The Journal Of The Acoustical Society Of America[JOURNAL]
Sun
200 papers
Orthogonal subspace approach for underwater acoustic wave-vector direction estimation based on acousto-optic sensing.
Li X, Wang B, Zhang R … +3 more , Ma Y, Yang X, Xue B
J Acoust Soc Am · 2026 May · PMID 42084283 · Publisher ↗
High-precision wave-vector direction estimation is critical for underwater acoustic positioning, target detection, and tracking. Traditional array-based methods typically require large apertures, whereas a single acousti...
High-precision wave-vector direction estimation is critical for underwater acoustic positioning, target detection, and tracking. Traditional array-based methods typically require large apertures, whereas a single acoustic vector sensor depends on inter-channel phase consistency and remains underexplored at mid-to-high frequencies. To overcome these limitations, we replace piezoelectric or electromagnetic principles with the acousto-optic effect for acoustic vector sensing, which provides multidimensional, high-order, non-contact sensing and is well suited to wave-vector sensing in the mid- to high- frequency range. Building on the classical MUSIC method and the acousto-optic sensing principle, we develop an orthogonal-subspace wave-vector direction estimation algorithm (named MUSIC-L) tailored to acousto-optic wave-vector sensing and validate it through simulations and experiments. Simulation results show that the proposed method is robust and that the theoretical error is almost independent of angle; at a signal-to-noise ratio of 10 dB with 80 snapshots (2 MHz sampling rate, 75 kHz source), the root mean square error is 1.4°. Finally, we design and fabricate an acousto-optic vector hydrophone prototype (0.5 m × 0.5 m × 0.185 m) and measure the wave-vector direction in an anechoic tank. The results show that, with 80 snapshots, the estimation error remains below 1°, with a standard deviation of approximately 0.23°.
Sound-field-projection synthesis using latent diffusion model for acousto-optic reconstruction.
Tanigawa R, Ishikawa K, Harada N … +1 more , Oikawa Y
J Acoust Soc Am · 2026 May · PMID 42084282 · Publisher ↗
Acousto-optic sensing (AOS) is a non-contact method for measuring sound using light, suitable for environments where microphones are impractical, such as confined spaces or within airflow. Despite its effectiveness, AOS...
Acousto-optic sensing (AOS) is a non-contact method for measuring sound using light, suitable for environments where microphones are impractical, such as confined spaces or within airflow. Despite its effectiveness, AOS captures line-integrated sound pressure along the optical path, resulting in signals that cannot be interpreted as those from point-wise microphones. To obtain the sound pressure distribution in three-dimensional (3D) space, volumetric sound-field reconstruction is required. Existing methods require multi-directional line-integrated projection data, typically obtained either by deploying multiple devices or by repeatedly exciting the sound source. The former requires deploying multiple high-cost devices, which is often impractical, whereas the latter is not applicable to sound sources that cannot be reproduced consistently. To overcome these limitations, we propose a task called sound projection synthesis, which synthesizes data in specified directions based on observation data. We achieve this by using a latent diffusion model conditioned on observed projections and view angles to generate new sound-field projections. A model pretrained on natural images was fine-tuned using sound-field data and optimized with a pixel-wise loss. Experiments demonstrate that the model can generate realistic projection data. Combining nine observed views with nine generated views improved 3D reconstruction accuracy compared with using only nine observed views.
Deep learning-based approach for linking microstructural and macroscopic acoustic properties of sound-absorbing polyurethane foam.
Jung WG, Kim DY, Lee JW … +1 more , Kang YJ
J Acoust Soc Am · 2026 May · PMID 42080554 · Publisher ↗
Sound absorption in porous materials is fundamentally governed by their microstructural morphology yet establishing a quantitative and design-oriented relationship between microstructure and acoustic behavior remains cha...
Sound absorption in porous materials is fundamentally governed by their microstructural morphology yet establishing a quantitative and design-oriented relationship between microstructure and acoustic behavior remains challenging. To address this challenge, a deep learning-based acoustic modeling framework is proposed for analyzing the microstructure of flexible polyurethane (PU) foam and predicting its acoustic performance. A microscopic analysis model is developed to semantically segment SEM images using a U-Net model and quantitatively extract the distribution parameters of microstructural properties, including cell size, pore size, pore shape factor, and strut thickness. An artificial neural network model is developed to model the relationship between these microstructural parameters and acoustic performance measured using an impedance tube, based on 210 flexible PU foam samples including thermally aged and non-aged materials from multiple manufacturers. Finally, the proposed approach is validated through comprehensive performance evaluation, comparison with experiments and alternative methods, and analysis of microstructural parameter contributions. The validated framework establishes a quantitative link between microscale morphology and acoustic performance, providing data-driven acoustic insights and practical guidance for acoustic material design, including feature selection, optimization of acoustic performance, and fabrication of sound-absorbing materials for applications such as automotive and construction.
Heart heard in motion.
Garcia D
J Acoust Soc Am · 2026 May · PMID 42066036 · Publisher ↗
Abstract loading — click title to view on PubMed.
Estimating band importance for environmental sound recognition using deep learninga).
Johnson EM, Healy EW
J Acoust Soc Am · 2026 May · PMID 42065403 · Publisher ↗
Environmental sound recognition (ESR) enables listeners to interpret complex acoustic environments, yet the frequency regions that support recognition are poorly understood. This study used deep learning to model ESR in...
Environmental sound recognition (ESR) enables listeners to interpret complex acoustic environments, yet the frequency regions that support recognition are poorly understood. This study used deep learning to model ESR in competing speech and estimate frequency band-importance functions (BIFs) underlying recognition performance. Trial-level responses were collected from 46 listeners who identified 25 everyday sounds mixed with speech across a wide range of target-to-masker ratios. Two model variants were evaluated: one trained to mimic human performance, which was trained on soft labels derived from listener responses, and one trained for maximum accuracy, which was trained on ground-truth correct sound labels, enabling a direct comparison between perceptually driven and task-optimal band-importance patterns. The human-trained model closely reproduced key features of human performance, whereas the ground-truth-trained model exceeded human accuracy and showed highly reliable performance across cross-validation folds. BIFs were estimated by bandstop filtering the target signal and quantifying the resulting drop in recognition accuracy. Both model variants yielded reproducible BIFs with five prominent peaks (∼0.43, 0.77, 1.46, 2.6, and 9.7 kHz), largely driven by subsets of sounds having sharply tuned spectral dependence. This convergence across training objectives suggests that human performance closely reflects the task-optimal frequencies for segregating environmental sounds from speech maskers.
A few-shot learning method for underwater acoustic target recognition based on generative data augmentationa).
Huang W, Zhang Q, Zhao B … +2 more , Xu T, Zhang H
J Acoust Soc Am · 2026 May · PMID 42065402 · Publisher ↗
The recognition of underwater acoustic targets (UATR) is of great significance for the protection of marine diversity and national defense security. The development of deep learning provides new opportunities for UATR, b...
The recognition of underwater acoustic targets (UATR) is of great significance for the protection of marine diversity and national defense security. The development of deep learning provides new opportunities for UATR, but faces challenges brought by the scarcity of reference samples and complex environmental interference. To tackle this problem, we propose a generative discriminative collaborative framework, a variational auto-encoder boosted learning framework based on latent space completion. Rooted in the core contradiction arising from the incompleteness of intra-class manifolds and the instability of discriminative boundaries, this framework incorporates the premise of latent space continuity. Leveraging a structure-preserving generative reconstruction mechanism, it implicitly supplements the original dataset, which in turn enables the reconstruction of intra-class distributions that are more continuous, integral, and discriminative at the feature level. In this paper, we construct a three-stage pipeline system consisting of auto-clean cut unified preprocessing, latent reconstruction variational auto-encoder multi-scale latent space reconstruction, and an acoustic identification model. Furthermore, by establishing a staged modeling workflow, data purification, latent space completion, and discriminative optimization converge on their individual objectives independently while maintaining overall synergy, thus forging a robust recognition paradigm tailored to few-shot learning scenarios.
Morphology modulation and its forefront implication of a three-dimensional particulate-fluid system by weaving acoustic field.
Qiao Z
J Acoust Soc Am · 2026 May · PMID 42065401 · Publisher ↗
Relied on three-dimensional metamaterial-based particulate-fluid system, an acoustic weave platform increases the sound pressure amplitude by frequency bandgap peak gain (Q-factor 4.5) in acoustically regular air media s...
Relied on three-dimensional metamaterial-based particulate-fluid system, an acoustic weave platform increases the sound pressure amplitude by frequency bandgap peak gain (Q-factor 4.5) in acoustically regular air media system with ways that a conventional method cannot. The interesting morphology alternations of the aggregation, fluidization, and trapping for numerous expanded-polystyrene particles (1-7.5 mm) were experimentally observed by engineering acoustic field in the low-frequency range of <0.8 kHz, improving the weak phenomenon in the absence of acoustic-metamaterial design. With vertical square-waveguide arrayed uniformly 12 of Helmholtz sound sources, the platform modulates the acoustic wave-packet movement and amplifies resonantly the time-spatial Y-shape-bifurcated-aggregation 54.7°-long-short-range-attraction wave phenomenon of complex macro soft-matter particles. Through experiment coinciding with simulation and theory, the main behaviors' phenomena were accurately explained by acoustic radiation force and secondary radiation force joint with the modulated three-dimensional acoustic field. The particle fluidization and trapping occur on contrary acoustic gradient fields at 220 and 250 Hz, respectively. There exist several vertically parallel "chiral" layer thin-film-aggregation stripes of millimeter-scale particles also obviously appearing at 220 Hz, more intuitively displaying the quasi-waves' constructive and destructive interferences of mm-scale particles themselves for wave-particle duality theory. The macro wavelike character helps to conveniently modulate collectively the environmental behaviors of fly-ash.
A systematic review of noise pollution in hospitals.
Pazani-Kohaki M, Yazdanfar SA, Lohrasbi L
J Acoust Soc Am · 2026 Apr · PMID 42053984 · Publisher ↗
The issue of noise pollution in hospitals has been discussed since 1851. Numerous studies have examined sound levels and sources of noise in hospitals, revealing that noise levels in hospitals often exceed the recommende...
The issue of noise pollution in hospitals has been discussed since 1851. Numerous studies have examined sound levels and sources of noise in hospitals, revealing that noise levels in hospitals often exceed the recommended standards from the World Health Organization (WHO). Noise pollution in hospitals has psychological and physical consequences for patients and staff. Identifying and implementing noise-reduction strategies in hospitals significantly improves acoustic conditions in these settings. This systematic review aims to identify and assess noise-generating sources across various hospital departments and the components that reduce noise in hospitals. Utilizing PRISMA guidelines (the Prisma checklist consists of 27 items related to the content of a systematic review and meta-analysis), data were gathered from five databases: Scopus database, Web of Science database, ScienceDirect database, Sage database, and Willey database. A total of 72 articles, dated between 2012 and 2024, were reviewed. Noise levels were recorded at 61-66 dB in intensive care units and 63 dB in inpatient wards. Most interventions for noise reduction have been managerial, with fewer engineering-based solutions. While interventions generally led to noise reduction, levels still did not meet WHO standards. Analysis of the articles identified patient-staff conversations and medical equipment alarms as the most frequently reported noise sources. Based on an extensive classification framework, noise-reducing components were grouped into six categories: noise management, equipment, materials, functional space, furniture, and cultural patterns, under three strategies: managerial, physical, and cultural.
Can adults and children use phonotactic knowledge to identify spectrally-degraded speech?
Mettler HM, Simeon KM, McCallick CR … +1 more , Grieco-Calub TM
J Acoust Soc Am · 2026 Apr · PMID 42048059 · Publisher ↗
Word recognition is facilitated by a listener's phonotactic knowledge-rules that govern how speech sounds combine to form words in language. Spectrally degrading speech disrupts the representations of speech sounds, howe...
Word recognition is facilitated by a listener's phonotactic knowledge-rules that govern how speech sounds combine to form words in language. Spectrally degrading speech disrupts the representations of speech sounds, however, making it more difficult to recognize words. Because phonotactic knowledge increases with language experience, individuals with more mature language may be better able to leverage phonotactic knowledge when attempting to recognize degraded speech. In this study, we tested the hypotheses that (1) adults are more sensitive to phonotactic information during word recognition than children; (2) children with larger vocabularies are more sensitive to phonotactic information; and (3) spectral degradation disrupts sensitivity to phonotactic cues during word recognition. To test these ideas, 36 adults and 36 school-aged children completed English Decision and Lexical Decision tasks. Both were presented with high-fidelity (unprocessed) and 16-channel noiseband vocoded (degraded) speech. Adults were more accurate and faster than children in the tasks. Performance declined in both groups when the speech stimuli were degraded. Among child participants, those with larger vocabularies were faster at making lexical decisions. The results provide insight into how adults and children leverage phonotactic information when processing speech and how those strategies are influenced when speech is degraded.
A modified angular spectrum method for rapid on-demand design of axisymmetric acoustic metalenses.
Wang S, Xiang X, Wang L … +1 more , Huang Y
J Acoust Soc Am · 2026 Apr · PMID 42048058 · Publisher ↗
The angular spectrum method (ASM) is rarely employed directly in the rapid design of acoustic metalenses, due to significant aliasing artifacts in long-distance diffraction calculations. For axisymmetric systems, althoug...
The angular spectrum method (ASM) is rarely employed directly in the rapid design of acoustic metalenses, due to significant aliasing artifacts in long-distance diffraction calculations. For axisymmetric systems, although ASM based on radial transforms holds promise for dimensionality reduction and acceleration, its application in on-demand focusing design remains underdeveloped. This paper presents a modified ASM that significantly enhances both the speed and accuracy of diffraction simulations. Compared to the finite-element method as the accuracy benchmark, the modified method reduces the relative error by a factor of 2-3 and is approximately 100 times faster than the direct Hankel transform ASM. Additionally, it produces results consistent with the Rayleigh-Sommerfeld integral, with a 95 times speed increase, enabling millisecond-scale computation at a grid size of λ/15 on a standard desktop. By integrating the method with an evolutionary algorithm, this method efficiently optimized binary acoustic metalenses and experimentally demonstrated complex focal shapes. This approach provides a practical, cost-effective, and efficient solution for the rapid on-demand design of axisymmetric acoustic focusing devices.
An active control approach for improvement of sound insulation performance of a double-panel structure.
Ning S, Jin Y, Jiang H … +2 more , Chu D, Liu Z
J Acoust Soc Am · 2026 Apr · PMID 42048057 · Publisher ↗
Double-panel structures (DPSs) have wide applications in noise control engineering, but their sound insulation (SI) performances deteriorate rapidly in the low-frequency range. To make up for this deficiency, an active c...
Double-panel structures (DPSs) have wide applications in noise control engineering, but their sound insulation (SI) performances deteriorate rapidly in the low-frequency range. To make up for this deficiency, an active control strategy through controlling the boundary conditions of the sound field is presented to improve the SI performance of DPSs. A controllable plate as a controllable boundary condition is incorporated into the boundaries of the interlayer sound field (ISF), and it can be actively controlled by applying control forces. The control forces are optimized by three control objectives, i.e., minimization of kinetic energy of the radiation panel, minimization of acoustical potential energy of the ISF, and minimization of radiation sound power of the low-order acoustic radiation modes. The control effects, control mechanism, and the influences of the geometrical parameters, number of controllable plates, and location of control force on the SI performance are investigated. The results indicate that in the low-frequency range, all three active control strategies can enhance the SI performance of a DPS. And the modal response calculations indicate that the control mechanism belongs to the acoustical and structural modal suppression and rearrangement. This research provides a solution for improving the low-frequency SI performance of engineering structures.
Improving low-frequency attenuation of passive earplugs using Helmholtz resonators.
Carillo K, Sgard F, Dazel O … +1 more , Doutres O
J Acoust Soc Am · 2026 Apr · PMID 42048056 · Publisher ↗
Earplugs are widely used in occupational and recreational settings to prevent noise-induced hearing loss. However, their effectiveness in real-world conditions is often limited by inconsistent use, imperfect fit, and int...
Earplugs are widely used in occupational and recreational settings to prevent noise-induced hearing loss. However, their effectiveness in real-world conditions is often limited by inconsistent use, imperfect fit, and intrinsic acoustic limitations, particularly in attenuating low-frequency sounds. Improving low-frequency attenuation is critical not only for increasing overall hearing protection but also for achieving a more balanced attenuation profile, which can enhance perceived sound quality and speech intelligibility. In this study, we investigate how the reflection coefficient at the earplug's medial surface facing the ear canal cavity influences low-frequency noise reduction (NR) in passive earplugs and how passive design can optimize it. From an analytical NR model, we derive the exact condition for maximum attenuation: a reflected wave in anti-phase with the incident wave, producing destructive interference within the occluded ear canal. Using meta-earplugs incorporating three Helmholtz resonators, we demonstrate, on both an acoustic test fixture and human participants, that inducing near anti-phase or quadrature-phase reflections increases low-frequency attenuation by up to 15 dB below 1 kHz. In addition, this passive strategy remains effective despite moderate acoustic leakage. Originally designed to reduce the occlusion effect, these meta-earplugs also show strong potential for enhancing low-frequency attenuation, thereby enabling more efficient and robust hearing protection.
An adaptive deep neural network for active road noise control.
Bai L, Xue JP, Lian SY … +5 more , He YM, Liang YY, Rao L, Wang SP, Lu J
J Acoust Soc Am · 2026 Apr · PMID 42041264 · Publisher ↗
Active road noise control (ARNC) systems have been widely used for low-frequency noise control in vehicle cabins. To address nonlinear distortions in ARNC systems, prior work introduced a fully causal deep neural network...
Active road noise control (ARNC) systems have been widely used for low-frequency noise control in vehicle cabins. To address nonlinear distortions in ARNC systems, prior work introduced a fully causal deep neural network (DNN)-based ANC framework [WaveNet-Volterra Neural Network (WaveNet-VNN)], which achieves superior performance over the ideal Wiener solution and conventional adaptive algorithms under rigorous and fair comparisons. However, real-world road noise exhibits significant distribution shifts caused by varying road conditions, traffic density, and vehicle speed, limiting the generalization of DNN-based ANC methods in practical ARNC systems, especially with scarce training data. This paper presents an adaptive neural network that integrates an online adaptation mechanism while preserving the high performance of causal neural networks. Specifically, lightweight Adapter modules are inserted into several layers of the pre-trained WaveNet-VNN model to serve as fine-tuning components during online adaptation, increasing model parameters, and computational cost by only 5%. Experimental results on a 42 × 2 × 2 ARNC system demonstrate that the proposed approach outperforms the ideal Wiener solution under cross-day distribution shifts and consistently delivers superior performance across different vehicle speeds. It also achieves convergence speed comparable to state-of-the-art traditional algorithms while demonstrating strong robustness across various scenarios, providing a feasible and effective solution for real-world DNN-based ARNC.
Underwater Doppler-bearing pulse source tracking based on an outlier-tolerant variational Bayesian adaptive Kalman filter.
Ye Y, Wang X, Cao H … +3 more , An L, Fan Q, Huang Z
J Acoust Soc Am · 2026 Apr · PMID 42041263 · Publisher ↗
Tracking underwater non-cooperative targets with Doppler-bearing target motion analysis relies on pulse signal parameters estimated via sonar arrays. However, successfully implementing a Kalman filter for this purpose re...
Tracking underwater non-cooperative targets with Doppler-bearing target motion analysis relies on pulse signal parameters estimated via sonar arrays. However, successfully implementing a Kalman filter for this purpose requires key parameters, including the noise covariance matrices, which cannot be known a priori in practical scenarios. A more critical challenge is that uncertainties in the marine environment and target-array geometry corrupt the parameter estimates with measurement outliers. Existing nonlinear variational Bayesian (VB) filters are not robust to this issue, as their fundamental reliance on Gaussian models and deterministic sampling degrades their convergence rate and tracking accuracy. For high-precision tracking under these conditions, this paper proposes an outlier-tolerant nonlinear VB adaptive Kalman filter that utilizes a hierarchical inverse-Wishart-gamma mixture distribution model to robustly identify outliers and more accurately approximate the measurement noise covariance matrix, while also employing the adaptive high-order cubature sampling method to improve the estimation accuracy of the expectation. The validity of this combined method is verified via rigorous numerical simulations and sea trials. Simulation results highlight the performance superiority of the proposed filter over traditional approaches utilizing inverse Wishart noise modeling and deterministic integral sampling. Furthermore, superior performance on sea trial data confirmed the filter's effectiveness and robustness.
Self-supervised speech quality assessment (S3QA): Leveraging speech foundation models for a scalable speech quality metric.
Ogg M, Bishop CA, Yi HG … +1 more , Robinson SR
J Acoust Soc Am · 2026 Apr · PMID 42041262 · Publisher ↗
Methods for automatically assessing speech quality in real world environments are critical for robust human language technologies and assistive devices. Behavioral ratings (e.g., mean opinion scores) are considered the g...
Methods for automatically assessing speech quality in real world environments are critical for robust human language technologies and assistive devices. Behavioral ratings (e.g., mean opinion scores) are considered the gold standard, but are susceptible to inter-rater variability, cannot be generalized, and are labor-intensive, thus limiting the acoustic challenges they can quantify. We present a scalable method for assessing speech quality: the self-supervised speech quality assessment model. First, we manipulated high-quality utterances, using acoustic challenges that emulate real-world degradation: filtering, reverberation, background noise, and digital compression. Second, we leveraged a pre-trained speech foundation model, WavLM, to derive cosine distances between the clean and degraded versions of each utterance in the embedding space. A transformer-based model was trained to predict these cosine distances, given only the degraded utterances. The trained model was evaluated on unseen corpora of synthetic mixtures, NISQA, and VOiCES. The self-supervised speech quality assessment model accurately predicts degradation cosine distances across a wide range of acoustic challenges and is aligned with behavioral ratings (mean opinion scores), automatic speech recognition performance and other important features (microphone distances). The model is available online.
Broadband noise reduction under grazing flow via coupled interference and resonance mechanisms.
Yan S, Tang L, Wu F … +3 more , Tong X, Huang L, Luo J
J Acoust Soc Am · 2026 Apr · PMID 42030181 · Publisher ↗
Achieving stable broadband noise reduction under grazing flow using a structurally simple configuration remains a challenge. To address this issue, a broadband noise control design based on the coupling of interference a...
Achieving stable broadband noise reduction under grazing flow using a structurally simple configuration remains a challenge. To address this issue, a broadband noise control design based on the coupling of interference and resonance mechanisms is proposed. In this study, side branch ducts generate destructive interference, whereas internally extended neck resonators provide resonant damping. The acoustic behavior of the structure was elucidated through analysis of the flow field and phase distribution using the k-ε turbulence model and the linearized Navier-Stokes equations. An analytical model formulated using the transfer matrix method was developed and validated against numerical simulations and experimental measurements conducted in a custom-designed grazing flow duct facility. The results demonstrate that the broadband performance originates from the interaction between interference and resonance effects. Specifically, the side-branch ducts generate stable broadband destructive interference that remains robust under grazing flow, whereas the extended-neck resonators introduce resonant dissipation at discrete frequencies. Notably, at a grazing flow velocity of 60 m/s and a sound pressure level of 140 dB, noise reduction is maintained over the frequency range of 700-2000 Hz. The proposed strategy therefore provides a practical design framework for broadband flow-noise control, with potential for implementation in complex duct systems.
Three-dimensional tracking using one-dimensional bearings from multiple horizontal linear arrays under isogradient sound speed profilea).
Zhao Y, Zheng C, Sun F … +2 more , Zhou F, Qiao G
J Acoust Soc Am · 2026 Apr · PMID 42030180 · Publisher ↗
Three-dimensional tracking that relies solely on one-dimensional bearing measurements from multiple horizontal linear arrays is particularly challenging in underwater environments. Depth-dependent variations in sound spe...
Three-dimensional tracking that relies solely on one-dimensional bearing measurements from multiple horizontal linear arrays is particularly challenging in underwater environments. Depth-dependent variations in sound speed cause acoustic rays to bend, leading to deviations in the incident bearing angles. Moreover, the relatively low sound speed in water makes the propagation delay non-negligible, a bearing angle measured at the current time instant typically corresponds to an earlier target state. To address these issues, we formulate a bearing measurement model under an isogradient sound speed profile that jointly accounts for ray bending and propagation delay. Bearing angles measured by multiple arrays at a common time correspond to target states at different emission times, which are coupled with unknown propagation delays through an implicit constraint. To handle this coupling, we propose a centralized particle filter with an iterative procedure to jointly estimate the propagation delay and the target state at the emission time for each array. The resulting emission-time state estimates are then marginalized to obtain the posterior estimate of the target state at the common measurement time. Finally, we derive the posterior Cramér-Rao lower bound to provide a performance benchmark. Simulation studies and sea-trial data demonstrate the effectiveness of the proposed approach.
Non-negative intensity formulations for stochastically excited finite structures.
Kha J, Karimi M, Maxit L
J Acoust Soc Am · 2026 Apr · PMID 42030179 · Publisher ↗
Non-negative intensity (NNI) has been previously proposed for identifying the surface areas of a vibrating structure that most significantly contributes to the sound power. Formulations of finite radiating structures und...
Non-negative intensity (NNI) has been previously proposed for identifying the surface areas of a vibrating structure that most significantly contributes to the sound power. Formulations of finite radiating structures under a deterministic load using the boundary element method and infinite radiating structures excited by a stochastic excitation have been developed in the literature. However, these formulations cannot permit the calculation of the NNI of a finite structure excited by a stochastic excitation, which has engineering applications for structures exposed to flow-induced vibrations or machinery noise, typically represented by a turbulent boundary layer excitation or diffuse acoustic field. To fill this gap, two formulations are proposed: one that is an analytical approach based on the wavenumber space formulation and another that is useful for numerical approaches using the realizations of deterministic wall pressure fields. Numerical results for both baffled and unbaffled finite panels under a turbulent boundary layer excitation are presented and compared against a relevant previous investigation.
Erratum: Mathematical modeling of vowel perception by users of analog multichannel cochlear implants: Temporal and channel-amplitude cues. [J. Acoust. Soc. Am. 107(3), 1521-1529 (2000)].
Svirsky MA
J Acoust Soc Am · 2026 Apr · PMID 42023945 · Publisher ↗
Abstract loading — click title to view on PubMed.
Spectrogram-derived graphs and inductive learning for multi-label avian vocalization detection in field recordings.
A N, Rajan R
J Acoust Soc Am · 2026 Apr · PMID 42023944 · Publisher ↗
This paper presents a methodology that employs inductive spatial geometric deep learning networks to detect multiple avian vocalizations from field recordings. Initially, a graph is constructed from the Mel-spectrogram o...
This paper presents a methodology that employs inductive spatial geometric deep learning networks to detect multiple avian vocalizations from field recordings. Initially, a graph is constructed from the Mel-spectrogram of each audio file using a trained deep convolutional neural network (Deep CNN). The extracted features are used to build a node-feature graph, which is then processed by two spatial inductive graph-based models: graph sample and aggregation (GraphSAGE) and the graph attention network (GAT), for multi-label classification. To enhance the robustness and generalization of the Deep CNN, SpecAugment is applied to generate additional Mel-spectrograms via data augmentation. The proposed framework is evaluated on the Xeno-canto bird sound database and compared against state-of-the-art methods. The results demonstrate that the proposed inductive spatial graph-based approach outperforms existing techniques, achieving macro F1-scores of 0.90 with GraphSAGE and 0.92 with GAT. We further replaced Deep CNN with AudioProtoPNet-20 and evaluated GAT on the Xeno-canto dataset, obtaining a macro F1-score of 0.93.
About
- Frequency
- Sun
- Papers found
- 200
- RSS feed
- Subscribe