Multi-microphone Array Research Articles

Drone exploration of bat echolocation: A UAV-borne multimicrophone array to study bat echolocation.

Multimicrophone array techniques offer crucial insight into bat echolocation, yet they severely undersample the environments bats operate in as they are limited in geographic placement and mobility. UAVs are excellent candidates to greatly increase the environments in which such arrays can be deployed, but the impact of UAV noise on recording quality and the UAV's behavioral impact on the bats may affect usability. We developed a UAV-borne multimicrophone setup capable of recording bat echolocation across diverse environments. We quantify and mitigate the impact of UAV noise on the recording setup and test the recording capability of the array by recording four common Danish bat species: Pipistrellus pygmaeus, Myotis daubentonii, Eptesicus serotinus, and Nyctalus noctula. The UAV produces substantial noise at ultrasonic frequencies relevant to many bat species. However, suspending the array 30m below the UAV attenuates the noise to levels below the self-noise of our recording system at 20 kHz and above, and we successfully record and acoustically localize all four bat species. The behavioral impact of the UAV is minimal as all four species approached the array to within 1m and all emitted recordable feeding buzzes. UAV-borne multimicrophone arrays will allow us to quantify bat echolocation in hitherto unexplored habitats and provide crucial insight into how bats operate their sonar across their entire natural habitat.

Drone exploration of bat echolocation: a UAV-borne multi-microphone array to study bat echolocation

Drone exploration of bat echolocation: a UAV-borne multi-microphone array to study bat echolocation

Rookognise: Acoustic detection and identification of individual rooks in field recordings using multi-task neural networks

Individual-level monitoring is essential in many behavioural and bioacoustics studies. Collecting and annotating those data is costly in terms of human effort, but necessary prior to conducting analysis. In particular, many studies on bird vocalisations also involve manipulating the animals or human presence during observations, which may bias vocal production. Autonomous recording units can be used to collect large amounts of data without human supervision, largely removing those sources of bias. Deep learning can further facilitate the annotation of large amounts of data, for instance to detect vocalisations, identify the species, or recognise the vocalisation types in recordings. Acoustic individual identification, however, has so far largely remained limited to a single vocalisation type for a given species. This has limited the use of those techniques for automated data collection on raw recordings, where many individuals can produce vocalisations of varying complexity, potentially overlapping one another, with the additional presence of unknown and varying background noise. This paper aims at bridging this gap by developing a system to identify individual animals in those difficult conditions. Our system leverages a combination of multi-scale information integration, multi-channel audio and multi-task learning. The multi-task learning paradigm is based the overall task into four sub-tasks, three of which are auxiliary tasks: the detection and segmentation of vocalisations against other noises, the classification of individuals vocalising at any point during a sample, and the sexing of detected vocalisations. The fourth task is the overall identification of individuals. To test our approach, we recorded a captive group of rooks, a Eurasian social corvid with a diverse vocal repertoire. We used a multi-microphone array and collected a large scale dataset of time-stamped and identified vocalisations recorded, and found the system to work reliably for the defined tasks. To our knowledge, the system is the first to acoustically identify individuals regardless of the vocalisation produced. Our system can readily assist data collection and individual monitoring of groups of animals in both outdoor and indoor settings, even across long periods of time, and regardless of a species’ vocal complexity. All data and code used in this article is available online.

Three Tools for Sonic Rupture: Translating Ambiance, Biophilic Sound Design and More-Than-Human Listening

Sounds of the city shape everyday perception; thus, as sounds change, so do our moods and responses to the environments in which we live. This article is part of a discourse trying to recognise the essential role that soundscape design should play in urban planning to improve quality of life experiences. It aims to discuss three tools for creating sonic ruptures in urban environments: translating ambiance, biophilic sound design, and more-than-human listening. The three tools are related to the sonic rupture concept, introduced as a form of soundscape design focused on improving the quality of life for urban dwellers and creating new possible approaches for urban design. Translating ambiance leverages ambiance theory to explore the prospects of translating the affective qualities of the natural world into cities. The biophilic sound design tool combines biophilic design and field recording practices to discover innovative ways to bring the sounds of nature into the city. The more-than-human listening tool explores the possible recordings that multi-microphone arrays can make in natural environments that, while impossible to hear with the human ear, can be integrated into urban installations. In combination, this suite of tools presents new ways of thinking about the roles artistic research can play in urban soundscape design as a means to expand the range of human experience – and thus the quality of life – in urban environments. The nature term, as used here, refers to those expanses of land free of urban development that present unique sonic experiences and expressions to the artistic researcher, which can be applied to urban design. The described approach has been tested in the creation of the Sonic Gathering Place, an installation that integrates the three post-sonic rupture tools, which will be touched on briefly in this paper.

Dynamically localizing multiple speakers based on the time-frequency domain

In this study, we present a deep neural network-based online multi-speaker localization algorithm based on a multi-microphone array. Following the W-disjoint orthogonality principle in the spectral domain, time-frequency (TF) bin is dominated by a single speaker and hence by a single direction of arrival (DOA). A fully convolutional network is trained with instantaneous spatial features to estimate the DOA for each TF bin. The high-resolution classification enables the network to accurately and simultaneously localize and track multiple speakers, both static and dynamic. Elaborated experimental study using simulated and real-life recordings in static and dynamic scenarios demonstrates that the proposed algorithm significantly outperforms both classic and recent deep-learning-based algorithms. Finally, as a byproduct, we further show that the proposed method is also capable of separating moving speakers by the application of the obtained TF masks.

Generalized Sidelobe Canceller Beamforming with Combined Postfilter and Sparse NMF for Speech Enhancement

A multi-microphone array speech enhancement method using Generalized Sidelobe Canceller (GSC) beamforming with Combined Postfilter (CP) and Sparse Non-negative Matrix Factorization (SNMF) is proposed in this paper. GSC beamforming with CP and SNMF is implemented to reduce directional noise, diffuse noise, residual noise and to separate interferences in adverse environment. In this paper, the directional noise is reduced using GSC beamforming, whereas the diffuse noise in each subband is reduced with a combined postfilter using Unconstrained Frequency domain Normalized Least Mean Square (UFNLMS) algorithm. Finally, the residual noise at the output of CP is eliminated by SNMF which optimizes the noise. The performance of the proposed method is evaluated using parameters like PESQ, SSNR, STOI, SDR and LSD. The noise reduction for four and eight microphones is compared and illustrated in spectrograms. The proposed method shows better performance in terms of intelligibility and quality when compared to the existing methods in adverse environments.

Detection of nearby UAVs using a multi-microphone array on board a UAV

In this work, we address the problem of UAV detection flying nearby another UAV. Usually, computer vision could be used to face this problem by placing cameras onboard the patrolling UAV. However, visual processing is prone to false positives, sensible to light conditions and potentially slow if the image resolution is high. Thus, we propose to carry out the detection by using an array of microphones mounted with a special array onboard the patrolling UAV. To achieve our goal, we convert audio signals into spectrograms and used them in combination with a CNN architecture that has been trained to learn when a UAV is flying nearby, and when it is not. Clearly, the first challenge is the presence of ego-noise derived from the patrolling UAV itself through its propellers and motor’s noise. Our proposed CNN is based on Google’s Inception v.3 network. The Inception model is trained with a dataset created by us, which includes examples of when an intruder UAV flies nearby and when it does not. We conducted experiments for off-line and on-line detection. For the latter, we manage to generate spectrograms from the audio stream and process it with the Nvidia Jetson TX2 mounted onboard the patrolling UAV.

A Sound Source Localization Device Based on Rectangular Pyramid Structure for Mobile Robot

A sound source localization device based on a multimicrophone array with the rectangular pyramid structure is proposed for mobile robot in some indoor applications. Firstly, a time delay estimation method based on the cross-power spectral phase algorithm and a fast search strategy of peak value based on the geometric distribution of microphones are developed to estimate the sound propagation delay differences between two microphones. Moreover, a rejection strategy is presented to evaluate the correctness of the delay difference values. And then, the device’s geometric equations based on the time-space mapping relationship are established to calculate the position of the sound source. For fast solving the equations, the multimicrophone array space is divided into several subspaces to narrow the solution range, and Newton iteration algorithm is introduced to solve the equations, while its solution is evaluated by an evaluation mechanism based on coordinate thresholds. Finally, some experiments are carried out to verify the performance of the device, of which the results show that the device can achieve sound source localization with a high accuracy.

On the Possibility of Using a Single Time Realization for Investigating Noise from Vortex Rings

The possibility of using a single time realization of a vortex ring to study its aeroacoustic properties is considered. For each realization, averaging is performed over six microphones located around the path of the ring at the same distance from its axis. Spectra obtained as a result of such averaging are presented. It is shown that the specific features of noise from an individual vortex ring agree well with earlier results when averaging over an ensemble of realizations. It is proposed for the first time to localize the sources of noise produced by a vortex ring during its motion using a multimicrophone array with a beamforming algorithm adapted for the considered problem. The position of the noise sources on the localization map is in good agreement with the trajectory measurements of the vortex ring. A beamforming array was used to obtain the averaged spectra of single realizations corresponding to localization of sources for different delays from the moment of vortex launch.

Development of the Strategy of Active Control of Instability Waves in Unexcited Turbulent Jets

The possibility of controlling instability waves in the mixing layer of a subsonic unexcited jet is studied. These waves can be noise sources in both free jets and jets as parts of configurations. In the study the method of experimental diagnostics of the instability waves in the near field of a jet using an azimuthal multimicrophone array is realized. The data on the near field fluctuations are used for testing the control strategy proposed by the authors. The strategy consists in narrowband sliding filtration of the original signal and the formation of a narrowband controlling action on the basis of the linear principle of signal superposition. The results of the study represent the next step toward the realization of an active control system suppressing natural instability waves in turbulent jets.

Oilbirds produce echolocation signals beyond their best hearing range and adjust signal design to natural light conditions.

Oilbirds are active at night, foraging for fruits using keen olfaction and extremely light-sensitive eyes, and echolocate as they leave and return to their cavernous roosts. We recorded the echolocation behaviour of wild oilbirds using a multi-microphone array as they entered and exited their roosts under different natural light conditions. During echolocation, the birds produced click bursts (CBs) lasting less than 10 ms and consisting of a variable number (2–8) of clicks at 2–3 ms intervals. The CBs have a bandwidth of 7–23 kHz at −6 dB from signal peak frequency. We report on two unique characteristics of this avian echolocation system. First, oilbirds reduce both the energy and number of clicks in their CBs under conditions of clear, moonlit skies, compared with dark, moonless nights. Second, we document a frequency mismatch between the reported best frequency of oilbird hearing (approx. 2 kHz) and the bandwidth of their echolocation CBs. This unusual signal-to-sensory system mismatch probably reflects avian constraints on high-frequency hearing but may still allow oilbirds fine-scale, close-range detail resolution at the upper extreme (approx. 10 kHz) of their presumed hearing range. Alternatively, oilbirds, by an as-yet unknown mechanism, are able to hear frequencies higher than currently appreciated.

Predictive Acoustic Tracking with an Adaptive Neural Mechanism

Tracking an acoustic signal in motion is pertinent in several domains such as human-robot interaction and search-and-rescue robotics. Conventional approaches to acoustic tracking acquire time-of-arrival-difference signals from multi-microphone arrays and localise the acoustic signal using Kalman or particle filtering, generalised cross-correlation or steered response power techniques. The authors have previously developed a biologically-inspired mechanism that utilises two microphones to reactively track an acoustic signal in motion. The mechanism leverages the directional response of an mathematical model of the lizard peripheral auditory system to extract information regarding sound direction. This information is utilised by a neural machinery to learn the acoustic signal's velocity through fast and unsupervised correlation-based learning adapted from differential Hebbian learning. This approach has previously been validated in simulation and via robotic trials to track a continuous pure tone acoustic signal with a semi-circular motion trajectory and a constant but unknown angular velocity. The neural machinery has been shown to be able to learn different target angular velocities in independent trials. Here we extend our previous work by demonstrating that an identical instance of the mechanism can be used to successfully predict the future spatial location of an acoustic signal with an identical semi-circular motion trajectory and a constant but unknown angular velocity. We evaluate the prediction performance of the simulated mechanism in independent trials for three different angular velocities.

The diverse environments multi-channel acoustic noise database: A database of multichannel environmental noise recordings

Multi-microphone arrays allow for the use of spatial filtering techniques that can greatly improve noise reduction and source separation. However, for speech and audio data, work on noise reduction or separation has focused primarily on one- or two-channel systems. Because of this, databases of multichannel environmental noise are not widely available. DEMAND (Diverse Environments Multi-channel Acoustic Noise Database) addresses this problem by providing a set of 16-channel noise files recorded in a variety of indoor and outdoor settings. The data was recorded using a planar microphone array consisting of four staggered rows, with the smallest distance between microphones being 5 cm and the largest being 21.8 cm. DEMAND is freely available under a Creative Commons license to encourage research into algorithms beyond the stereo setup.

Echolocation intensity and directionality of perching and flying fringe-lipped bats, Trachops cirrhosus (Phyllostomidae)

The Neotropical frog-eating bat, Trachops cirrhosus, primarily hunts stationary prey, either by gleaning on the wing, or in a sit-and-wait mode hanging from a perch. It listens passively for prey-generated sounds, but uses echolocation in all stages of the hunt. Like other bats in the family Phyllostomidae, T. cirrhosus has a conspicuous nose leaf, hypothesized to direct and focus echolocation calls emitted from the nostrils. T. cirrhosus is highly flexible in its cognitive abilities and its use of sensory strategies for prey detection. Additionally, T. cirrhosus has been observed to echolocate both with closed and open mouth. We hypothesize that its flexibility extends to echolocation call design. We investigated the effect of hunting mode, perching or flying, as well as the effect of mouth opening, on the acoustic parameters and directionality of the echolocation call. We used a multi-microphone array, a high-speed video camera, and a microphone-diode-video system to directly visualize the echolocation sound beam synchronized with the bat's behavior. We found that T. cirrhosus emits a highly directional sound beam with half amplitude angle (HAM) of 12–18° and DI (directionality index) of ~17 dB, among the most directional bat sonar beams measured to date. The directionality was high both when flying and when perching. The emitted intensity was low, around 88 dB SPL at 10 cm from the mouth, when hanging, but higher, around 100 dB SPL at 10 cm, when flying or just before take-off. Our data suggests that the limited search volume of T. cirrhosus sonar beam defined by the high directionality and the rather low intensity of its echolocation calls is adapted to the highly cluttered hunting habitat and to the perch hunting mode.

Impact of rigid sphere scattering on measurement of acoustic shocks

Multi-microphone arrays embedded in a rigid spherical housing have been used to estimate field quantities such as vector intensity. However, the measured pressure waveforms are modified by the scattering of the incident pressure wave on the sphere. The frequency response function and the corresponding impulse response function for microphones in a sphere can be created using Mie scattering theory. This permits analysis of the scattering effect on pressure measurements for shock-containing broadband waveforms, such as those produced by solid rocket motors. The abrupt pressure rises associated with shocks are seen to be either overestimated or underestimated depending on the angle of the measuring microphone. These shocks are the most affected portion of the waveform due to their high frequency content. In addition, deconvolution of measured rocket signals using the predicted impulse responses of the microphones in the array provides an apparently consistent estimate of the free-field signal at the probe center.

An experimental comparison of the focused beamformer and the inverse method for the characterisation of acoustic sources in ideal and non-ideal acoustic environments

The current availability and affordability of multi-microphone array systems has awakened a strong interest in sound source location and characterisation in many fields of experimental acoustics and noise control. Although the theory behind the design of such arrays has been known for some time, the algorithms used to process the microphone signals are the subject of on-going research and development. The beamformer algorithm is well-known and relatively simple to implement. It is useful for ‘scanning’ an area to find the position of sound sources but has limited ability to characterise fully the spatial distribution of the strength of an acoustic source. The inverse method, on the other hand, requires prior information regarding the position of sources or an appropriate discretisation of the source strength distribution. The method is more difficult to implement but can yield more useful source characterisation data. This paper presents a comparison between the two methods based on experimental data. The results show that, in the presence of more than one source, the beamformer cannot yield reliable estimates of the source strength of individual sources since the output from the beamformer is shown to be dependent upon the degree of correlation between multiple sources as well as the source strengths themselves. The inverse method, on the other hand, is shown to yield reliable estimates of source strength when more than one source is present, regardless of the correlation between the sources, although the results presented here are restricted to the case of a relatively small number of sources. It is clearly demonstrated, both theoretically and through carefully controlled experiments, that either method can be used effectively under reverberant conditions through the use of measured Green functions in place of the simple geometrically-derived free-space values of the Green function. This greatly improves the possibility for the successful use of these methods in many important industrial applications.

Frequency alternation and an offbeat rhythm indicate foraging behavior in the echolocating bat, Saccopteryx bilineata

The greater sac-winged bat, Saccopteryx bilineata (Emballonuridae), uses two distinct echolocation call sequences: a 'monotonous' sequence, where bats emit ~48 kHz calls at a relatively stable rate, and a frequency-alternating sequence, where bats emit calls at ~45 kHz (low-note call) and ~48 kHz (high-note call). The frequencies of these low-high-note pairs remain stable within sequences. In Panama, we recorded echolocation calls from S. bilineata with a multi-microphone array at two sites: one a known roosting site, the other a known foraging site. Our results indicate that this species (1) only produces monotonous sequences in non-foraging contexts and, at times, directly after emitting a feeding buzz and (2) produces frequency-alternating sequences when actively foraging. These latter sequences are also characterized by an unusual, offbeat emission rhythm. We found significant positive relationships between (1) call intensity and call duration and (2) call intensity and distance from clutter. However, these relationships were weaker than those reported for bats from other families. We speculate on how call frequency alternation and an offbeat emission rhythm might reflect a novel strategy for prey detection at the edge of complex habitat in this ancient family of bats.

The Algorithm Research of Adaptive Noise Canceller based on Multi-Microphone Arrays and Beamforming

This paper presents a microphone array with delay and sums beamforming(DASB) while using an adaptive noise canceller(ANC) with reference microphone array to suppress coherent as well as incoherent noise in disturbed speech signals.The algorithm does not rely on any signal model and dispenses with assumption of statistical properties of noise.It can be seasoned with changes of background and used for target speech recognition.The algorithm is of good real-time performance and robustness.The simulation results demonstrate that the proposed method works well in the environment with different noise.

<p>Book Review: Digital Hearing AIDS</p>

SYNOPSIS: This valuable resource on digital signal processing (DSP) principles and approaches is designed to inform and guide anyone interested in hearing sciences. The text focuses on the complexities of signal processing algorithms used in modern digital hearing aids and therein presumes the reader has a basic understanding of hearing aid fitting algorithms and their mathematical concepts and representations. Common features found in modern hearing aids such as feedback cancellation, wind protection screens, dynamic range compression, multiple and single microphones arc described with clarity in the text and additionally illustrated with mathematical representations.
 The textbook contains 13 chapters that range from the basics of analog to digital Hearing Aid Technologies, in Chapter 1 through the specific algorithms required to balance Bilateral Signal Processing in the final Chapter 13. In between these chapters, the reader is confronted with the myriad of insights and challenges to the hearing aid user including “the Cocktail Party” and other effects that compromise desired signal to noise ratios. Chapter 2 introduces Signal Processing Basics; Chapters 3 - 7 deal with the DSP hardware and algorithms; 3) the Electro-acoustic System; 4) Directional Microphones; 5) Adaptive and Multi-microphone Arrays; 6) Wind Noise-, and 7) Feedback Cancdlation.Thc remaining chapters cover signal processing related to physiological and perception dynamics: 8) Single Microphone Noise Suppression; 9) Spectral Subtraction; 10) Spectral Modifications; II) Sound Modifications (to improve speech intelligibility); and 12) Sound Classification to help determine parameters for algorithms in challenging listening environments.

Vespertilionid bats control the width of their biosonar sound beam dynamically during prey pursuit

Animals using sound for communication emit directional signals, focusing most acoustic energy in one direction. Echolocating bats are listening for soft echoes from insects. Therefore, a directional biosonar sound beam greatly increases detection probability in the forward direction and decreases off-axis echoes. However, high directionality has context-specific disadvantages: at close range the detection space will be vastly reduced, making a broad beam favorable. Hence, a flexible system would be very advantageous. We investigated whether bats can dynamically change directionality of their biosonar during aerial pursuit of insects. We trained five Myotis daubentonii and one Eptesicus serotinus to capture tethered mealworms and recorded their echolocation signals with a multimicrophone array. The results show that the bats broaden the echolocation beam drastically in the terminal phase of prey pursuit. M. daubentonii increased the half-amplitude angle from approximately 40 degrees to approximately 90 degrees horizontally and from approximately 45 degrees to more than 90 degrees vertically. The increase in beam width is achieved by lowering the frequency by roughly one octave from approximately 55 kHz to approximately 27.5 kHz. The E. serotinus showed beam broadening remarkably similar to that of M. daubentonii. Our results demonstrate dynamic control of beam width in both species. Hence, we propose directionality as an explanation for the frequency decrease observed in the buzz of aerial hawking vespertilionid bats. We predict that future studies will reveal dynamic control of beam width in a broad range of acoustically communicating animals.