Dr Randall Ali

In studying acoustic signal processing, students are faced with seemingly abstract concepts such as continuous-time and discrete-time representations, time-frequency transformations, convolution, vector and inner product spaces, orthogonality, and optimal and adaptive filtering. While mathematical development of the theory and illustration through simple examples during teaching can help shed light on some of these concepts, they are not always sufficient for students to achieve a full level of understanding. As a means to improve this situation, we have consequently developed several interactive demonstrations in acoustic signal processing, covering a range of topics, from the fundamentals of sampling and quantisation to the more advanced optimal and adaptive filtering. These demonstrations have been made with Jupyter notebooks and offer several advantages. Apart from being open source, students immediately get to work with realistic data as they can generate and record acoustic signals using their computer’s loudspeaker and microphone. Using the live coding aspect of the notebooks, they can also quickly process, visualise, and listen to the results from the processing of these acoustic signals, as well as repeat this workflow for various parameter sweeps. We will present some of these demonstrations and show how they can serve as an effective teaching aid.

The task of removing sinusoidal components from observed signals can be accomplished by using a notch filter with a specific attenuation at a particular frequency. In some applications, however, such as acoustic feedback control, the frequency at which attenuation is required is unknown and possibly time-varying, and hence an adaptive notch filter is a more appropriate solution. Transitioning from a fixed notch filter to an adaptive one is by no means trivial and involves the understanding of a range of digital signal processing (DSP) topics from pole-zero placement techniques for designing infinite impulse response filters to optimal and adaptive filtering algorithms. In the signal processing algorithms and implementation graduate course taught at KU Leuven (Belgium), we study the design of an adaptive notch filter, which is based on a constrained biquadratic IIR representation, andwhose parameters are updated using a least-mean-square algorithm. Students also have to implement the algorithm on a 16-bit DSP TMS320C5515. In this presentation, we will discuss the design and implementation challenges of this adaptive notch filter and how it serves as an illustrative example/homework problem where several aspects of DSP are interwoven.

In massive distributed microphone arrays, such as in conference systems, the use of all sensors leads to unnecessary energy consumption and limited network life, as some sensors may have a limited contribution to specific estimation tasks, such as sound source localization (SSL). In this work, we propose two pre-processing methods for microphone pair selection in steered response power (SRP) based SSL using massive arrays of spatially distributed microphones. The first method is based on thresholding which results in a selection of broadside microphone pairs, that is, microphone pairs that are oriented broadside to the source direction, and only the signals from those microphones are collected for use in the SRP algorithm. The second method is based on sparseness of the cross-correlation in which the pairs which have a high sparsity factor are selected for use in localization. These methods are further improved by leveraging orientational diversity. Simulations show that we achieve a localization performance only using the selected microphone pairs that is comparable to the performance when all microphones are employed.

In this work, we propose to utilize a recently developed covariance matrix interpolation technique in order to improve noise reduction in multi-microphone setups in the presence of a moving, localized noise source. Based on the concept of optimal mass transport, the proposed method induces matrix interpolants implying smooth spatial displacement of the noise source, allowing for physically reasonable reconstructions of the noise source trajectory. As this trajectory is constructed as to connect two observed, or estimated, covariance matrices, the proposed method is suggested for offline applications. The performance of the proposed method is demonstrated using simulations of a speech enhancement scenario.

The use of an external microphone in conjunction with an existing local microphone array can be particularly beneficial for noise reduction tasks that are critical for hearing devices, such as cochlear implants and hearing aids. Recent work has already demonstrated how an external microphone signal can be effectively incorporated into the common noise reduction technique of using a Minimum Variance Distortionless Response (MVDR) beamformer. In this paper, we provide a further extension, whereby an external microphone signal can be incorporated into an existing framework of a Generalised Sidelobe Canceller (GSC) that has been designed for a local microphone signal array. It will be shown that the resulting GSC with an external microphone results in an easily implementable addition to the existing GSC framework for a local microphone array, and can exhibit an improved noise reduction performance.

In this paper, we propose a distributed cross-relation-based adaptive algorithm for blind identification of single-input multiple-output (SIMO) systems in the frequency domain, using the alternating direction method of multipliers (ADMM) in a wireless sensor network (WSN). The network consists of a fixed number of nodes each equipped with a processing unit and a sensor that represents an output channel of the SIMO system. The proposed algorithm exploits the separability of the cross-channel relations by splitting the multichannel identification problem into sub-problems containing a subset of channels, in a way that is determined by the network topology. Each node delivers estimates for the subset of channel frequency responses, which are then combined into a consensus estimate per channel using general-form consensus ADMM in an adaptive updating scheme. Using numerical simulations, we show that it is possible to achieve convergence speeds and steady-state misalignment values comparable to fully centralized low-cost frequency-domain algorithms.

The purpose of this study is to characterize the intelligibility of a corpus of Vowel-Consonant-Vowel (VCV) stimuli recorded in five languages (English, French, German, Italian and Portuguese) in order to identify a subset of stimuli for screening individuals of unknown language during speech-in-noise tests. The intelligibility of VCV stimuli was estimated by combining the psychometric functions derived from the Short-Time Objective Intelligibility (STOI) measure with those derived from listening tests. To compensate for the potential increase in speech recognition effort in non-native listeners, stimuli were selected based on three criteria: (i) higher intelligibility; (ii) lower variability of intelligibility; and (iii) shallower psychometric function. The observed intelligibility estimates show that the three criteria for application in multilingual settings were fulfilled by the set of VCVs in English (average intelligibility from 1% to 8% higher; SRT from 4.01 to 2.04 dB SNR lower; average variability up to four times lower; slope from 0.35 to 0.68%/dB SNR lower). Further research is needed to characterize the intelligibility of these stimuli in a large sample of non-native listeners with varying degrees of hearing loss and to determine the possible effects of hearing loss and native language on VCV recognition.

A minimum variance distortionless response (MVDR) beamformer can be an effective multi-microphone noise reduction strategy' provided that a vector of transfer functions from the desired speech signal at a reference microphone to the other microphones, i.e. a vector of the relative transfer functions (RTFs), is known. When using a local microphone array (LMA) and an external microphone (XM), this RTF vector has two distinct parts: an RTF vector for that of only the LMA and a single RTF component for the XM, with the reference microphone on the LMA. Whereas a priori assumptions can be made for the RTF vector for the LMA, the RTF for the XM must be estimated as the XM position is generally unknown. This paper investigates a procedure for estimating this unknown RTF by making use of the a priori RTF vector for the LMA, thereby completing the RTF vector for use of the MVDR beamformer. It is shown that such a procedure results in an Eigenvalue Decomposition (EVD) of a 2×2 matrix for a system of M microphones in the LMA and one XM. The resulting performance is evaluated within the context of a monaural MVDR beamformer.

Conventionally, the single constraint of the minimum variance distortionless response (MVDR) beamformer for speech enhancement has been defined using one of two approaches. Either it is based on a priori assumptions such as microphone characteristics, position, speech source location, and room acoustics, or on a relative transfer function (RTF) vector estimate using a data dependent method. Each approach has its respective merits and drawbacks and a decision usually has to be made between one of the approaches. In this paper, an alternative approach of using an integrated MVDR beamformer is investigated, where both the hard constraints from the two conventional approaches are softened to yield two tuning parameters. It will be shown that this integrated MVDR beamformer can be expressed as a convex combination of the conventional MVDR beamformers, a linearly constrained minimum variance (LCMV) beamformer, and an all-zero vector, with real, positive-valued coefficients. By analysing how the tuning parameters affect these coefficients, two tuning rules for a practical implementation of the integrated MVDR are subsequently proposed. An evaluation with simulated and recorded data demonstrates that the integrated MVDR beamformer can be beneficial as opposed to relying on either of the conventional MVDR beamformers.

Distributed signal-processing algorithms in (wireless) sensor networks often aim to decentralize processing tasks to reduce communication cost and computational complexity or avoid reliance on a single device (i.e., fusion center) for processing. In this contribution, we extend a distributed adaptive algorithm for blind system identification that relies on the estimation of a stacked network-wide consensus vector at each node, the computation of which requires either broadcasting or relaying of node-specific values (i.e., local vector norms) to all other nodes. The extended algorithm employs a distributed-averaging-based scheme to estimate the network-wide consensus norm value by only using the local vector norm provided by neighboring sensor nodes. We introduce an adaptive mixing factor between instantaneous and recursive estimates of these norms for adaptivity in a time-varying system. Simulation results show that the extension provides estimation results close to the optimal fully-connected-network or broadcasting case while reducing inter-node transmission significantly.

A method for estimating the relevant quantities in a multi-channel Wiener filter (MWF) for speech dereverberation is proposed for a microphone system consisting of a local microphone array (LMA) and a single external microphone (XM). Typically these MWF quantities can be estimated by considering pre-whitened correlation matrices with a dimension equal to the number of microphones in the system. By following another procedure involving a pre-whitening-transformation operation, it will be demonstrated that when a priori knowledge of the relative transfer function (RTF) vector pertaining to only the LMA is available and when the reverberant component of the signals received by the LMA is uncorrelated with that of the XM, the MWF quantities may be alternatively estimated from a 2 x 2 matrix. Simulations confirm that using such an estimate results in a similar performance to that obtained by using the higher-dimensional correlation matrix.

In the development of acoustic signal processing algorithms, their evaluation in various acoustic environments is of utmost importance. In order to advance evaluation in realistic and reproducible scenarios, several high-quality acoustic databases have been developed over the years. In this paper, we present another complementary database of acoustic recordings, referred to as the Multi-arraY Room Acoustic Database (MYRiAD). The MYRiAD database is unique in its diversity of microphone configurations suiting a wide range of enhancement and reproduction applications (such as assistive hearing, teleconferencing, or sound zoning), the acoustics of the two recording spaces, and the variety of contained signals including 1214 room impulse responses (RIRs), reproduced speech, music, and stationary noise, as well as recordings of live cocktail parties held in both rooms. The microphone configurations comprise a dummy head (DH) with in-ear omnidirectional microphones, two behind-the-ear (BTE) pieces equipped with 2 omnidirectional microphones each, 5 external omnidirectional microphones (XMs), and two concentric circular microphone arrays (CMAs) consisting of 12 omnidirectional microphones in total. The two recording spaces, namely the SONORA Audio Laboratory (SAL) and the Alamire Interactive Laboratory (AIL), have reverberation times of 2.1s and 0.5s, respectively. Audio signals were reproduced using 10 movable loudspeakers in the SAL and a built-in array of 24 loudspeakers in the AIL. MATLAB and Python scripts are included for accessing the signals as well as microphone and loudspeaker coordinates. For a detailed description, please refer to the paper (preprint, published). Two files are provided, containing two different versions of the database: MYRiAD_V2.zip The full version of the database (31.3 GB). MYRiAD_V2_econ.zip The economy-sized version, containing source signals and RIRs only (201.7 MB). If you use the database, please cite the paper as follows: @article{dietzen2023myriad, author = {Dietzen, T. and Ali, R. and Taseska, M. and van Waterschoot, T.}, title = {{MYRiAD}: A Multi-Array Room Acoustic Database}, journal = {EURASIP J. Audio Speech Music Process.}, volume = {2023, article no. 17}, number = {}, month = {Apr.}, year = {2023}, pages = {1--14} } ___________________________________________________________________________________________________________ Change log (as compared to Version 1.0): Fixed erroneous file names in /audio/AIL/SU1/P2/. In the full version, applied a time shift to some of the speech, noise, and music recordings in the SAL (at most 2 samples, compensating for a slow phase drift, see manuscript for further details). Created an economy-sized version of the database containing source signals and RIRs only. Adjusted the following scripts for the economy-sized version: - /tools/MATLAB/load_audio_data.m - /tools/Python/load_audio_data.py

In wireless acoustic sensor networks (WASNs), the conventional steered response power (SRP) approach to source localization requires each node to transmit its microphone signal to a fusion center. As an alternative, this paper proposes two different fusion strategies for local, single-node SRP maps computed using only the microphone pairs within a node. In the first fusion strategy, we sum all single-node SRP maps in a fusion center, requiring less communication than the conventional SRP approach because the single-node SRP maps typically have less parameters than the raw microphone signals. In the second fusion strategy, the single-node SRP maps are distributively averaged without using a fusion center, requiring communication amongst connected nodes only. Simulations show that we achieve a good trade-off between communicational load and localization performance.

While substantial noise reduction and speech enhancement can be achieved with multiple microphones organized in an array, in some cases, such as when the microphone spacings are quite close, it can also he quite limited. This degradation can, however, be resolved by the introduction of one or more external microphones (XMs) into the same physical space as the local microphone array (LMA). In this paper, three methods of extending an LMA-based generalized sidelobe canceller (GSC-LMA) with multiple XMs are proposed in such a manner that the relative transfer function pertaining to the LMA is treated as a priori knowledge. Two of these methods involve a procedure for completing an extended blocking matrix, whereas the third uses the speech estimate from the GSC-LMA directly with an orthogonalized version of the XM signals to obtain an improved speech estimate via a rank-i generalized eigenvalue decomposition. All three methods were evaluated with recorded data from an office room and it was found that the third method could offer the most improvement. It was also shown that in using this method, the speech estimate from the GSC-LMA was not compromised and would be available to the listener if so desired, along with the improved speech estimate that uses both the LMA and XMs.

The Minimum Variance Distortionless Response (MVDR) beam-former is a popular multi-microphone noise reduction and speech enhancement strategy that can be implemented either as a fixed-constraint MVDR beamformer, with a pre-defined Relative Transfer Function (RTF) or based on a Multi-channel Wiener Filter (MWF) estimate. However, each implementation is not fully robust within a dynamic acoustic environment. For instance, performance degradations exist for the fixed-constraint MVDR beamformer when the source is not in the constraint direction and also for the MWF when the estimated RTF is poor. In this paper, we propose a contingency noise reduction strategy that uses a Linearly Constrained MWF (LC-MWF) to combine the positive aspects of both implementations. We proceed to derive the LC-MWF in relation to the MVDR beam-former implementations and demonstrate through simulations that the LC-MWF is indeed an intermediary solution that encompasses a wider range of acoustic conditions.

With distributed signal processing gaining traction in the audio and speech processing landscape through the utilization of interconnected devices constituting wire-less acoustic sensor networks, additional challenges arise, including optimal data transmission between devices. In this paper, we extend an adaptive distributed blind system identification algorithm by introducing a residual-based adaptive coding scheme to minimize communication costs within the network. We introduce a coding scheme that takes advantage of the convergence of estimates, i.e., van-ishing residuals, to minimize information being sent. The scheme is adaptive, i.e., tracks changes in the estimated system and utilizes entropy coding and adaptive gain to fit the time-varying residual variance to pretrained codebooks. We use a low-complexity approach for gain adaptation, based on a recursive variance estimate. We demonstrate the approach's effectiveness with numerical simulations and its performance in various scenarios.

Estimating the position of animals over time provides useful additional information for understanding animal behavior and for ecology studies in general. A common approach for this task is to deploy microphone arrays (nodes) and use the acoustic signals to estimate the direction of arrival (DOA) of the sound source. DOAs from different nodes are then intersected to find the source's position. However, when multiple sources are active, the DOA association problem (AP) arises as it becomes unclear which DOAs correspond to the same source. This problem is further exacerbated in bioacoustical scenarios where large distances increase the error in the DOA estimates, and sounds often overlap in both time and frequency. In this paper, we propose a method to tackle the DOA AP in such challenging environments. In particular, we beamform to each of the estimated DOAs and extract features that characterize each of the detected sources, then, we associate features from different nodes based on their similarity, resulting in groups of DOAs that belong to the same source. Preliminary simulations suggest the potential of the proposed method for scenarios with missed detections and unknown number of sources, even when the number of microphones available at each node is limited.

Modeling late reverberation in real-time interactive applications is a challenging task when multiple sound sources and listeners are present in the same environment. This is especially problematic when the environment is geometrically complex and/or features uneven energy absorption (e.g. coupled volumes), because in such cases the late reverberation is dependent on the sound sources' and listeners' positions, and therefore must be adapted to their movements in real time.We present a novel approach to the task, named modal decomposition of acoustic radiance transfer (MoD-ART), which can handle highly complex scenarios with efficiency. The approach is based on the geometrical acousticsmethod of acoustic radiance transfer, fromwhich we extract a set of energy decaymodes and their positional relationships with sources and listeners. In this paper, we describe the physical and mathematical significance of MoD-ART, highlighting its advantages and applicability to different scenarios. Through an analysis of the method's computational complexity, we show that it compares very favorably with ray-tracing.We also present simulation results showing thatMoD-ART can capture multiple decay slopes and flutter echoes.

Room acoustic simulation using physically motivated sound propagation models are typically separated into wave-based methods and geometric methods. While each of these methods has been extensively studied, the question on when to transition from a wave-based to a geometric method still remains somewhat unclear. Towards building greater understanding of the links between wavebased and geometric methods, this paper investigates the transition question by using the method of stationary phase. As a starting point, we consider an elementary scenario with a geometrically interpretable analytic solution, namely that of an infinite rigid boundary mirroring a single monopole sound source, and apply the stationary phase approximation (SPA) to the wave-based boundary integral equation (BIE). The results of the analysis demonstrate how net boundary contributions give rise to the geometric interpretation offered by the SPA and provide the conditions when the SPA is asymptotically equal to the analytical solution in this elementary scenario. Although the results are unsurprising and intuitive, the insights gained from this analysis pave the way for investigating relations between wave-based and geometric methods in more complicated room acoustics scenarios.

Room impulse responses (RIRs) between static loudspeaker and microphone locations can be estimated using a number of well-established measurement and inference procedures. While these procedures assume a time-invariant acoustic system, time variations need to be considered for the case of spatially dynamic scenarios where loudspeakers and microphones are subject to movement. If the RIR is modeled using image sources, then movement implies that the distance to each image source varies over time, making the estimation of the spatially dynamic RIR particularly challenging. In this paper, we propose a procedure to estimate the early part of the spatially dynamic RIR between a stationary source and a microphone moving on a linear trajectory at constant velocity. The procedure is built upon a state-space model, where the state to be estimated represents the early RIR, the observation corresponds to a microphone recording in a spatially dynamic scenario, and time-varying distances to the image sources are incorporated into the state transition matrix obtained from static RIRs at the start and end points of the trajectory. The performance of the proposed approach is evaluated against state-of-the-art RIR interpolation and state-space estimation methods using simulations, demonstrating the potential of the proposed state-space model.

In rooms with complex geometry and uneven distribution of energy losses, late reverberation depends on the positions of sound sources and listeners. More precisely, the decay of energy is char-acterised by a sum of exponential curves with position-dependent amplitudes and position-independent decay rates (hence the name common slopes). The amplitude of different energy decay components is a particularly important perceptual aspect that requires efficient modeling in applications such as virtual reality and video games. Acoustic Radiance Transfer (ART) is a room acoustics model focused on late reverberation, which uses a pre-computed acoustic transfer matrix based on the room geometry and materials , and allows interactive changes to source and listener positions. In this work, we present an efficient common-slopes approximation of the ART model. Our technique extracts common slopes from ART using modal decomposition, retaining only the non-oscillating energy modes. Leveraging the structure of ART, changes to the positions of sound sources and listeners only require minimal processing. Experimental results show that even very few slopes are sufficient to capture the positional dependency of late reverberation, reducing model complexity substantially.

The steelpan is a percussive idiophone whose sound is generated by striking an arrangement of dome-shaped notes within the sunken top of a steel drum (open on the other end) with a mallet. The instrument originated in Trinidad and has its roots in the early twentieth century that can be traced back to post-emancipation traditions. As the steelpan has permeated the wider cultural space of the Caribbean in complex ways, it is important, within Caribbean cultural heritage, that it is safeguarded and its associated traditions are appropriately documented. In particular, tuning techniques of pioneer steelpan tuners from the early to mid-twentieth century are being lost to time as the majority of these tuners are already (or soon to be) deceased. This paper discusses a mathematical model of the steelpan that can be discretized and digitally synthesized, highlighting the role that sound synthesis models can play within the domain of cultural heritage. It is shown how the model can be used to understand the timbres of various steelpans and also demonstrates its potential as a point of reference for defining the object of the steelpan as a form of intangible cultural heritage.

When developing an auralization for acoustic scenarios involving moving sources and receivers, one key feature is the ability to simulate the Doppler shift, i.e., the changing frequency content from the receiver's perspective. As the time-varying delay between a source and receiver is what accounts for the Doppler shift, an approximation of this delay is required to successfully render the changes in frequency content at the receiver. Depending on the signal-processing strategy chosen to accomplish this task, there is, however , a potential to introduce audible artifacts due to frequency folding (aliasing), frequency replication (imaging), and broadband noise. In this paper we discuss the manifestation of such artifacts and propose a method to eliminate them, which can be integrated into the digital signal processing chain of larger aural-ization schemes. The method is built upon a source-time dominant approach and uses a combination of oversampling, interpolation, and time-varying filtering to predict and eliminate frequency regions at the receiver that are vulnerable to aliasing and imaging. We demonstrate the strengths and weaknesses of the method using a circularly moving source with a fixed receiver.