Journal of the Acoustical Society of America最新文献

Sound field reproduction uses multiple loudspeakers to create a desired sound field within a target area. The placement of the secondary sources is a particularly critical factor influencing reproduction performance. Iterative optimization is a common class of placement optimization methods, but existing methods tend to either only select or only remove secondary sources unidirectionally. This characteristic often causes the search-algorithm to get stuck in a local optimum. This study proposes a secondary source placement optimization method based on bidirectional stepwise iteration. In each iteration, the proposed method first selects the two loudspeakers from the candidate set with the highest contribution to sound reproduction to become secondary sources. It then removes the secondary source with the lowest contribution from the entire set of selected sources and returns it to the candidate set. The proposed method was validated by simulations and a public experimental dataset. Results show that this method is less likely to get stuck in a local optimum compared to unidirectional iterative optimization methods, thus achieving better performance. This study also explores how factors like the number of secondary sources and room reverberation affect the performance and compares the computational complexity of our method with that of unidirectional optimization methods.

Conventional hearing aid (HA) algorithms, developed based on computational and non-differentiable auditory processing models, do not typically compensate for cochlear synaptopathy (CS). Traditional HAs apply fixed or rule-based gain adjustments within predefined frequency bands and compression ratios, instead of model-based fitting that numerically optimizes processing from neural representations of normal and impaired hearing. To compensate for combined CS and outer-hair-cell (OHC) loss, deep neural network (DNN)-based closed-loop HA systems are gaining traction. Here, we present several DNN-based HA algorithms that embed personalized, differentiable DNN-based auditory models (dCoNNear) inside a closed-loop system to train personalized HA algorithms compensating for OHC damage and/or CS. The HA algorithms were trained using backpropagation to minimize differences between hearing-impaired and normal-hearing auditory nerve (AN) responses. Performance was evaluated using speech and standard auditory stimuli. Results showed enhanced temporal-envelope (TENV) processing of modulated pure tones, particularly for CS, where sharpening of the TENV led to stronger AN onset responses. Transfer functions indicated that DNN-based HA algorithms applied adaptive level-dependent and frequency-specific gain aligned with OHC damage. The algorithms improved the normalized root mean square error of AN responses compared to NAL-NL2 for certain TIMIT phoneme categories. This evaluation offers insights into how machine-learning approaches outperform traditional HA strategies.

Auditory temporal resolution plays a critical role in the everyday experience of listening to complex acoustic patterns. The two most commonly used methods for obtaining measures of auditory temporal resolution are gap detection and amplitude modulation (AM) detection. In an attempt to develop a standardized clinical test of auditory temporal resolution, we used Zippy Estimation by Sequential Testing (ZEST), a Bayesian threshold estimation procedure, to measure gap and AM detection thresholds. In the present study, we collected gap and AM detection thresholds from more than 140 normal-hearing and hearing-impaired participants, using ZEST and a standard 1-up 2-down procedure for comparison. The results showed that the thresholds obtained with ZEST were very close to those obtained with the 1-up 2-down procedure and that the threshold estimates of ZEST approached the asymptote within 10 to 20 trials, indicating the reliability and efficiency of ZEST as a method for measuring auditory temporal resolution. In addition, our data revealed significant correlations between the gap detection thresholds and the peak sensitivity, but not the cut-off frequency, of the temporal modulation transfer function estimated from the AM detection thresholds, when the participants' ages and hearing levels were partialed out.

This work presents a direct comparison of time reversal (TR) and cross correlation (CC) for wave refocusing, with application to shear wave elastography. Standard theoretical equivalence relies on idealized assumptions, such as diffuse fields, that are rarely met in practice. A unified framework based on a reciprocal configuration to enable a direct, assumption-free comparison is proposed. A key distinction is drawn between active TR, requiring separate forward and backward wavefield acquisitions, and passive TR, computed from a single forward acquisition. It is demonstrated that TR and CC are two sides of the same refocusing process, connected through a decomposition of the CC field into diagonal and off diagonal (OD) components. The performance of passive CC is governed by the source characteristics: for impulsive sources, refocusing is maximized when the inter-source delay exceeds the signal duration, suppressing OD interference; for continuous noise, performance is limited by the noise autocorrelation. It is validated that high-fidelity shear wave speed maps can be reconstructed via direct wave inversion from TR and CC focusing fields, with a maximum difference of 4%, confirming that equivalent mechanical information is extracted.

We report the repeated in situ observation of nonlinear mesoscopic elasticity in an ∼760 m3 unstable rock column undergoing drilling for bolting reinforcement. Analysis of the anomalous nonlinear fast dynamics at the onset of drilling showed a clear correlation between top-of-column acceleration and small (a few percent) drops in resonance frequency. Slow dynamics were also observed, appearing as reproducible recovery of the fundamental resonance frequency after both drilling stopped and earthquake shaking occurred. Characteristic relaxation times ranged from tens to several hundred seconds, with recovery in  log10(t) spanning about 1-2 orders of magnitude. Significant nonlinear effects affecting the column's fundamental mode arose for strains as low as 10-9-10-8, with the proportion of the affected column volume depending on drilling location. The magnitude of the observed nonlinearities exceeded that found in laboratory tests on intact rock, revealing extensive, multi-scale cracking within the column. These results demonstrate that coincidental passive seismic surveys can effectively probe nonlinearity at the geophysical scale, with potential applications in geotechnical and civil engineering and in monitoring internal fracturing of rock structures.

To address the low element utilization of biconical arrays in ship noise measuring systems, this study proposes a logarithmically distributed biconical array configuration and a corresponding constant beam width beamforming (CBB) method. The method begins by calculating optimal element positions using the continuous array beam pattern equation. Subsequently, a compensation function is derived to achieve a frequency-invariant beam pattern from the modal-domain beam pattern equation. Finally, the compensation function is converted into a filter to achieve CBB. The effectiveness of the proposed method is validated through simulation experiments. Compared to the traditional nested array method, the proposed approach achieves at least 6 dB of additional array gain. The proposed beamforming method is decoupled from the weighted vector design in traditional beamforming problems. This allows users to design a beam pattern based on engineering requirements while ensuring frequency invariance. Furthermore, the proposed method can be applied to vector arrays, resulting in better performance than traditional scalar arrays. The proposed method reduces broadband processing errors by over 45% compared to the nested method, making it suitable for various broadband array signal processing scenarios. These include measuring the radiated noise of underwater targets, estimating the arrival direction of broadband waves, and identifying underwater targets.

In this study, a method for readily and inexpensively generating intense midair ultrasound fields that are reconfigurable in real-time is proposed for applications of midair nonlinear acoustic effects. To generate such ultrasound fields, specifically designed ultrasound sources or phased arrays of ultrasound transducers are conventionally used. The former can be more readily fabricated but cannot drastically reconfigure the generated ultrasound field, and the latter can create electronically controllable ultrasound fields but is much more expensive and difficult to implement. The proposed method utilizes a planar ultrasound source with a fixed surface vibration pattern and a newly designed amplitude mask, which partially covers source emission to form a predetermined ultrasound field with a corresponding specific spatial pattern when placed at a specific position. This new mask allows for switching of the generated fields among several presets by changing the mask's position on the source. The proposed technique requires only slight mechanical translation of the mask over the source to instantly reconfigure the resulting midair ultrasound field. This method enables the creation of a reconfigurable ultrasound field with a large source aperture in a significantly inexpensive, practical setup, potentially extending the workspace of current midair ultrasound applications to a whole-room scale.

Convergence zone (CZ) propagation is characterized by a small number of eigenrays, each with one deep ocean turning point and no more than two upper ocean turning points. Because of this geometry, these rays undergo a small amount of random internal wave scattering relative to the CZ range of ∼ 60 km. It is hypothesized that CZ multipath will therefore be correlated over limited observation intervals. The theory of wave propagation in random media is applied to CZ propagation through a Garrett-Munk internal wave field to predict the correlation between CZ eigenrays at mid-frequency over a finite observation interval. The theoretical results predict correlated multipath over intervals between 100 and 200 s. This prediction is then compared to data collected during an experiment in the Philippine Sea. The complex amplitudes of multipath arrivals in the CZ from 5.5 kHz sinusoidal transmissions are estimated using beamforming with a vertical line array and seen to be significantly correlated over approximately 27 s.

Conventional passive sound absorbers are limited in low-frequency performance due to the constraints of long wavelengths and structural dimensions. This paper proposes a self-sensing acoustic impedance control method that achieves broadband low-frequency absorption without relying on external sensors. The method utilizes the electromechanical coupling of a loudspeaker for both sensing and actuation, while a finite impulse response filter (FIR) is adopted as the control algorithm, offering greater flexibility compared with shunt loudspeaker designs. A lumped-parameter model is first established to analyze the interaction between the controller and the diaphragm surface, from which the equivalent acoustic impedance is derived as a function of the FIR filter coefficients. Theoretical analysis shows that the first-order characteristics of the FIR filter modify the effective mass and damping of the diaphragm, shifting the absorption peak, whereas the higher-order characteristics suppress the acoustic reactance, enabling broadband absorption in the low-frequency range. Numerical simulation is employed to analyze the tunable acoustic performance of the device. Experimental validation in an impedance tube verifies the effectiveness of the proposed method. With different parameter settings, the absorber achieves absorption coefficients above 0.6 in the 150-300 Hz range, and it is consistent with the analysis results.

Acoustically induced cell deformation is receiving increasing attention as a contactless and biocompatible technique for single-cell mechanical phenotyping. Existing studies have extracted cell mechanical properties from acoustic deformation experiments using a thin elastic shell cell model, which describes only membrane elasticity. To address the need to quantify the whole-cell mechanical properties regarding intracellular structures, this work models cells as homogeneous viscoelastic spheres and presents an effective computational model for predicting the cell deformation dynamics in a viscous fluid driven by ultrasound. A perturbation approach is employed to clarify the first-order linear acoustic and second-order nonlinear acoustic effects. To circumvent extreme challenges in numerical simulation due to the separated length scales between acoustic boundary layer thickness and cell radius, the acoustic boundary layer is incorporated into effective fluid-cell coupling boundary conditions via boundary layer analysis. The cell deformation in an ideal standing wave, calculated using the effective model, agrees well with full boundary-layer-resolved simulations while achieving a nearly sixfold reduction in memory consumption. The Young's modulus of Michigan Cancer Foundation-7 (MCF-7) cells is extracted by fitting their deformation to reported experimental data in a focused acoustic beam and an acoustic squeezer. The obtained value of 90 ± 30 Pa is consistent with the accepted range.