Sound Source Detection Research Articles

Holographic method of under-water noise source localization in shallow water

The results of a high-frequency experiment for the detection and direction estimation of underwater sound noisy source are presented. The experiment was conducted in the shallow waters of the Black Sea coast. The noise emission of the source was received by three vector-scalar receivers located on the bottom. By using holographic processing, the detection and direction estimation of a moving underwater source against the background of intense shipping in the experiment region were carried out. Estimates of the input signal-to-noise ratio are presented.

An acoustic pressure-gradient MEMS vector hydrophone system based on piezoelectric micromachined ultrasonic transducer array

This article presents an acoustic pressure-gradient MEMS vector hydrophone system, which consists of a 2 × 2 piezoelectric scalar hydrophone array and a microcontroller unit (MCU)-based signal-conditioning circuit. The reported scalar hydrophone is fabricated based on a scandium-doped aluminum nitride (Sc0.2Al0.8N) piezoelectric micromachined ultrasonic transducer (PMUT) array. Four identical scalar hydrophones are adopted and arranged in a 2 × 2 array for detecting the acoustic pressure gradient along different incident angles. The received electrical signals in each scalar hydrophone are collected and processed by the MCU directly. Experimental results show that the vector hydrophone system has a receive sensitivity of −168 dB (re: 1 V/μPa) in the omnidirectional receive mode, and a receive sensitivity of −198 dB (re: 1 V/μPa) at 10 kHz in the dipole receive mode, respectively. With the aid of an advanced data preprocessing approach, the reported vector hydrophone system achieves the direction-of-arrival (DOA) estimations of underwater acoustic signals with an error of less than 2°. The measurement results show that the reported MEMS vector hydrophone system has the potential for the detection of underwater sound sources.

GLFER-Net: a polyphonic sound source localization and detection network based on global-local feature extraction and recalibration

Polyphonic sound source localization and detection (SSLD) task aims to recognize the categories of sound events, identify their onset and offset times, and detect their corresponding direction-of-arrival (DOA), where polyphonic refers to the occurrence of multiple overlapping sound sources in a segment. However, vanilla SSLD methods based on convolutional recurrent neural network (CRNN) suffer from insufficient feature extraction. The convolutions with kernel of single scale in CRNN fail to adequately extract multi-scale features of sound events, which have diverse time-frequency characteristics. It results in that the extracted features lack fine-grained information helpful for the localization of sound sources. In response to these challenges, we propose a polyphonic SSLD network based on global-local feature extraction and recalibration (GLFER-Net), where the global-local feature (GLF) extractor is designed to extract the multi-scale global features through an omni-directional dynamic convolution (ODConv) layer and multi-scale feature extraction (MSFE) module. The local feature extraction (LFE) unit is designed for capturing detailed information. Besides, we design a feature recalibration (FR) module to emphasize the crucial features along multiple dimensions. On the open datasets of Task3 in DCASE 2021 and 2022 Challenges, we compared our proposed GLFER-Net with six and four SSLD methods, respectively. The results show that the GLFER-Net achieves competitive performance. The modules we designed are verified to be effective through a series of ablation experiments and visualization analyses.

A Survey of Sound Source Localization and Detection Methods and Their Applications.

This study is a survey of sound source localization and detection methods. The study provides a detailed classification of the methods used in the fields of science mentioned above. It classifies sound source localization systems based on criteria found in the literature. Moreover, an analysis of classic methods based on the propagation model and methods based on machine learning and deep learning techniques has been carried out. Attention has been paid to providing the most detailed information on the possibility of using physical phenomena, mathematical relationships, and artificial intelligence to determine sound source localization. Additionally, the article underscores the significance of these methods within both military and civil contexts. The study culminates with a discussion of forthcoming trends in the realms of acoustic detection and localization. The primary objective of this research is to serve as a valuable resource for selecting the most suitable approach within this domain.

One size fits all? Adaptation of trained CNNs to new marine acoustic environments

Convolutional neural networks (CNNs) have the potential to enable a revolution in bioacoustics, allowing robust detection and classification of marine sound sources. As global Passive Acoustic Monitoring (PAM) datasets continue to expand it is critical we improve our confidence in the performance of models across different marine environments, if we are to exploit the full ecological value of information within the data. This work demonstrates the transferability of developed CNN models to new acoustic environments by using a pre-trained model developed for one location (West of Scotland, UK) and deploying it in a distinctly different soundscape (Gulf of Mexico, USA). In this work transfer learning is used to fine-tune an existing open-source ‘small-scale’ CNN, which detects odontocete tonal and broadband call types and vessel noise (operating between 0 and 48 kHz). The CNN is fine-tuned on training sets of differing sizes, from the unseen site, to understand the adaptability of a network to new marine acoustic environments. Fine-tuning with a small sample of site-specific data significantly improves the performance of the CNN in the new environment, across all classes. We demonstrate an improved performance in area-under-curve (AUC) score of 0.30, across four classes by fine-training with only 50 spectrograms per class, with a 5% improvement in accuracy between 50 frames and 500 frames. This work shows that only a small amount of site-specific data is needed to retrain a CNN, enabling researchers to harness the power of existing pre-trained models for their own datasets. The marine bioacoustic domain will benefit from a larger pool of global data for training large deep learning models, but we illustrate in this work that domain adaptation can be improved with limited site-specific exemplars.

ORCA-SPY enables killer whale sound source simulation, detection, classification and localization using an integrated deep learning-based segmentation

Acoustic identification of vocalizing individuals opens up new and deeper insights into animal communications, such as individual-/group-specific dialects, turn-taking events, and dialogs. However, establishing an association between an individual animal and its emitted signal is usually non-trivial, especially for animals underwater. Consequently, a collection of marine species-, array-, and position-specific ground truth localization data is extremely challenging, which strongly limits possibilities to evaluate localization methods beforehand or at all. This study presents ORCA-SPY, a fully-automated sound source simulation, classification and localization framework for passive killer whale (Orcinus orca) acoustic monitoring that is embedded into PAMGuard, a widely used bioacoustic software toolkit. ORCA-SPY enables array- and position-specific multichannel audio stream generation to simulate real-world ground truth killer whale localization data and provides a hybrid sound source identification approach integrating ANIMAL-SPOT, a state-of-the-art deep learning-based orca detection network, followed by downstream Time-Difference-Of-Arrival localization. ORCA-SPY was evaluated on simulated multichannel underwater audio streams including various killer whale vocalization events within a large-scale experimental setup benefiting from previous real-world fieldwork experience. Across all 58,320 embedded vocalizing killer whale events, subject to various hydrophone array geometries, call types, distances, and noise conditions responsible for a signal-to-noise ratio varying from -14.2 dB to 3 dB, a detection rate of 94.0 % was achieved with an average localization error of 7.01^circ. ORCA-SPY was field-tested on Lake Stechlin in Brandenburg Germany under laboratory conditions with a focus on localization. During the field test, 3889 localization events were observed with an average error of 29.19^circ and a median error of 17.54^circ. ORCA-SPY was deployed successfully during the DeepAL fieldwork 2022 expedition (DLFW22) in Northern British Columbia, with a mean average error of 20.01^circ and a median error of 11.01^circ across 503 localization events. ORCA-SPY is an open-source and publicly available software framework, which can be adapted to various recording conditions as well as animal species.

Acoustic Monitoring of Professionally Managed Marine Mammals for Health and Welfare Insights.

Research evaluating marine mammal welfare and opportunities for advancements in the care of species housed in a professional facility have rapidly increased in the past decade. While topics, such as comfortable housing, adequate social opportunities, stimulating enrichment, and a high standard of medical care, have continued to receive attention from managers and scientists, there is a lack of established acoustic consideration for monitoring the welfare of these animals. Marine mammals rely on sound production and reception for navigation and communication. Regulations governing anthropogenic sound production in our oceans have been put in place by many countries around the world, largely based on the results of research with managed and trained animals, due to the potential negative impacts that unrestricted noise can have on marine mammals. However, there has not been an established best practice for the acoustic welfare monitoring of marine mammals in professional care. By monitoring animal hearing and vocal behavior, a more holistic view of animal welfare can be achieved through the early detection of anthropogenic sound sources, the acoustic behavior of the animals, and even the features of the calls. In this review, the practice of monitoring cetacean acoustic welfare through behavioral hearing tests and auditory evoked potentials (AEPs), passive acoustic monitoring, such as the Welfare Acoustic Monitoring System (WAMS), as well as ideas for using advanced technologies for utilizing vocal biomarkers of health are introduced and reviewed as opportunities for integration into marine mammal welfare plans.

MEMS co-vibration combined hydrophone

This paper proposes a MEMS co-vibration combined hydrophone (CCH) to position the weak acoustic source of small-volume UUV acoustic self-guided cabins. With an external diameter of 36 mm, it adopts the MEMS triaxial accelerometer to form the vector channel and the MEMS capacitive omnidirectional microphone to form the scalar channel. This paper first analyzes the acoustic receiving model of an elastic-free sphere in water, designs a MEMS co-vibration combined hydrophone packaging and then simulates its acoustic performance. The average density of the encapsulated CCH is 1.01 g/cm3. The vector channel bandwidth is 20 Hz−1,000 Hz, and the sensitivity is −173 dB @ 1 kHz (re 1 V/ μPa) with good cosine directivity. The scalar channel sensitivity is −171 dB @ 1 kHz (re 1 V/ μPa) with omnidirectional directivity. The error of the orientation test in the standing wave tube is less than 1 degree. The sound source-level test results are mostly consistent with the theoretical calculation based on the actual distance. The above results indicate that the hydrophone can be mounted to UUV to perform sound source detection, which can meet the need for precise positioning, and has a specific application value.

AI-based sound source localization system with higher accuracy

This paper focuses on the research of sound source detection technology. Combining the artificial intelligence (AI) and 6th generation (6G) communications, it is a hot topic to build a centralized microphone array to locate the sound source in a particular scene. However, degrees of vagueness still exist concerning the correlations and influences between the array establishment, localization implementation and its performance. To this end, we propose a sound source localization system (SSLS) with a high accuracy, ultra reliability and low latency communication. Meanwhile, an AI-based back propagation (BP) neural network enables the SSLS to predict the sound source position (SSP). SSLS is mainly composed of a data acquisition terminal and a data processing terminal. After relying on the Visual Studio and adopting the C/C#, a reliable SSP measurement is provided by the SSLS. When SSLS is running, sound source signals (SSSs) are collected through the array to obtain the time difference of arrival (TDOA) values in real time. Different from existing methods, this SSLS changes the SSS into a pulse signal by detecting the signal with a fixed energy intensity. At the same time, an intensity-based trigger interruption is designed to complete the TDOA acquisition. In this way, besides a reduction in TDOA calculation, the signal recognition is also improved. Further, while using TDOAs to locate the sound source, localization steps with a BP network are presented. Considering extreme sound source angles, a multi-plane based data fusion method is proposed. With the setting of different angles, especially extreme angles, we assess both quantitative and qualitative results of the proposed method, demonstrating that our method can achieve high-accuracy localization results regardless of angle degrees and the number of planes. Finally, comparison experiments confirm that the proposed SSLS significantly outperforms existing methods.

Theoretical investigation of active listening behavior based on the echolocation of CF-FM bats.

Bats perceive the three-dimensional environment by emitting ultrasound pulses from their nose or mouth and receiving echoes through both ears. To determine the position of a target object, it is necessary to know the distance and direction of the target. Certain bat species that use a combined signal of long constant frequency and short frequency modulated ultrasounds synchronize their pinnae movement with pulse emission, and this behavior has been regarded as helpful for localizing the elevation angle of a reflective sound source. However, the significance of bats' ear motions remains unclear. In this study, we construct a model of an active listening system including the motion of the ears, and conduct mathematical investigations to clarify the importance of ear motion in direction detection of the reflective sound source. In the simulations, direction detection under rigid ear movements with interaural level differences was mathematically investigated by assuming that bats accomplish direction detection using the amplitude modulation in the echoes caused by ear movements. In particular, the ear motion conditions required for direction detection are theoretically investigated through exhaustive simulations of the pseudo-motion of the ears, rather than simulations of the actual ear motions of bats. The theory suggests that only certain ear motions, namely three-axis rotation, allow for accurate and robust direction detection. Our theoretical analysis also strongly supports the behavior whereby bats move their pinnae in the antiphase mode. In addition, we suggest that simple shaped hearing directionality and well-selected uncomplicated ear motions are sufficient to achieve precise and robust direction detection. Our findings and mathematical approach have the potential to be used in the design of active sensing systems in various engineering fields.

More than a whistle: Automated detection of marine sound sources with a convolutional neural network

The effective analysis of Passive Acoustic Monitoring (PAM) data has the potential to determine spatial and temporal variations in ecosystem health and species presence if automated detection and classification algorithms are capable of discrimination between marine species and the presence of anthropogenic and environmental noise. Extracting more than a single sound source or call type will enrich our understanding of the interaction between biological, anthropogenic and geophonic soundscape components in the marine environment. Advances in extracting ecologically valuable cues from the marine environment, embedded within the soundscape, are limited by the time required for manual analyses and the accuracy of existing algorithms when applied to large PAM datasets. In this work, a deep learning model is trained for multi-class marine sound source detection using cloud computing to explore its utility for extracting sound sources for use in marine mammal conservation and ecosystem monitoring. A training set is developed comprising existing datasets amalgamated across geographic, temporal and spatial scales, collected across a range of acoustic platforms. Transfer learning is used to fine-tune an open-source state-of-the-art ‘small-scale’ convolutional neural network (CNN) to detect odontocete tonal and broadband call types and vessel noise (from 0 to 48 kHz). The developed CNN architecture uses a custom image input to exploit the differences in temporal and frequency characteristics between each sound source. Each sound source is identified with high accuracy across various test conditions, including variable signal-to-noise-ratio. We evaluate the effect of ambient noise on detector performance, outlining the importance of understanding the variability of the regional soundscape for which it will be deployed. Our work provides a computationally low-cost, efficient framework for mining big marine acoustic data, for information on temporal scales relevant to the management of marine protected areas and the conservation of vulnerable species.

Pressure amplification and directional acoustic sensing based on a gradient metamaterial coupled with space-coiling structure

The detection and localization of weak sound sources are a topic of great interest in a wide range of applications, but generally require a pressure amplification and directivity acoustic system. Although acoustic technologies have been advanced in recent years, the realization of pressure amplification and directivity acoustic systems remains a slippery subject. Here, we propose a gradient acoustic metamaterial (GAM) coupled with space-coiling structure (GAM-SCS) for the detection and localization of weak sound sources. The GAM-SCS is evaluated both numerically and experimentally. Compared with the GAM without space-coiling structure, the GAM-SCS has a better pressure amplification performance which enables the sound signals to be pre-amplified. Meanwhile, the induced space-coiling structure, making use of subwavelength features, also enables the GAM-SCS to work in a lower frequency. Furthermore, the directivity of the GAM-SCS also shines lights on directional acoustic sensing to resist noise interference. We believe that this work opens up a new vista to realize the detection and localization of weak sound sources based on acoustic metamaterials.

Prediction model formulations for detection, enumeration, and localization of multiple sound sources using spherical harmonics

A spherical microphone array is used to detect and localize sound sources in terms of model-based machine learning (ML). In this application, it is crucial to establish parametric models to distinguish background sound environment from presence of sound sources. In the presence of sound sources, the parameter models are also used to localize an unknown number of potentially multiple sound sources. In this work, a model-based Bayesian learning framework is presented for localizing an unknown number of sound sources. Among them, a no-source scenario needs to be accounted for. The model-based machine learning applies the model comparison between the no-source model and the one-source model for sound source detection. After detecting sound sources, the machine learning needs to involve sound source enumeration and localization in order to correctly localize potential multiple sound sources. Specifically, sound environment is analyzed using Bayesian model comparison of two different models accounting for absence and presence of the sound sources for source detection. Upon a positive detection, potentially multiple source models are involved to analyze direction of arrivals (DoAs) for far-field and to localize sound sources for near-field including source distances, amplitudes, and DoAs using Bayesian model selection and parameter estimation.

A generalized network based on multi-scale densely connection and residual attention for sound source localization and detection.

Sound source localization and detection (SSLD) is a joint task of identifying the presence of individual sound events and locating the sound sources in space. However, due to the diversity of sound events and the variability of sound source location, SSLD becomes a tough task. In this paper, we propose a SSLD method based on a multi-scale densely connection (MDC) mechanism and a residual attention (RA) mechanism. We design a MDC block to integrate the information from a very local to exponentially enlarged receptive field within the block. We also explored three kinds of RA blocks that can facilitate the conductivity of information flow among different layers by continuously adding feature maps from the previous layers to the next layer. In order to recalibrate the feature maps after convolutional operation, we design a dual-path attention (DPA) unit that is largely embodied in MDC and RA blocks. We firstly verified the effectiveness of the MDC block, RA block, and DPA unit, respectively. We then compared our proposed method with another four methods on the development dataset; finally, with SELDnet and SELD-TCN on another five datasets, experimental results show the generalization of our proposed method.

The Effect of Attenuation from Fish on Passive Detection of Sound Sources in Ocean Waveguide Environments

Attenuation from fish can reduce the intensity of acoustic signals and significantly decrease detection range for long-range passive sensing of manmade vehicles, geophysical phenomena, and vocalizing marine life. The effect of attenuation from herring shoals on the Passive Ocean Acoustic Waveguide Remote Sensing (POAWRS) of surface vessels is investigated here, where concurrent wide-area active Ocean Acoustic Waveguide Remote Sensing (OAWRS) is used to confirm that herring shoals occluding the propagation path are responsible for measured reductions in ship radiated sound and corresponding detection losses. Reductions in the intensity of ship-radiated sound are predicted using a formulation for acoustic attenuation through inhomogeneities in an ocean waveguide that has been previously shown to be consistent with experimental measurements of attenuation from fish in active OAWRS transmissions. The predictions of the waveguide attenuation formulation are in agreement with measured reductions from attenuation, where the position, size, and population density of the fish groups are characterized using OAWRS imagery as well as in situ echosounder measurements of the specific shoals occluding the propagation path. Experimental measurements of attenuation presented here confirm previous theoretical predictions that common heuristic formulations employing free space scattering assumptions can be in significant error. Waveguide scattering and propagation theory is found to be necessary for accurate predictions.

A point sound source location and detection method based on 19-element hemispheric distributed acoustic pressure sensor array

To solve the key problem of diagnosing the operating condition of an oil transfer pump unit in a 3D closed space, this paper presents an approach for a point sound source location and detection method based on a hemispheric distributed sound pressure sensor array. The array model consists of 19 sound pressure sensors acting in the radial direction and uniformly distributed over the hemispherical surface. A spatial rectangular coordinate system is established by taking the projection point of the central sensor arranged at the apex of the hemisphere to the ground as the origin of the spatial coordinates. With reference to the central sensor, the point sound source is located by selecting the maximum measured sound level and its spatial coordinate in each of the three layers of sensors surrounding it as parameters and using a triangular or a quadrilateral area location algorithm based on virtual instrument technology. According to the location of the source, the A-weighted sound level of the sound source point is derived by the inversion of the sound field distribution law. Results show that the triangular and quadrilateral area location algorithms are both effective. The errors in location become larger for a measured sound source far from the centre.

Source detection and localization in the ocean with a Bayesian approach

Source localization and detection of sound sources in the ocean are investigated from a Bayesian Matched Field Processing perspective. Unknown parameters comprise the source spectrum along with source location; the spectrum is estimated within the localization and detection process. We develop incoherent and coherent processors, integrating the source spectrum estimation employing a Gibbs sampler. Using synthetic data we find that the coherent processor is superior to the incoherent processor and the standard Matched Field approach both in terms of source location estimates and uncertainty. The coherent and incoherent techniques are also applied to real data from the Hudson Canyon experiment with the same conclusion. Employing Receiver Operating Characteristic (ROC) curves, the coherent and incoherent processors are evaluated and compared in the task of joint detection and localization as well. The coherent detector/localization processor is superior to the incoherent one. Joint detection and localization performance is evaluated with Localization-ROC curves. [Work supported by ONR.]

Vertical and slanted sound propagation in the near-ground atmosphere: Amplitude and phase fluctuations.

Sound propagation along vertical and slanted paths through the near-ground atmosphere impacts detection and localization of low-altitude sound sources, such as small unmanned aerial vehicles, from ground-based microphone arrays. This article experimentally investigates the amplitude and phase fluctuations of acoustic signals propagating along such paths. The experiment involved nine microphones on three horizontal booms mounted at different heights to a 135-m meteorological tower at the National Wind Technology Center (Boulder, CO). A ground-based loudspeaker was placed at the base of the tower for vertical propagation or 56 m from the base of the tower for slanted propagation. Phasor scatterplots qualitatively characterize the amplitude and phase fluctuations of the received signals during different meteorological regimes. The measurements are also compared to a theory describing the log-amplitude and phase variances based on the spectrum of shear and buoyancy driven turbulence near the ground. Generally, the theory correctly predicts the measured log-amplitude variances, which are affected primarily by small-scale, isotropic turbulent eddies. However, the theory overpredicts the measured phase variances, which are affected primarily by large-scale, anisotropic, buoyantly driven eddies. Ground blocking of these large eddies likely explains the overprediction.

The Design of IoT Device System for Disaster Prevention using Sound Source Detection and Location Estimation Algorithm

The Design of IoT Device System for Disaster Prevention using Sound Source Detection and Location Estimation Algorithm

Design of a Low-Cost Configurable Acoustic Sensor for the Rapid Development of Sound Recognition Applications

Concerned about the noise pollution in urban environments, the European Commission (EC) has created an Environmental Noise Directive 2002/49/EC (END) requiring Member states to publish noise maps and noise management plans every five years for cities with a high density of inhabitants, major roads, railways and airports. The END also requires the noise pressure levels for these sources to be presented independently. Currently, data measurements and the representations of the noise pressure levels in such maps are performed semi-manually by experts. This process is time and cost consuming, as well as limited to presenting only a static picture of the noise levels. To overcome these issues, we propose the deployment of Wireless Acoustic Sensor Networks with several nodes in urban environments that can enable the generation of real-time noise level maps, as well as detect the source of the sound thanks to machine learning algorithms. In this paper, we briefly review the state of the art of the hardware used in wireless acoustic applications and propose a low-cost sensor based on an ARM cortex-A microprocessor. This node is able to process machine learning algorithms for sound source detection in-situ, allowing the deployment of highly scalable sound identification systems.