Journal of the Acoustical Society of America最新文献

Diffusion models have recently shown strong generative capabilities in inverse imaging problems. This paper introduces the first diffusion-based framework for acoustic source mapping that directly solves the deconvolution approach for the mapping of acoustic sources inverse problem. Supervised regression-based learning methods cannot well model the sparse and peak-shaped source distributions, but the proposed generative model explicitly learns the structural prior of source maps and thus avoids blurry artifacts in the output map. During training, the diffusion model is conditioned on both the delay-and-sum beamforming map and multi-scale point spread function features extracted by an autoencoder. The beamforming map provides coarse spatial cues on source positions and strengths, while the point spread function features provide frequency-aware information. The target map is a smoothed form of sparse source labels to help the model capture structural priors, and a time-weighted loss is proposed to help model better exploit the conditions. During inference, the model can generate high-resolution source distribution maps in only 20 sampling steps. Experimental results on three generalization tasks, i.e., unseen frequencies, unseen numbers of sources, and real-world transfer functions, demonstrate that the proposed method outperforms existing traditional and supervised regression-based deep learning-based approaches.

Passive detection and recognition capabilities of Unmanned Underwater Vehicles (UUVs) are significantly degraded by propulsion system self-noise, characterized by pronounced modulation interference and low signal-to-noise ratios. Existing denoising methods commonly produce spectral holes becse of over-suppression and insufficiently mitigate modulation interference. To overcome these limitations, this paper proposes a two-stage denoising-inpainting framework. In the first stage, a mask-based denoising network rapidly attenuates prominent self-noise to obtain a preliminarily enhanced signal. In the second stage, the Spectrum Inpainting Network (SINet) is introduced to precisely reconstruct the target spectrogram. To restore spectral holes and suppress modulation interference, SINet integrates a Modulation-Hole Restoration module to better capture modulation and contextual information. Furthermore, the framework incorporates a Shaft-Frequency Suppression Loss to guide the network focusing toward residual components within the shaft-frequency band in the detection of envelope modulation on noise spectrum. Extensive experiments on the ShipsEar dataset and collected UUV self-noise data demonstrate that the proposed framework can effectively suppress modulation interference and enhance target signal fidelity. The interference shaft-frequency peak-to-average ratio and spectral mean squared error are reduced by 75% and 22%, leading to a notable 6.87% improvement in target recognition accuracy.

Ultrasound computed tomography (USCT) is a non-invasive quantitative imaging technique that estimates the reflectivity, sound speed, and attenuation properties of the imaged medium by transmitting and receiving ultrasound waves through multiple transducers, offering a cost-effective approach to medical imaging. Time-reversal imaging (TRI), based on the full two-way wave equation, provides effective illumination for steeply dipping structures and complex regions with strong vertical and lateral velocity contrasts. Applied to USCT, TRI enables the recovery of sub-Fresnel-scale fine structures without requiring iterative inversion or MHz-level high-frequency wavelets, thereby improving both image clarity and computational efficiency. However, the conventional zero-lag cross correlation imaging condition introduces low-frequency, high-amplitude artifact noise, which obscure lesion boundaries and reduce diagnostic reliability. To address this issue within the USCT geometry, we propose a modified imaging condition based on implicit full-wavefield decomposition using the Hilbert transform. This approach separates the forward and adjoint wavefields into upgoing/downgoing and leftgoing/rightgoing components, while avoiding the need to explicitly store these large datasets. As a result, it enables effective computation, improves imaging quality, and enhances signal-to-noise ratio. Synthetic experiments with breast and brain models demonstrate the potential applicability of this method for high-resolution USCT.

Noisy industrial environments are challenging for effective communication. To facilitate communication between workers in noisy environments, active systems featuring interpersonal radio communication may be used despite the fact that these systems do not enable users to dynamically address specific individuals or perceive the directionality in speech. A promising approach to these challenges is an emerging technology called radio acoustical virtual environment (RAVE). RAVE aims to improve communication in two ways: (1) allowing users to dynamically address specific individuals based on vocal effort, and (2) transforming speech signals to convey spatial directionality of its origin. To explore its potential, a mockup version of RAVE incorporating key algorithms was developed and evaluated with 18 participants. The evaluation demonstrated RAVE's potential to improve communication. Additionally, the results offer valuable insights into user interaction and inform future design decisions for active hearing protection devices with integrated communication systems.

This work presents a methodology to study the noise generated by confined flows, i.e., flows enclosed within a device, which therefore significantly influence both the flow characteristics and the transmission of noise. A coupling workflow is implemented to link computational fluid dynamics simulations and structural dynamic analysis, where the pressure field computed in the fluid domain provides the forcing terms for the structural vibration equations. Outside the solid structure, the Ffowcs Williams-Hawkings acoustic analogy is applied to compute the noise radiated by the vibrating structure. The numerical framework is implemented using a combination of open-source and in-house software. The underlying modeling assumptions are discussed and validated. The results highlight the critical role of structural vibrations and demonstrate that relying solely on fluid-dynamic simulations can lead to misleading noise predictions. To isolate fluid-dynamic-induced vibrations and the associated structure-borne noise, and to exclude contributions from the driving motor and mechanical components such as bearings, an idealized device is considered, representing a simplified turbomachine configuration.

The widely held assumption that closing eyes enhances auditory sensitivity has been supported by auditory attention experiments. However, the visual effects on auditory thresholds of detecting target sounds masked by noise remain unexplored. We investigated the participants' detection thresholds (n = 25) of target sounds (canoe paddle, drum, lark chirping, train, and keyboard) masked by 70 dB(A) pink noise under four visual conditions (eyes closed, eyes open with blank board, static visual stimulation, and dynamic visual stimulation). Taking blank visual stimulation as the baseline, eye closure elevated detection thresholds by 1.32 dB on average, whereas dynamic and static relevant visual stimulation lowered them by 2.98 and 1.60 dB, respectively, contrary to conventional belief. Electroencephalogram recordings (n = 27) demonstrated avalanche critical index reduction of 22.3%-45.2% across five auditory stimuli under eye closure compared with blank stimulation, revealing that non-visual states preferentially stabilize neural dynamics near critical states. We propose a unified auditory-cortical framework based on the brain dynamics theory to explain both the enhanced auditory target detection during visual engagement in noisy environments and optimized auditory segregation via visual disengagement in quiet settings, advancing our understanding of visual effects on auditory perception in complex noisy soundscapes.

This study experimentally, numerically, and theoretically investigates the cavity/bubble dynamics and radiated acoustics during the water entry of a centimeter-scale cylindrical projectile with a conical nose. Experiments were conducted in a laboratory tank, employing synchronized high-speed imaging and hydrophone measurements to characterize the cavity closure modes and their resultant acoustic signatures across a range of Froude numbers. The acoustic signal features a weak radiated signal upon impact, followed by significant pressure oscillations spanning more than 20 cycles in the flow field after cavity elongation and pinch-off. A numerical model based on the Finite Volume Method successfully captures these physical processes. Subsequently, a semi-theoretical model that incorporates the projectile's boundary effect is developed from potential flow theory. The model not only yields a dominant cavity oscillation frequency that agrees well with experimental data, but also reveals that the boundary effect leads to a cavity oscillation frequency markedly higher than the Minnaert frequency of an equivalent-volume ellipsoidal bubble containing an internal rigid core. The dominant cavity frequency falls nearly linearly with Fr, governed by nose geometry and projectile inertia. This study clarifies the underlying physics connecting cavity dynamics during water entry to underwater acoustic radiation.

In a recently published paper, we established an N-beam approach to describe acoustical fields, with the prospect of using this approach to develop an acoustical localization procedure allowing one to speed up the computations of acoustical beam shape coefficients. The present paper explicitly deals with the development of such an acoustical localization procedure.

Data from headphone experiments with human listeners indicate a small peak near 1000 Hz in the threshold for the detection of interaural level difference (Grantham effect). The peak has been observed in at least six different laboratories. This article discusses this peak in three parts. The first part reviews eight articles on the peak effect. The second part begins a new explanation for the effect. It proposes that the peak originates through the unique "in-back" perceptual localization of tones or narrowband noise in the 1000-Hz region, evidently caused by an anomalous physical front-back level difference. The angular extent of this physical anomaly is here determined through extensive free-field front/back measurements with a KEMAR manikin, indicating that the anomalous level difference is unique and extends over a wide range of source azimuths at two elevations. The third part suggests that listeners have learned to discount the left-right localization cues in the 1000-Hz frequency region in favor of more reliable frequency regions because individual sounds that appear in back are much less well localized in the left-right dimension than sounds that appear in front. These facts may be the origin of the Grantham Effect.