Applied Acoustics最新文献_第4页

Going deeper with log-graph Fourier transform-based feature extraction for playback speech detection 更深入地研究基于对数图傅立叶变换的特征提取，用于回放语音检测

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2026-01-02 DOI: 10.1016/j.apacoust.2025.111212

Tingting Wang , Marvin Borsdorf , Qiquan Zhang , Longting Xu , Xi Shao

Replay spoofing speech poses a great concern to automatic speaker verification (ASV) systems, with the advent of high-quality recording and playback devices. Recent studies have explored graph Fourier transform (GFT)-based features and graph frequency cepstral coefficients for replay speech detection. However, these methods often suffer from unstable graph spectral differences between genuine and replayed speech, caused by sensitivity in the graph Fourier basis (i.e., GFT with eigenvalue decomposition). This instability limits the effectiveness of graph frequency cepstral coefficient features for reliable replay attack detection. To address this, we propose constructing robust undirected graph topologies using an exponential function and graph shift operator. Based on these graph topologies, we derive stable graph Laplacian matrices with singular value decomposition to define the logarithmic graph Fourier transform and logarithmic joint graph Fourier transform. These enable extraction of enhanced (joint) graph frequency cepstral coefficient features for replay attack detection. Experimental results on the ASVspoof dataset show that ASV systems utilizing our (joint) graph frequency cepstral coefficient features significantly outperform current state-of-the-art methods on the ASVspoof datasets. We release our source code at: https://github.com/Wangfighting0015/GFT_project.

随着高质量录音和重放设备的出现，语音重放欺骗成为自动说话人验证（ASV）系统关注的焦点。最近的研究探索了基于图傅里叶变换（GFT）特征和图频率倒谱系数的重放语音检测。然而，这些方法在真实语音和重放语音之间往往存在不稳定的图谱差异，这是由于图傅里叶基（即带有特征值分解的GFT）的敏感性造成的。这种不稳定性限制了图频倒谱系数特征在可靠的重放攻击检测中的有效性。为了解决这个问题，我们提出使用指数函数和图移位算子构造鲁棒无向图拓扑。在此基础上，利用奇异值分解导出稳定的图拉普拉斯矩阵，定义对数图傅里叶变换和对数联合图傅里叶变换。这使得提取增强（联合）图频率倒谱系数特征用于重放攻击检测。在ASVspoof数据集上的实验结果表明，利用我们的（联合）图频率倒谱系数特征的ASV系统在ASVspoof数据集上的性能明显优于当前最先进的方法。我们的源代码发布在：https://github.com/Wangfighting0015/GFT_project。

{"title":"Going deeper with log-graph Fourier transform-based feature extraction for playback speech detection","authors":"Tingting Wang , Marvin Borsdorf , Qiquan Zhang , Longting Xu , Xi Shao","doi":"10.1016/j.apacoust.2025.111212","DOIUrl":"10.1016/j.apacoust.2025.111212","url":null,"abstract":"<div><div>Replay spoofing speech poses a great concern to automatic speaker verification (ASV) systems, with the advent of high-quality recording and playback devices. Recent studies have explored graph Fourier transform (GFT)-based features and graph frequency cepstral coefficients for replay speech detection. However, these methods often suffer from unstable graph spectral differences between genuine and replayed speech, caused by sensitivity in the graph Fourier basis (i.e., GFT with eigenvalue decomposition). This instability limits the effectiveness of graph frequency cepstral coefficient features for reliable replay attack detection. To address this, we propose constructing robust undirected graph topologies using an exponential function and graph shift operator. Based on these graph topologies, we derive stable graph Laplacian matrices with singular value decomposition to define the logarithmic graph Fourier transform and logarithmic joint graph Fourier transform. These enable extraction of enhanced (joint) graph frequency cepstral coefficient features for replay attack detection. Experimental results on the ASVspoof dataset show that ASV systems utilizing our (joint) graph frequency cepstral coefficient features significantly outperform current state-of-the-art methods on the ASVspoof datasets. We release our source code at: <span><span>https://github.com/Wangfighting0015/GFT_project</span><svg><path></path></svg></span>.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111212"},"PeriodicalIF":3.4,"publicationDate":"2026-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884077","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Urban noise pollution prediction using traffic patterns and AI models in Sahloul Road, Sousse City, Tunisia 突尼斯苏塞市Sahloul路使用交通模式和人工智能模型进行城市噪音污染预测

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-29 DOI: 10.1016/j.apacoust.2025.111218

Najah Kechiche , Jawher Bouaziz , Nader Boumrifeg , Walid Hassen , Tarek Salem Abdennaji , Chemseddine Maatki , Badr M. Alshammari , Lioua Kolsi

This study evaluates multiple machine learning and deep learning approaches for forecasting A-weighted equivalent continuous sound levels (LAeq) and investigates spatiotemporal fluctuations in road traffic noise at the Sahloul Road in Sousse City, Tunisia. Road traffic noise levels (dB)) were measured at five urban sites with differing traffic intensities using TES 1352A sound level meter, with data collected at both hourly and 15-minute intervals. Classification of traffic composition into motorcycles, light vehicles, and heavy vehicles is carried out using the video-monitored traffic count. Results show that heavy vehicles and total traffic volume present the strongest correlation with noise levels. Maximum noise occurred during the morning (07:00–09:00) and evening (16:00–18:00) hours, which exceeded the World Health Organization (WHO) guideline by more than 50 % for the entire measurement study. Four machine learning algorithms (XGBoost, Random Forest, LightGBM, and LSTM) are applied, utilizing vehicle counts as factors indicating traffic volume for the prediction of LAeq. On an hourly timescale, the performance of the XGBoost algorithm was best (R2 = 0.952, MAE = 0.18 dB). However, the performance of the algorithm decreased progressively for smaller temporal resolutions, which showed a marked difference among the two timescales, indicating higher variability of noise at the smaller timescale. As such, the LSTM algorithm indicated poor performance, specifically at the temporal resolutions (15 min). this study’s findings have brought out the effectiveness of ensemble tree-based methods for predicting urban noise levels, as well as the importance of considering the time resolution when structuring measures for reducing urban noise pollution.

本研究评估了用于预测a加权等效连续声级（LAeq）的多种机器学习和深度学习方法，并调查了突尼斯苏塞市Sahloul路道路交通噪声的时空波动。使用TES 1352A声级计在五个不同交通强度的城市地点测量道路交通噪音水平（dB），每隔一小时和15分钟收集一次数据。利用视频监控的交通统计，将交通构成分为摩托车、轻型车辆和重型车辆。结果表明，重型车辆和总交通量与噪声水平的相关性最强。最大噪音发生在早上（07:00-09:00）和晚上（16:00-18:00），超过了整个测量研究中世界卫生组织（WHO）指南的50%以上。应用了四种机器学习算法（XGBoost、Random Forest、LightGBM和LSTM），利用车辆计数作为指示交通量的因素来预测LAeq。在小时尺度上，XGBoost算法的性能最好（R2 = 0.952, MAE = 0.18 dB）。然而，在较小的时间分辨率下，算法的性能逐渐下降，这在两个时间尺度之间表现出明显的差异，表明在较小的时间尺度下噪声的变异性更高。因此，LSTM算法表现出较差的性能，特别是在时间分辨率（15分钟）下。本研究的结果表明了基于集合树的方法预测城市噪声水平的有效性，以及在制定减少城市噪声污染的措施时考虑时间分辨率的重要性。

{"title":"Urban noise pollution prediction using traffic patterns and AI models in Sahloul Road, Sousse City, Tunisia","authors":"Najah Kechiche , Jawher Bouaziz , Nader Boumrifeg , Walid Hassen , Tarek Salem Abdennaji , Chemseddine Maatki , Badr M. Alshammari , Lioua Kolsi","doi":"10.1016/j.apacoust.2025.111218","DOIUrl":"10.1016/j.apacoust.2025.111218","url":null,"abstract":"<div><div>This study evaluates multiple machine learning and deep learning approaches for forecasting A-weighted equivalent continuous sound levels (LAeq) and investigates spatiotemporal fluctuations in road traffic noise at the Sahloul Road in Sousse City, Tunisia. Road traffic noise levels (dB)) were measured at five urban sites with differing traffic intensities using TES 1352A sound level meter, with data collected at both hourly and 15-minute intervals. Classification of traffic composition into motorcycles, light vehicles, and heavy vehicles is carried out using the video-monitored traffic count. Results show that heavy vehicles and total traffic volume present the strongest correlation with noise levels. Maximum noise occurred during the morning (07:00–09:00) and evening (16:00–18:00) hours, which exceeded the World Health Organization (WHO) guideline by more than 50 % for the entire measurement study. Four machine learning algorithms (XGBoost, Random Forest, LightGBM, and LSTM) are applied, utilizing vehicle counts as factors indicating traffic volume for the prediction of LAeq. On an hourly timescale, the performance of the XGBoost algorithm was best (R2 = 0.952, MAE = 0.18 dB). However, the performance of the algorithm decreased progressively for smaller temporal resolutions, which showed a marked difference among the two timescales, indicating higher variability of noise at the smaller timescale. As such, the LSTM algorithm indicated poor performance, specifically at the temporal resolutions (15 min). this study’s findings have brought out the effectiveness of ensemble tree-based methods for predicting urban noise levels, as well as the importance of considering the time resolution when structuring measures for reducing urban noise pollution.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111218"},"PeriodicalIF":3.4,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884046","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Quantification of spectral shift and broadening induced by out-of-band modes in the transient process of room sound field 室内声场瞬态过程中带外模式引起的谱移和频宽的量化

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-29 DOI: 10.1016/j.apacoust.2025.111210

Yoshinari Yamada

The spectral structures of sound fields in rooms differ between steady-state conditions and transient processes. This study presents a methodology to enable analysis of the spectral characteristics of the transient processes of room sound fields. The energy spectra appearing during the decay and growth processes for sinusoidal and white noise excitations are formulated by redefining energy-time functions in reference to previous findings. The derived formulae present a unified approach for the interrupted sine and noise methods. The energy spectrum is used to examine the spectral shift and broadening induced by out-of-band modes during transient processes. Frequency statistics, spectral bandwidth, and band energy ratios are introduced as physical indices. The proposed method was applied to analyze the decay process of actual room sound fields following preliminary examinations using room impulse responses generated by numerical simulations. The results indicate that the effect of out-of-band modes was insignificant when the product of the excitation bandwidth and reverberation time exceeded 18 and 36, respectively, in cases with and without acoustic modal overlap. Band limiting of the received signal may not always be necessary under these conditions. In concert halls, this condition was satisfied even when the excitation bandwidth was a 1/12 octave. Thus, the physical indices introduced in this study can characterize sound fields in different rooms from an unconventional perspective.

室内声场的谱结构在稳态和瞬态过程中是不同的。本研究提出了一种分析室内声场瞬态过程频谱特性的方法。参考前人的发现，通过重新定义能量-时间函数，给出了正弦和白噪声激发在衰减和生长过程中出现的能谱。导出的公式为中断正弦法和噪声法提供了统一的方法。利用能谱分析了瞬态过程中带外模式引起的谱移和谱宽。引入频率统计、频谱带宽和频带能量比作为物理指标。利用数值模拟产生的房间脉冲响应，对实际房间声场的衰减过程进行了初步分析。结果表明，在有声模态重叠和无声模态重叠的情况下，当激励带宽和混响时间的乘积分别超过18和36时，带外模态的影响不显著。在这些条件下，接收信号的频带限制可能并不总是必要的。在音乐厅中，即使激励带宽为1/12倍频程，也满足这一条件。因此，本研究中引入的物理指标可以从一个非常规的角度来表征不同房间的声场。

{"title":"Quantification of spectral shift and broadening induced by out-of-band modes in the transient process of room sound field","authors":"Yoshinari Yamada","doi":"10.1016/j.apacoust.2025.111210","DOIUrl":"10.1016/j.apacoust.2025.111210","url":null,"abstract":"<div><div>The spectral structures of sound fields in rooms differ between steady-state conditions and transient processes. This study presents a methodology to enable analysis of the spectral characteristics of the transient processes of room sound fields. The energy spectra appearing during the decay and growth processes for sinusoidal and white noise excitations are formulated by redefining energy-time functions in reference to previous findings. The derived formulae present a unified approach for the interrupted sine and noise methods. The energy spectrum is used to examine the spectral shift and broadening induced by out-of-band modes during transient processes. Frequency statistics, spectral bandwidth, and band energy ratios are introduced as physical indices. The proposed method was applied to analyze the decay process of actual room sound fields following preliminary examinations using room impulse responses generated by numerical simulations. The results indicate that the effect of out-of-band modes was insignificant when the product of the excitation bandwidth and reverberation time exceeded 18 and 36, respectively, in cases with and without acoustic modal overlap. Band limiting of the received signal may not always be necessary under these conditions. In concert halls, this condition was satisfied even when the excitation bandwidth was a 1/12 octave. Thus, the physical indices introduced in this study can characterize sound fields in different rooms from an unconventional perspective.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111210"},"PeriodicalIF":3.4,"publicationDate":"2025-12-29","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884083","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Predicting the age effects on concurrent vowel scores using a temporal jitter computational model 使用时间抖动计算模型预测年龄对并发元音分数的影响

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-27 DOI: 10.1016/j.apacoust.2025.111215

Harshavardhan Settibhaktini , Rithik Rathi , Ananthakrishna Chintanpalli

The difference in fundamental frequency (F0) among vowels is a crucial cue for detecting concurrent vowels. Normal hearing (NH) listeners have higher percent identification scores for both vowels as the F0 difference increased, reaching an asymptote at ∼ 3 Hz. In complex listening environments, aged hearing (AH) listeners exhibit a reduction in the overall concurrent vowel scores across F0 differences, which may be related to age-related loss of neural synchrony in the auditory system. To understand these age effects, the current modeling study predicts the concurrent vowel scores across F0 differences for both NH and AH subjects. The NH model used the neural responses of an auditory-nerve (AN) model with the neurogram similarity index measure (NSIM) metric, instead of the previous F0-guided segregation, to predict concurrent vowel scores. Previous behavioral studies have shown that temporal jitter in the acoustic domain can cause an age-related decrease of neural synchrony, resulting in reduced identification scores. Thus, a temporal-jitter concurrent vowel was implemented in the AH model to obtain the neurograms from the AN model. The NSIM metric was applied on these neurograms to predict the concurrent vowel scores of AH subjects. Both models qualitatively predicted the concurrent vowel scores pattern across F0 differences, obtained from the previous behavioral data. A chi-square test analysis has shown that the scores correlated well with the concurrent vowel data. The temporal-jitter AH model predictions showed neural asynchrony, reducing the concurrent vowel scores among AH listeners. These model predictions suggest that the neural asynchrony in the temporal-jitter AH model might contribute to the reduced concurrent vowel scores across F0 differences.

元音之间基频（F0）的差异是检测并发元音的关键线索。正常听力（NH）听者对两个元音的识别分数百分比随着F0差异的增加而增加，在~ 3hz处达到渐近线。在复杂的听力环境中，老年听力（AH）听者在F0差异中表现出整体并发元音得分的降低，这可能与听觉系统中与年龄相关的神经同步性丧失有关。为了理解这些年龄影响，目前的建模研究预测了NH和AH受试者在F0差异中的并发元音得分。NH模型使用听觉神经（an）模型的神经反应与神经图相似指数测量（NSIM）度量，而不是之前的f0引导分离，来预测并发元音得分。先前的行为研究表明，声域的时间抖动会导致与年龄相关的神经同步性下降，从而导致识别分数下降。因此，在AH模型中实现了一个时间抖动并发元音，从而从AN模型中获得神经图。在这些神经图上应用NSIM度量来预测AH受试者的并发元音得分。这两个模型都定性地预测了从先前的行为数据中获得的跨F0差异的并发元音得分模式。卡方检验分析表明，这些分数与并发元音数据有很好的相关性。时间抖动AH模型预测显示神经不同步，降低了AH听众的并发元音得分。这些模型预测表明，时间抖动AH模型中的神经异步性可能导致F0差异中并发元音分数的降低。

{"title":"Predicting the age effects on concurrent vowel scores using a temporal jitter computational model","authors":"Harshavardhan Settibhaktini , Rithik Rathi , Ananthakrishna Chintanpalli","doi":"10.1016/j.apacoust.2025.111215","DOIUrl":"10.1016/j.apacoust.2025.111215","url":null,"abstract":"<div><div>The difference in fundamental frequency (F0) among vowels is a crucial cue for detecting concurrent vowels. Normal hearing (NH) listeners have higher percent identification scores for both vowels as the F0 difference increased, reaching an asymptote at ∼ 3 Hz. In complex listening environments, aged hearing (AH) listeners exhibit a reduction in the overall concurrent vowel scores across F0 differences, which may be related to age-related loss of neural synchrony in the auditory system. To understand these age effects, the current modeling study predicts the concurrent vowel scores across F0 differences for both NH and AH subjects. The NH model used the neural responses of an auditory-nerve (AN) model with the neurogram similarity index measure (NSIM) metric, instead of the previous F0-guided segregation, to predict concurrent vowel scores. Previous behavioral studies have shown that temporal jitter in the acoustic domain can cause an age-related decrease of neural synchrony, resulting in reduced identification scores. Thus, a temporal-jitter concurrent vowel was implemented in the AH model to obtain the neurograms from the AN model. The NSIM metric was applied on these neurograms to predict the concurrent vowel scores of AH subjects. Both models qualitatively predicted the concurrent vowel scores pattern across F0 differences, obtained from the previous behavioral data. A chi-square test analysis has shown that the scores correlated well with the concurrent vowel data. The temporal-jitter AH model predictions showed neural asynchrony, reducing the concurrent vowel scores among AH listeners. These model predictions suggest that the neural asynchrony in the temporal-jitter AH model might contribute to the reduced concurrent vowel scores across F0 differences.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111215"},"PeriodicalIF":3.4,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Tailoring vibration modes in piezoelectric metamaterial beams for sound directivity control 用于声指向性控制的压电超材料梁的振动模式裁剪

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-27 DOI: 10.1016/j.apacoust.2025.111214

Camila Sanches Schimidt , Leopoldo Pisanelli Rodrigues de Oliveira , Virgilio Junior Caetano , Carlos De Marqui Junior

This work presents a reconfigurable piezoelectric metamaterial beam designed to manipulate acoustic radiation and control sound directivity through tailored vibrational modes. The proposed architecture enables the induction of specific modes at arbitrary frequencies via an electromechanical coupling mechanism. The beam incorporates periodically distributed piezoelectric elements, each connected to digitally programmable synthetic impedance shunts that allow adaptive tuning. The metamaterial is divided into three regions: a central section that amplifies a selected target mode shape and two side sections that operate within a programmed band gap to suppress bending waves. A modal-analysis-based modeling framework captures the combined effects of band gap tuning and mode induction, guiding the selection of shunt parameters to achieve the desired vibration patterns. Experimental measurements show excellent agreement with numerical predictions, confirming the model’s accuracy and the effectiveness of the reconfiguration strategy in reproducing similar vibration shapes at different frequencies. Coupled structural–acoustic simulations further demonstrate the steering of radiated sound energy by modulating the beam’s vibration profile. The results highlight the metamaterial’s ability for on-demand wave manipulation and adaptive sound field shaping, exhibiting similar directivity patterns at distinct frequencies. Overall, the findings establish a versatile platform for adaptive wave manipulation and sound directivity control.

这项工作提出了一种可重构的压电超材料梁，旨在通过定制的振动模式来操纵声辐射和控制声指向性。所提出的架构能够通过机电耦合机制在任意频率下感应特定模式。该光束包含周期性分布的压电元件，每个元件都连接到数字可编程合成阻抗分流器，允许自适应调谐。该超材料分为三个区域：中心部分放大选定的目标模式形状，两个侧部分在编程带隙内工作以抑制弯曲波。基于模态分析的建模框架捕获带隙调谐和模态感应的综合效应，指导分流参数的选择以实现所需的振动模式。实验结果与数值预测结果非常吻合，证实了模型的准确性和重构策略在不同频率下再现相似振动形状的有效性。结构声耦合仿真进一步证明了通过调制光束的振动剖面来控制辐射声能。结果强调了这种超材料在按需波操纵和自适应声场塑造方面的能力，在不同频率下表现出相似的指向性模式。总的来说，这些发现为自适应波操纵和声指向性控制建立了一个通用的平台。

{"title":"Tailoring vibration modes in piezoelectric metamaterial beams for sound directivity control","authors":"Camila Sanches Schimidt , Leopoldo Pisanelli Rodrigues de Oliveira , Virgilio Junior Caetano , Carlos De Marqui Junior","doi":"10.1016/j.apacoust.2025.111214","DOIUrl":"10.1016/j.apacoust.2025.111214","url":null,"abstract":"<div><div>This work presents a reconfigurable piezoelectric metamaterial beam designed to manipulate acoustic radiation and control sound directivity through tailored vibrational modes. The proposed architecture enables the induction of specific modes at arbitrary frequencies via an electromechanical coupling mechanism. The beam incorporates periodically distributed piezoelectric elements, each connected to digitally programmable synthetic impedance shunts that allow adaptive tuning. The metamaterial is divided into three regions: a central section that amplifies a selected target mode shape and two side sections that operate within a programmed band gap to suppress bending waves. A modal-analysis-based modeling framework captures the combined effects of band gap tuning and mode induction, guiding the selection of shunt parameters to achieve the desired vibration patterns. Experimental measurements show excellent agreement with numerical predictions, confirming the model’s accuracy and the effectiveness of the reconfiguration strategy in reproducing similar vibration shapes at different frequencies. Coupled structural–acoustic simulations further demonstrate the steering of radiated sound energy by modulating the beam’s vibration profile. The results highlight the metamaterial’s ability for on-demand wave manipulation and adaptive sound field shaping, exhibiting similar directivity patterns at distinct frequencies. Overall, the findings establish a versatile platform for adaptive wave manipulation and sound directivity control.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111214"},"PeriodicalIF":3.4,"publicationDate":"2025-12-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840712","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A multi-channel active road noise control system using incremental SVD based virtual references and local-clustered control strategy 基于增量奇异值分解和局部聚类控制策略的多通道道路噪声主动控制系统

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-24 DOI: 10.1016/j.apacoust.2025.111216

Ziheng Xia, Yansong He, Zhifei Zhang, Hao Chen, Xiaoyu Fu

Active road noise control (ARNC) systems use multiple acceleration signals from the chassis as the references to attenuate vehicle interior road noise. However, due to the coupling effect among the chassis structures, the predetermined reference signals cannot always ensure their independence and coherence to the interior noises. In addition, the computational complexity of the conventional centralized ARNC systems increase rapidly as the growth of the channels. To address the above issues, this study proposes a multi-channel ARNC system using virtual references and local-clustered control strategy. Firstly, a virtual reference method based on incremental singular value decomposition is developed to generate an orthogonal virtual reference set, which improves the system’s noise reduction performance. The conventional centralized system is then decomposed into two independent active headrest subsystems based on the local-clustered control strategy to reduce the computational complexity. Numerical simulations and on-board experiments are conducted to validate the effectiveness of the proposed ARNC system. The results demonstrate that the proposed system can effectively attenuate the interior road noises with lower computational effort compared to the conventional centralized ARNC system: when the vehicle is driving at the speed of 60 km/h, the overall noise reduction values improved by 1.5 dB(A), and when the vehicle is driving at the speed of 80 km/h, the values improved by 1.0 dB(A).

主动道路噪声控制（ARNC）系统利用来自底盘的多个加速度信号作为参考来衰减车辆内部道路噪声。然而，由于底盘结构之间的耦合效应，预定的参考信号不能总是保证其对内部噪声的独立性和相干性。此外，传统集中式ARNC系统的计算复杂度随着信道的增加而迅速增加。为了解决上述问题，本研究提出了一种采用虚拟参考和本地聚类控制策略的多通道ARNC系统。首先，提出了基于增量奇异值分解的虚拟参考方法，生成正交虚拟参考集，提高了系统的降噪性能；基于局部聚类控制策略，将传统的集中式系统分解为两个独立的主动头枕子系统，以降低计算复杂度。通过数值仿真和车载实验，验证了该ARNC系统的有效性。结果表明，与传统的集中式ARNC系统相比，该系统可以有效地衰减车内道路噪声，且计算量更少：当车辆以60 km/h行驶时，总体降噪值提高了1.5 dB(A)，当车辆以80 km/h行驶时，总体降噪值提高了1.0 dB(A)。

{"title":"A multi-channel active road noise control system using incremental SVD based virtual references and local-clustered control strategy","authors":"Ziheng Xia, Yansong He, Zhifei Zhang, Hao Chen, Xiaoyu Fu","doi":"10.1016/j.apacoust.2025.111216","DOIUrl":"10.1016/j.apacoust.2025.111216","url":null,"abstract":"<div><div>Active road noise control (ARNC) systems use multiple acceleration signals from the chassis as the references to attenuate vehicle interior road noise. However, due to the coupling effect among the chassis structures, the predetermined reference signals cannot always ensure their independence and coherence to the interior noises. In addition, the computational complexity of the conventional centralized ARNC systems increase rapidly as the growth of the channels. To address the above issues, this study proposes a multi-channel ARNC system using virtual references and local-clustered control strategy. Firstly, a virtual reference method based on incremental singular value decomposition is developed to generate an orthogonal virtual reference set, which improves the system’s noise reduction performance. The conventional centralized system is then decomposed into two independent active headrest subsystems based on the local-clustered control strategy to reduce the computational complexity. Numerical simulations and on-board experiments are conducted to validate the effectiveness of the proposed ARNC system. The results demonstrate that the proposed system can effectively attenuate the interior road noises with lower computational effort compared to the conventional centralized ARNC system: when the vehicle is driving at the speed of 60 km/h, the overall noise reduction values improved by 1.5 dB(A), and when the vehicle is driving at the speed of 80 km/h, the values improved by 1.0 dB(A).</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111216"},"PeriodicalIF":3.4,"publicationDate":"2025-12-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840710","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Measuring occupational exposure to very high frequencies noise: Assessing measurement variability under controlled conditions 测量职业性接触甚高频噪声：在受控条件下评估测量变异性

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-22 DOI: 10.1016/j.apacoust.2025.111170

Jonathan Terroir

Many occupational sectors are exposed to noise from very high audible frequencies (VHF) or low-frequency ultrasound (LFUS). Despite numerous studies on VHF/LFUS exposure, the question of the impact of several parameters in the on-site measurement protocol on the representativeness of the measurement remains open. Laboratory measurements were carried out to assess the variability of exposure measurements under controlled conditions as a function of the presence/absence of the microphone’s protection grid, the location of the microphone (shoulder, ear entrance, etc.), the side of the dummy where the microphone is located, and the frequency and angle of the source. Particular attention has been paid to the possibility of positioning the microphone at temple level, thanks to a specific mounting system adapted to on-site use.

许多职业部门都暴露于极高可听频率（VHF）或低频超声波（LFUS）的噪声中。尽管有许多关于甚高频/低频带频率暴露的研究，但现场测量方案中的几个参数对测量代表性的影响问题仍然存在。进行实验室测量以评估受控条件下暴露测量的可变性，作为麦克风保护网格存在/不存在的函数，麦克风的位置（肩膀，耳朵入口等），麦克风所在的假人侧面，以及源的频率和角度。由于采用了适合现场使用的特殊安装系统，因此特别注意了将麦克风定位在寺庙水平的可能性。

引用次数: 0

Minimum determinant estimate for source bearing estimation in shallow-water waveguides 浅水波导中源方位估计的最小行列式估计

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-22 DOI: 10.1016/j.apacoust.2025.111211

Qian Ma , Mingyang Li , Chao Sun

Most prevalent source bearing estimation methods are based on the plane-wave assumption, which might lead to biases in the estimates due to multimode propagation in waveguide channels, especially when the source is near the endfire direction of the sensor array. While the environment-dependent method, such as matched field processing (MFP), can provide unbiased estimates, it requires either a computationally expensive 3-D search over range, depth and bearing or prior knowledge of the source range and depth. The recently-developed subspace intersection method (SIM) circumvents these limitations by exploiting the alignment between the signal vector and the modal subspace, enabling accurate bearing estimation via a 1-D search. However, it requires prior knowledge of the number of sources. This work reformulates source bearing estimation as a matrix similarity measurement problem. It is demonstrated that the maximum likelihood estimate (MLE) for the source bearing can be derived by minimizing a tailored Euclidean distance between the sampled covariance matrix and the modal subspace projection matrix. Furthermore, a novel minimum determinant estimate (MDE) is proposed based on the Jensen-Bregman LogDet divergence, which minimizes the determinant of the sum of the data sampled covariance matrix and the modal subspace projection matrix. Numerical simulations in a shallow water waveguide demonstrate that the MDE achieves accurate bearing estimation of multiple sources without requiring information on the number of sources, and produces ambiguity surfaces with a clean background. The proposed method is also validated using experimental data from the SWellEx-96 trial.

大多数常用的源方位估计方法都是基于平面波假设，这可能会导致估计偏差，因为波导通道中的多模传播，特别是当源靠近传感器阵列的端射方向时。虽然与环境相关的方法，如匹配场处理（MFP），可以提供无偏估计，但它需要对范围、深度和方位进行计算上昂贵的三维搜索，或者需要事先了解源的范围和深度。最近开发的子空间相交方法（SIM）通过利用信号矢量和模态子空间之间的对齐来克服这些限制，通过一维搜索实现准确的方位估计。然而，它需要事先知道源的数量。本文将源方位估计重新表述为矩阵相似度测量问题。通过最小化采样协方差矩阵与模态子空间投影矩阵之间的定制欧氏距离，可以得到源方位的最大似然估计（MLE）。在此基础上，提出了一种基于Jensen-Bregman LogDet散度的最小行列式估计（MDE），使数据抽样协方差矩阵和模态子空间投影矩阵的和的行列式最小。在浅水波导中进行的数值模拟表明，该方法在不需要源数量信息的情况下实现了多源的准确方位估计，并产生了具有干净背景的模糊曲面。采用swelex -96试验的实验数据验证了所提出的方法。

{"title":"Minimum determinant estimate for source bearing estimation in shallow-water waveguides","authors":"Qian Ma , Mingyang Li , Chao Sun","doi":"10.1016/j.apacoust.2025.111211","DOIUrl":"10.1016/j.apacoust.2025.111211","url":null,"abstract":"<div><div>Most prevalent source bearing estimation methods are based on the plane-wave assumption, which might lead to biases in the estimates due to multimode propagation in waveguide channels, especially when the source is near the endfire direction of the sensor array. While the environment-dependent method, such as matched field processing (MFP), can provide unbiased estimates, it requires either a computationally expensive 3-D search over range, depth and bearing or prior knowledge of the source range and depth. The recently-developed subspace intersection method (SIM) circumvents these limitations by exploiting the alignment between the signal vector and the modal subspace, enabling accurate bearing estimation via a 1-D search. However, it requires prior knowledge of the number of sources. This work reformulates source bearing estimation as a matrix similarity measurement problem. It is demonstrated that the maximum likelihood estimate (MLE) for the source bearing can be derived by minimizing a tailored Euclidean distance between the sampled covariance matrix and the modal subspace projection matrix. Furthermore, a novel minimum determinant estimate (MDE) is proposed based on the Jensen-Bregman LogDet divergence, which minimizes the determinant of the sum of the data sampled covariance matrix and the modal subspace projection matrix. Numerical simulations in a shallow water waveguide demonstrate that the MDE achieves accurate bearing estimation of multiple sources without requiring information on the number of sources, and produces ambiguity surfaces with a clean background. The proposed method is also validated using experimental data from the SWellEx-96 trial.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111211"},"PeriodicalIF":3.4,"publicationDate":"2025-12-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840707","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Highly efficient design of sound attenuation and ventilated metamaterial via machine learning and genetic algorithm 通过机器学习和遗传算法高效设计消声和通风超材料

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-21 DOI: 10.1016/j.apacoust.2025.111205

Shaoji Zhang , Bo Song , Lei Zhang , Aiguo Zhao , Cheng Shen , Xiangyan Meng , Jiajie Luo , Yucheng Yuan , Hao Li , Liang Gao , Yusheng Shi

Ventilated acoustic metamaterials possess dual functionalities of ventilation and noise suppressing, which meets the needs of many scenarios. This work proposes a unit cell with high parametric tunability that allows flexible control of transmission loss peak quantities, facilitating customized ventilated metamaterials through cascaded unit cell configurations. To achieve the desired acoustic performance within specific frequency ranges, a partitioned optimization strategy was employed, targeting individual unit cells for distinct frequency sub-bands. By integrating machine learning and genetic algorithms, the geometrical parameters of the unit cell can be rapidly optimized to meet the acoustic performance target. In this work, we designed dual-band ventilated metamaterials and broadband ventilated metamaterials to demonstrated the framework’s effectiveness. Both ventilated metamaterials were investigated via finite element method and experiments. The transmission loss performances of experiments are perfect agreement with simulations. Machine learning model surrogate approach bypasses the repetitive process of modeling and finite element analysis, addressing the time-consuming and labor-intensive limitations of traditional trial–error and exhaustive methods, thereby establishing an accelerated pathway for ventilated metamaterial design.

通风声学超材料具有通风和降噪双重功能，可以满足多种应用场合的需要。这项工作提出了一种具有高参数可调性的单元电池，可以灵活地控制传输损耗峰值量，通过级联单元电池配置促进定制通风超材料。为了在特定频率范围内获得理想的声学性能，采用了分区优化策略，针对不同频率子带的单个单元格。通过结合机器学习和遗传算法，可以快速优化单元格的几何参数，以满足声学性能目标。在这项工作中，我们设计了双波段通风超材料和宽带通风超材料来证明该框架的有效性。通过有限元法和实验对两种通风材料进行了研究。实验结果与仿真结果吻合较好。机器学习模型替代方法绕过了建模和有限元分析的重复过程，解决了传统试错法和穷举法耗时费力的局限性，从而为通风超材料设计建立了一条加速途径。

{"title":"Highly efficient design of sound attenuation and ventilated metamaterial via machine learning and genetic algorithm","authors":"Shaoji Zhang , Bo Song , Lei Zhang , Aiguo Zhao , Cheng Shen , Xiangyan Meng , Jiajie Luo , Yucheng Yuan , Hao Li , Liang Gao , Yusheng Shi","doi":"10.1016/j.apacoust.2025.111205","DOIUrl":"10.1016/j.apacoust.2025.111205","url":null,"abstract":"<div><div>Ventilated acoustic metamaterials possess dual functionalities of ventilation and noise suppressing, which meets the needs of many scenarios. This work proposes a unit cell with high parametric tunability that allows flexible control of transmission loss peak quantities, facilitating customized ventilated metamaterials through cascaded unit cell configurations. To achieve the desired acoustic performance within specific frequency ranges, a partitioned optimization strategy was employed, targeting individual unit cells for distinct frequency sub-bands. By integrating machine learning and genetic algorithms, the geometrical parameters of the unit cell can be rapidly optimized to meet the acoustic performance target. In this work, we designed dual-band ventilated metamaterials and broadband ventilated metamaterials to demonstrated the framework’s effectiveness. Both ventilated metamaterials were investigated via finite element method and experiments. The transmission loss performances of experiments are perfect agreement with simulations. Machine learning model surrogate approach bypasses the repetitive process of modeling and finite element analysis, addressing the time-consuming and labor-intensive limitations of traditional trial–error and exhaustive methods, thereby establishing an accelerated pathway for ventilated metamaterial design.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111205"},"PeriodicalIF":3.4,"publicationDate":"2025-12-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145840709","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

An inverse method for source identification in rectangular waveguides with reverberation 含混响矩形波导中声源识别的逆方法

IF 3.4 2区物理与天体物理 Q1 ACOUSTICS

Applied Acoustics

Pub Date : 2025-12-19 DOI: 10.1016/j.apacoust.2025.111209

Xin Liu, Xiaodong Jing

Closed-section wind tunnels represent a typical reverberant environment where microphone arrays are frequently used to identify noise sources. However, wall reflections can be a serious problem for source localization, especially at lower frequencies or when a source locates close to reflecting walls. To address this problem, a novel inverse method is developed for source localization in a hard walled rectangular duct, which incorporates wall reflections into the proposed algorithm by using an appropriate rectangular waveguide Green’s function. No flow is considered in this study, focusing solely on wall reflection effects. Numerical and experimental results obtained with the conventional beamforming (CBF), the image source model (ISM) and the present inverse method (IM) are presented and analyzed. Pronounced spurious sidelobes appear on the CBF maps, causing reduced resolution or erroneous source location at low frequencies. The ISM can remove most of the spurious side lobes, but it still suffers from low resolution at low frequencies and its spatial resolution is direction dependent. By comparison, the IM shows considerably improved performance in terms of mainlobe width, localization accuracy and sidelobe level, with its resolution close to omnidirectional similar to that obtained under anechoic conditions. It is effective for both coherent and incoherent sound sources. At the low frequency of 500 Hz, it achieves subwavelength resolution by exploiting evanescent modes. Furthermore, the robustness of the IM to noise is examined through both simulations and experiments, showing that it outperforms the other two methods at an SNR of 10 dB.

封闭风洞是一种典型的混响环境，在这种环境中，传声器阵列经常被用来识别噪声源。然而，墙壁反射对于源定位来说可能是一个严重的问题，特别是在较低频率或源位于反射壁附近时。为了解决这一问题，本文提出了一种新的硬壁矩形管道中源定位的逆方法，该方法通过使用合适的矩形波导格林函数将壁反射集成到算法中。本研究不考虑流动，只关注壁面反射效应。给出并分析了传统波束形成（CBF）、像源模型（ISM）和逆方法（IM）的数值和实验结果。明显的伪副瓣出现在CBF图上，导致低频率分辨率降低或错误的源定位。该方法可以去除大部分杂散旁瓣，但在低频时分辨率较低，且空间分辨率与方向有关。相比之下，该方法在主瓣宽度、定位精度和副瓣电平方面都有明显改善，其分辨率接近全向，与消声条件下的分辨率相似。它对相干声源和非相干声源都有效。在500hz的低频下，利用倏逝模式实现亚波长分辨率。此外，通过仿真和实验验证了IM对噪声的鲁棒性，表明它在信噪比为10 dB时优于其他两种方法。

{"title":"An inverse method for source identification in rectangular waveguides with reverberation","authors":"Xin Liu, Xiaodong Jing","doi":"10.1016/j.apacoust.2025.111209","DOIUrl":"10.1016/j.apacoust.2025.111209","url":null,"abstract":"<div><div>Closed-section wind tunnels represent a typical reverberant environment where microphone arrays are frequently used to identify noise sources. However, wall reflections can be a serious problem for source localization, especially at lower frequencies or when a source locates close to reflecting walls. To address this problem, a novel inverse method is developed for source localization in a hard walled rectangular duct, which incorporates wall reflections into the proposed algorithm by using an appropriate rectangular waveguide Green’s function. No flow is considered in this study, focusing solely on wall reflection effects. Numerical and experimental results obtained with the conventional beamforming (CBF), the image source model (ISM) and the present inverse method (IM) are presented and analyzed. Pronounced spurious sidelobes appear on the CBF maps, causing reduced resolution or erroneous source location at low frequencies. The ISM can remove most of the spurious side lobes, but it still suffers from low resolution at low frequencies and its spatial resolution is direction dependent. By comparison, the IM shows considerably improved performance in terms of mainlobe width, localization accuracy and sidelobe level, with its resolution close to omnidirectional similar to that obtained under anechoic conditions. It is effective for both coherent and incoherent sound sources. At the low frequency of 500 Hz, it achieves subwavelength resolution by exploiting evanescent modes. Furthermore, the robustness of the IM to noise is examined through both simulations and experiments, showing that it outperforms the other two methods at an SNR of 10 dB.</div></div>","PeriodicalId":55506,"journal":{"name":"Applied Acoustics","volume":"245 ","pages":"Article 111209"},"PeriodicalIF":3.4,"publicationDate":"2025-12-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145798064","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0