media with density inherited from the parent solid and longitudinal wave speed inferred from the elastic modulus. The same hierarchical discretization machinery is therefore used for fluid layers, equivalent acoustic solids, and perfectly matched layers. A transfer-matrix model is implemented both for classical plane-wave loading and for a discrete DAF extension based on the same random plane-wave realization adopted at the finite-element inlet. Three hierarchical models, namely LE4/B2, LE9/B3, and LE16/B4, are assessed over 2000 to 16000 Hz for an air-aluminum-air configuration; the DAF extension highlights the role of full three-dimensional modeling and outlet absorption. The study extends recent vibroacoustic literature based on CUF from eigenanalysis and wave propagation to the prediction of forced transmission loss in finite acousticized structures.

Speaker: Dario Magliacano (Politecnico di Torino)

Full paper (PDF)

743 Non-contact low-frequency acoustic methods for detection of wall-thinning/delamination in pipes

Ageing water infrastructure that includes water and sewer pipes often suffers from defects which are difficult to detect with water industry state of the art techniques. Wall thinning or delamination (external/internal) in pipe walls provides a challenging boundary value problem in 3D cylindrical system. Non-contact low frequency acoustic methods offer robust and easy to deploy technology capable of detecting in gases or liquids hidden changes in the pipe conditions. In the present work, a Green’s function approach under low-frequency assumption has been adapted to deal with a waveguide (pipe) problem with localized elastic interface backed by fluid cavity, which resembles wall thinning or delamination. The developed methodology is integrated into non-contact pipe condition monitoring tools used in the following related research projects: (i) AI:LINER, an EU multi-institutional project that combines novel acoustic solution, CCTV data, AI-based failure detection, and in-situ monitoring techniques to enhance the asset management life cycle of sewer networks; and (ii) an EPSRC-funded project, to develop bio-inspired micromachine sensors for measuring acoustic quantities.

Speaker: Mriganka Shekhar Chaki (University of Sheffield)

744 Hybrid Modeling of Air Suspension for Chassis Module Force Transmissibility Prediction in NVH Analysis

Air suspension modeling is critical for estimating the NVH characteristics of chassis modules and full vehicles. However, accurate analytical modeling remains challenging because air suspensions exhibit inherently nonlinear behavior, and their force transmissibility varies with excitation amplitude. In previous research, a modeling technique was developed to predict the component-level behavior of an air suspension. Although the model showed reasonable agreement in component-level analysis, its predictive capability was limited when applied to chassis-level transmissibility prediction under varying loading conditions.To address this limitation, this study proposes a hybrid modeling approach that incorporates component test results into an analytical model for chassis module analysis. By integrating experimentally identified characteristics into the model, the proposed approach more effectively captures the load-dependent force transmissibility of the air suspension than conventional analytical modeling alone. The developed model was applied to the prediction of chassis module transmissibility, and its validity was confirmed through correlation with experimental results. The results indicate that the proposed approach improves the agreement between predicted and measured NVH characteristics of chassis modules equipped with air suspension.

Speaker: JOONGHWAN JANG (Hyundai Mobis)

Full paper (PDF)

A03.05 Structure-borne sound and noise from building equipment: P462

Conveners: Berndt Zeitler (Hochschule für Technik Stuttgart), Jochen Scheck (HFT Stuttgart), Andreas Mayr (Rosenheim Technical University of Applied Sciences)

745 Web-application PreNoise® for predicting noise from building service equipment

PreNoise® allows users to predict noise levels of service equipment in buildings, based on the standards of EN 12354-5. It provides an early-stage approximation of the planned object and allows users to quickly select suitable systems with the required noise protection level in the planning phase and helps to assess the impact of changes to the building's structure. PreNoise® is developed by the Fraunhofer-Institute for Building Physics IBP in Stuttgart. The method has been validated against laboratory measurements of drainage systems conducted according to EN 14366-1. For the tested configurations in massive constructions, typical deviations between predicted and measured noise levels are within ± 1 - 3 dB . The current implementation covers massive constructions, and extensions to lightweight and timber buildings are planned. The main practical constraint of this approach is the availability of source data, which is usually derived from laboratory measurements. For drainage systems, IBP additionally provides the source database generated by numerical simulations, while for other technical sources the source characteristics must be measured (e.g. EN 15657).PreNoise® is available as a web application with a free version (limited functionality) and an extended R&D-version (offline).The current web application enables the prediction of wastewater systems for both the source room and the receiving room directly behind the installation wall. Within certain limits, users can adjust the room size as well as the thickness and material of the installation wall. The full version offers extended features, including diagonal transmission and multiple sources.

Speaker: Sven Öhler (Fraunhofer Institute, IBP)

A03.08/A05.07 Heat pump acoustics in residential environments: P463

Conveners: Jack Havie-Clark, Christoph Reichl (AIT Austrian Institute of Technology)

746 A Field Study of Air Source Heat Pump Noise Levels and Propagation in a UK Housing Development

Net Zero targets in the UK require a shift to low or zero carbon emission technology, which is supported by the UK government funding to support the installation of Air Source Heat Pumps (ASHPs) for domestic heating provision. ASHPs offer better thermal efficiency than traditional boiler systems and have no direct carbon emissions. Research and planning policy is needed, however, to ensure that noise generation stays within accepted levels. For ASHP installation in existing dwellings, the MCS 020a ‘Air Source Heat Pump Sound Calculation’ standard is used by installers to validate if an installation can proceed under permitted developments rules, or whether a full planning assessment is required. Having to undertake the latter is a barrier to adoption, so there is a need to ensure MCS 020a is not unduly stringent. But on the other hand, if it were not stringent enough then that could lead to noise complaints. There is, therefore, a need to validate that MCS 020A is fit for purpose, and this field study tests this objectively by comparing it to operational ASHP noise measurements. This poster will report the design of the study, the analyses undertaken, and signpost the key results and conclusions.

Speaker: Katie Salter (Acoustics Innovation Institute)

A03.08/A05.07 Heat pump acoustics in residential environments: S020

Conveners: Jack Havie-Clark, Christoph Reichl (AIT Austrian Institute of Technology)

747 Toward Standardized Directivity-Resolved Acoustic Data for Heat Pump Placement: Progress and Results from IEA HPT Annex 63

Accurate prediction of heat pump noise requires acoustic source data that go beyond A-weighted overall sound power levels. For outdoor units of air source heat pumps, frequency-dependent directivity is particularly relevant, since fan openings, heat exchangers, blind faces and top surfaces can radiate differently and therefore influence receiver levels depending on installation orientation.This contribution presents recent progress in IEA HPT Annex 63, “Placement Impact on Heat Pump Acoustics”, and builds on earlier work presented at Forum Acusticum. The open Annex 63 database has been expanded with additional measurements from several laboratories and combines frequency- and face-resolved sound power data with metadata on unit geometry, operating conditions and measurement methodology. Each entry contains one-third-octave-band partial sound power levels for five standardized unit faces, enabling comparisons between units and laboratories. Data from ISO 9614 sound intensity measurements, ISO 3744 sound pressure measurements using cubic enveloping surfaces, and ring- or dome-array measurements in free field over a reflecting plane are harmonized through defined conversion procedures.Particular emphasis is placed on acoustic directivity and its relevance for placement. A cross-laboratory analysis of five comparable units at A7W55, with fan and heat exchanger on opposite faces and three blind faces, shows systematic face-dependent radiation patterns. Blind faces provide broadband shielding of approximately 5 dB below 2 kHz, while heat-exchanger faces radiate about 10 dB more than fan faces above 2 kHz.With Annex 63 concluding in September, the presentation summarizes the database status, harmonization methods, conversion challenges, key acoustic findings and the outlook toward generic source models for improved heat pump noise prediction.

Speaker: Christoph Reichl (AIT Austrian Institute of Technology)

748 Towards Modelling Guidelines for Domestic Air Source Heat Pumps

The joint IOA / CIEH practitioner note refers to “appropriate sound propagation modelling carried out by an SQA”. The industry and international standard model for outdoor sound propagation modelling is ISO 9613-2, despite not being proven to be suited to modelling complex urban morphologies. Variability in how practitioners build and configure these models can lead to materially different predicted sound levels for a specific scenario. This paper outlines the modelling features that may be useful to include in modelling guidelines for ISO 9613-2 based predictions of sound from domestic air source heat pumps (ASHP), developed in the context of planning for domestic ASHPs.Guidelines are recommended for ground absorption, spectral data, reflection order, and building geometry to work towards establishing a consistent, robust framework. The advantages and constraints of point source or box sources are also discussed. Practical guidance is proposed for corner placements, partial barriers, and receiver positioning.

Speaker: Jack Harvie-Clark (Apex Acoustics Ltd)

Full paper (PDF)

749 Linking 2D Noise Mapping and 3D Augmented Reality for Heat Pump Acoustic Planning

The increasing deployment of air-to-water heat pumps and other outdoor technical systems in residential areas introduces new challenges for environmental noise, acoustic comfort, and public acceptance. Conventional planning approaches often rely on single-number sound power levels and simplified distance-based conversions to sound pressure levels. However, these methods typically neglect frequency-dependent emission characteristics, source directivity, operating conditions, and interactions with the built environment.This contribution presents the continued development of an integrated framework linking a web-based 2D sound propagation tool with a 3D augmented reality system developed within the RAARA project. The 2D tool enables users to position heat pump outdoor units in real geospatial environments and visualize calculated sound pressure levels on an interactive map. Its simplified ISO 9613-based model considers geometric spreading, façade reflections, ground effects, directivity, and barrier attenuation, allowing rapid preliminary assessment of residential installation scenarios.The 3D augmented reality tool extends this assessment into a spatial, in-situ experience. Noise sources, future façades, and mitigation measures can be virtually placed on site before installation, with acoustic effects visualized at points and lines of interest. Future developments include the visualization of sound pressure levels on surrounding surfaces and real-time auralization, considering frequency content, directional radiation, operating state, and automatically recognized environmental geometry.By combining analytical 2D noise mapping with immersive 3D visualization and future auralization, the proposed approach supports a more intuitive interpretation of sound propagation and improves communication between experts, planners, residents, and decision-makers.

Speaker: Christoph Reichl (AIT Austrian Institute of Technology)

A05.04 Outdoor sound propagation: P455

Conveners: Christian Adams (Graz University of Technology), Timothy Van Renterghem (Ghent University), Martin Czuka (AIT Austrian Institute of Technology)

750 Database of Scale-Model Validation Measurements for Urban Sound Propagation

The simulation of urban sound propagation is challenged by the strong impact of diffraction around building edges.Although several geometric propagation models can account for diffraction, their validation is hindered, because real-world measurements are inherently noisy or do not exist.For this reason, this work presents a database of scale-model measurements specifically designed for urban environments with an emphasis on diffraction.The database includes measurements of different urban scenarios, ranging from simple terraced houses to more complex configurations with multiple buildings, varying source-receiver placements and different scale factors.Due to the reduced scale, the measurements could be performed in an anechoic chamber, reducing the influence of noise and other environmental factors.The database is made publicly available to facilitate the validation of sound propagation models in urban environments and to encourage further research in this area.

Speaker: Pascal Palenda (IHTA, RWTH Aachen University)

A09.00 Machine learning and artificial intelligence in acoustics: P509

Conveners: Mirco Pezzoli, Alois Sontacchi (University of Music and Performing Arts), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Franz Pernkopf (Signal Processing and Speech Communication Laboratory)

751 murenn: Multi-Resolution Neural Networks in PyTorch

Multi-Resolution Neural Networks (MuReNN) are a new generation of models for deep learning for speech and audio processing. Compared to convolutional networks (convnets), they are more resource-efficient, less sensitive to initialization, and may generalize better across recording conditions. The key idea behind MuReNN is to learn a filterbank in which each filter is factorized between a non-learnable component (a complex-valued discrete wavelet) and a learnable component (a dilated convnet kernel). During this talk, we will present a differentiable and GPU-accelerated implementation of MuReNN in the PyTorch framework for Python. We will give a quick tutorial on how to build and train MuReNN layers and integrate them into full-fledged deep learning pipelines. Our open-source library is available at: https://github.com/kymatio/murenn

Speaker: Xiran Zhang (Nantes University)

A09.01/A17.01 Machine learning for array processing: P520

Conveners: Maximo Cobos (Universitat de València), Thushara Abhayapala

752 Online Estimation of Virtual Sensing Observation Filter Coefficients with Real-World Sounds for Local Active Noise Control

Adaptive local active noise control (ANC) requires accurate estimates of the sound pressure at the point of cancellation. These are often obtained using nearby microphones via virtual sensing methods such as the remote microphone technique (RMT). Obs-TasNet was recently proposed as a neural network–based method for online estimation of RMT observation filter coefficients. In this work, we evaluate Obs-TasNet's performance in synthesized reverberant environments with stochastic as well as nonstationary sounds sampled from the FSD50K dataset. Overall, Obs-TasNet exhibits only minor degradation of estimation error when applied to nonstationary sounds compared to stochastic noise sources. In reverberant environments, performance declines toward higher frequencies approaching the array's aliasing frequency. Nevertheless, Obs-TasNet consistently outperforms an inverse-distance-weighting baseline. These results indicate that Obs-TasNet is robust across source types and can handle acoustic conditions relevant to practical deployment.

Speaker: Felix Holzmüller (Institute of Electronic Music and Acoustics)

Full paper (PDF)

A10.02 Metamaterials for noise and vibration reduction: applications and experimental methods: P475

Conveners: Jacques Cuenca (Test Division, Siemens Industry Software NV), Elena Shabalina

753 Broadband Sound Absorption Using Weakly Coupled Parallel- Arranged Helmholtz Resonators in Acoustic Metasurfaces

Broadband, low to mid-frequency sound absorption with thin, compact structures presents a significant challenge in acoustical engineering. This work presents a novel acoustic metasurface based on weakly coupled parallel Helmholtz resonators (HRs). The design begins with an optimized dual-cavity HR unit, where a small inter-cavity perforation bridges two resonance peaks to create a continuous absorption band. By extending this principle to a network of several cavities interconnected via carefully designed perforations, multiple overlapping resonance modes can be created, effectively covering a continuous and wider frequency range. This scaled-up weak-coupling network allows for superior impedance matching across a broader spectrum, maintaining the near-optimal condition where dimensionless reactance approaches zero, and resistance approaches unity over an expanded interval. Based on this mechanism, a nine-HRs acoustic metamaterial unit cell (four perforated pairs and one single resonator), achieving an average absorption coefficient of 0.82 between 750 Hz to 3000 Hz, with individual peaks above 0.9. By coupling five such unit cells, a mixed multi-cavity parallel-arranged HRs based acoustic metasurface is realized, attaining an average absorption of 0.87 from 650 Hz to 3000 Hz with only 57 mm total thickness. The design was optimized via finite element analysis (FEM), fabricated using Low-Force Stereolithography (LFS), and validated experimentally with an impedance tube based on the two-microphone method. Excellent agreement between simulation and measurement confirms the design's accuracy and robustness, demonstrating a compact, high-performance solution for practical noise-control applications.

Speaker: FAISAL RAFIQUE (SUSTech, Shenzhen)

A10.03/A23.06 Metamaterials and acoustic black holes in vibro-acoustics and air: P477

Conveners: Wonju Jeon, Florian Toth (Institute of Mechanics and Mechatronics), Jae Yeun Lee

754 Integrating Acoustic Black Holes into satellite assemblies for vibration reduction

Vibrations and shock loads pose challenges for aerospace components, potentially causing brittle fractures or electronic failures. Conventional countermeasures like dampers or absorbers add mass and reduce the available payload. Integrating Acoustic Black Holes (ABHs) offers a solution without additional mass.Within a structure, an ABH is a local geometric taper with a near-zero-thickness tip. In this way, the propagation of bending waves is reduced. In theory, reducing the thickness of a beam to zero stops the wave propagation and prevents the reflection of the waves. In practice, finite reduction is compensated by applying damping material to high-strain areas, enabling efficient dissipation. This achieves significant structural damping with minimal material usage.In this work, ABH structures are integrated into an iso-static mounting structure for satellites, which serves as an isolating connection for optical components. Design studies determine the optimal positions, shapes, and dimensions of the ABHs, without compromising structural integrity. A promising concept with a circular ABH in a second layer plate, as well as lateral ABH strips around the circumference, is simulated to estimate the dynamic behavior and the potential of improvement. It is shown that even challenging low-frequency in-plane and bending vibrations are successfully reduced by the ABHs.

Speaker: Jannik Krohn (Fraunhofer LBF)

Full paper (PDF)

A10.13/A12.07 Numerical methods for wave propagation in complex media: P513

Conveners: Andrés Prieto (CITMAga, Universidade da Coruña), Paulo Amado-Mendes (University of Coimbra, Dep. Eng. Civil)

755 Nalu-Acoustics: A GPU-Accelerated High-Order Finite- Difference Solver for Acoustic Propagation with Application to Offshore Wind Farm Noise

This study presents Nalu-Acoustics, a GPU-accelerated high-order finite-difference solver for acoustic propagation analysis over large domains. The solver employs dispersion-relation-preserving (DRP) spatial discretization and low-dissipation low-dispersion Runge–Kutta (LDDRK) time integration to minimize numerical dispersion and dissipation. Perfectly matched layers (PML) are applied at domain boundaries to suppress spurious reflections and reduce the computational domain size. The solver is implemented in CUDA with multi-GPU support via domain decomposition, currently utilizing up to eight GPUs. A single-GPU configuration achieved an 11-fold speedup over the equivalent CPU implementation. To validate the solver, the authors conducted transmission loss measurements in shallow waters near the Southwest Offshore Wind Farm demonstration site. Over the 100–1000 Hz one-third octave bands, the predicted transmission loss agreed with the measurements within a mean deviation of 1.0 dB, comparable to the measurement standard deviation. The validated solver was applied to noise propagation from the 20-turbine demonstration site, using source data from an offshore wind turbine at rated power. The simulation resolved multi-source interference and reflection at the sea surface and seabed. The results indicate that Nalu-Acoustics can be used to estimate noise impact zones around offshore wind farms and to support environmental assessment of underwater noise exposure to marine life.

Speaker: Eunkuk Son (Korea Institute of Energy Research)

Full paper (PDF)

A17.04/A24.07 Spatial Audio Signal Processing: P486

Conveners: Jens Ahrens (Chalmers University of Technology), Franz Zotter (University of Music and Performing Arts), Orchisama Das, Jung-Woo Choi (KAIST), Antonio Figueroa-Duran (Universidad Politécnica de Madrid)

756 Discrete-time IIR filter design for radial filters by numerically optimised pole/zero placement

For a certain mode in spherical wave field expansions, a radial filter characterises the radially dependent diffraction phenomenon due to a spherical obstacle. Recently, a band-limited impulse invariance method (BLIIM) was proposed. It allows a more precise design of discrete-time filters for improved numerical simulations of such diffraction phenomena. However, for frequencies close to the Nyquist frequency and/or for high modal orders, BLIIM is numerically not straightforward or not feasible. Then, an optimisation-based infinite impulse response filter design with the continuous-time frequency response as target function, solving for the coefficients of the z-domain transfer function is one obvious workaround. An optimisation problem is exemplarily discussed for the radial filter occurring in the velocity-to-pressure spherical exterior expansion. This high-pass filter exhibits a steep slope and an extensively rippled pass band for a high modal order. The numerical experiments show that precise magnitude and phase responses of such radial filters can be achieved, requiring only some informed heuristics for the filter design parameters. The proposed approach outperforms IDFT-based frequency sampling, but not the BLIIM. The approach can be used where BLIIM is not feasible.

Speaker: Frank Schultz (University of Rostock)

Full paper (PDF)

757 Estimating Spatial Sound Field Feature Maps from a Reduced Set of Measurements

Acoustic features derived from binaural and spherical harmonic representations of a sound field can be used for objective analysis of sound field properties as well as for perceptual modeling. When a sound field with a particular spatial extent is considered, such as in multichannel loudspeaker reproduction, the position-dependent variation of sound field features may be used to define a sweet area where the reproduction is regarded as sufficiently similar to the loudspeaker setup's sweet spot. Because measuring spatially dense sound field feature maps requires considerable practical effort, reconstructing the spatial variation of sound field features from a reduced number of measurements is desirable. To explore the possibility of reconstructing spatial feature maps from spatially sparse measurements, sound field parameters were measured in a listening room using an automated rover carrying a dummy head or a spherical microphone array. Different strategies for spatial subsampling of the sound field feature maps obtained from the measurements are investigated to arrive at a minimal measurement setup required for the application of sweet area analysis in multichannel loudspeaker reproduction.

Speaker: Roman Kiyan (IKT - Leibniz Universität Hannover)

Full paper (PDF)

758 Sound Separation and Classification With Object and Semantic Guidance

The spatial semantic segmentation task focuses on separating and classifying sound objects from multichannel signals. To achieve two different goals, conventional methods fine-tune a large classification model cascaded with the separation model and inject classified labels as separation clues for the next iteration step. However, such integration is not ideal, as fine-tuning over a smaller dataset loses the diversity of large classification models; features from the source separation model are different from the inputs of the pretrained classifier, and injected one-hot class labels lack semantic depth, often leading to error propagation. To resolve these issues, we propose a Dual-Path Classifier (DPC) architecture that combines object features from a source separation model with semantic representations acquired from a pretrained classification model without fine-tuning. We also introduce Semantic Embedding Conditioning (SEC), which enriches the semantic depth of injected clues. Our system achieves a state-of-the-art 11.19 dB CA-SDRi and enhanced semantic fidelity on the DCASE 2025 task4 evaluation set, surpassing the top-ranked performance of 11.00 dB. These results highlight the effectiveness of integrating separator-derived features and rich semantic clues.

Speaker: Younghoo Kwon (KAIST)

Full paper (PDF)

A17.04/A24.07 Spatial Audio Signal Processing: S123

Conveners: Jens Ahrens (Chalmers University of Technology), Franz Zotter (University of Music and Performing Arts), Orchisama Das, Jung-Woo Choi (KAIST), Antonio Figueroa-Duran (Universidad Politécnica de Madrid)

759 Compensation of WFS Driving Gains for Large Front-Fill Loudspeaker Arrays

Sound spatialization techniques have proved effective for voice reinforcement in theaters, as they enable the synthesis of virtual sources that are co-located with the real sources on stage, thereby preserving the spatial coherence of the scene. However, in the case of wide front-fill arrays and off-center source positions, this approach generally concentrates the energy toward the side of the source, which compromises the listening comfort for a part of the audience seated on the opposite side, particularly those in the vicinity of the front-fill array. This is the case, for instance, with Wave Field Synthesis (WFS), whose driving functions yield low driving gains for loudspeakers on the opposite side of an off-center source. To address this limitation, we present a Large-Stage Compensation (LSC) of loudspeaker driving gains, tuned using computational perceptual models. It empirically increases the gains of loudspeakers located on the opposite side while keeping delay values unchanged to maintain the desired wavefront curvature. Acoustic and perceptual metrics (such as energetic coverage, speech transmission index, localization accuracy, and coloration) are simulated to evaluate and optimize the method, finding a balance between improved intelligibility and limited degradation in spatial and timbral quality. The LSC is modulated according to the source’s position on stage, and it is gradually applied only for critical off-center positions. The proposed method is evaluated using a simulation of a 30-meter-wide front-fill system at the 'Palais des Papes' in Avignon. Results show that the LSC improves energetic coverage and speech intelligibility, with moderate, quantifiable increases in coloration and localization errors.

Speaker: Hector Teyssier (STMS Lab, IRCAM, Sorbonne Université)

760 Efficient Diffuse Field Decorrelation: Spatialized Velvet-Noise

Rendering diffuse sound fields of late reverberation with computational efficiency and accurate spatial impression is a challenging task. This paper introduces spatialized velvet-noise, a framework that efficiently generates a decorrelated sound field from a single monaural reverberation in real time. The method utilizes a single velvet-noise sequence that is distributed spatially. The approach is compared to other computationally efficient decorrelation strategies, including allpass filters, simple delays, and alternative velvet-noise configurations. This comparison involves a technical analysis of spectral deviations and interaural coherence, alongside a formal listening experiment. Results reveal that spatialized velvet-noise significantly outperforms most of the tested configurations in providing an enveloping sensation or maintaining a neutral timbre. Notably, none of the tested method was found to surpass spatialized velvet-noise in the listening experiment. While simple delays are the only computationally more efficient option, spatialized velvet-noise significantly outperforms them in the listening test.

Speaker: Leon Kaiser (University of Music and Performing Arts)

Full paper (PDF)

761 A physics-based low-order filter model for near-field binaural rendering

Accurate binaural rendering of nearby sound sources is key for perceptual plausibility and reliable distance cues in virtual acoustics environments. Achieving this requires modeling of near-field effects, which exhibit a complex dependence on source distance and incidence angle. These effects can be introduced into measured or simulated far-field head-related transfer functions (HRTFs) by applying a near-field transfer function (NFTF). The NFTF may be defined as the ratio of the near-field to the far-field transfer function for rigid sphere scattering as a first approximation for the human head. However, efficient computation of the NFTF remains challenging in interactive virtual acoustics scenes. In this work, we propose a physics-based, low-order filter model for the NFTF comprising two first-order shelving filters. The model directly integrates analytical derivations for the asymptotic low-frequency and high-frequency filter gains based on rigid sphere scattering, while the filters' cutoff frequencies are parameterized by source distance and incidence angle. Because the proposed model is grounded in the underlying physics of rigid sphere scattering with arbitrary radius and distances of source and receiver, it is suited for extensions to broader applications, such as modeling the directional characteristics of sources and the scattering of objects. Objective evaluation demonstrates that the filter model provides an accurate representation for the NFTF with an average root-mean-square error (RMSE) of 0.1 dB and a maximum RMSE of 0.6 dB. These results indicate that the proposed model offers a practical and precise solution for incorporating near-field effects in interactive binaural rendering.

Speaker: Yuqing Li (Medizinische Physik, Universität Oldenburg)

762 Perceptual Evaluation of Acoustic-Center Alignment in Spherical Harmonic Encoding of Loudspeaker Directivity for Binaural Reproduction

In virtual and augmented reality (VR/AR), spatial audio rendering often relies on limited-order spherical-harmonic (SH) representations. While SH truncation in head-related transfer functions (HRTFs) is known to introduce high-frequency phase errors that degrade perceptual quality, analogous issues for loudspeaker directivity have received less attention. This work investigates preprocessing strategies for transforming measured loudspeaker directivity into SH form in the presence of direction-dependent time-of-arrival (TOA) inconsistencies, particularly pronounced in multi-way loudspeakers with frequency-dependent acoustic centers. Such inconsistencies, when encoded at finite SH order, can cause spectral artifacts. We compare simple magnitude-oriented approaches (peak alignment, minimum-phase) with methods that aim to preserve physically meaningful phase via broadband and subband acoustic-center compensation, as well as per-driver encoding with crossover reconstruction. A MUSHRA-like listening test assessed perceptual similarity to a reference based on the original measurements for virtualized stereo reproduction in anechoic and semi-diffuse conditions. Results show that the simpler preprocessing approaches generally yield higher perceptual similarity than acoustic-center-preserving approaches, while per-driver encoding performs comparably to peak alignment at higher orders. An ERB-band spectral error correlates with the perceptual trends, underscoring the primary importance of magnitude fidelity under SH truncation.

Speaker: Yasen Velchev (University of Music and Performing Arts)

Full paper (PDF)

A21.01 Automotive noise and vibration: P435

Conveners: Manfred Haider (AIT Austrian Institute of Technology GmbH), Martin Czuka (AIT Austrian Institute of Technology)

763 A Comparative study of transfer function between Target and Error microphone due to seat position changes for ARNC performance development

This study examines the influence of seat‑position changes on the transfer functions between the target point, representing the occupant’s ear location, and the error microphones installed at each seat to optimize Active Road Noise Control (ARNC) performance. In ARNC systems, error microphones remain fixed at predefined locations, while the target position varies with seat adjustment. This positional variation may affect the intended control performance, making it necessary to quantitatively evaluate its impact during the early system design phase. This study aims to evaluate the Virtual Transfer Path (VTP) by measuring and comparing the transfer functions between the occupant ear location and the error microphones for various driver and co‑driver seat positions within their allowable adjustment ranges. The comparison of the VTPs obtained from in‑vehicle measurements indicates that the transfer functions remain largely unchanged in the booming band (20-150Hz), whereas statistically meaningful differences emerge above 200Hz in the cavity/rumble frequency range. These observations indicate that the influence of seat-position variations on ARNC performance becomes more pronounced at frequencies above 200Hz, reflecting increased sensitivity of the VTP at higher frequencies.These findings further imply that, in the control of higher‑frequency in‑cabin noise sources beyond Road Noise (e.g., Wind and HVAC noise), a seat‑position grouping strategy with representative transfer functions can be an effective approach to ensuring robust ANC performance.

Speaker: Kwihyun Kang (HYUNDAI MOBIS)

764 Auditory Arrival Time Estimation for Approaching Vehicles With and Without Sound Level Cues and Azimuthal Angle Cues

When a vehicle approaches, there are several acoustic cues that can inform about the arrival time of the vehicle (time-to-collision, TTC), such as the change in intensity at the pedestrian’s position or the change in azimuthal angle. But which cues do we actually make use of? To answer this question, we manipulated the availability of sound level and angle cues. In the normal condition, both the sound intensity and the azimuthal angle of the simulated sound sources changed dynamically as the vehicle approached. In a level-[DO1.1]fixed condition, the intensity at the participants position did not change dynamically but instead corresponded to a static vehicle position, while the angle of the simulated sound sources changed during the approach. In an angle-fixed condition, the simulated sound sources remained at a static position, while the sound level changed during the approach. Within each condition, we varied the presented TTC as well as the size, speed, and sound power of the vehicle, and the presence of a reflecting ground surface. When angle cues were removed, participants showed an increased overestimation of the TTC, but the estimated TTC clearly followed the presented TTC. In contrast, when the level cues were removed, participants’ estimates varied only little with the presented TTC, showing a strong central tendency effect. At short TTCs, where the largest changes in angle occur, the estimated TTCs still reflected some consideration of the presented TTC. This demonstrates that in the absence of level cues, participants could make some use of angle cues, but level cues most strongly contribute to the arrival time estimation.

Speaker: Thirsa Huisman (Johannes Gutenberg University Mainz)

A21.03 Tyre/road noise: P436

Convener: Manfred Haider (AIT Austrian Institute of Technology GmbH)

765 Long-Term Changes in Road Surface Noise Measured by the CPX Method, National Specification 270

Road traffic noise is significantly influenced by road surface properties, in particular by the surface type and its condition. The noise generated at the tyre–road interface is commonly assessed using the Close Proximity (CPX) method according to ISO 11819-2, which enables direct source related measurements.This contribution addresses long term monitoring of the noise emission of different road surface types in the Czech Republic, with a strong international relevance. At the national level, a new road surface noise classification system (TP 270) is being introduced in connection with the implementation of the CNOSSOS EU calculation methodology for road surface noise. The system enables detailed assessment of the acoustic condition of road surfaces, tracking of acoustic degradation over time and optimization of road design and maintenance aimed at reducing traffic noise exposure. Long term results confirm consistency with international datasets, providing a reliable basis for the assessment and prediction of road surface noise within the EU.

Speaker: Petra Markova (Transport Research Centre)

Full paper (PDF)

A22.04 Propagation, Modeling and Simulation of Underwater Sound Fields: S159

Conveners: María Campo-Valera (Universidad Internacional de la Rioja), Dídac Diego-Tortosa (Institut de Ciències del Mar (ICM-CSIC))

766 Small leak detection in viscoelastic water pipes using the acoustic pressure gradient

Detecting small, quiet leaks in viscoelastic water pipes (e.g. polyethylene, polyvinylchloride) using acoustic techniques is challenging due to the high rate of sound attenuation in these pipes and presence of background noise. This research explores monitoring the radial acoustic pressure gradient within the water column as a novel method of detecting small leaks in water distribution pipes. Time-dependant CFD simulations are carried out to predict the sound pressure gradient generated in the vicinity of a leak. A bespoke full-scale laboratory facility with realistic pipe sections is developed and used to verify the proposed method. The resilience of the approach to interference from bulk turbulent flow is explored. This research supports the development of novel micromachine (MEMS) hydroacoustic sensors for leak detection to work with modern plastic pipe materials.

Speaker: Edward John (University of Sheffield)

767 Adaptive Chip-Level Delay--Doppler Tracking for DSSS-Based Underwater Acoustic Platforms

High-rate delay and motion estimation is important for maneuvering underwater platforms using direct-sequence spread-spectrum signals. Conventional frame- or symbol-level estimators provide limited temporal resolution, while direct differentiation of noisy delay measurements amplifies velocity-estimation errors. This paper proposes an adaptive delay-Doppler tracking method based on chip-level DSSS correlation observations. Sliding correlation and three-point interpolation are used to obtain sub-sample delay measurements. A coarse frame-level Doppler estimate is first obtained from the synchronization signal, and the residual Doppler variation is then extracted from the phase differences between consecutive complex correlation peaks after removing the BPSK data-dependent phase. A constant-acceleration Kalman tracker is then developed to jointly estimate propagation delay, radial velocity, and acceleration. Its measurement covariance is adaptively adjusted according to the peak-to-sidelobe ratio and localpeak curvature. Simulations show that the proposed method reduces thedelay and velocity RMSE compared with a fixed-covariance Kalman filter.Results from a Songhua Lake experiment further demonstrate that theestimated radial velocity is consistent with the GPS-derived reference. The proposed method enables DSSS communication waveforms to provide high-update-rate motion estimates without an additional acoustic ranging signal.

Speaker: Jie Wu (Guangzhou Marine Laboratory)

Full paper (PDF)

768 Operational Underwater Noise from a Monopile-Supported Offshore Wind Turbine: A vibroacoustic Analysis

Offshore wind turbines radiate continuous underwater noise during operation, yet predictive tools for assessing this structure-borne component are scarce. This study presents a physics-based vibroacoustic framework for a 10 MW monopile-supported turbine. Time.domain aero-hydro-servo-elastic simulations (OpenFAST) provide nodal accelerations, which are converted into equivalent dipole sources using a frequency-dependent radiation correction. Acoustic propagation is computed via analytical Green's functions with free-surface and seabed image reflections. A synthetic drivetrain excitation model captures high frequency generator forces missing in standard aeroelastic simulations.Results show that low-frequency radiation exhibits a dipolar patern aligned with the fore-aft bending mode, while drivetrain excitations above ~10 Hz produce more axisymmetric directivity. In the shallow-water environment (30 m depth), propagation is partially confined, with distance decay rates intermediate between cylindrical and spherical spreading. The emitted spectrum overlaps with the hearing ranges of several marine species in the mid-to-high frequency bands, primarily in the near field.The framework enables early-design quantification of underwater acoustic footprint and can support environmental impact assessmentes and monitoring strategies for monopile offshore wind farms.

Speaker: Raúl Sanz-Ramírez (Universidad Politécnica de Madrid)

Full paper (PDF)

769 Automatic Three-Dimensional Acoustic Field Mapping for Water Tank-Based Bioacoustic Experiments

Laboratory water tanks are widely used to investigate the effects of underwater noise on aquatic organisms. However, small bounded volumes generate complex and highly reverberant acoustic fields, and the spatial distribution of sound inside experimental tanks is often insufficiently characterized prior to behavioural or physiological studies. As a result, the acoustic exposure experienced by organisms may vary significantly within the tank, introducing uncertainty in controlled noise experiments. This work presents an automated methodology for three-dimensional acoustic field mapping in laboratory tanks based on broadband impulse response measurements. A spatial scan is performed throughout the tank, and at each measurement position a broadband excitation signal is recorded and processed through deconvolution to estimate the impulse response. The methodology is applied to an experimental tank previously used for low-frequency and virtual reality exposure studies and the resulting dataset allows the spatial variability of the sound field within the tank to be assessed and provides a reproducible framework for tank calibration prior to bioacoustic experiments. This approach contributes to improving the reliability and interpretability of controlled underwater noise studies conducted in laboratory environments, focusing on fish or other marine organisms.

Speaker: Ignasi Nou-Plana (La Salle - URL)

Full paper (PDF)

A23.02 Modelling & Simulation Techniques: P426

Conveners: Giuseppe Petrone (University of Naples Federico II), Marcus Maeder (Technical University of Munich)

770 Developing a Measurement System to Determine Human Ear Canal Wall Mechanical Impedance in-situ: Effect and Suppression of Torsional Motions in the Prototype

When wearing hearing aids, a mechanical interaction between the earmould and the skin of the outer ear canal occurs. The extent of this interaction, its influence on wearing comfort and long-term effects caused by continuously wearing a hearing aid are largely unknown. To quantify this interaction we develop a measurement system to measure the mechanical point impedance inside the human ear canal. The measurement system consists of an acceleration sensor and a vibration exciter that are both mounted on a cantilever beam. An electromechanically motivated model is used to estimate the attached mechanical load impedance. Because the model neglects torsional motion, a finite‑element (FE) model of the device was used to evaluate the induced rotational velocity for ideal and misaligned sensor/actuator placements. Furthermore, a tuning‑fork‑shaped beam was investigated in the simulation as a means to suppress torsional components. The FE analysis shows that the tuning‑fork geometry reduces rotational velocity compared to a conventional rectangular beam. Prototypes incorporating the new beam were subsequently built and analyzed using Laser-Doppler-Vibrometry, indicating a modest reduction in torsional motion for the new design.

Speaker: Marius Benkert (Jade University of Applied Sciences)

Full paper (PDF)

A10.08 Adaptive, Time-Varying, and Topological Acoustic and Elastic Metamaterials: S067

Conveners: Fabio Nistri (Politecnico di Torino), Vinicius Dal Poggetto, Gregory Chaplain (University of Exeter)

771 Diatom-inspired phononic pseudocrystal insulators

Diatoms are unicellular microorganisms whose silica exoskeletons, known as frustules, exhibit extraordinary hierarchical porosity across micro- and nanoscales. These natural structures combine mechanical strength, lightweight characteristics, and photonic functionalities. Despite the extensive investigation of their optical behavior, the elastodynamic properties of frustules and their potential as templates or bio-inspiration for elastic-wave manipulation remain largely unexplored.Guided by scanning electron microscopy (SEM) analysis of several diatom species, we identify key structural motifs—such as radial gradients and locally resonant features—and leverage them to design a family of cyclically symmetric phononic pseudocrystals capable of broadband elastic-wave attenuation, tonotopicity, and elastic energy confinement. The proposed structures are first modeled numerically and then fabricated at the microscale through two-photon polymerization (2PP). Experimental validation is currently ongoing and is carried out via laser-based vibrometric measurements.Overall, this work demonstrates that diatom frustules represent a compelling source of inspiration not only in photonics, but also for the control of elastoacoustic waves. Their inherent hierarchicity and grading of structural features—and its natural extension to three-dimensional architectures—has the potential to unlock compact, broadband phononic insulators of practical relevance for vibration control and wave filtering applications.

Speaker: Chiara Gazzola (Politecnico di Milano)

772 Elastic waves in continuosly defomring media

This work revisits wave propagation in deformable solids by extending the classical acoustoelastic framework to regimes where deformation evolves in both space and time. In standard acoustoelasticity, waves propagate through a pre-stressed but static medium, and their speed is modified by the underlying deformation. However, this approach breaks down when the material configuration itself becomes dynamic.Here, we consider media in which deformation actively modulates the material properties experienced by the wave. This leads to a fully coupled space–time problem, where wave propagation and kinematics interact. An analytical model is developed to describe this interplay, revealing effects that go well beyond classical acoustoelastic predictions, including frequency conversion, asymmetric transmission, and Doppler-like shifts induced by moving deformation fronts.Numerical simulations are used to validate the theoretical framework and to explore regimes where analytical solutions are not accessible. Crucially, the study is supported by experimental validation on soft elastomeric materials, where large, controllable deformations occur at wave speeds of the same order as the deformation rate. These experiments confirm the key predictions of the model, demonstrating the impact of space–time modulation on wave behavior.Overall, this work establishes acoustoelasticity as a limiting case of a broader, dynamically evolving framework for wave propagation in deformable media.

Speaker: Fabio Nistri (Politecnico di Torino)

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:00 Coffee break

10:20 Coffee break

10:20 Coffee break

A03.04 Acoustics of wooden buildings: S393

Conveners: Jean-Luc Kouyoumji (FCBA Institute of Technology), Heinz Ferk

773 Critical Issues during ISO 10848-1 Measurements of CLT Junctions with Resilient Interlayers

ISO 10848‑1 provides a well‑established methodology for characterising the vibration reduction index Kij of building junctions. However, its current version does not include specific guidance for configurations involving resilient interlayers, which are commonly used in Cross Laminated Timber (CLT) constructions. This raises the question of how to apply ISO 10848‑1 reliably when a resilient interlayer is present within a CLT junction. This article outlines practical and methodological considerations for this issue, addressing aspects not currently covered by the standard. Through a series of laboratory measurements on two different CLT mock-ups with resilient interlayers, the influence of several parameters was investigated by varying one parameter at a time, enabling direct comparisons of the resulting changes in Kij. Parasitic airborne noise, the effects of the fastening system (always present in CLT structures), and load effects - already examined in the literature - are briefly discussed. Additional effects specific to resilient interlayers are considered, including the influence of material creep, hardness, thickness, contact surface, and the position of the interlayer within the junction. These factors can lead to substantial variations in the measured Kij values, potentially resulting in inaccurate calculations and limiting the comparability between products from different manufacturers.

Speaker: Vincenzo Pettoni Possenti (Department of Industrial Engineering, University of Bologna)

Full paper (PDF)

774 The effect of static preloading on flanking sound transmission through cross-laminated timber junctions with resilient elements

The adoption of cross-laminated timber (CLT) is hindered by its poor acoustic performance. A critical factor is flanking sound, where vibrational energy is transmitted between two building components across a common junction. Resilient strips or pads can be added at the junction to decouple these components and reduce flanking sound. While effective, the performance of these solutions depends largely on the type of junction that is being considered. The attenuation of flanking sound between two building components i and j is quantified by the vibration reduction index Kij, which can be experimentally determined on-site or in a laboratory in accordance with ISO 10848. This work summarizes the findings of an experimental campaign on a horizontal X-junction made of CLT in a laboratory setting. Specific attention is paid to the effect of external static preloading, leading to added structural rigidity of the junction. In practice, this is the case for load-bearing walls supporting multiple floor levels. This effect is simulated by an external line load on the structure provided by a hydraulic press. In this study, multiple configurations are considered with and without resilient elements such as continuous strips and discrete acoustic pads at varying external loads.

Speaker: Paulo Pinto (CDM Stravitec)

Full paper (PDF)

775 Flanking sound transmission reduction through a decoupled CLT floor

Cross-laminated timber (CLT) is widely appreciated for its fast installation, sustainability, and lightweight properties. However, its high sound radiation efficiency makes it particularly susceptible to flanking transmission. At the same time, its aesthetic appeal often leads architects to leave surfaces exposed, avoiding additional linings.In standard design solutions, suspended ceilings are commonly installed to limit direct vertical sound transmission from the floor above while also concealing building service installations. This study explores an alternative strategy that introduces a resilient layer between floor elements to reduce lateral sound transmission between adjacent rooms via the upper floor.An experimental campaign was carried out on a horizontal X-junction mock-up. First, the vibration-reduction index (Kij) was measured on a continuous floor slab. The slab was then divided into two sections, separated by a resilient material, and the Kij for the floor-to-floor transmission path was measured again.The results demonstrate that floor decoupling with resilient material is an effective strategy, particularly at low and high frequencies, achieving improvements of up to 8 dB in the single-number Kij. The performance of this solution was further assessed through calculations in accordance with ISO 12354, based on the experimental data. These findings highlight new opportunities for acoustic design in CLT structures.

Speaker: Vincenzo Pettoni Possenti (Department of Industrial Engineering, University of Bologna)

Full paper (PDF)

776 Analysis of flanking sound transmission at floor– wall junctions in timber buildings

This paper presents an overview of the results from a multi‑year collaborative research project investigating structure‑borne sound transmission across floor–wall junctions in typical Swiss multistory residential buildings. The study focuses on experimentally characterizing flanking transmission paths that significantly influence airborne and impact sound insulation in lightweight and hybrid construction systems.A comprehensive experimental campaign was conducted on full‑scale laboratory specimens representing common floor constructions, including hollow box floors, mass timber floors, wood joist floors, and concrete–wood composite floors. These floor systems were combined with both wood‑frame and mass timber wall constructions to reflect current building practice. Sound insulation measurements were performed to quantify the contribution of individual transmission paths across floor–wall junctions.The results allow a systematic comparison of sound insulation performance between the investigated floor systems and wall type combinations as well as connectors. Particular emphasis is placed on identifying dominant structure‑borne transmission paths and on assessing the influence of floor and wall construction types on flanking sound transmission. The findings provide improved insight into junction behavior in timber‑based buildings and support the development of more reliable acoustic designs for modern lightweight and hybrid building systems.

Speaker: Stefan Schoenwald (Empa)

777 Kij measurement in flanking transmission test facility to predict sound insulation for wooden multistorey buildings

In 2023, 4 types of junctions were measured in an in-situ measurement mockup: the “Maquette Acoustique Bois”. The mockup is a three floors CLT based building, with four rooms on each level. The measurement strategy is Statistical Energy Analysis (SEA). In a previous paper the mockup was used to measure vibration reductions levels indexes Kij, in different configurations, where smart building systems junctions were implemented to reduce flanking transmissions. In this article authors demonstrate how measured results performed on the multistory mockup can be further implemented in the standard EN 12354 model or in a full physical SEA model. Engineer prediction tool proposed by EN 12354-serie standard is based on SEA, it is also based on measured Kij characteristics to predict flanking transmissions. Comparison between the different types of junctions measured show efficient strategies of building systems. Future acoustic prediction strategies will be based on measured Kij of smart junctions that include resilient channels and dampers. Industry is very creative in this domain and the only differentiation solution is to make available appropriately characterized products, with Kij measurements in open access.

Speaker: Jean-Luc Kouyoumji (FCBA Institute of Technology)

Full paper (PDF)

A03.05 Structure-borne sound and noise from building equipment: S017

Conveners: Berndt Zeitler (Hochschule für Technik Stuttgart), Jochen Scheck (HFT Stuttgart), Andreas Mayr (Rosenheim Technical University of Applied Sciences)

778 Predicting Noise from Building Service Equipment - Review of the Current Situation and Future Research Needs

The calculation method to predict noise from building service equipment according to EN 12354-5 has seen significant improvements and gained acceptance in recent years. In some countries (e.g. Germany), it is also finding its way into national building regulations as a means to show compliance with requirements. Still, there is more research needed for the prediction of noise from building service equipment to reach the same level of acceptance and usage as the prediction of noise from airborne or impact sound sources. In this contribution, the current situation of prediction methods for noise from building services is reviewed, focusing on technical as well as normative aspects. An overview is provided on the relevant standards (e.g. EN 12354-5, EN 15657, EN ISO 10848) and their relationship. Based on this overview, the needs for future research are identified. These include research questions regarding the sources, the receivers, their coupling, as well as questions concerning the predicted quantities and their uncertainties. A companion contribution by Scheck et al. focuses on characterization methods for noise sources from sanitary installations and the relevant standards (e.g. EN 14366-1, future EN 14366-2, EN ISO 3822).

Speaker: Christoph Höller (OTH Regensburg)

Full paper (PDF)

779 Source characterisation for all building-like source-receiver conditions - use case heatpump

EN 12354-5 describes methods for predicting noise from technical equipment based on both airborne and structure-borne sound emissions. With respect to structure-borne sound, laboratory methods for source characterisation are specified in EN 15657, providing the necessary input data for prediction models.In particular, for timber constructions, a complete characterisation of the source, including its mobility and the single-equivalent free velocity, is essential for reliable predictions of building-services noise.This paper presents a use case for the structure-borne sound characterization of heat pumps using indirect measurement methods on both high- and low-mobility reception plates.Furthermore, comparisons of source mobility obtained using the two-stage method and the direct measurement method according to EN 15657 are provided.Measurements on high-mobility reception plates require particular attention to mounting conditions, as the static load of the source may influence the results if not properly controlled. A key advantage of indirect measurement methods is the determination of blocked force levels and free velocity levels under realistic installation conditions. This is especially relevant for distributed contact interfaces between the source and the receiving structure, where direct methods are only of limited applicability.

Speaker: Andreas Mayr (Rosenheim Technical University of Applied Sciences)

Full paper (PDF)

780 Force-based transfer functions for force-source conditions – Where simplification pays off

Prediction of structure-borne sound from building service equipment (BSE) in heavyweight buildings, as outlined in EN 12354-5, requires knowledge of the normalized impact sound pressure level of floors or walls that the BSE is connected to. Standard tapping machine measurements provide a straightforward method of obtaining this quantity for floors, but are not applicable to walls. For such cases, EN 17823 describes the characterization of stairs and stair isolating elements using complex, narrowband transfer functions according to ISO 10848-1 to determine the impact sound pressure level of the wall. A similar transfer function method forms the basis of EN 14366-1 for the acoustic characterization of piping systems, where injected power and the resulting sound pressure in the receiving room are linked following the power substitution method of EN 15657. However, typical piping systems installed on heavyweight structures typically fulfill the force-source condition. Under this condition, the high effort of narrowband transfer function measurements is not required: RMS values of force and sound pressure — or velocity — in third-octave bands are shown to be sufficient, significantly reducing measurement effort and reducing uncertainties. Furthermore, the proposed simplified approach is suitable for in-situ characterization of BSE.

Speaker: Lukas Däuble (HFT Stuttgart)

Full paper (PDF)

781 Direct Measurement of Structure-Borne Transfer Functions in One-Third- Octave Bands

Structure-borne transfer functions enable the analysis of the frequency-dependent transfer behaviour of linear, time-invariant structures. While complex narrowband transfer functions are typically evaluated to assess specific system properties, building acoustics classically describes the behaviour of components using spectra in one-third-octave bands and spatial averaging. For structure-borne sound transmission, the normalized impact sound pressure level is the established metric. EN 12354-5 defines transfer functions as the ratio of the sound pressure levels to the excitation to describe this transmission. Although this quantity is usually determined using narrowband data and subsequently reduced to one-third-octave bands for prediction purposes, this study evaluates the feasibility of direct one-third-octave band measurements.The methodology compares narrowband reference measurements with the direct acquisition using a two-channel sound level analyser in laboratory conditions across a separating ceiling. The investigations cover the direct relationship HpF between sound pressure level and force level as well as the determination of the installed structure-borne power level LW,inst as a reference. Various signal processing modes were tested, including the equivalent continuous sound pressure level Leq and the maximum sound pressure level using both Fast and Impulse time weightings LF,max and LI,max.Initial investigations suggest that evaluating single impact excitation correlates well with integrated narrowband data. The results of direct measurement in one-third-octave bands indicate good agreement, though frequency-dependent deviations, particularly in higher frequency ranges, are observed and discussed. Overall, the study demonstrates that direct one-third-octave analysis can provide an efficient framework for evaluating transmission paths in practical building acoustics.

Speaker: Gina Boschatzke (Rosenheim Technical University of Applied Sciences)

Full paper (PDF)

782 On the Room Correction of Transient Sound Pressure Levels Generated by Building Services

Noise from building service equipment is often neither ideally stationary nor ideally impulsive, but exhibits a complex temporal structure. Requirements for sound insulation against noise from building service equipment in Germany are given in terms of LAF,max,n. This raises the question of how such noise should be treated in the prediction method according to EN 12354-5 and EN 15657, which is presently formulated for stationary excitation. While levels of stationary sounds are commonly normalized using the equivalent absorption area of the receiving room, EN ISO 10140-3 provides an alternative correction for ideal impulsive excitation based on receiving room volume and reverberation time. Building on previous work, this paper investigates how EN 12354-5 and EN 15657 can be extended towards transient excitation. A universal room-correction method is proposed that accounts for both room properties and signal transience. Validation measurements in a building-like test stand and in a reverberation chamber with varied reverberation times show that the proposed universal (transience-adaptive) correction reduces the influence of the room condition and improves the agreement between measured and predicted maximum sound pressure levels for transient signals.

Speaker: Anna Rieger (OTH Regensburg)

Full paper (PDF)

A03.08/A05.07 Heat pump acoustics in residential environments: S382

Conveners: Jack Havie-Clark, Christoph Reichl (AIT Austrian Institute of Technology)

783 Towards an In-Situ Measurement Protocol for Sound from Domestic ASHPs

In the planning of noise impacts from domestic air source heat pumps (ASHPs), compliance criteria typically refer to a single value in dB(A). Manufacturers provide laboratory test data based on reference thermal conditions, defined flow and return temperatures and single operating points, but in-situ environmental conditions vary widely and continuously, while system characteristics, settings and demands lead to wide variations in operational patterns. To demonstrate compliance with design criteria, a protocol for the measurement and assessment of in-situ sound emissions is required. A framework based on the highest daily LAeq,1hr sound emission is proposed; regression against the daily minimum temperature enables the emission to be expressed at 7°C, the laboratory reference condition. Spatial, temporal and environmental contamination factors are discussed. Pilot measurements at eight installations, comprising more than 1,200 hours of monitoring, illustrate three distinct operational sound emission profiles – unimodal, bimodal and continuously variable – and the potential to use a short-term measurement of a hot water cycle to infer longer-term emission is explored. Two dimensions along which in-situ operation may diverge from the part-load laboratory test condition are discussed qualitatively. A complementary qualitative protocol, based on narrative interviews, is described for use where the effect of an installation on residents is in question.

Speaker: Jack Harvie-Clark (Apex Acoustics Ltd)

Full paper (PDF)

784 Annoyance Ratings of Heat Pump Noise in Different Auralised Scenarios

Air-to-water heat pumps are increasingly used for heating, making their acoustic emissions an important factor in public acceptance. The perceived annoyance of heat pump noise depends not only on the sound characteristics of the unit but also on the installation location. The aim of this study was to investigate how simulated propagation scenarios (garden, living room with tilted window, and free-field reference), loudness conditions (original loudness and equal loudness), and contextual framing (own heat pump vs neighbour’s heat pump) affect the short-term annoyance of heat pump noise. A listening experiment was conducted in which participants rated the annoyance of heat pump stimuli using an 11-point rating scale. The stimuli are from recordings of one heat pump in two different operating conditions. The results show clear differences in the annoyance ratings between the auralised scenarios, with lower ratings observed in the living room scenario than in the garden scenario. Equalising the loudness of the stimuli and the contextual framing had only minor or no effects on the ratings, and no clear influence of the operating condition was observed. Overall, the results indicate that the auralised scenario has the strongest influence on the perceived short-term annoyance of heat pump noise.

Speaker: Lara Stürenburg (IHTA, RWTH Aachen University)

Full paper (PDF)

785 Varying disturbance of a recreational task during an air-source heat pump defrosting cycle

At low temperatures, air-source heat pumps (ASHPs) regularly perform a so-called defrosting cycle where the compressor of the ASHP is switched off, the circuit reversed, and the compressor switched on to defrost the heat exchanger. After a brief period the circuit is reversed back and the ASHP returns to regular operation. During this cycle the characteristics of the noise produced by the ASHP may change a number of times, e.g., when the compressor is switched off or on. The main interest of our work was to investigate how these different types of noise disturb people. For this, a recording of an ASHP in a climate chamber during a defrosting cycle were used. Six partial and shortened defrosting cycles with relatively stationary sections in between were concatenated to a single continuous recording of 27 minutes duration. Three overall sound pressure level settings ranging over 20 dB were used to scale the various sections. 20 listeners performed a low-level recreational task, i.e. coloring a Mandala, while listening to the continuous ASHP recording and to report whenever they felt disturbed by the noise. Pooled overall responses normalized by the duration of the state or event showed increased disturbance during phases in which the sound changes, in particular the restart of regular operation and the first valve switch. The relation between normalized responses and (psycho)acoustic quantities such as loudness varied also with the ASHP state suggesting that defrosting cycles may need to be considered separately when assessing the annoyance of ASHPs.

Speaker: Christian Kasess (Acoustics Research Institute, OeAW)

A05.05 Advances in noise mapping: S034

Conveners: Christian Kirisits, Eoin King (The Galway Sound Lab, School of Engineering,University of Galway)

786 New Road Traffic Emission Values for CNOSSOS in the Netherlands

Current emission factors used in the Dutch implementation of CNOSSOS are based on measurements from campaigns conducted approximately 17 and 30 years ago. Road vehicles have changed in the meantime, and these values are no longer representative for the current vehicle fleet in the Netherlands. To address this, new measurements were conducted in 2020 to determine updated emission values for road vehicles. These new values were incorporated into the national calculation method in 2024. No new emission factors for CNOSSOS were derived at that time. To keep the national method and CNOSSOS aligned we have derived new emission factors for CNOSSOS based on these more recent measurements. In this paper, we describe the methodology used to derive the new emission factors. We also analyse the impact of using these updated values on noise maps produced in the upcoming round of noise mapping, compared to the previous round. The effects vary by vehicle speed: at low speeds, emission values are substantially lower than before, while at high speeds, they are slightly higher. For context, we also compare these new Dutch CNOSSOS emission factors with those used in other countries for noise mapping.

Speaker: Arnaud Kok (National Institute for Public Health and the Environment)

787 Measurement-Based Environmental Noise Modelling of a Football Stadium

Assessment of environmental noise emissions from football stadiums is a complex task due to dynamically varying crowd noise and the contribution of high-power sound reinforcement systems. Reliable prediction of stadium-related environmental noise requires source characterization based on real measurement data and properly calibrated calculation models.This study presents a measurement-based environmental noise modelling approach applied to a medium-sized football stadium in Budapest, Hungary, with a capacity of approximately 13,000 spectators. The measurement campaign was conducted during a sold-out Hungarian top-division football match.Continuous synchronized measurements were carried out simultaneously inside and outside the stadium using multiple sound level meters. Several fixed noise monitoring stations were installed in the surrounding area, while additional mobile measurement positions were used to investigate spatial variations in sound propagation. Inside the stadium, six simultaneously operating measurement systems recorded the acoustic contribution of the crowd and the stadium’s public address system throughout the event.The primary objective of the study was to derive representative acoustic emission data suitable for environmental noise modelling purposes. Based on the measured results, sound power levels were estimated for the audience, treated as incoherent surface sources, and for the public address system, represented as discrete point sources. The environmental noise model was subsequently calibrated and validated using measured sound pressure levels at multiple receiver positions.The presented methodology provides a practical framework for realistic stadium noise prediction and environmental impact assessment under real operating conditions.

Speaker: Miklos Markus (FONOR Ltd.)

788 Mobile Acoustic Sensor network for Automatic NOise Mapping: Project MASANOM

The objective of the project MASANOM is to automatically plot noise maps with georeferenced data sent from low-budget autonomous noise measurements stations installed in low-background noise vehicles riding around the city. This method reduces significantly the cost of generating noise maps, thus, it is specially relevant for small cities or developing countries.However, there are several difficulties to overcome and computations that have to be performed to apply this system. First, the noise measurement station needs to have a GPS and a communication system, while at the same time being energetically autonomous. This is achieved by greatly optimizing all of the computations to reduce its power use, and by using a solar panel and a vibration energy harvester to power the station. Second, the noise measurements taken at a position in the road have to be corrected and transformed to façade levels at 4 m height according to the European Directive 2002/49/EC, which is achieved with transfer functions obtained with fits from numerical simulations. Finally, the vehicle fleet size and routes have to be computed with traffic simulations to grant the spatio-temporal coverage required to produce accurate annual noise levels to plot a city noise map.

Speaker: Andreu Balastegui (Universitat Politècnica de Catalunya)

Full paper (PDF)

789 Beyond noise contours: simulating psychoacoustic metrics and population-based noise impacts of individual flights

Noise maps of individual flights conventionally show contours of A-weighted levels: they mark where a given exposure is exceeded, but not how the noise is perceived nor how many people are affected. To move beyond this, the aircraft noise simulation model sonAIR was extended with the psychoacoustic single-event metrics Effective Perceived Noise Level (EPNL) and Zwicker's loudness, and with exposure-response-based population impact indices, all computed from simulated one-third-octave band spectra. EPNL, normally obtained from certification measurements, is computed after realigning the band grid, with a dampened spectral roll-off above 5kHZ that avoids spurious tone-correction penalties; the time-varying loudness is approximated by evaluating the stationary Zwicker method per time step. Verification against calibrated microphone recordings of two A320neo flyovers shows that the simulated metrics are as accurate as the underlying, previously validated emission and propagation model. Linking the resulting maps with population data yields maps of highly annoyed people and expected awakening reactions. Applied to a conventional and a low-noise approach procedure from the SESAR project DYN-MARS, the absolute number of affected people differs strongly between metrics, whereas the relative benefit of the low-noise procedure remains stable.

Speaker: Stefan Schucker (Empa Acoustics and Noise Control 509)

Full paper (PDF)

790 Predicting the Impact of Future Light-Rail Expansion on Urban Park Soundscapes Using CNOSSOS-EU and the TRAPT Model

Potential environmental noise from future transportation expansion, driven by urbanization and traffic growth, poses significant risks to human health, making it urgent to predict and quantify. In contrast, urban parks serve as important spaces for escaping the increasingly noisy urban environment. Previous studies have applied the Tranquillity Rating Prediction Tool (TRAPT), integrating acoustic indicators with visual factors, to assess tranquillity in urban parks. However, with the rapid expansion of light-rail systems and the widespread adoption of theCNOSSOS-EU noise modelling framework in Europe, there is an urgent need to evaluate future soundscape conditions under projected scenarios. This study conducted sound pressure level measurements and photographic surveys at 57 locations within the park. The CNOSSOS-EU framework was applied to predict noiselevels associated with the planned Dublin Luas light-rail extension in Tolka Valley Park. Tranquillity mapping was then performed using the TRAPT and Inverse Distance Weighting (IDW) under different traffic volume scenarios for 2035 and 2057. The results indicate a general decline in tranquillity ratings across the park. The most affected areas decrease by 5.33 in 2035 and 5.48 in 2057. The median sound pressure level increases from 56.9 dB (2026) to 59.4 dB (2035) and 59.8 dB (2057), with peaks reaching 90.0 dB.

Speaker: Guo Ye (Trinity College Dublin)

Full paper (PDF)

A09.00 Machine learning and artificial intelligence in acoustics: S386

Conveners: Mirco Pezzoli, Alois Sontacchi (University of Music and Performing Arts), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Franz Pernkopf (Signal Processing and Speech Communication Laboratory)

791 Dynamic Sound Field Regression via Physics-informed Temporal- Frequency Residual Learning

Dynamic sound sources, such as footsteps or other moving sources, are common in real life but difficult to model because their characteristics and trajectory are often unknown. This paper presents a machine learning algorithm for sound field construction based on a temporal-frequency residual neural network. The algorithm is designed to capture how sound evolves across both time and frequency, effectively reconstructing the field even though the source trajectory and characteristics are unavailable. A key advantage of this algorithm is its ability to incorporate physical priors during training. By embedding acoustic propagation constraints as a regularization term, the neural network can better guide the reconstruction process toward physically consistent results. We tested the algorithm in simulated 2D and 3D scenarios; the results show that the proposed algorithm consistently outperforms the standard baseline in terms of reconstruction accuracy, while the physical prior significantly enhances stability under complex motion patterns.

Speaker: Yuanxin Xia (DTU Electro)

792 Real-Time Broadband Sound Field Regression via Massively Parallel Frequency Stacking Neural Networks

Neural implicit fields are a powerful framework for sound field regression, but their application to broadband problems has been challenged by the prohibitively expensive computational cost of training neural networks. To break this barrier, this paper proposes a massively parallel strategy via frequency stacking. This architecture factorizes the broadband problem into independent, per-frequency branches that are trained in parallel, leveraging efficient GPU batching and parallel execution. Our results show that the proposed stacking architecture outperforms the conventional frequency conditioning architecture, with accelerations greater than two orders of magnitude and lower regression errors, which demonstrates a real-time potential for broadband sound field regression in small spaces. It is also demonstrated that incorporating a partial differential equation (PDE) residual loss further enhances the accuracy, particularly in the low to mid-frequency regime, despite at a cost of slightly longer training times. These findings offer practical guidelines for efficient, potentially real-time, yet high-fidelity sound field regression.

Speaker: Yuanxin Xia (DTU Electro)

Full paper (PDF)

793 Topological Features for Environmental Sound Classification

In this work, we investigate whether topological features derived from persistent homology of Takens embeddings can complement conventional acoustic features for environmental sound classification on the ESC-50 dataset. A compact statistical summary of persistence diagrams is extracted from short overlapping segments, aggregated at the recording level, and combined with conventional acoustic features. Experiments show that fusing topological and acoustic descriptors consistently improves classification performance across multiple classifiers. With a Random Forest classifier, the proposed fusion achieves 48.5% accuracy, an absolute improvement of 7.6 percentage points over the baseline. With a multilayer perceptron, the improvement reaches 12.1 percentage points, yielding 55.1% accuracy. A dimensionality-matched control experiment confirms that these gains reflect informational content, not extended dimensions. This work provides empirical evidence that persistent homology captures structural properties of environmental sounds that are complementary to conventional acoustic features.

Speaker: Jiwon Seo (Fraunhofer IDMT)

Full paper (PDF)

794 AI-Driven Detection and Acoustic Modeling of Anthropogenic Underwater Noise Across European Marine Basins

Anthropogenic underwater noise is becoming an increasing concern for marine ecosystems, especially in coastal and semi-enclosed areas exposed to intense commercial, industrial, and recreational vessel traffic. Within the DeuteroNoise project, we present an AI-driven framework to detect, characterize, and model vessel-generated underwater noise across five European marine basins: the North Adriatic Sea, the Lagoon of Venice, the Barcelona coast, the North Sea, and the Black Sea. Our approach combines passive acoustic monitoring, in situ measurements, and simulation-oriented analyses to identify representative soundscapes, describe their spatial and temporal variability, and support the controlled reproduction of realistic acoustic environments in laboratory conditions. Particular emphasis is placed on advanced signal processing techniques for denoising, feature extraction, spectral and temporal analysis, event detection, and acoustic pattern characterization from complex underwater recordings. These descriptors are then incorporated into artificial intelligence and machine learning pipelines for vessel-noise identification, acoustic scene classification, and predictive modeling in diverse marine environments. The proposed framework aims to improve the interpretation of large volumes of raw acoustic data, reduce uncertainty in soundscape assessment, and enable the development of computational models that can be transferred across basins with different environmental and traffic conditions. In addition, these results provide a solid basis for controlled experimental studies on marine invertebrate deuterostomes exposed to anthropogenic noise. By bringing together monitoring, signal processing, and AI-based predictive tools, this work contributes to a more comprehensive assessment of underwater noise pollution and supports future mitigation strategies, environmental management, and evidence-based decision-making for healthier and more sustainable marine ecosystems.

Speaker: Rosa Ma Alsina-Pagès (La Salle, Universitat Ramón Llull)

Full paper (PDF)

795 Application of Siamese networks to Anomaly Detection in Multibeam Sonograms For CCS Monitoring

Effective anomaly detection in active sonar data is critical for the long-term monitoring of offshore Carbon Capture and Storage (CCS) sites, where early identification of leaks is essential for environmental safety and regulatory compliance. Multibeam sonar systems offer high-resolution imaging of the water column but produce large amounts of data, rendering conventional reconstruction-based anomaly detection methods computationally demanding and unsuitable for bandwidth-limited deployments.This study investigates the application of Siamese convolutional neural networks for anomaly detection in multibeam sonar imagery. Instead of reconstructing input images, the proposed method learns a similarity measure from feature embeddings derived from pairs of sonar observations, enabling anomaly detection based on distance comparisons. A dataset comprising labelled anomalous and non-anomalous multibeam sonar images was collected at an open-water test facility, with anomalies generated using controlled scattering and wake-producing events.This study explores how model configuration and training data composition influence Siamese-network-based anomaly detection in multibeam sonar imagery. Experiments were carried out using different network sizes and varying fractions of the available pairwise training combinations to examine their impact on model behaviour and training efficiency. By considering how changes in architectural complexity and combination count affect learning stability and generalisation, the work seeks to identify configurations that are well suited to resource-constrained, edge-computing environments typical of long-term CCS monitoring deployments

Speaker: Jonathan Farr (University of Southatmpon)

Full paper (PDF)

A14.01 New approaches for improving and assessing outcomes with hearing aids: S379

Conveners: Hendrik Husstedt (Deutsches Hörgeräte Institut GmbH), Inga Holube, Florian Denk (German Institute of Hearing Aids)

796 Bone-conduction-induced skin-vibrations - Effect of measurement site, sex, and soft-tissue thickness

Purpose: Bone-conduction hearing systems (BCHS) transmit sound via mechanical vibration through soft-tissue and skull. The output and the sound quality (THD, Attenuation) of bone-conduction implants and skin-drive hearing systems (that work with low contact force) cannot be directly quantified via conventional measurement devices like artificial mastoid and skull-simulator. Methods: LDV-measurements were performed on 16 participants (age 20.5-29.5 years; 10 female, 6 male) at multiple positions on the head. Pure-tone stimuli (500–4000 Hz) were presented via a Radioear B81 transducer at intensities from −10 to 50 dB HL. Soft-tissue thickness at each position was assessed by ultrasound at low and regular contact-force.Results: Vibration amplitudes exhibited frequency-dependent attenuation with distance from the stimulation site. We measured the highest amplitudes and variability (SD: 7–14 dB) closest to the transducer at the mastoid. Measured amplitude was lower at the concha by 23.2 dB, at the earfront by 38.8 dB, and at the zygomatic arch by 32.3 dB. The lowest variability (SD: 2–6 dB) was measured at the zygomatic arch. Measurements at the concha provided a balance between signal strength and reproducibility. At higher frequencies (1000-4000 Hz) females exhibited 7-10 dB higher vibration amplitudes than males at identical stimulation levels, correlating with thinner soft-tissue. Under regular contact force, tissue thickness decreases by 23-34% with the soft-tissue at the mastoid showing the largest decrease (34%).Conclusion: In-vivo LDV measurements revealed correlations between tissue thickness at the measurement site and vibration amplitude. These findings underscore the importance of tissue properties and coupling conditions in clinical and technical evaluation.

Speaker: Dawid Brüning (Martin Luther University Halle-Wittenberg)

797 Clinical Use of the revoloud Loudness Validation Method: A Survey of Hearing Care Professionals

Currently, a typical hearing aid fitting procedure starts with amplification settings prescribed by audi-ogram-based fitting methods, followed by a subse-quent, often iterative, fine-tuning phase. While tech-nical procedures such as real-ear measurements are widely used to verify amplification, they do not di-rectly assess perceptual dimensions. The method revoloud was developed to enable a rapid clinical assessment of aided loudness perception. The method presents 60 natural signals covering three presenta-tion levels and four frequency categories. Patients rate the perceived loudness of each signal, and the results are visualized as a loudness map relative to a normal-hearing reference. To investigate the meth-od’s use in routine clinical practice, a survey among 125 hearing care professionals using revoloud was conducted. The results indicate frequent use of the method in daily practice, with 83 % of respondents reporting usage at least three times per week. A large majority of respondents reported perceived benefits, including fewer loudness-related complaints with hearing aids in daily life and improved patient understanding and involvement during the fitting process. These findings suggest that perceptual loud-ness assessment can support outcome-oriented hear-ing aid fitting in clinical practice.

Speaker: Kirsten C Wagener (Hörzentrum Oldenburg gGmbH)

Full paper (PDF)

798 Self-Directed Gain Selection Based on Speech MCL Measurements

In clinical settings, hearing-aid gain is typically prescribed based on the audiogram. In out-of-clinic settings, audiograms are difficult to measure because controlled stimulus presentation is generally impracticable. The most comfortable level (MCL) for speech listening could offer an alternative basis for hearing-aid gain prescription, but current speech MCL measurement methods have some shortcomings.This study developed and evaluated a new method for measuring speech MCLs based on pairwise stimulus comparisons and a Bayesian backend. An extended version of this method was then used for determining individual most-preferred gain (IPG) settings for speech listening.Fifteen normal-hearing and 13 hearing-impaired participants were tested. The speech MCLs were compared to data obtained with a standard bracketing method. The IPG settings were determined based on a set of candidate gain settings derived for standard audiograms, without a priori knowledge about the participants’ own audiograms. The IPG settings were determined twice – once for listening comfort, and once for perceived speech clarity.The speech MCLs measured with the new and standard method were strongly correlated, with comparable test durations. The IPG settings obtained for listening comfort showed less high-frequency gain than those obtained for speech clarity. Overall, the developed methods show promise for out-of-clinic applications.

Speaker: Michal Fereczkowski (University of Southern Denmark)

799 Objective Analysis of DNN-based Hearing Loss Compensation and Comparison to Listening-test Results

Recent advances in auditory modelling and deep neural networks have enabled the development of new hearing loss compensation methods. In these approaches, the parameters of a neural network are optimized so that the auditory representation of its output for a hearing-impaired listener is as close as possible to the normal-hearing auditory representation of clean speech. Previous studies have demonstrated the potential benefits of such systems, for example Leer et al. (2025) and Drakopulos (2025), using physiological auditory models during training.In this work, we propose a hearing loss compensation method based on the psychoacoustical loudness perception model AUDMOD (Bramsløw, 2004). The method was evaluated both objectively, using HASPI, and in a listening test with ten hearing-impaired participants, , using the Danish Sentence Test (DAST) and an AB preference test. The tested algorithms included different combinations of the proposed hearing loss compensation method and neural-network-based noise reduction, and their performance was compared with NAL-NL2.The resulting individual output signals were analyzed using several additional approaches, including spectral and temporal insertion-gain analysis and the assessment of modulation-frequency preservation. The findings reveal discrepancies between HASPI predictions and speech intelligibility measured in the listening tests.

Speaker: Szymon Drgas (Poznan University of Technology)

800 Performance Assessment for DNN-based Speech Enhancement Algorithms

Speech understanding in noise remains the primary challenge for individuals with hearing loss and a central target for hearing aid technology. Over the past forty years, advances in directional microphones and statistical noise cancellation have delivered substantial improvements in signal to noise ratio (SNR). More recently, deep neural network (DNN)–based speech enhancement has enabled a fundamental shift in algorithmic capabilities, offering robust performance even in complex and dynamic acoustic environments.Despite these technological advances, outcome measures used to assess hearing aid performance have changed little. Metrics such as speech intelligibility, SNR improvement, and user preference remain dominant. In addition, test setups often rely on static noise sources, limited spatial complexity, and favorable SNRs. While well suited to demonstrate the benefits of classical signal processing approaches, simplified test setups tend to overestimate algorithmic benefit and fail to reflect the benefit in real world listening.In this talk, I will demonstrate how the estimated SNR improvement depends on the spatial complexity, noise type, and input SNR of the background noise. Under simplified conditions, classical noise cancellation approaches can perform comparably to or better than DNN-based speech enhancement. However, in more complex environments, DNN based approaches provide superior benefit. The findings demonstrate that simplified test paradigms obscure the conditions under which DNN-based algorithms outperform classical methods. To keep pace with technological progress, the field is moving toward more ecologically valid test methods, advanced modeling approaches, and outcome measures that extend beyond SNR and speech intelligibility and toward quality of life relevant measures.

Speaker: Stefan Raufer (Sonova AG)

A18.06 Hospital Soundscape, Hospital Acoustics, Soundscape design: S370

Conveners: Elif Özcan (Erasmus MC / TU Delft), Ela Fasllija (Erasmus MC / TU Delft), Thomas Hampton

801 Noise in the outdoor areas of a University Hospital: a survey method

Noise in hospitals can significantly affect patients and healthcare staff, disrupting physiological processes and contributing to stress, annoyance, and reduced well-being. Although indoor hospital acoustics have been widely investigated, the acoustics of outdoor healthcare spaces remain under-explored, despite their potential to support healing environments. As part of a three-year project, the acoustic environment of Sant'Orsola University Hospital in Bologna, Italy, is being investigated. An integrated methodology combined acoustic monitoring, numerical modelling, and two participatory soundwalks conducted in accordance with ISO 12913-2. Forty-two participants assessed locations using validated Italian soundscape descriptors. Results show a soundscape characterized by anthropogenic sources, particularly internal traffic, service vehicles, ambulances and construction activities, whereas natural sounds are limited to courtyards, green areas, and locations near water features. The Circumplex Model establishes baselines for ISO pleasantness and ISO eventfulness. This supports defining target values and evaluating whether design interventions move the soundscape towards the desired perceived state. This study demonstrates how integrating perceptual and physical data can support the identification of critical areas and inform human-centred design strategies. By emphasising the importance of considering the soundscape from the earliest planning stages, these findings contribute to the development of multidisciplinary guidelines to improve the restorative and supportive outdoor environments for healthcare.

Speaker: Alessia Nora (Department of Industrial Engineering, University of Bologna)

Full paper (PDF)

802 Methods, acoustic measures, and sound sources in hospital soundscape research: a scoping review

The present study aims to systematically review existing research on hospital soundscapes with a specific focus on studies incorporating user-based assessments. The objective is to identify methodological approaches, acoustic indicators, and dominant sound sources, in order to highlight current gaps and inform future research directions.The research method involved a preliminary search in Scopus and Web of Science databases using combination of keywords related to the acoustic environment, hospital setting and user perception. All retrieved records were imported into Covidence for screening and management of the review process. The studies were reported in accordance with PRISMA 2020 guidelines. From an initial corpus of 148 studies, 16 met the eligibility criteria and were included for detailed analysis. The selected studies were evaluated according to research design, acoustic indicators, and reported sound source types and the extracted data were synthesized thematically. Four main methodological categories emerged: experimental, mixed-method, perception-based, and computational. Across these studies, acoustic characterization was predominantly based on sound level metrics, with 13 out of 16 reporting indicators such as LAeq, Lmax, Lden, L10,90 followed by event/frequency-based metrics, psychoacoustic, speech-related, composite or index based, and room acoustic parameters as less frequently reported. Regarding sound sources, human-related sounds were reported in 13 out of 16 studies making them the most frequently identified sound source. Medical equipment and alarms were reported in 10 studies, followed by mechanical systems, while natural and environmental sounds were rarely represented. The review provides a structured overview of current research practices, identifies methodological gaps in user-based hospital soundscape research, and offers directions for future methodological development.

Speaker: Enkela Alimadhi (AARE-SA)

Full paper (PDF)

803 Exploring the Effect of Sound Augmentation on Soundscape Perception in a Hospital Outpatient Waiting Area

Prior research has shown that human-based sounds dominate healthcare waiting environments and are associated with negative perceptual qualities. Yet acoustic design interventions targeting soundscape quality remain limited. This study explores the effect of sound augmentation on soundscape perception in a hospital outpatient waiting area based on an in situ binaural recording using Odeon’s auralization tool. A treated condition incorporated simulated wall and ceiling absorption based on measured impulse responses. Five scenarios were constructed: IR0: original; IR1: with acoustical treatment; IR2: augmented with a bird sound; IR3: augmented with a water sound; and IR4: augmented with a music sound. The augmentation sounds were selected based on findings from a previous healthcare soundscape literature. Five experts evaluated all scenarios using the 79-item Turkish soundscape attribute set. Pleasantness and Eventfulness values were calculated using the loadings of a validated Turkish soundscape circumplex model to compare the perceptual effects of the intervention scenarios. Findings revealed that sound augmentation shifted soundscape perception toward more positive affective dimensions. While bird and water sounds showed the greatest perceptual improvement, music yielded comparatively weaker effects. This study is intended to inform further research involving laypeople and guide acoustic design interventions in outpatient healthcare settings.

Speaker: Oya Yildiz (Bilkent University)

Full paper (PDF)

A21.03 Tyre/road noise: S405

Convener: Manfred Haider (AIT Austrian Institute of Technology GmbH)

804 Acoustic Performance of road pavements and the role of pavement characteristics on the brisa motorway network

BRISA, Portugal’s leading highway concessionaire, is committed to place sustainability at the core of its road infrastructure management, with a particular focus on mitigating traffic noise through efficient pavement management. In this context, BRISA has been investigating the acoustic impact of replacing end-of-life pavements, analysing both the immediate noise reduction and the long-term evolution of acoustic performance.To this end, tyre–road noise has been monitored on several motorway sections using Close-Proximity (CPX) and Statistical Pass-By (SPB) methods. Measurements were conducted prior to pavement renewal and repeated periodically after.The results show consistent reductions in traffic noise following pavement renewal. However, the acoustic performance does not evolve monotonically over time, exhibiting fluctuations as pavements age and their surface characteristics change.These findings highlight the need for further research on the relationship between pavement surface properties and acoustic behaviour throughout the pavement life cycle. The outcomes of this work aim to support the refinement of predictive noise models, including improvements to the CNOSSOS’ pavement database, and to contribute towards optimisation of both pavement design and maintenance strategies for reducing environmental noise, underscoring the critical role of pavement management in the development of quieter and more sustainable road networks.

Speaker: Maria Ramos (BRISA)

Full paper (PDF)

805 Effects of Speed and Acceleration on OBSI Tyre--Road Noise for Electric Vehicles

With the increasing adoption of electric vehicles, tyre--road noise is becoming the dominant source of vehicle exterior noise at low and medium speeds. This study analyses on-board sound intensity (OBSI) data collected under constant-speed and acceleration conditions for five production tyres on three electric vehicles, measured on an ISO~10844:1994 reference surface.The baseline speed dependence of OBSI levels is described using a log-linear mixed-effects model, confirming speed as the dominant factor governing tyre--road noise (9.9dB per speed doubling, RMSE = 0.24dB). Acceleration-related contributions, evaluated using a two-stage residual approach, show substantial variability across tyre--vehicle combinations, with no consistent overall effect and indications of confounding with vehicle ch aracteristics.The trailing--leading level difference exhibits a tyre-dependent speed sensitivity that shows a strong association with rolling resistance (r = 0.95, p = 0.014), supported by within-vehicle comparisons. These findings suggest a potential link between rolling resistance and the balance of leading- and trailing-edge noise generation mechanisms.

Speaker: Jiawen Meng (Karlsruhe Institute of Technology)

Full paper (PDF)

806 A numerical model for calculating the forced response of a tyre rolling on a rough surface

The understanding of tyre/road interactions is of great interest in traffic noise reduction and tyre design. This paper deals with the forced response of a slick rolling tyre in contact with a rough road surface. The calculations are performed in quasi-statics in a fixed contact zone using a Eulerian coordinate system, assuming no coupling between structural vibrations and the contact conditions for the sake of time cost. In the contact model, the tyre tread is simplified by an elastic half-space and the contact problem is solved using a direct numerical method. The resulting contact forces are firstly post-processed by means of time Fourier transform and angular Fourier decomposition. Then, they are injected into a waveguide finite element model developed for rotating and inflated circular structures under small perturbations. The vibrational fields along the tyre tread center line are obtained at frequencies up to 4 kHz. Four cases are studied, including two vehicle speeds (30 km/h and 90 km/h) and two real road surfaces with different macro-textures. The energy distribution of the excited modes and the effect of rotation are observed in dispersion diagrams. While based on a different methodology than existing approaches, the results show good agreement with the literature.

Speaker: Jinhong Li (Université Gustave Eiffel - UMRAE)

807 An Axisymmetric Modelling Approach for Predicting Tyre Vibration

Road traffic noise is a major environmental concern in urban areas due to its adverse health effects. At a steady driving speed of 35-120 km/h, tyre noise is known to be the dominant noise source of car noise. The tyre has a multi-layer structure of rubber, steel and textile components. Therefore, a full 3D model is computationally expensive and prevents repeated simulations. In contrast, simplified tyre wave models show limited accuracy above 300Hz, which corresponds to the dominant frequency range of pass-by noise. To solve this limitation, this research proposes an axisymmetric FEM approach for the tyre capable of predicting the frequency response in high-frequency bands. This model consists of two equivalent layers representing the rubber and carcass materials, which enables a significant reduction in computational cost compared with conventional 3D models. In addition, by retaining the fundamental tyre structure, the model can predict frequency responses in the high-frequency range. Validation against experimentally measured tyre frequency responses demonstrates improved agreement in the frequency band above 300 Hz, where previous approaches showed limited accuracy. Therefore, the model provides an efficient tool for evaluating tyre surface vibrations in road-excitation-based tyre noise prediction.

Speaker: Kazuma Saito (University of Southampton)

Full paper (PDF)

808 Optimizing Microperforated Metamaterial Panels in the Wheel-Arch for Tire/Road Noise Reduction: Numerical approach

This study proposes the use of an acoustic metamaterial in the wheel arch to reduce the noise generated by tire-road contact. The method employed is based primarily on the use of a simplified numerical model of a wheel arch that combines finite elements with boundary elements to evaluate the impact of an acoustic absorption treatment in this area. Based on this assessment, an acoustic metamaterial consisting of parallel micro-perforated panels is selected and optimized to achieve the target performance in terms of absorption, dimensions, and frequency range. Finally, the expected absorption coefficient is verified by several ways, using an analytical model, a numerical model, and a Kundt tube based measurements of the 3D-printed model. The obtained results pave the way for full-scale evaluation of such metamaterial device in a wheel arch.

Speaker: Robin Mafféïs (Université Gustave Eiffel)

Full paper (PDF)

A21.05 Noise barriers and mitigation techniques for road traffic and railway: S155

Convener: Andreas Fuchs (AIT Austrian Institute of Technology GmbH)

809 From European Compliance to National Approval: New Legal Requirements for Noise Barriers in Germany

During the last 20 years new in-situ methods have been developed for testing the acoustic properties of noise reducing devices (mainly noise barriers) under real-life condition (e.g. on-site, and not only in the laboratory). The main relevant properties are called sound reflexion and airborne sound insulation, and specific standards were developed by the European Committee for Standardisation CEN/TC 226/WG 6 and finally published with several updates during the last 5 years (e.g. EN 1793-5 and EN 1793-6 in der road sector and EN 16272-5 and EN 16272-6 in the railway sector). Nowadays these European standards are mainly used for type approval testing and certification of new and older products in the laboratory (mainly in the frame of the CE marking according to the standard EN 14388), for acoustic approval of newly built noise barriers on site and for long-term monitoring of installed noise barriers. The present paper discusses the current situation regarding the implementation of the standards mentioned before, highlighting the differences between European countries, mainly in Germany and Austria, with special focus on the acoustic approval and the long-term monitoring of noise barriers, both in the road and in the railway sector.

Speaker: Marco Conter (TÜV SÜD Rail GmbH)

810 Technically effective, socially acceptable – new approaches to highly absorbent noise barriers with a clear view

Noise protection measures are a central component of modern infrastructure planning; however, they are increasingly subject to a tension between acoustic effectiveness and societal acceptance. While conventional noise barriers meet the technical requirements for sound, they often give rise to conflicts with urban planning, landscape integration, and social demands. Massive structural designs impair visual connections, create barriers within urban spaces, and encounter public resistance, particularly in sensitive environments.The approach presented demonstrates that this conflict of objectives can be resolved through a new system architecture that combines transparency with highly absorbent and strongly reflection reducing acoustic properties. By employing a modular design and innovative technology whose resonance behavior can be specifically tuned to defined frequency ranges, acoustic performance can be adapted to different sound spectra. The solution fulfills the technical and regulatory requirements for both rail and road applications.This opens new design potential for noise reducing infrastructure that equally addresses functional, aesthetic, and societal requirements. Such systems contribute to reducing planning resistance, increasing public acceptance of large scale infrastructure projects, and enabling the socially compatible integration of noise protection measures into public space.

Speaker: Jan Lederer (DB Bahnbau Gruppe GmbH)

811 Integrated Photovoltaic Noise Barriers For Infrastructure Applications

Mitigating environmental noise from transport infrastructure remains a major challenge in densely populated areas. At the same time, the transition towards low-carbon energy systems requires large-scale deployment of solar photovoltaics (PV). In this context, Photovoltaic Noise Barriers (PVNBs) offer a multifunctional solution combining noise mitigation and renewable electricity generation within existing infrastructure corridors.This paper presents the development and assessment of innovative PVNB concepts within the European projects SEAMLESS and INCREASE. Both projects investigate the integration of photovoltaic modules into acoustic barriers through different design approaches. SEAMLESS focuses on optimizing conventional PV noise barriers, balancing acoustic performance, structural behavior, and energy yield. In contrast, INCREASE explores micro-perforated panel (MPP) solutions to enhance broadband sound absorption and reduce reflections associated with rigid PV surfaces.A multidisciplinary design methodology was adopted, combining acoustic, mechanical, and photovoltaic requirements. FEM simulations were used to optimize the proposed designs and assess the acoustic performance of the developed concepts. In addition, a full-scale SEAMLESS prototype was experimentally validated through in situ reflection and transmission measurements in a free-field environment using the Adrienne method. The results confirm the feasibility and potential of PVNBs as sustainable solutions for transportation noise mitigation while simultaneously contributing to renewable energy generation.

Speaker: Itziar Aspuru Soloaga (Fundación Tecnalia Research & Innovation)

Full paper (PDF)

812 High-Fidelity Wave-Based Simulation of Traffic Noise in Complex Urban Environments

Reliable prediction of environmental noise in urban areas remains challenging due to the combined effects of atmospheric variability, irregular terrain, buildings, and noise barriers. Most engineering models currently used in practice are based on ray-tracing or simplified energetic approaches, which may have limited accuracy at low frequencies and in geometrically complex environments.In this work, we present a high-fidelity three-dimensional simulation framework for outdoor sound propagation based on direct numerical solution of the acoustic wave equation. The numerical method employs Summation-By-Parts (SBP) finite-difference operators combined with the Simultaneous Approximation Term (SAT) technique for stable boundary treatment. The solver is implemented in CUDA and incorporates realistic terrain and atmospheric data, enabling efficient kilometer-scale simulations on GPUs.Particular focus is placed on low-frequency traffic noise propagation in urban environments containing buildings and noise barriers. Efficient reduced-order representations of barriers and buildings are introduced to significantly reduce computational cost while retaining the dominant low-frequency acoustic effects. Simulations are compared with predictions obtained using the Nord2000 model as implemented in SoundPLAN.The results demonstrate that wave-based simulations capture diffraction, interference, and shielding effects that are difficult to represent using traditional ray-based methods. In several complex urban configurations, substantial differences in predicted sound pressure levels are observed, particularly at low frequencies and in shadow regions behind buildings and barriers. The proposed framework provides a high-fidelity complement to existing engineering prediction tools for environmental noise assessment.

Speaker: Ken Mattsson (Uppsala University)

813 Evaluation of a Noise Barrier Project: Achieved Noise Reductions and Residents’ Perceived Benefit

Noise barriers are widely used to reduce road traffic noise, yet the relationship between calculated noise reduction and residents’ perceived benefit remains unclear. This paper presents a before/after evaluation of a noise barrier project comprising approx. 8 km of 6–9 m high barriers along the E45 (North Jutland Motorway) in Aalborg, affecting more than 2,000 dwellings. The evaluation combines noise calculations at dwellings (Lden), questionnaires collected before and after construction, and qualitative interviews with selected respondents. The results indicate an average noise reduction across all dwellings of approximately 4–5 dB. The most noise-exposed dwellings, with façade noise levels above 68 dB, experienced the largest effect, with average reductions of up to 9–12 dB. Survey results show, among other findings, that the share of respondents who are highly annoyed decreases from ~60% to ~42%, and the share considering the noise level unacceptable decreases from 57% to 38%. The barrier nonetheless divides local opinion: approx. 23% of residents report liking the barrier, while 30% report disliking it. A positive association is observed between calculated and perceived noise reduction, but the effect size appears small, suggesting that multiple factors beyond decibel reduction shape residents’ experience. One likely contributor is overly high expectations from residents. The expected noise reduction before the project averaged 6.8 on a 1–10 scale, whereas the perceived reduction after the project averaged 2.5 – underscoring the importance of clear communication about anticipated outcomes.

Speaker: Jakob Fryd (Danish Road Directorate)

Full paper (PDF)

A23.02 Modelling & Simulation Techniques: S390

Conveners: Giuseppe Petrone (University of Naples Federico II), Marcus Maeder (Technical University of Munich)

814 Listening to the polymer melt flow – vibroacoustic process analysis

Polymer melt flow in injection moulding is a black box phenomenon. During the process, the access to the flow field inside the steel tool is limited while high temperatures and pressures occur. A small number of sensors can be placed and flow simulations help understanding the process with numerous sensitive material and process parameters.To gain additional information, the injection moulding tool of the graded rib structures has been equipped with an accelerometer gaining structure borne sound data during the process. Data analysis and feature extraction will help to get additional information about the filling. The results are compared to sensor data and corresponding flow simulation results.The setup is implemented in the tool of the graded rib structures for local fibre migration. This process is sensitive to different material and process parameters. Prospects and limits of the vibroacoustic measurements will be shown in comparison with process and simulation data.

Speaker: Matthias Klaerner (TU Chemnitz)

815 Perceptual Impact of Particle Dampers on Electric Vehicle Powertrain Noise

Electric vehicle powertrains exhibit prominent tonal and structure borne noise components originating from power electronic modules and electromagnetic excitations. Although particle dampers have been widely investigated for vibration attenuation, their application in electric vehicle powertrains with respect to structure-borne acoustic radiation remains limited. To the authors’ knowledge, psychoacoustic investigations of particle dampers in electric vehicle powertrains have not yet been reported. This study investigates the integration of a particle damper within the lid of a power electronic module to mitigate acoustic and psychoacoustic parameters. A cavity integrated into the module lid was partially filled with rubber granulate and experimentally evaluated on an electric drive test bench. Acoustic measurements were conducted under representative operating conditions, and psychoacoustic metrics were calculated according to ECMA 418-2, including loudness, tonality, and fluctuation strength. Results demonstrate a broadband reduction in sound pressure level. More importantly, the particle damper significantly reduces time varying loudness and tonal prominence. Since both sensations adversely affect the sound quality, it is likely that the integration of a particle damper leads to measurable improvements in perceived acoustic quality. The findings highlight the potential of particle dampers as an effective passive solution for noise vibration and harshness optimization in electric powertrains.

Speaker: Braj Bhushan Prasad (Otto von Guericke University Magdeburg)

Full paper (PDF)

816 Nonlinear passive absorbers for low-frequency broadband attenuation: design rules and experimental validation

Noise mitigation is a challenging societal issue. To overcome the limitations of conventional sound-absorbing materials and classical Tuned Mass Oscillators for broadband low-frequency control, this work proposes the use of nonlinear absorbers, known as Nonlinear Energy Sinks (NES), with stiffness governed by a nonlinear law. The acousto-mechanical studied NES is made of a thin viscoelastic rubber membrane.When these systems are coupled to a primary system and excited at a sufficient level to activate their nonlinear behavior, an irreversible energy transfer occurs: energy is transferred from the primary system to the NES, where it is dissipated. These absorbers are limited by their high excitation threshold required for activation.This study focuses on using the NES to attenuate the first acoustic mode of a circular duct. A theoretical model is developed to optimize the energy transfer phenomenon and to lower the membrane’s activation threshold. The theoretical prediction model is compared to experimental data obtained through a vibro-acoustic testing bench for different membranes. Additionally, an analytical and experimental parametric study is carried out.

Speaker: Rita Moussa (Université de technologie de Compiègne)

817 Dynamic Opto-Mechanical Response of Optical Fibers: Modeling and Interferometric Measurements

Optical fibers used in high-precision interferometric systems are highly sensitive to external mechanical deformations. These environmental perturbations induce unwanted optical phase shifts, which severely limit the stability of the system. Therefore, accurately predicting this strain-induced phase shift is critical for designing vibration-insensitive optical components. This work investigates the opto-mechanical sensitivity of multilayered optical fibers. While the response of these fibers has been modeled and validated under static conditions, real-world environmental perturbations are inherently dynamic. Building upon this static framework, this study extends to the dynamic opto-mechanical behavior of the fiber. We present dynamic interferometric measurements conducted on a straight fiber subjected to controlled lateral mechanical excitations. To evaluate our predictive capabilities, the measured dynamic phase response is compared with predictions derived from a numerical modeling of the mechanical behavior coupled with an analytical modeling of the induced optical phase shift. This process is applied to series of interferometer configurations including a straight tense fiber transversely excited at one hand, a fiber glued to a 2D vibrated plate following different types of pathways, targeting an optimal configuration for which the optical phase shift is minimized.

Speaker: Eliott Breton (Laboratoire d'Acoustique de l'Université du Mans (LAUM))

A10.02 Metamaterials for noise and vibration reduction: applications and experimental methods: S061

Conveners: Jacques Cuenca (Test Division, Siemens Industry Software NV), Elena Shabalina

818 Parametric Investigation of Tesla Valve-inspired Acoustic Metamaterials for Tunable Low-frequency Sound Absorption in Cementitious Composites

Low-frequency noise remains difficult to mitigate in urban environments because conventional absorbers require large thicknesses or compromise structural performance. This study investigates a Tesla-valve-inspired acoustic metamaterial embedded within high-strength cementitious composites and evaluates how geometric parameters control acoustic behavior. A parametric investigation was conducted to examine the influence of coiling angle, channel length, valve count, and Helmholtz neck diameter on sound absorption in the 200–2500 Hz frequency range. Numerical simulations were performed using the thermoviscous acoustics interface in COMSOL Multiphysics, combined with the Johnson–Champoux–Allard (JCA) model, to capture viscous and thermal dissipation within the metamaterial channels and the surrounding mortar matrix. The results indicate that increasing the coiling angle increases tortuosity and the effective acoustic path length, shifting the resonance frequency from approximately 830 Hz to 420 Hz. Channel length enables near-linear tuning of the resonance position, whereas increasing the valve count introduces multiple standing-wave modes that broaden the absorption bandwidth. The Helmholtz neck diameter further tunes mid- and high-frequency absorption behavior. Experimental measurements using an in-situ PU-probe method confirm the simulation trends, with peak absorption coefficients reaching approximately 0.90–0.95. The findings provide practical design guidelines for integrating load-bearing cementitious acoustic metamaterials for tunable low-frequency sound control.

Speaker: Alemayehu Moges Kebede (unist)

Full paper (PDF)

819 Labyrinth Type Gossembrot-Inspired Metamaterial for Low- Frequency Sound Attenuation

Low-frequency sound insulation remains challenging due to the long sound wavelengths, making conventional acoustic materials largely ineffective at low frequencies. This study proposes a novel single-path, space-coiled metamaterial designed for efficient low-frequency sound attenuation. The metamaterial is a 28 mm thick cylinder with a 100 mm diameter, comprising a top cover with an input vent, a bottom cover with an output vent, and a central labyrinthine space-coiled region inspired by the Labyrinth Type Gossembrot. Acoustic waves propagate along a continuous, 1.26 m long path spiraling around an 22 mm diameter central core, with channel widths of 3–3.5 mm, significantly extending the effective acoustic path. Sound attenuation in the proposed structure is governed by Fabry–Pérot resonances along the single continuous path. Sound transmission loss (STL) was computed in COMSOL Multiphysics considering the impedance tube setup according to ASTM E2611-17, including thermoviscous effects. The metamaterial was fabricated using 3D printing, showing excellent agreement between numerical and experimental results. The first resonance occurs at 132 Hz, demonstrating a substantial shift of the resonance behavior toward the low-frequency range. The proposed metamaterial enables integration into building wall panels to reduce transmitted noise.

Speaker: Erfan Asgari (InnoRenew CoE, University of Primorska)

Full paper (PDF)

820 Design of a Phase Gradient Metamaterial with perforated foams

In recent years, phase gradient metamaterials have attracted significant attention for their ability to steer incident waves in specific directions. This behavior follows the generalized Snell’s law, which relates the directions of reflected and transmitted waves to the phase gradient of their coefficients along the surface. By imposing a specific linear phase profile, these materials can convert reflected and transmitted waves into surface waves, which can then be completely suppressed by adding an absorbing layer, making them promising for high-performance absorbers. In practice, such metamaterials consist of periodic arrays of discrete cells incorporating structures like porous layers, space-coiling geometries, or locally resonant elements, each designed to produce a targeted phase response. Recent studies aim to combine strong absorption with high sound transmission loss, requiring control over both reflection and transmission phase gradients. However, achieving these profiles often leads to bulky and impractical designs. This study proposes a metamaterial based on high tortuous perforated foam considered normally for outdoor use, resembling a double-porosity material. Its tunable perforation parameters, along with excellent low-frequency absorption, make it a strong candidate for generating the required phase gradients within a subwavelength structure. The paper outlines the full design process, from initial phase-control analysis demonstrating sufficient flexibility through perforation tuning to the optimization of the metamaterial. This includes selecting suitable foam properties and geometries to obtain a thin structure with high absorption and transmission loss. A sensitivity analysis is also performed to evaluate the impact of variations in phase and amplitude due to discrepancies between theoretical and actual material properties.

Speaker: Adrien Guthapfel (KU Leuven, Department of Mechanical Engineering)

Full paper (PDF)

A10.03/A23.06 Metamaterials and acoustic black holes in vibro-acoustics and air: S476

Conveners: Wonju Jeon, Florian Toth (Institute of Mechanics and Mechatronics), Jae Yeun Lee

821 Sonic Black Holes for Enhancing Detection of Weak Signals

Acoustic wave propagation inside sonic black holes (SBHs) is characterized by a decrease in the speed of sound and a gradual increase in pressure magnitude. Beyond their application as broadband sound absorbers, this paper explores their use as broadband pressure enhancement devices to improve the detection of weak signals by exploiting this gradual increase in pressure. We numerically and experimentally demonstrate an enhancement of up to 20 dB in signal detection over a broad frequency range by placing a microphone inside the SBH and comparing its performance with that of a microphone in a free field. This approach paves the way for passive acoustic wave amplification of weak signals, enabling low-power devices with enhanced detection capability.

Speaker: Semere B. Gebrekidan (Institute of Mechanics and Mechatronic)

Full paper (PDF)

822 A Scattering Model of a Two-dimensional Sonic Black Hole

Two-dimensional sonic black holes (2D SBHs) based on radially graded waveguide profiles have recently emerged as compact structures for wave focusing and energy trapping. In this work, we develop a semi-analytical scattering model that fully characterizes the total field response of under cylindrical- and plane-wave incidence. The SBH is treated as a metafluid with radially varying effective density and sound speed, both derived from the local power-law thickness profile. Inside the SBH, the acoustic pressure is expanded in angular Fourier series, and the radial problem is discretized using the Gaussian expansion method (GEM). By enforcing pressure and velocity continuity at the SBH boundary, the interior modal coefficients and the external scattering coefficients are obtained. The model is validated against finite-element simulations, showing excellent agreement. From the computed modal scattering amplitudes we extract the total scattering cross section, the differential scattering cross section, and the backscattering strength. Results demonstrate that the 2D SBH can significantly suppress the reflected wave over a broad frequency range. This work provides a rigorous and efficient framework to predict and interpret the scattering signature of 2D SBHs, and highlights their potential for acoustic stealth, reflection reduction, and wave manipulation in planar environments.

Speaker: Jie Deng (La Salle, Universitat Ramon Llull)

823 Increasing damping in composite panels with locally graded acoustic black holes

In aerospace engineering, primary structural components such as fuselage skins and engine cowlings must withstand rigorous dynamic loads and high-frequency vibrations while maintaining high thermal and mechanical stability. Traditional vibration mitigation often relies on auxiliary damping treatments that add parasitic mass, a critical disadvantage in weight-sensitive airframe design. This research proposes a novel approach using functionally graded panels composed of carbon fiber-reinforced epoxy resin to achieve intrinsic passive vibration control. The methodology includes the acoustic black hole effect by locally tailoring the fiber volume fraction rather than utilizing traditional geometric tapering. Therewith, a localized gradient in flexural stiffness is created, facilitating the focalization and subsequent dissipation of mechanical energy. Crucially, this material-based gradient maintains a uniform specimen thickness, thereby preserving the structural robustness and thermal shielding capabilities required for aerospace service. The work opens with the description of the production process using a resin transfer moulding. Vibroacoustic investigations are carried out on beam and plate specimens with a laser-scanning vibrometer to detect mode shapes on the one hand and the related damping function.

Speaker: Jan Boysen (Graz University of Technology)

824 Optimizing passive acoustic filters in tubes using a tracking formulation

Acoustic topology optimization is a broad topic, with one sub-group being structural optimization. In this discipline, we aim to distribute soundhard material to achieve specific acoustic properties, primarily through a density-based approach. One typical example is the transition zone between two waveguides, soundhard structures that restrict wave propagation and often enforce a unidirectional field.In this work, we assume that the cross-sections of both waveguides are equal while designing our soundhard structures to achieve a given broadband transmission frequency response. Using lowpass, highpass, bandpass, and bandstop target functions, we synthesize one design for each transmission profile. To avoid modelling thermoviscous losses and ensure manufacturability with standard 3D printing methods, we use a robust filter approach that combines erode, center and dilate filters in a min/max optimization problem.The obtained designs mostly matched the target response within the specified range. The inherent property of structural filters to always transmit low-frequency waves made it hard for the optimizer to converge for highpass and bandpass designs. Lowpass and bandstop filters performed very well, even below the optimization's frequency range. The initial studies showed promising results, motivating future research and parameter studies to further solidify the understanding of the filter devices' working principles.

Speaker: Felix Huber (Institute of Mechanics and Mechatronic)

Full paper (PDF)

A17.04/A24.07 Spatial Audio Signal Processing: S400

Conveners: Jens Ahrens (Chalmers University of Technology), Franz Zotter (University of Music and Performing Arts), Orchisama Das, Jung-Woo Choi (KAIST), Antonio Figueroa-Duran (Universidad Politécnica de Madrid)

825 Spatially Adaptive Neural Signal Processing for Binaural Rendering and Crosstalk Cancellation in Personal Sound Zones

This paper presents the Binaural Spatially Adaptive Neural Network (BSANN), a spatially adaptive neural signal-processing framework for loudspeaker-based binaural rendering in personal sound zones (PSZs). BSANN generates ear-wise stereo loudspeaker filters directly from listener coordinates, enabling position-dependent updates without per-pose matrix inversions and independent control of the acoustic field around the left and right ears of multiple listeners. This ear-wise formulation improves performance balance across listeners under asymmetric room conditions and also provides passive interaural separation. To improve practical realism without requiring in-situ room measurements, the network is trained using physically informed acoustic transfer functions that combine simulated room responses with measured anechoic loudspeaker responses, analytic piston directivity, and rigid-sphere head scattering. A second-stage active crosstalk cancellation (XTC) refinement further suppresses contralateral leakage while preserving sound-zone isolation through teacher-anchored adaptation. The method is evaluated in a fixed two-listener setup using a 24-channel loudspeaker array and in-ear measurements in a reverberant environment. Over 100 Hz to 20 kHz, the full framework achieves log-frequency-weighted inter-zone isolation (IZI) of 10.23/10.03 dB, inter-program isolation (IPI) of 11.11/9.16 dB, and XTC of 10.55/11.13 dB for the two listeners. These results indicate that coordinate-conditioned neural signal processing can effectively unify multizone reproduction and loudspeaker-based binaural rendering in practical spatial-audio systems.

Speaker: Hao Jiang (Princeton University)

826 Analysis of Plant Matrix Perturbations on Crosstalk Cancellation Performance Based on Singular Value Decomposition

This work presents a formulation for predicting Crosstalk Cancellation (CTC) performance in the presence of stochastic additive perturbations to the plant matrix. The reproduction of binaural signals over loudspeakers using CTC is highly sensitive to plant modelling errors, arising from simplified models (e.g., neglecting loudspeaker directivity or head scattering) as well as unpredictable factors such as environmental changes or listener–loudspeaker misalignment. These effects introduce a mismatch between the assumed and actual plant matrix, degrading the accuracy of the reproduced binaural signals and the perceived spatial audio. Regularisation provides a means to stabilise the inverse solution, but it may reduce the achievable CTC performance in ideal conditions. This work derives a frequency-domain formulation, based on Singular Value Decomposition, to predict both the expected loss in CTC performance and the maximum achievable CTC level under stochastic additive perturbations when filters are designed using Tikhonov regularisation. Numerical simulations demonstrate the accuracy of the proposed approach and its validity under realistic perturbation conditions.

Speaker: Francesco Veronesi (University of Southampton)

Full paper (PDF)

827 On the Relationship Between Tikhonov Regularization and Butterworth Filters in Radial Filters of Spherical Microphone Arrays

Radial filters are required to transform the spherical harmonic coefficients of the pressure measured by a rigid-sphere microphone array into those of the plane-wave density representation of the measured sound field. Ideal radial filters exhibit poles at the origin, resulting in strong low-frequency amplification and numerical instability. To address this issue, regularization schemes are commonly applied, the most widely used being Tikhonov regularisation. It is shown that, at low frequencies, the Tikhonov regularisation filter for the n-th spherical harmonic order is equivalent to a zero-phase system obtained by applying an n-th order Butterworth high-pass filter in forward and reverse directions. Furthermore, explicit relationships are derived between the regularization parameter β, the corresponding Butterworth cut-off frequency, and the maximum magnitude of the resulting radial filter.

Speaker: Filippo Maria Fazi (Institute of Sound and Vibration Research)

828 Implementation of the Cardioid Equatorial Microphone Array Based on Omnidirectional Microphones

Cardioid equatorial microphone arrays are baffleless ambisonic capture devices that use outward facing cardioid microphones arranged on a circle. The microphone mounts are straightforward to manufacture, and the compact form factor produces minimal visual disturbance. But low-cost cardioid microphones with low mismatch are not always readily available. We present an implementation that uses pairs of omnidirectional microphones to create virtual cardioid sensors. This solution poses challenges at low frequencies. Our implementation therefore uses omnidirectional sensors at low frequencies and the virtual cardioid sensors at mid- and high frequencies

Speaker: Jens Ahrens (Chalmers University of Technology)

Full paper (PDF)

A18.04 Indoor soundscape: S374

Conveners: Simone Torresin (University of Trento, Dept. of Civil, Env., Mech. Eng.), Papataya Nur Dokmeci Yorukoglu

829 Investigating Perception of Children in Primary School Classrooms: A Thematic Analysis

Classroom acoustics research has traditionally focused on objective acoustic measurements and speech intelligibility, while comparatively limited attention has been given to the subjective experience of the children in the classroom sound environments. This study explores primary school children’s perceptions through a qualitative analysis and examines how these experiences relate to the conceptual framework proposed in ISO 12913-1 for soundscape research. Data were collected through focus group interviews with 22 students (ages 6-12) in Pakistan using semi-structured open-ended questions. The interviews were analyzed using a thematic analysis, conducted through a theoretically informed, deductive approach guided by concepts from ISO 12913-1 and the study’s research questions, combined with the inductive identification of patterns emerging from the children’s responses. The analysis revealed several key themes related to children’s experiences of the classroom acoustic environment, which were connected within a conceptual framework, including the interpretation of auditory sensation, affective and physiological responses, behavioral adaptation, and teacher-led interventions. Children also described characteristics of an ideal classroom sound environment that supports concentration and well-being. The findings provide an opportunity to compare children’s perspectives with previously proposed frameworks based on adult viewpoints, hence contributing to a more user-centered understanding of classroom acoustic environments.

Speaker: Shomaila Syed (University of Ferrara)

Full paper (PDF)

830 From Framework to System: Caregiver Driven Soundscape Augmentation for Indoor Dementia Care Environments

Building on a Functional Block Diagram framework previously introduced to link dementia-specific auditory processing deficits to soundscape design principles. This contribution reports on the translation of that framework into the architecture of an indoor soundscape augmentation system. The system is designed around caregiver workflows in dementia care units. Prior deployments of the soundscape augmentation platform in nursing home resident rooms demonstrated the feasibility of closed-loop soundscape delivery. Yet these deployments highlighted that caregivers (nurses, personal support workers, and families) remain the principal mediators of residents’ sonic environments during care activities. They are also the most reliable observers of non-verbal behavioural responses. This study revisits the platform’s sound selection and scheduling logic through the Functional Block Diagram’s deficit-to-intervention pathways. It pays particular attention to the care activities most affected by auditory disruption, including morning care, bathing, and evening routines. The contribution discusses system-level features that support caregiver observation, feedback, and adjustment. It recognizes caregivers as both end-users of the indoor soundscape and co-producers of its effectiveness. Sound interventions are most effective when integrated into existing dementia care routines rather than implemented as separate additions. This work examines how soundscape augmentation can be embedded within the daily operations of care practice.

Speaker: Arezoo Talebzadeh (Ghent University)

831 Home Soundscapes: Perceived Control Under the Interaction Effect with Thermal Stimuli

Climate change led to increased temperatures during summer. Many homes in Europe and India are naturally ventilated. Opening or closing windows is one of the perceived controls to adapt to acoustic and/or thermal environments. When occupants open windows to cool indoor temperatures, they may also be exposed to sounds from outside, which might be unwanted sounds like traffic. Occupants react differently: some tolerate the noise to maintain thermal comfort, while others close the window to maintain acoustic well-being. Influences on acoustic well-being including multi-domain effects have not been extensively researched, primarily because soundscape data collection methods have mainly focused on the acoustic environment. We must shift our focus from single-domain effects to multi-domain effects. Therefore, this research aims to explore home soundscapes considering the interactions with thermal stimuli as another environmental factor. This will be done by surveying home occupants in Germany and India to cover different cultural and climate zone contexts using an affective response questionnaire for acoustic well being to examine perceived control over indoor soundscapes during work, relaxation, and sleep under combined sound and temperature conditions. The results will contribute developing adaptive models of acoustic well being that reflect the interplay of thermal and acoustic environments.

Speaker: Amneh Hamida (RWTH Aachen University)

832 Crossing boundaries: the role of architectural acoustics in shaping the design process

Architectural objectives fall into three main categories: structural performance, physical environmental performance, and aesthetic and cultural performance. The acoustic performance relies in the second category with an important, yet not fully explored, impact on the other two. To further push the boundaries between categories in the design workflow, research on acoustic-performance based design has shown the benefits of the acoustic knowledge and quantifiable measures used to inform the design process from its earliest phases.This plenary aims to highlight how this approach strengthens the aesthetic and qualitative objectives leading to a more human-centric design. Moreover, it explores how the integration of acoustics with other environmental factors (e.g., thermal, lighting, air quality) that are conventionally considered more “mainstream”, improves multidomain decisionmaking during conceptual design, ensuring that quantifiable criteria are integrated from the outset through the systematic use of performance feedback.This talk will also be an opportunity to highlight the limitations of the current cultural and technical approaches by presenting some of the acoustic integration challenges in the design of acoustic environments through examples at different building scales. The aim is to illustrate possible future directions of research in architectural acoustics of everyday life spaces, which still presents important gaps to fill in within a multidisciplinary integration framework in the current architectural practice and architectural acoustic education in the context of growing sustainability demand and human-centric approach.

Speaker: Louena Shtrepi (Politecnico di Torino)

13:00 Lunch break

A03.04 Acoustics of wooden buildings: S394

Conveners: Jean-Luc Kouyoumji (FCBA Institute of Technology), Heinz Ferk

833 Fast three-dimensional analysis of the airborne and impact sound insulation of finite-sized cross-laminated timber panels

Cross-laminated timber (CLT) panels are gaining popularity as load-bearing wall and floor elements due to their low surface mass, potential for sustainable sourcing, and speed of installation on site. However, accurately and efficiently predicting their airborne and impact sound insulation remains challenging. This difficulty arises from their complex vibroacoustic behavior, driven by strongly orthotropic layers, significant shear deformation, thickness resonances, and low impact impedance. To reduce computational effort, the vibroacoustic response is often approximated using an equivalent homogeneous orthotropic plate or solid panel, with elastic properties derived from measurements or a rule-of-mixtures approach. Such models, however, are typically semi-empirical and lose accuracy at higher frequencies. In this work, an alternative prediction method is proposed. The three-dimensional elastic deformation of each orthotropic layer is modeled exactly in the frequency–wavenumber domain, yet spatial discretization is avoided, such that the method remains computationally efficient. Finite-size effects are accounted for through equivalent anti-symmetric periodic loading and a baffled radiation analysis. For impact sound insulation, the floor’s point impedance is explicitly included. A series of validation examples illustrate that, with this method, accurate airborne and impact sound insulation predictions can be attained when directly employing the elastic properties of the constituent wood.

Speaker: Edwin P.B. Reynders (KU Leuven, Department of Civil Engineering)

834 Challenges in modelling plate contact conditions in cross- laminated timber junctions: a numerical-experimental investigation

Existing numerical models reliably predict flanking sound transmission in cross-laminated timber (CLT) junctions when the stiffness of the connection approaches either the ideally rigid condition or that of an elastic layer interposed between plates. Nevertheless, it is still challenging to deterministically define coupling stiffness values for modelling the direct mechanical contact at the interface between plates. In this contribution, a numerical-experimental investigation is presented with the aim of highlighting fundamental issues for the determination of such coefficients. The vibration reduction index Kij was measured for two configurations of a CLT L-junction mock-up. In the first case, the plates were in direct contact and connected by angle brackets. In the second, they were left in contact with each other, but without any fasteners. The comparison of finite element method simulations against experiments highlighted that the plates' coupling stiffness conditions are influenced by the simultaneous action of both the connectors and the static load on the junction. This is consistent with findings from previous studies on the effects of junction loading on flanking sound transmission. Implications of this study include the necessity of deriving an explicit relationship between the constraint/load conditions of the junction and the coupling stiffness coefficients at the plates' interface.

Speaker: Antonio Esposito (Empa)

Full paper (PDF)

835 Raw earth brick walls and wood elements - Junctions characterization and sound insulation

The building sector is confronted with two key challenges: minimizing the environmental impacts while ensuring occupants well-being. In this context, the use of sustainable materials with high thermal and acoustical performances appears as a necessity. Therefore, raw earth appears as strategic low-carbon material, as it is naturally abundant, locally available and renewable. It also offers interesting thermal properties to improve summer comfort, thus impacting the energy efficiency and comfort of buildings. Raw earth construction often includes wood elements as structural components.The main objective of the CarAc’Terre project is to remove regulatory and normative obstacles to raw earth construction techniques relative to acoustics. A multi-scale methodology for characterizing the acoustic performance of raw earth construction systems is proposed using both a performance-based and a perceptual approach. A small-size building was constructed out of raw earth bricks in order to characterize vibration reduction index of junctions. The effect of the presence of wood post at the junction was examined, as well as the presence of wood-based cladding to limit acoustic leakage at wood-brick interface. Furthermore, CLT panels covered the construction and the associated junctions were also evaluated. Some rooms were closed using a wood-based stud walls in order to measure airborne sound insulation performance. This paper describes the small-size building, the junction characterization measurement results as well as measured and predicted sound insulation.

Speaker: Catherine Guigou-Carter (CSTB)

Full paper (PDF)

836 Rain Sound Insulation of Timber Roofs — Influence of Roof Cladding, PV Panels and Roof Windows

Laboratory measurements in accordance with EN ISO 10140-5:2021 are used to analyse the rain sound insulation of timber roof constructions with varying structural configurations. Rain induced sound intensity levels are examined for exposed-rafter roofs and timber frame roofs with aluminium sheet, zinc sheet, and two types of tile cladding (sliding tiles and beaver-tail tiles). The influence of additional underlays beneath the metal cladding — a bituminous membrane and a structured mat — is investigated. The effects of PV elements on tile- and aluminium-sheet-clad roofs are also examined, as is the influence of a roof window with and without a roller blind.

Speaker: Bernd Nusser (Holzforschung Austria)

Full paper (PDF)

A03.05 Structure-borne sound and noise from building equipment: S258

Conveners: Berndt Zeitler (Hochschule für Technik Stuttgart), Jochen Scheck (HFT Stuttgart), Andreas Mayr (Rosenheim Technical University of Applied Sciences)

837 Modal Transfer Matrix Modeling of Multilayered Plates under Point Force Excitation in Building Acoustics

Compromises among cost efficiency, sustainability, and occupant comfort require efficient prediction concepts in building acoustics. For structure-borne sound problems, the driving-point mobility at the excitation position is a key quantity for estimating the power introduced into the structure by the excitation source.Previous studies have shown that approximating the driving-point mobility by a constant characteristic mobility can lead to substantial deviations, particularly at low frequencies, near boundaries, and for layered building components. For such components, the Transfer Matrix Method is a well-established approach for modeling sound transmission. Additional modal extensions allow the finite dimensions and simply supported boundary conditions of rectangular plates to be included.The present work applies the modal Transfer Matrix Method (mTMM) to evaluate the driving-point mobility of a multilayered, simply supported plate subjected to a point force acting on the top layer, thereby representing a structure-borne excitation.For the investigated reference cases, the proposed mTMM approach shows agreement with conventional finite element simulations, while validation against experimental results is discussed. The approach retains the computational efficiency and low modeling effort previously demonstrated for the mTMM compared with full finite element models.However, the restriction to simply supported plates limits the method's applicability to practical building structures. To address this limitation while preserving computational efficiency, a truncated sine-basis extension is proposed for alternative boundary conditions. This extension enables elastically restrained boundaries to be represented within the modal Transfer Matrix framework.

Speaker: Korbinian Schwab (Technical University of Munich, Chair of Building Physics)

Full paper (PDF)

838 Characterisation of Sanitary Installations as Sound Sources – what do we have, what do we need?

Sanitary installations are among the most disruptive sources of noise in buildings. At the same time, they are some of the most complex examples of building service equipment since they usually consist of several active and passive components such as toilets, washbasins, taps, valves, pipes and pipe clamps etc. and complex interactions. To design systems that meet target values given as requirements in national building codes, a prediction based on EN 12354-5 is required. This in turn requires laboratory measurement methods that provide input data for the prediction. The standard EN 14366-1, which covers wastewater installations, was recently revised. The development of a future standard EN 14366-2 covering freshwater installations is underway in a CEN working group. Both standards refer to the general methods for the characterisation of building service equipment in EN 15657, which are reviewed in a companion paper by Höller et al. A significant challenge in developing EN 14366-2 is the role of appliances such as taps or valves that have so far been targeted by EN ISO 3822, but not in a way that enables predictions. This contribution provides an overview of the current situation regarding the acoustical laboratory characterisation of sanitary installations and identifies needs for future research and standardization work.

Speaker: Jochen Scheck (HFT Stuttgart)

Full paper (PDF)

839 Experimental Study on Noise Reduction of Drainage Pipes using Acoustic Materials: Comparison between Simplified Laboratory and In-situ Measurements

As high-density residential living becomes a global trend, managing bathroom drainage noise from prevalent under-slab piping is vital for indoor acoustic comfort. This study investigates the noise reduction effects of various drainage pipe materials and structural acoustic claddings to mitigate noise at the source. To evaluate performance efficiently during development, a simplified laboratory measurement methodology is proposed. Unlike complex full-scale tests like EN 14366-1, this method evaluates a straight pipe point-supported at both ends using a pink noise source, with a sound intensity probe positioned at a 100 mm distance following the ISO 9614 series. These laboratory results were compared against in-situ maximum noise levels. In-situ refers to field tests in a full-scale residential mock-up building, measuring noise in the receiving room below during actual toilet flushes according to the KS F 2871 standard. The analysis indicated that while laboratory noise reduction tended to be slightly lower than in-situ results, overall noise patterns across test cases remained consistent. Similar trends were observed in the 50 Hz to 80 Hz and 200 Hz to 1600 Hz ranges, with moving average trend lines closely aligned. Furthermore, tests on seven additional acoustic materials revealed that combining damping materials and sound absorbers yielded the most significant noise reduction across eleven measured frequency bands. These findings suggest the proposed simplified method is an effective screening tool for evaluating low-noise pipes and claddings, serving as a reliable basis for material selection prior to field implementation.

Speaker: So-young Kim (Korea Conformity Laboratories)

Full paper (PDF)

840 Acoustic performance of wastewater installation systems in heavy and lightweight buildings considering source- receiver mobility interaction

Drainpipe noise can be a prominent annoyance and privacy problem in modern housing. Heavy pipe systems with low vibration mobility are standard practice for achieving acoustic comfort in solid construction buildings. However, it is not clear whether the vibro‑acoustic performance of the pipe systems in massive constructions can be transferred to lightweight construction. This study examines the influence of source-receiver mobility interaction on drainpipe noise transmission. Laboratory measurements according to EN 14366-1 were performed for both a heavy, low mobility pipe system and a lighter standard system, each mounted on both heavy and lightweight wall types. The measurements are complemented by prediction calculations based on EN 12354-5, which enable a systematic variation of the building construction. For the investigated configurations, the results indicate that the acoustic performance of the pipe systems is influenced by the source-receiver mobility interaction and the building transfer functions. On the heavy wall, the two pipe systems showed comparable noise levels, and hence, similar single number quantity (SNQ) values. In contrast, on the lightweight constructions, the comparable performance of the two pipes was not preserved: the lower-mobility heavy pipe showed clearly lower SNQ in one lightweight construction, while the lighter standard pipe showed lower SNQ in the other lightweight construction. These results indicate that the acoustic performance of wastewater systems determined in one construction is not directly applicable to other constructions: the performance depends on the combined effect of the source-receiver mobility interaction and the building transfer function.

Speaker: Sven Öhler (Fraunhofer Institute, IBP)

Full paper (PDF)

841 Experimental Verification of Simulated Blocked Forces in Wastewater System using the Empirical Model of Water‑Flow Excitation

Structure‑borne sound generated by wastewater pipesystems often dominates noise emissions from sanitaryinstallations in buildings. This noise is mainly caused byturbulent annular flow and by water impacts on inlets, tees,and bends. The water excitation, and thus the resultingstructural response, is treated as a stochastic process anddescribed by its power spectral density (PSD). A suitablemathematical model of PSD for the individual excitationsub‑sources of the water‑induced loads on a wastewaterpipe is developed in a systematic, stepwise manner bycombining experimental data with a numerical model. Bothexperimental and numerical configurations comply withEN 14366-1:2023 [1], with the vibration response quantifiedby the blocked forces at the wall fixing points. The turbulentflow is represented by a linear function of frequency indouble logarithmic scale, whereas the water impact at thebasement bend is modelled by a Gaussian‑shaped function.The parameters of PSD are identified by minimizing thedifference between measured and simulated blockedforces. The quantified excitation terms are incorporatedinto the numerical models, and the simulated blockedforces are compared with the measurements. The overallagreement is satisfactory and supports the applicability ofthe proposed method for predicting structure‑borne noisefrom wastewater systems.

Speaker: Yohko Aoki (Fraunhofer Institute, IBP)

Full paper (PDF)

A10.02 Metamaterials for noise and vibration reduction: applications and experimental methods: S516

Conveners: Jacques Cuenca (Test Division, Siemens Industry Software NV), Elena Shabalina

842 Molded acoustic metamaterial for automotive door noise reduction: experimental validation in real vehicle

Road noise transmitted through vehicle door structures remains a critical challenge in automotive noise, vibration, and harshness (NVH) performance, particularly in the low-frequency range below 1 kHz. This study presents a mass-producible acoustic metamaterial designed for integration into automotive door trims to reduce structure-borne sound transmission without increasing weight or manufacturing complexity.The proposed design consists of a single-material, periodically patterned plate fabricated using a one-step molding process, ensuring compatibility with industrial production. The periodic protrusion geometry is designed to induce a flexural wave bandgap in the sub-1000 Hz range, targeting the dominant frequency components of road noise. Dispersion analysis confirms the formation of a polarization-specific bandgap that suppresses out-of-plane wave propagation, which is the primary contributor to sound radiation.Experimental validation demonstrates that the molded metamaterial plate improves sound transmission loss (STL) by approximately 2 dB in the 400–800 Hz range compared to a flat plate of equivalent mass. When implemented in an automotive door trim, frequency response measurements show a reduction in vibro-acoustic transfer from the door to the cabin. Furthermore, on-road driving tests reveal a reduction in interior sound pressure levels by approximately 1.4 dB at 40 km/h, where road noise is dominant.These results demonstrate that simple, monolithically molded acoustic metamaterials can effectively enhance vibro-acoustic performance in practical vehicle applications. The proposed approach provides a scalable and cost-effective pathway for integrating metamaterial-based noise reduction into mass-produced automotive components.

Speaker: Sunao Tomita (Toyota Central R&D Labs., Inc.)

843 Metamaterial design and optimization methods for a low-height noise barrier for railways

As part of the European LIFE SILENT project, which addresses noise mitigation in complex environments, this study focuses on improving the acoustic performance of low-height noise barriers (LHNBs) by optimizing their surface acoustic impedance. Specifically, it presents the inverse design of a broadband sound-absorbing metamaterial layer intended for the railway-facing side of an LHNB. The proposed methodology utilizes the optimization and numerical modeling of a combination of Neck Embedded Helmholtz Resonators (NEHR) and Fabry-Pérot (FP) channels. For the design of the unit elements composing the barrier's surface, the target sound absorption coefficient is specifically calibrated to the railway noise spectrum. The performance of this design was validated through laboratory testing of 3D-printed prototypes, demonstrating substantial potential for improved environmental noise reduction. Finally, the study highlights the practical manufacturing challenges associated with large-scale production for real-world deployment.

Speaker: Domenico De Salvio (University of Bologna)

844 Design and Modelling of an Acoustic Metasurface for Low- frequency Room Mode Absorption

Low-frequency room modes remain a major challenge in architectural acoustics, as they cannot be effectively mitigated using conventional absorbers or resonators without a significant thickness. As an alternative, acoustic metamaterials offer the possibility of achieving efficient low-frequency absorption with limited thickness. In this work, a numerical methodology for the design and optimisation of a spiral metasurface aimed at room mode control is presented. A finite element model of a rectangular enclosure is implemented in COMSOL Multiphysics and adjusted using measurements. Based on the modal characteristics of the room, a spiral labyrinth structure is designed and tuned to target a specific low-frequency mode. Analytical and numerical simulations are conducted to evaluate the absorption performance of the spiral metasurface and to optimise both its internal geometry and its spatial distribution within the enclosure. The selected geometries are then 3D printed and their performance is assessed experimentally using an impedance tube. The good agreement between numerical simulations and measurements exhibits the potential of this approach as a tool for acoustic metamaterial design for room-mode control. The observed amplitude reduction of the selected room mode shows that metasurfaces based on labyrinth structures are an effective and scalable solution for low-frequency modal control in enclosures, overcoming the limitationsof traditional absorptive materials.

Speaker: Juan Luis de la Torre Moral (Grupo Acústica Arquitectónica, Universidad Politécnica Madrid)

Full paper (PDF)

A12.11/A20.06 Fluid-structure-acoustic interaction in voice and speech generation: S087

Convener: Stefan Kniesburges (HNO-Klinik, UK Erlangen)

845 Design and Validation of a Cyber-Physical Laryngeal Model of Vocal Fold Posturing

Vocal fold (VF) posturing influences the fluid-structure–acoustic interactions that govern human phonation. However, the influence of VF posturing on phonotraumatic vocal hyperfunction is still unclear. This deficiency arises, in part, because current experimental modeling approaches rely on simplified VF and laryngeal geometries with limited degrees of freedom. Consequently, current approaches are unable to replicate different glottal orientations (pressed, breathy, etc.) that can arise due to abnormal VF posturing. In response, this work proposes a physiologically inspired, cyber-physical laryngeal model with three degrees-of-freedom that mimics VF posturing in the transverse glottal plane. Multi-layered synthetic VFs are cast inside a laryngeal structure and attached anteriorly and medially to a representative thyroid cartilage, and posteriorly to moveable arytenoid structures. The silicone VFs include a hollow cylindrical channel along the anterior-posterior direction that can be inflated to mimic VF thickening due to the thyroarytenoid muscle. The positioning and orientation of the arytenoid structure is physically activated with a three degree-of-freedom positioning system that controls medial-lateral and anterior-posterior displacement, and rotation about the cricoarytenoid joint. Kinematics are derived from a numerical model of laryngeal muscle activation. Validation is performed by comparing the model speech parameters (e.g., phonation threshold pressure, flowrate, fundamental frequency, and sound pressure level) with relevant clinical and computational studies, as a function of VF posturing. Good agreement in the dependency of aerodynamic, kinematic, and acoustic parameters on VF posturing is found. This novel approach provides the ability to systematically link VF posturing to phonatory outcomes via experiment, opening a new avenue for exploring the mechanics of phonation in a benchtop facility.

Speaker: Ishtiaq Ahmed (Rochester Institute of Technology)

846 Aeroacoustic Analysis of Prescribed Vocal Fold Motion in the SimVoice Benchmark using Wall-Modelled Large-Eddy Simulation

Voice disorders affect millions of people across Europe, with significant implications for quality of life. Understanding the physical mechanisms underlying phonation and their variation under disturbances remains a central challenge. While the fundamental principles of voice production are well established, accurately modelling the complex fluid–structure–acoustic interactions between the vocal folds and the laryngeal airflow remains challenging. To foster reproducibility and cross-validation within the phonation research community, the SimVoice benchmark was established as a shared reference for both numerical and experimental investigations.In this study, complementary numerical and experimental analyses of the benchmark configuration are performed to investigate its aeroacoustic characteristics and further support methodological development. A computationally efficient framework based on the MGLET solver is employed to conduct Large-Eddy Simulations (LES) with prescribed vocal fold motion, permitting parameter studies of subglottal pressure variations and deformation patterns. Additional measurements using both rigid (non-oscillating) and deformable vocal folds provide a robust extension of the dataset for model calibration and parameter tuning. Exploring multiple operating points enhances the benchmark's characterization, advancing the broader goal of reproducible, physically grounded voice simulation. Our findings give new insights into the physics of voice production and contribute to the continued validation of the SimVoice benchmark.

Speaker: Thomas Brunner (IGTE, TU Graz)

Full paper (PDF)

847 Aeroacoustic Simulation of Soft Phonation Employing a Kinematic Model of Mucosal Waves

This work presents a hybrid computational framework for the three-dimensional simulation of human phonation, combining prescribed vocal fold motion, computational fluid dynamics, and aeroacoustic wave propagation. The adopted kinematic mucosal wave model is derived from laboratory experiments of soft phonation, and its kinematics are imposed as dynamic boundary conditions in the incompressible Navier–Stokes solver. OpenFOAM is employed due to its high performance, open-source availability, and extensive capabilities, including advanced turbulence modeling and dynamic mesh functionality. The aeroacoustic model is based on the aeroacoustic wave equation, with preprocessing consisting of the conservative projection of sound sources from the CFD mesh onto the acoustic mesh, followed by wave propagation simulated with the program openCFS.This approach enables controlled investigation of flow-induced sound generation while providing significant computational savings compared to fully coupled fluid–structure interaction simulations. The influence of flow dynamics on acoustic source terms is analyzed, and numerical simulations of sound propagation in a simplified vocal tract geometry are presented.

Speaker: Jan Valášek (Institute of Mathematics, Czech Academ)

Full paper (PDF)

848 Experimental and simulation based investigation of mucosal waves

A recently started research project brings together experienced research teams from Czechia, Germany and Austria to expand the knowledge on the nature of the mucosal wave which is important for (1) improving the accuracy of the mathematical and physical models of voice production, and for (2) more insightful diagnosis of voice disorders. Based on experimental observations, kinematic laws will be derived to inform and drive mathematical models of VF oscillation, enabling them to accurately reproduce the observed mucosal behaviour. These models will subsequently be applied to investigate the aeroacoustic mechanisms underlying voice production. Furthermore, the ex vivo characterization of MW kinematics will guide the optimization of geometric and structural parameters of a novel self-oscillating silicone VF model incorporating an embedded liquid layer, designed to better replicate physiological conditions. In parallel, computational models of self-sustained VF oscillations incorporating fluid-structure-acoustic interaction will be refined to improve their ability to simulate MW dynamics. The presentation will include an overview of the research project and will highlight first results towards the numerical simulation of the generation of human phonation taking into account the full fluid-structure-acoustic interactions.

Speaker: Manfred Kaltenbacher (TU Graz)

849 Impact of Dental Morphology on the Aeroacoustic Production of the Fricative [s]: A Numerical Study Comparing Neanderthal and Modern Human Dentition

The sibilant fricative [s] is produced by a turbulent jet passing through a constriction formed by the tongue and the hard palate. Upon impinging on the upper incisors, this jet induces an acoustic dipole at the tooth edge, which constitutes the primary source of the frication noise characteristic of [s] (Stevens, 1971).In this study, we investigate the physical and aeroacoustic principles governing the production of the fricative [s], with a specific focus on a case study approximating Neanderthal oral anatomy. In this anatomical configuration, the vocal tract and tongue exhibit distinct morphological differences compared to those of modern humans (Alvarez, 2024). We specifically examine the potential role of shovel-shaped incisors, a distinctive dental trait associated with Neanderthals.Based on fossil data from Neanderthal specimens, a parametrized model of a shovel-shaped incisor is developed. This dental geometry is integrated into a simplified vocal tract model adapted from (Yoshinaga et al., 2018). Additional geometries are then generated by progressively flattening the incisor curvature to approximate modern human dentition.The aeroacoustic sources are determined using a hybrid computational approach. First, an incompressible Large Eddy Simulation (LES) resolves the turbulent flow field, including velocity and pressure distributions in the vocal tract. These results then serve as input for the second step, where the Perturbed Convective Wave Equation (PCWE) is solved using the open-source solver OpenCFS. This two-step methodology ultimately provides the acoustic pressure field as output.This comparative approach finally enables to quantify how variations in dental morphology influence both the turbulent airflow patterns and the resulting acoustic radiation.

Speaker: Honorine Bertrand (Sorbonne Université, CNRS, Institut Jean Le Rond d'Alembert)

A17.04/A24.07 Spatial Audio Signal Processing: S401

Conveners: Jens Ahrens (Chalmers University of Technology), Franz Zotter (University of Music and Performing Arts), Orchisama Das, Jung-Woo Choi (KAIST), Antonio Figueroa-Duran (Universidad Politécnica de Madrid)

850 Nonlinear-Phase Regularisation for Reducing Non-Causal Components in Radial Filters for Spherical Microphone Arrays

Spatial sound capture using rigid spherical microphone arrays typically involves radial filtering in the spherical harmonic domain to compensate for scattering effects caused by the rigid sphere. Due to the ill-conditioned nature of this inverse filtering problem, regularisation is indispensable for stabilising the filters, particularly at low frequencies and high spherical harmonic orders. Zero-phase regularisation is commonly used to avoid phase distortions, but this results in non-causal pre-ringing of the filters. A modelling delay is therefore required in online processing. Minimum-phase regularisation has also been investigated, where phase alignment is achieved using carefully designed all-pass filters. Nevertheless, the resulting overall phase distortion remains a challenge. In this contribution, we propose a radial filter design in which nonlinear-phase regularisation is selectively applied to the poles of the radial filters. Anticausal poles that are close to the imaginary axis in the Laplace domain, and thus exhibit slow decay, are reflected into the left half-plane using all-pass filters. This approach enables the design of radial filters with reduced phase distortion compared to the conventional minimum-phase design, while still achieving a significant reduction in non-causal components. The proposed approach provides a flexible design framework, allowing the number of zero-phase and nonlinear-phase regularised poles to be chosen based on both permissible phase distortion and acceptable processing latency.

Speaker: Nara Hahn (Institute of Sound and Vibration Research)

Full paper (PDF)

851 Ambisonic Multi-Direction Decomposition Method for Resolving Coincident Acoustic Reflections

Not all reflections in measured Ambisonic room impulse responses (ARIRs) arrive in distinct time intervals. In principle, compact spherical microphone arrays for higher-order Ambisonic recordings enable the decomposition of coincident reflections.However, existing approaches for directionally decomposing and upmixing higher-order ARIRs resolve fewer coincident reflections than the (𝑁+1)² components of an order-𝑁 ARIR would theoretically permit.They either rely on a fixed spatial sectorization (e.g., HO-SIRR), which may be insufficiently resolved or misaligned, or estimate a limited number of time-varying directions (e.g., REPAIR), whose associated signal extraction can become ill-conditioned when directions are not well separated.This contribution adopts the Ambisonic Multi-Direction Decomposition Method (AMDDM) to address these limitations in higher-order ARIR decomposition. It employs a fixed maximum-determinant grid that yields (𝑁+1)² well-separated directional signals, whose short-term covariance matrix is used to align the grid with the acoustic reflections by penalizing off-diagonal entries.In a numerical study with simulated ARIRs containing multiple reflections coinciding within a single time frame, we compare AMDDM with HO-SIRR and REPAIR. The results highlight the conditions under which AMDDM improves directional decomposition, as well as its limitations.

Speaker: Thomas Röck (University of Music and Performing Arts)

Full paper (PDF)

852 Perceptual Relevance of Scene Information Across Room Impulse Response Synthesis Paradigms

Convincing auralisation in extended reality (XR) relies less on physical accuracy and more on perceptual plausibility, where a listener perceives a simulation as equivalent to a real acoustic event. Room impulse response (RIR) synthesis has been widely investigated, yet the relationship between numerical accuracy and perceptual plausibility remains comparatively underexplored, particularly when simulations rely on incomplete or inferred data. This work presents a perceptual comparison of reverberation synthesis paradigms spanning stochastic, parametric, geometric, wave-based and hybrid approaches. Rather than arbitrarily selecting implementations, the evaluated methods are chosen to represent distinct assumptions about how room acoustics can be modelled or approximated. These range from frequency-shaped stochastic decay models and compact parametric reverberators to explicit geometric simulations and scattering delay networks. Each method is provided with the corresponding scene information required for simulation, including material properties inferred semantically from the environment, such as wood flooring or glass surfaces.To isolate perceptual differences arising specifically from reverberation, all conditions share an identical direct-sound rendering pipeline, including individual Head Related Transfer Functions (HRTFs), distance attenuation and Direct-to-Reverberant Ratio (DRR) with only the reverberant field varying across conditions. Perceptual evaluation is performed through controlled 3 Degrees-of-Freedom (3DoF) listening tests using a two-alternative forced-choice (2AFC) paradigm, where participants select which of two renderings is perceived as more acoustically plausible for the presented room. In addition, a measured RIR is presented as one of the conditions to serve as a perceptual reference anchor. The proposed framework investigates how perceptual plausibility relates to reverberation structure, model complexity, and scene-description requirements across fundamentally different synthesis paradigms.

Speaker: Rahul Roy Chowdhury (Imperial College London)

853 Annoyance-Based Beamforming using Perceptual Covariance Weighting

Adaptive beamforming in the spherical harmonics (SH) domain is widely used for spatial filtering and is often formulated as a minimum variance optimization, which suppresses interfering components based on directional energy. Perceptual disturbance, however, is not determined by energy alone but also depends on spectral and temporal characteristics, leading to situations in which energetically weak yet perceptually salient components remain insufficiently attenuated. This work proposes an annoyance-weighted beamforming formulation that incorporates psychoacoustic metrics directly into the beamformer design. Directional signals are obtained via plane-wave decomposition and evaluated using Di’s psychoacoustic model, combining loudness, sharpness, roughness, fluctuation strength, and tonality into a single annoyance measure. These directional estimates are then used to construct an annoyance-weighted spatial covariance matrix, replacing the conventional covariance in the minimum variance distortionless response (MVDR) formulation while retaining its closed-form solution. A power-weighted reference formulation is introduced for comparison between perceptual and energy-based spatial filtering alongside the baseline MVDR. The method was evaluated using simulated acoustic scenes with varying interferer characteristics and source distributions based on the spherical harmonics formulation. The results demonstrate clear differences between the spatial distributions of annoyance and energy. The proposed method yields distinct beamforming patterns and achieves improved selectivity of perceptually annoying interference relative to energy-based approaches. A complementary trade-off is observed with conventional MVDR beamforming, which more effectively reduces overall residual energy.

Speaker: Konstantin Fontaine (Technische Universität Berlin)

Full paper (PDF)

A18.04 Indoor soundscape: S376

Conveners: Simone Torresin (University of Trento, Dept. of Civil, Env., Mech. Eng.), Papataya Nur Dokmeci Yorukoglu

854 Impact of Hospital Soundscapes on Cognitive Performance and Physiological Stress Responses: A Pilot Study

This pilot study aimed to explore the relationship between objective acoustic conditions, perceived soundscape quality, and short-term cognitive and physiological responses in hospital environments. Continuous sound measurements were conducted over seven consecutive days in the emergency department and intensive care unit of Ataşehir Medical Park Hospital using a Type 1 sound level meter (HD2010). Equivalent continuous sound pressure levels (LAeq) and maximum A-weighted levels (LAFmax) were recorded. Audio excerpts were selected using a stratified sampling approach across four daily time periods (morning, afternoon, evening, and night), capturing both typical and peak noise conditions. A total of ten nurses evaluated the selected recordings using perceived affective quality attributes based on ISO 12913, including pleasant, chaotic, eventful, and calm, rated on a Likert scale. To assess short-term cognitive and stress-related responses, participants completed a Stroop task and rated their perceived stress before and after exposure to each audio excerpt. In addition, physiological responses were monitored through heart rate measurements to provide an objective indicator of autonomic arousal associated with stress. Preliminary analyses indicated higher LAeq and LAFmax levels in the emergency department compared to the intensive care unit. These elevated sound levels are anticipated to be associated with higher ratings of “eventful” and “chaotic” and lower ratings of “calm” and “pleasant.” Exposure to more negatively perceived soundscapes is expected to correspond with increased self-reported stress, elevated heart rate, and reduced cognitive performance, reflected in longer reaction times and decreased accuracy on the Stroop task.

Speaker: Donya Dalir (Yeditepe University)

Full paper (PDF)

855 Across Experiment Mediums: Influence of Biophilic Interventions to Mediate Adaptive Acoustic Comfort for Intrusive Traffic Noise

Outcomes of three complementary experiments investigating the influence of biophilic interventions and their mediating effects on adaptive acoustic comfort in office environments are presented. The experiments were conducted across three settings: a non-immersive laboratory, an immersive visualization laboratory (IVL), and a real-world field environment. Indoor greenery (examined in the non-immersive and IVL settings) and birdsong masking (examined in the IVL and field settings) were investigated as key biophilic variables influencing the perception of traffic noise intrusion through open windows. Statistical analyses, structural equation modeling and repeated-measures ANOVA, consistently indicate that both visual and auditory biophilic elements enhance perceived soundscape pleasantness. Indoor greenery, particularly at coverage levels of 10–15%, significantly improved pleasantness evaluations. In field conditions, birdsong masking enhanced pleasantness without increasing overall sound pressure levels by more than 3 dBA. Furthermore, perceptible open-window view increased the perceived appropriateness of intrusive traffic noise, with biophilic elements acting as mediators of adaptive acoustic comfort. Although soundscape assessment scales based on ISO/TS 12913-3 and indoor soundscape constructs demonstrated strong correlations, notable differences emerged between the Eventfulness and Content dimensions in their relationships with overall satisfaction. The consistency of findings across experimental settings highlights the robustness of the identified perceptual pathways. Moreover, the convergence of results across different levels of experimental fidelity supports the validity of non-immersive methods for investigating underexplored aspects of indoor soundscapes. Overall, as adaptation conditions for acoustic comfort remain underexplored, this study contributes to the understanding of these mechanisms and demonstrates the significant role of biophilic design strategies as adaptation factors in enhancing adaptive acoustic comfort in office environments.

Speaker: Merve Esmebasi (National University of Singapore)

856 Soundscape Evaluation in Library Environments: a Comparison of Different Acoustic Scenarios and the Influence of Visual Cues

Soundscape research has highlighted the importance of acoustic environments for human perception and experience. However, most studies have focused on outdoor environments, while few have investigated indoor settings such as libraries. The evolution of libraries from quiet to multifunctional spaces has introduced diverse acoustic conditions. This study investigates soundscape perception in two libraries using controlled ambisonic audio reproduction. The study included two experimental phases. The first compared the perceived soundscape of a simulated contemporary multifunctional library with that of a recorded soundscape in a traditional quiet library. The second assessed the influence of immersive video reproduced with META Quest 2 on soundscape perception using the same audio recorded in the traditional library. Twenty-four participants evaluated sound environments representing typical full-occupancy conditions of the contemporary reverberant multifunctional New Civic Central Library of Torino and the traditional reverberant quiet Saint Geneviève Library in Paris. The former was simulated through auralisation at 50.2 dB(A), while the latter was reproduced from two in-situ recordings at 50.4 dB(A) and 50.3 dB(A). Preliminary results showed that the two libraries differed significantly in the perceived presence of individual verbal and non-verbal sounds, while adding immersive visual cues enhanced perceived soundscape quality and increased Calm and Pleasant ratings.

Speaker: Ioana Tsankova Grozeva (Politecnico di Torino)

A18.05 Soundscape and Inclusion, from Theory to Practice: S132

Conveners: Arezoo Talebzadeh (Ghent University), Kirsten van den Bosch (University of Groningen), Brigitte Schulte-Fortkamp (TU Berlin, Department of Engineering Acoustics)

857 “Where Does The Sound of War Fit in the Circumplex Model?” The Shifting Geopolitical Landscape and the Epistemology of Audible Safety in Soundscape Research

In an era marked by profound geopolitical transformation, for many of us, the specter of war looms closer than ever. As its repercussions permeate daily discourse, critical questions arise regarding the adequacy of dominant frameworks in capturing the nuances of such environments. Does the circumplex model truly encapsulate the complexity of soundscape perception, including that of horror and terror caused by conflicts, or does its binary framing risk oversimplifying the underlying mechanisms that shape perception? Rather than focusing solely on the what, such as the identification of soundsources that drive soundscape perception, we inquire about the why: why is birdsong often perceived as pleasant? Or the sound of warplanes as unpleasant? The concept of audible safety could serve as a foundational epistemology and vocabulary for understanding soundscape perception, including contexts of trauma.

Speaker: Kirsten van den Bosch (University of Groningen)

858 Aural diversity in soundscape perception: Comparing ADHD, autistic, and neurotypical listeners on the ISO 12913 circumplex

Soundscape research generally assumes a uniform listener, yet people differ in how they process sound. This study tested whether neurotype shapes soundscape perception by comparing ADHD, Autistic and Neurotypical listeners (𝑛 = 30 per group) on ten field-recorded soundscapes, delivered online as binaural stereo and rated on the ISO 129132 perceptual circumplex through a calibrated interactive task. Across the pooled soundscapes, ADHD listeners rated the soundscapes as significantly less eventful than Neurotypical listeners (𝑝 = 0.043), while pleasantness was unaffected; no significant differences were found when all three groups were compared together. The ADHD effect was small but consistent, appearing in the same direction in eight of the ten soundscapes rather than being driven by any single soundscape. ADHD listeners also showed a non-significant trend towards more extreme responding. Consistent with recent evidence of preserved global auditory processing in autism, the Autistic group closely resembled the Neurotypical group on every measure (SPI = 92.4). The findings provide empirical support for an aural diversity framework for soundscape research, showing that neurodivergent groups differ from one another and should not be treated as a single category.

Speaker: Paul Paul A Magrath (The University of Salford)

Full paper (PDF)

859 First Step Toward a Sound Categorization Reflecting Autistic Atypical Auditory Perception

Autistic individuals exhibit specific positive and negative auditory experiences related to their atypical auditory perception, resulting in decreased or heightened reactivity to everyday sounds. Although negative auditory experiences substantially hinder their daily life, the sounds involved remain poorly characterized with respect to context-dependent factors (e.g., situation) and context-independent parameters (e.g., acoustic). To address this gap, a two-step methodology is proposed: (i) identifying the most reported sound items through a systematic literature review and (ii) categorizing the sound items through thematic analysis. A resulting structured categorization of sound items related to positive and negative auditory experiences is proposed. This categorization can guide the selection of ecologically relevant sounds in future research. The next research step involves validating the categorization with autism stakeholders.

Speaker: Valentin BAUER (IRCAM)

Full paper (PDF)

860 Students’ Viewpoint On Classroom Soundscapes: Development Of A Child-Centred Pictorial Questionnaire

Designing supportive and comfortable learning spaces requires an understanding of how children perceive their classroom sound environment. While research on classroom acoustics has traditionally focused on objective parameters, studies addressing children’s perceptual responses remain limited and methodologically unclear, due to the frequent use of attributes adapted from adult-oriented questionnaires, along with heterogeneous scales. This paper presents the development of a child-centered questionnaire for assessing acoustic perception in primary school classrooms. The study employed a multi-stage approach to ensure age appropriateness. A set of perceptual attributes was initially identified through a systematic literature review and subsequently refined through focus group discussions with children. This process ensured that the attributes reflected children’s own language and experiential understanding. A semantic differential scale was created by organizing the final set of attributes into 32 bipolar pairs. A 9-point rating scale was developed, combining visual anchors, verbal descriptors, and illustrative characters to support comprehension across different age groups. The use of pictorial elements enhances engagement and facilitates more reliable responses from younger children. The resulting questionnaire provides the basis for the development of a classroom soundscape model, serving as a structured and accessible tool for capturing children’s perceptions of classroom sound environments.

Speaker: Shomaila Syed (University of Ferrara)

Full paper (PDF)

861 Safe and sound in the city: gendered pedestrian experiences in an industrial-residential neighbourhood

Calls for a more diverse soundscape approach have recently been raised, to investigate a variety of urban experiences. In this paper, we take a look at how gender and soundscape co-influence urban experiences. Through commented walks, 12 women and 11 men were interviewed about their experience walking in an industrial-residential neighbourhood. We analyzed spontaneous mentions of sound and safety as a function of gender. Sound was linked to safety but implicitly and in two ways: sound informs about 1. road traffic risks (e.g., sound of a large truck passing close), or 2. presence of others in a space and time perceived as unsafe (e.g., knowing there are people around in case something happens). Women and men talked about road traffic risks in similar proportions and with similar themes (e.g., danger, stress, mitigation strategies). However, only women mentioned sound as indicating the (reassuring) presence of others. These results indicate implicit connections between safety and sound: sound can reflect danger (traffic risks) but also life and safety (presence of others), emphasizing that quiet is not always the goal, especially in the public space. We also discuss how different genders experience sound and safety while walking in public spaces.

Speaker: Cynthia Tarlao (Université du Québec à Montréal)

Full paper (PDF)

A21.01 Automotive noise and vibration: S151

Conveners: Manfred Haider (AIT Austrian Institute of Technology GmbH), Martin Czuka (AIT Austrian Institute of Technology)

862 SPEAKeR-Project: Simulation-based Product Development of Acoustic Systems for Electric Vehicles for the Conceptual Implementation of Adjustable Directivity

As road traffic becomes increasingly electrified, Acoustic Vehicle Alerting Systems (AVAS) are gaining importance. Their development involves a fundamental trade-off: vulnerable road users must be reliably warned and protected, requiring compliance with legal specifications such as UN R138 and FMVSS 141 regarding frequency range and minimum sound pressure levels. At the same time, noise exposure and unnecessary sound emissions, particularly in urban environments, should be minimized.Today, AVAS are often designed and evaluated only in late development phases using physical prototypes, since precise and computationally efficient simulation of the acoustic transfer path from the loudspeaker membrane to the receiver position is not yet available. Within the SPEAKeR project, Novicos and Dortmund University of Applied Sciences and Arts are developing an efficient full-system simulation for the virtual design of acoustic pedestrian warning systems. The approach enables vehicle-specific AVAS development already in early stages of the product development process.Key methodological elements include near-field effects, acoustic-mechanical feedback of the loudspeaker membrane, and frequency-dependent impedance surfaces for the efficient description of reflection, absorption, and transmission. In addition, segmented transfer paths allow only modified subsections to be recalculated after design changes. The simulation results are integrated into the creative sound design process, so that vehicle-specific transfer behaviour can be considered during auralisation.The methodology will be validated using a Smart AVAS demonstrator with situation-dependent adjustable directivity. Unlike conventional multi-loudspeaker concepts, targeted sound guidance will be realized with a single loudspeaker system. The SPEAKeR project combines improved local detectability, regulatory compliance, and reduced overall noise emissions, contributing to quieter, safe, and accepted electric mobility.

Speaker: Tim Kliem (Fachhochschule Dortmund)

Full paper (PDF)

863 EEG-Based Evaluation of AVAS Sound Signatures for Perceptual Assessment in Urban Environments

The rapid adoption of electric vehicles (EVs) has led to regulatory mandates for Acoustic Vehicle Alerting Systems (AVAS) to ensure pedestrian safety at low speeds. While AVAS effectively compensate for reduced propulsion noise, their widespread deployment introduces new challenges for urban soundscapes, particularly when sound design neglects perceptual integration and contextual appropriateness.This study proposes a neuro-acoustic framework for evaluating AVAS sound signatures using electroencephalography (EEG) to capture objective auditory perception. A controlled listening experiment was conducted with 30 participants (15 male, 15 female). AVAS sound signatures were developed and recorded binaurally under multiple operational conditions at both driver and pedestrian positions to ensure spatial realism.Participants were exposed to these stimuli in a laboratory setting, while EEG responses were recorded and analysed across key frequency bands, including Alpha, Theta, Beta, and Gamma. Variations in these bands were used to infer cognitive and affective states such as attention, engagement, stress, and annoyance under different acoustic conditions.The results demonstrate that EEG-based metrics provide a sensitive and objective means to differentiate between AVAS sound signatures beyond conventional acoustic descriptors. Distinct neural response patterns were observed across conditions, highlighting the influence of sound design on perceptual outcomes for both drivers and pedestrians.The study establishes the potential of integrating neurophysiological assessment into AVAS evaluation frameworks. Such an approach can support the development of context-sensitive and perceptually optimised AVAS sounds, enabling manufacturers to move beyond compliance-driven design towards solutions that enhance safety while preserving the quality of urban soundscapes.

Speaker: Manish Manohare (Transportation Research and Injury Prevention Centre, IIT Delhi)

864 Order-Based Spatial Decomposition in Higher-Order Ambisonics for Vehicle Interior Noise Analysis

In multi-motor battery electric vehicles the loss of combustion masking makes the order-structured tonal noise of each drive unit perceptually salient, and units that share the same machine and single-speed gear ratio, such as the two wheel motors of one axle or the twin units of a symmetric dual-motor vehicle, hold their same-numbered orders within about one percent of each other whenever both are active. This spectral degeneracy prevents tracking-filter separation for any small number of channels, including the binaural pair used in standard NVH practice. We propose a spatial decomposition framework for third-order Ambisonic recordings. Phase-coherent tracked-harmonic extraction, driven by one instantaneous phase per shaft, is applied independently to every spherical-harmonic channel, and full-sphere direction-of-arrival maps attribute each separated component to its source. Where the order families are degenerate, a constrained adaptive spatial prefilter, whose spherical-harmonic steering vectors are independent of frequency, places nulls on the interfering motor and on the background before extraction. The framework is validated on a real anechoic-chamber recording and on two-motor scenes constructed from measured electric-drive order data and measured room impulse responses, and is benchmarked against binaural processing of the same fields. Where per-channel tracking, binaural processing and even exact-phase joint Vold-Kalman estimation remain limited, with weaker-motor correlations of 0.60–0.84, spatial-null preprocessing restores the separation to 0.94, and the recovered components retain their directions of arrival.

Speaker: Zhenxian Li (INSA Lyon, LVA UR677)

Full paper (PDF)

865 Enhancement of composite firewall airborne insulation through locally resonant low frequency tuned vibration absorbers

Improving low‑frequency airborne insulation in automotive firewall panels is challenging because structure‑borne transmission dominates below 500 Hz, where traditional acoustic treatments are less effective. Measurements performed on a dash panel from a battery-electric-vehicle (BEV) revealed a pronounced transmission‑loss (TL) dip near 500 Hz, indicating a structural weakness linked to the dash panel dynamics. This work investigates tuned vibration absorbers (TVAs) as lightweight, locally resonant solutions to mitigate such narrowband deficiency. TVAs introduce sub‑wavelength resonances that increase mechanical impedance and locally suppress bending‑wave mobility, thereby enhancing low‑frequency TL. A validated structural finite element (FE) model forms the basis for assessing TVA performance. Airborne TL predictions were further validated using a coupled vibro‑acoustic FE model and measurements in coupled rooms. A first TVA layout, constrained by realistic mass and packaging limitations, was evaluated and yielded 3–6 dB TL improvement at the 500 Hz dip frequency without negative effects on adjacent bands. Finally, a Genetic Algorithm optimization step was performed to explore multi‑frequency TVA arrangements under design constraints, demonstrating additional potential for targeted TL enhancements. These results confirm the feasibility of TVA‑based countermeasures for automotive firewall panels and motivate further optimization-driven layout refinements.

Speaker: Benjamin Morin (Autoneum Management AG)

Full paper (PDF)

A21.06 Aircraft noise: S156

Conveners: Michael Quaglia, Vincent Clair (LMFA - Ecole centrale de Lyon)

866 New Formulation of the exact Green’s Function for the Diffraction Field around a Rigid Wedge for efficient Calculations

Green’s functions for edge diffraction are often used in analytical modeling of installed jet-noise or trailing edge noise. A new formulation of the exact Green’s function for the problem of diffraction of a point source by a rigid wedge of infinite and finite length is derived. The derivation is based on a variable change in the diffraction integral of the exact Green’s function. The transformation recasts the integral into a Riemann Stieltjes form, with the integration performed over a normalized edge length parameter ranging from 0 to 1 (corresponding to the actual edge length from minus infinity to infinity). For wedges of finite length, the limits of integration match the limits of the finite edge.The new formulation offers two key advantages over existing approaches:i) The singularity of the integrand that exists in all formulations of the exact Green’s function is handled automatically without requiring special techniques such as local analytical approximations or increased number of quadrature points. The new formulation acts as adaptive sampling, focusing on the singular area, which in turn improves computational efficiency. ii) The new Riemann Stieltjes form allows the use of a simple quadrature formula without sacrificing accuracy across all frequencies and source-receiver configurations.When compared to the state-of-the-art exact integral formulation, the new method is substantially faster for all wedge angles, all frequencies and all source-receiver configurations. Compared to an approximate method restricted to high frequencies, the new approach is slower but still computationally competitive.

Speaker: Petros Nikolaou (University of Patras)

867 Analytical Investigation of Sound Diffraction by Straight and serrated Trailing Edges

This work investigates the sound field produced when a point source interacts with the trailing edge of an airfoil, modeled as a half-plane. A newly presented Green’s function for sound diffraction by infinitely long and finite-length diffracting edges is employed. First, it is shown that the diffracted field dominates the total sound field, when the source is on the airfoil surface or very close to the trailing edge. An equation is presented that provides, a priori, the relative importance of the diffracted field in the total sound field. Second, the solution is modified to predict the diffracted field around serrated edges and the geometrical acoustics field is added to compute the total field. Next, the premise that serrated trailing edges offer advanced noise mitigation compared to straight edges is tested. Serration profiles with varying segment lengths and inclination angles are compared to the corresponding straight-edge profiles on various receiver planes and for different source locations (modeling the location of different noise sources such as jet noise or airfoil self-noise). Serrated profiles offer increased noise protection (compared to straight-edge profiles) to some receivers, but reduced protection to others. The optimal serration profile that maximizes the overall noise reduction on a receiver plane depends on source location and receiver plane. However, a serration profile with segment length of 1.5 wavelengths and a 75-degrees inclination angle performs well in most cases. The benefits of serrated edges are maximized in the airfoil’s shadow zone, except when the source is at the trailing edge itself.

Speaker: Nafsika Papantonopoulou (University of Patras)

Full paper (PDF)

868 Structure-Borne Sound of an Electrified Powertrain - Identification of Load Condition-Based Local Maxima

The European Union is aiming for significant greenhouse gas reductions in aviation by 2050. In the course of this, electrified drive concepts are also becoming increasingly important, resulting in new challenges with regard to the noise emission of electric machines. In order to gain a better understanding of the generation of noise, a coupled drive train consisting of two permanent magnet axial flux machines connected in series was investigated experimentally. The rotational speed and torque as well as the sound pressure level in the free field and surface velocity of the structure were measured at various stationary operating points and run-up scenarios using microphones, triaxial acceleration sensors and the information provided by the inverter. In order to draw conclusions about the excitation sources, spectral maxima of the measured variables are compared with characteristic frequencies from theory. In addition, information from product data sheets, such as the number of pole pairs of the machine, is used as input parameters for this calculation. The method compares the spectra of different load-dependent test setups simulating a multitude of flight scenarios, which, considering the potential sources of the local maxima, may enable the design of low-noise concepts.

Speaker: Lewin Weber-Karpinski (DLR Institute of Electr. Aero-Engines)

Full paper (PDF)

869 Estimation of Flight and Geometry Parameters for low-Boom Aircraft Design using the Unscented Kalman Filter

In this work the flight characteristics and aircraft geometry required to produce a certain sonic boom signature are estimated using the Unscented Kalman filter (UKF). The Kalman filter is an algorithm that estimates state spaces by combining model predictions and observational data. The UKF extends the Kalman filter to nonlinear systems. Instead of linearizing the system, it uses sampling points (sigma points) to capture the mean and covariance of the state distribution. For the calculations the UKF requires a state vector xk , a measurement vector zk, and a mathematical model. Initially, sigma points are calculated by the mean of the state vector and its covariance. Next a prediction is made using the mathematical model. After that the UKF updates the prediction using the Kalman gain and measurements. The work is organized in two phases. First, UKF is validated by using it to estimate the flight parameters, Mach number and altitude, required by a certain aircraft geometry to achieve measured sonic boom overpressure and duration. Second, UKF is used to solve the inverse design problem where the flight parameters and aircraft shape are estimated so that a target low-boom signature is produced. For the first phase the state vector includes flight parameters. For the second phase the shape and lift of the aircraft are included. The measurement vector includes sonic boom overpressure and duration in both phases. Rise time is added for the second phase. Two mathematical models are employed: the Carlson empirical formula and a non-linear propagation model.

Speaker: Ioanna Anagnostaki (University of Patras)

Full paper (PDF)

870 Autonomous Aircraft Noise Recording: Developing Datasets for Machine Learning

The advancement of deep learning in environmentalacoustics is hindered by a lack of high-fidelity, labelleddatasets. While aircraft noise is a major environmental concern, collecting annotated data at scaleremains a logistical hurdle; attended measurementslack volume and permanent monitors lack geographicdiversity. This paper presents an autonomous aircraft noise recording station that integrates a noisemonitoring terminal with a custom solar-power subsystem and ADS-B (Automatic Dependent Surveillance–Broadcast) receiver. This architecture enablesautomatic alignment of acoustic events with flighttelemetry, including aircraft type, location and altitude in order to reduce manual annotation effort. Thispaper details a hardware architecture, energy budgetanalysis for year-round autonomy, and field protocolsfor data integrity. The resulting methodology enablesthe repeatable creation of ’AI-ready’ acoustic datasets.

Speaker: Gary Isherwood (University of Sussex)

Full paper (PDF)

A23.00 Vibro-Acoustics: S161

Convener: Christian Adams (Graz University of Technology)

871 Effects of Distorted Geometric Scaling on Rectangular Plates

Geometric scaling is widely employed in structural dynamics to predict the behaviour of systems that are too large, costly, or impractical to test at full scale. Under uniform scaling, where all spatial dimensions are scaled by the same factor, well-established and global similitude relationships allow a direct correspondence between modal properties. However, when different scaling factors are applied along orthogonal directions, distorted geometric similitude arises, leading to non-uniform modifications of the dynamic response. This work investigates the effect of such distortion on the modal characteristics of simply supported rectangular plates. Analytical formulations based on the Kirchhoff-Love plate theory are used to compute natural frequencies and mode shapes for both parent and distorted configurations (avatars). Particular attention is given to the evolution of modal ordering, which deviates from the original sequence observed under uniform scaling. The results also show that a global scaling law remains applicable up to a certain level of geometric distortion, which is characterised through different distortion measures.

Speaker: Giada Cardellino (Graz University of Technology)

872 Acoustic Classification of Visually Identical Materials Using Impact Emissions

Human hearing can reliably distinguish materials based on impact sounds, and machine learning methods can replicate this ability when acoustic differences are pronounced. However, discriminating between materials that sound highly similar even to the human ear remains challenging.In this study, we demonstrate that glass and metal can be effectively separated using simple thresholding of acoustic features. In contrast, classifying eight types of optically identical plastic parts of identical geometry proves significantly more difficult. For this eight-class task, models trained on 48 kHz signals achieve an accuracy of 25%, exceeding the random baseline of 12.5%. Increasing the sampling rate to 192kHz improves performance to 36%, while cascading classification strategies further raise accuracy to 47%, with some plastic types exceeding 60% class-wise accuracy.These results show that acoustically distinguishing visually identical plastic materials is feasible with moderate accuracy. The experimental setup ensures that observed signal variations arise solely from intrinsic material properties.

Speaker: Anna Maly (Joanneum Research Forschungsgesellschaft mbH)

Full paper (PDF)

873 A Seismic Barrier Strategy Using an Improved Ring-Resonator Periodic Material (RRPM) Based on 2D WFEM

With the rapid expansion of urbanization, an increasing number of high-rise structures require protection from earthquake-induced damage through seismic isolation systems. However, current technologies lack effective isolation barriers capable of mitigating horizontal seismic waves. This study proposes an improved ring-resonator Unit Cell (UC) for two-dimensional (2D) periodic materials is proposed based on Bloch–Floquet theory and locally resonant mechanisms. The Wave Finite Element Method (WFEM) and Finite Element Method (FEM) are employed to investigate the Band Gap (BG) characteristics and various wave modes of the UC. Furthermore, the Transmission Loss (TL) of the proposed 2D Ring-Resonator Periodic Materials (RRPM) is validated through numerical simulations. Results demonstrate that the proposed ring-resonator UC exhibits a significantly wider and more effective band gap compared to conventional circular resonator designs.

Speaker: Yunhe XI (ISAE-Supméca)

Full paper (PDF)

874 Discrete Approximation of Continuous Relaxation Spectra in Vibrating Structures

Continuous relaxation spectra accurately describe how vibrating materials dissipate energy across scales. The fractional wave equation models this behaviour, yet engineering practice often collapses damping into a single structural parameter, obscuring its intrinsic multi-scale nature. Herein, we present a discrete approximation of continuous relaxation spectra based on a superposition of relaxation functions with characteristic time scales. The associated stress–strain relation is derived from Ludwig Boltzmann’s Theory of Elastic After-Effects, which was published during his professorship in experimental physics at the University of Graz in 1876. We establish the conditions that ensure convergence of the discrete spectra to their continuous counterparts in a theoretical consideration. Variational mode decomposition and Hilbert-Huang spectral analysis are used to determine the discrete relaxation functions. Free vibration measurements on a cantilever beam validate the approach using nonlinear regression analysis to determine the decay behaviour of the relaxation functions. The methods enable robust time-domain damping identification, including nonlinear damping, where frequency-domain techniques are inapplicable. Finally, we demonstrate applicability to more complex materials and structures, exemplified by a wooden rod, highlighting the method’s practicality for the engineering of vibrating structures.

Speaker: Christian Adams (Graz University of Technology)

A03.08/A05.07 Heat pump acoustics in residential environments: S387

Conveners: Jack Havie-Clark, Christoph Reichl (AIT Austrian Institute of Technology)

875 Exploring the Roles of Loudness and Sound Character in Heat Pump Noise Annoyance

This study presents a detailed analysis of how sound character influences the annoyance ratings of heat pump noise, based on data from a listening experiment conducted as part of the IEA/HPT Annex 63 on Placement Impact on Heat Pump Acoustics. This large-scale listening experiment was originally designed to investigate the impact of heat pump placement in residential environments, considering factors such as the number of units, their spatial arrangement, operating conditions, background noise, and noise barriers.Using this dataset, the present work performs a secondary analysis focusing on perceptual sound characteristics. In particular, it examines the hypothesis that loudness provides a more suitable description of annoyance than conventional sound level metrics, including A-weighted sound pressure level. In addition, the study explores whether further psychoacoustic attributes, such as tonal components, roughness, and sharpness, contribute to annoyance beyond what can be explained by level or loudness alone.To this end, sounds are systematically compared under controlled conditions of similar level or loudness, allowing the influence of specific sound characteristics to be investigated. The dataset, while originally developed for a different research focus, provides a suitable basis for an initial assessment of these perceptual factors.The results provide indications that both hypotheses are supported: perceived loudness appears to be more closely related to annoyance than level-based measures, and additional sound characteristics further influence listener responses. These findings highlight the importance of considering both the intensity and the character of sound when investigating heat pump noise annoyance, while also pointing to the need for dedicated studies to confirm these effects.

Speaker: Julian Becker (HEAD acoustics GmbH)

Full paper (PDF)

876 Impact of Acoustic Character on Human Response to Heat Pump Noise

Air Source Heat Pumps (ASHPs) represent an essential technology for the decarbonisation of home heating across Europe. As ASHP deployment becomes more widespread, the introduction of this relatively new noise source may increase community noise annoyance. Current standards for assessing ASHP noise often do not consider acoustic character, for example, tonality (the prominence of discrete frequencies or narrow frequency bands). This study investigates the impact of tonality on human perception of ASHP noise. Real recordings of ASHPs with differing tonality were presented to participants in a laboratory environment. The annoyance response of participants was evaluated using a paired comparison test between the ASHP sounds and reference sounds. The reference sounds were spectrally similar to the ASHP sounds, but with minimal tonal content. The findings of this study quantify the impact of tonality on perception of ASHP noise. This evidence may support a refinement of ASHP noise assessment standards, enabling acoustic character to be incorporated. As a consequence, communities may have an improved experience of ASHP noise.

Speaker: Katie Salter (Acoustics Innovation Institute)

Full paper (PDF)

877 Metamaterial‑Based Noise Control Strategies for Residential Heat Pump Ventilation Ducts: Methods and Design Principles

Residential heat pumps introduce significant low‑frequency aerodynamic and mechanical noise into indoor and outdoor environments, especially through compact ventilation ducts. Conventional duct silencers are often too bulky, prone to degradation, or introduce unacceptable pressure drops. Acoustic metamaterials (AMMs) offer a promising alternative due to their subwavelength behaviour, design flexibility, and ability to maintain ventilation while delivering substantial attenuation in the 31.5–2000 Hz range relevant to heat pumps. This paper presents a focused synthesis of metamaterial mechanisms, modelling approaches, and performance data framed specifically for applications in residential heat pump acoustics. The discussion emphasises the performance of resonant cavities, micro‑perforated panels, membranes and interference‑based structures under realistic flow conditions, including the influence of Mach number, turbulence, and duct geometry. The paper further outlines recommended modelling practices, integration strategies, and research gaps that must be addressed for widespread adoption of AMM‑based silencers in heat pump ventilation systems.

Speaker: Gioia Fusaro (University of Bologna)

Full paper (PDF)

A09.00 Machine learning and artificial intelligence in acoustics: S388

Conveners: Mirco Pezzoli, Alois Sontacchi (University of Music and Performing Arts), Martin Hagmüller (Signal Processing and Speech Communication Laboratory), Franz Pernkopf (Signal Processing and Speech Communication Laboratory)

878 Acoustic classification of packaging waste based on ultrasonic and audible sound wave emissions in reverberant enclosures

Reverse vending machines encourage recycling, but require automatic container classification. This is typically achieved by reading barcodes or using computer vision, which can fail due to unreadable codes or poor lighting. Recently, an active acoustic approach using sound waves has been proposed as an alternative, offering low computational cost and minimal maintenance. However, its performance depends on the acoustic environment. This study investigates how enclosure geometry and wall absorption influence classification accuracy. Acoustic impulse responses were measured using exponential sine sweeps, with an omnidirectional parametric loudspeaker generating ultrasonic and audible waves via the parametric acoustic array effect. The data trained machine learning and deep learning models to classify plastic, metal cans, glass, and cartboard-tetrabrick. The tests were performed in a mini-reverberation chamber and in a shoebox enclosure with and without wall-mounted absorbers. The highest accuracy was achieved in the mini-reverberation chamber (96%), highlighting the benefits of diffuse and reverberant environments.

Speaker: Marc Arnela (La Salle, Universitat Ramón Llull)

Full paper (PDF)

879 Interpretable Machine Learning for Assessing Environmental and Mounting Effects on Aeronautical Glass Wool Absorption

The in-service acoustic performance of fibrous absorbers in aircraft fuselage linings is affected by installation variability and environmental exposure, yet their relative contributions remain poorly quantified. This work reframes the problem as a feature detectability task, where the sound absorption coefficient is used to infer the physical parameters governing its variability. Two experimental datasets are constructed from impedance tube measurements under controlled mounting conditions and humidity exposure, including cyclic moisture loading representative of flight operations. Supervised classification models are employed as qualitative discriminators: classification accuracy is used to assess whether a given feature leaves a detectable signature in the absorption curves. SHAP-based explainability is then used to quantify the frequency-dependent contribution of each feature. Results show that material type, protective layers, and humidity-related parameters are consistently detectable, with classification accuracies up to 99% and 85%, respectively. In contrast, installation effects remain indistinguishable from the random baseline under controlled mounting conditions. Moisture effects are strongly frequency dependent, with dominant contributions grouped in three bands, and exhibit progressive degradation under cyclic exposure. The proposed framework provides a methodology able to isolate and quantify the physical drivers of acoustic variability in porous materials under realistic operating conditions.

Speaker: Alfonso Caiazzo (University of Naples Federico II)

Full paper (PDF)

880 SPSC-HCM-16C: A Multi-Channel Lung Sound Benchmark

This paper presents SPSC-HCM-16C, a benchmarkdataset1for patient-level respiratory disease classification from synchronized 16-channel lung soundrecordings collected in a real clinical environment.The dataset contains recordings from 183 subjects,including healthy controls and four respiratory diseasegroups, and supports three diagnostic settings: 2-class,3-class, and 5-class classification. The aim is to establish a standardized and reproducible benchmark forfuture research in multi-channel computational auscultation. To this end, we define subject-independentevaluation protocols, including a fixed 60/20/20 splitand an 80/20 split with 5-fold cross-validation on thetraining-validation subset, together with a transparent reference baseline. The baseline uses MFCC features, a modified 16-channel ResNet101 backbone withpartial fine-tuning of Layer3, and mean pooling forpatient-level aggregation. Under the fixed 60/20/20split, the baseline achieves Macro-F1 scores of 93.68%,77.06%, and 44.61% for the 2-class, 3-class, and 5-classtasks, respectively, with consistent trends observedunder 5-fold cross-validation. These results establish areproducible reference for future comparison and highlight the challenges posed by weak supervision, classimbalance, and inter-class similarity in multi-channellung sound analysis.

Speaker: Benedikt Mayrhofer (Signal Processing and Speech Communication Laboratory)

Full paper (PDF)

A10.13/A12.07 Numerical methods for wave propagation in complex media: S072

Conveners: Andrés Prieto (CITMAga, Universidade da Coruña), Paulo Amado-Mendes (University of Coimbra, Dep. Eng. Civil)

881 Application of Time-Explicit Discontinuous Galerkin Method for Modelling of Small and Finite Amplitude Acoustic Propagation in Dissipative Fluids

The time-explicit discontinuous Galerkin finite element framework has been a powerful tool used for modelling of acoustic wave propagation in the time domain. Its applications stretch from room and underwater acoustics to medical and non-destructive inspection ultrasound and can be based on acoustic waves of small and finite amplitudes. The effective simulation of acoustic propagation in dissipative fluids requires accurate modelling of the sound dispersion and absorption. This can be challenging for the materials that obey a frequency power law attenuation, which is the case for many biological tissues.This work focuses on the development of a numerical approach based on the time-explicit discontinuous Galerkin framework for modelling of sound propagation in media that exhibit a power law attenuation. First, we derive the system of governing equations for small and finite amplitude propagation in media with relaxation starting from the constitutive equation and the conservation laws. Then, when the attenuation is given through a frequency power law, we apply the adaptive Antoulas-Anderson algorithm to get an "equivalent" relaxation process as a rational approximation of the dispersion relation within the given frequency range.

Speaker: Kirill Shaposhnikov (COMSOL A/S)

882 A Timoshenko beam-based methodology to estimate the vibrational fatigue in pressurized industrial pipelines

Vibrational fatigue is a leading cause of pipeline failure in industrial facilities, yet accurately predicting it remains a significant engineering challenge due to the complex interplay among structural dynamics, fluid-induced loading, and geometric nonlinearities. This work presents a numerical procedure to estimate the vibrational fatigue of piping systems subjected to dynamic excitation in industrial environments.The mathematical model discretizes each pipeline element as a three-dimensional Timoshenko beam, accounting for bending, shear, and torsional deformation, including inertia effects, all of which are critical in thick-walled or short-span pipe configurations where the classical Euler–Bernoulli assumptions introduce non-negligible errors. A key feature of the present approach consists of the presence of a deformed prestress equilibrium configuration at steady state due to the internal hydrodynamic pressure. The inner pipelining pressure and the magnitude of the pipe radial deformation are computed through an asymptotic one-dimensional hydrodynamic model based on the assumption of the flow of an incompressible viscous fluid and serve as input data for the subsequent time-harmonic dynamic response evaluation, enabling accurate capture of the pressure-stiffening effect on the natural frequencies and mode shapes of the system.Fatigue damage accumulation is estimated by combining the dynamic stress response, obtained via modal superposition, with an appropriate cycle-counting method and available S-N curve data at the pipeline joints. The tool is validated against both analytical benchmarks and experimental measurements reported in the literature, demonstrating reliable accuracy across representative pipe geometries and loading conditions.

Speaker: Andrés Prieto (CITMAga, Universidade da Coruña)

883 Fourier-Neumann Numerical Models of Acoustic Propagation

Computationally-efficient single-frequency acoustic solvers have many applications. Our interest is in therapeutic and diagnostic biomedical ultrasound. I will describe recent work on a class of numerical acoustic models of wave propagation through heterogeneous soft tissue media based on unconditionally convergent Neumann series, made efficient through the use of the Fast Fourier Transform. Comparisons will be give to the well-validated pseudospectral time domain model k-Wave, and the factors affecting the rate of convergence will be discussed.

Speaker: Ben T Cox (University College London)

A14.01 New approaches for improving and assessing outcomes with hearing aids: S380

Conveners: Hendrik Husstedt (Deutsches Hörgeräte Institut GmbH), Inga Holube, Florian Denk (German Institute of Hearing Aids)

884 Evaluating hearing devices in interactive conversations using audiovisual virtual reality

Hearing device performance assessed via traditional Speech Reception Thresholds (SRT) may not fully reflect real-world communication effectiveness, as SRT measurements often occur in controlled, non-ecological conditions. To address this limitation, we introduce two novel methods for evaluating device benefit in realistic, dynamic communication environments that incorporate natural head movements and interactive social behavior using low-delay, interactive virtual audiovisual simulations. The first method employs two spatially separated, interleaved tracks of an adaptive sentence test under passive listening conditions, yielding SRT-based measures of device benefit. The second method directly assesses device benefit through interactive, free-form triadic conversations, measuring performance in terms of signal-to-noise ratio (SNR) improvement. Results demonstrate that SNR-based performance metrics in naturalistic conversations are feasible and yield values comparable to those obtained from SRT testing. Importantly, input SNRs in free conversation can be systematically adjusted to reflect typical real-world speech communication environments, whereas adaptive SRT testing tends to produce unrealistically low SNRs. Furthermore, the inclusion of visual stimuli in the interleaved SRT paradigm promotes more natural participant movement and engagement, enhancing ecological validity. These findings suggest that integrating SNR and SRT measurements within complex, interactive virtual environments—where auditory and visual cues are synchronized with natural conversation—significantly improves the ecological validity of hearing device evaluation.

Speaker: Giso Grimm (Carl von Ossietzky Universität Oldenburg)

885 Adaptation of a 3D-printed head surrogate for photoacoustic bone-conduction studies

For the investigation of bone conduction mechanisms, cadaveric materials are often used. However, their application is limited by ethical concerns and logistics (specialized laboratories, availability). Furthermore, different preservation methods compromise reproducibility. For our research we adapted the head model developed by Irwansyah et al., aiming to accurately replicate the mechanical impedance ratio between soft tissues and bone by using PETG for the skull and Ecoflex 00-30 silicone for soft tissues. To utilize this head surrogate for contactless, quantitative assessment of the output of photoacoustic bone conduction systems measured by Laser-Doppler-Vibrometry, the light absorption properties of the soft tissues were adjusted using pigments to match those of human skin. When stimulated with a photoacoustic bone conduction hearing device at the mastoid, the measured skin-vibration velocities at the zygomatic arch at frequencies >3000 Hz were within 1 to 6 dB of values obtained from living participants. When stimulated using a conventional B81 bone conduction transducer at the mastoid the measured velocities at 500 Hz, 1000 Hz and 4000 Hz were within 8 dB of values obtained from living participants. Although the model shows a pronounced anti-resonance at 2000 Hz, where velocities deviated by approx. 15 dB from living-subject data, the surrogate reliably replicates a vibratory response under photoacoustic stimulation above 3000 Hz. The origin of the 2000 Hz anti-resonance can likely be attributed to structural simplifications of the surrogate. Future iterations will incorporate layered soft tissue geometries as well as a stiffer skull material and revised skull geometry to better resolve the anti-resonance and extend the validated frequency range.

Speaker: Dawid Brüning (Martin Luther University Halle-Wittenberg)

886 Sound Preference Predicts Real-Life Hearing Aid Preference

Clinical audiology aims to enable seamless participation in everyday life, including effective communication, enjoyment of music, and awareness of surrounding environments. Central to these goals is the user’s ability to adapt comfortably to hearing aids. However, achieving an optimal hearing experience is challenging, as no single hearing aid sound design meets the preferences of all individuals. To address this challenge, a sound preference tool was developed to capture individual sound preferences and support hearing aid recommendations. The present study investigated whether preferences identified by this tool predicted real-life hearing aid preference. Seventeen participants completed the study, which involved two clinic visits and a two-week home trial. During the first visit, participants used the sound preference tool to compare two hearing aid sound profiles across everyday listening scenarios. Based on participant responses, the tool identified a preferred profile corresponding to one of two hearing aid sound designs. Participants then alternated daily between the two hearing aid sound designs during the home trial. At the second visit, real-life hearing aid preference was assessed. Results showed a correspondence between tool-based sound preference and preferred hearing aid in real life. These findings suggest that the sound preference tool can predict real-life hearing aid preference.

Speaker: Frederic Marmel (ORCA Labs)

Full paper (PDF)

16:00 Closing Ceremony