Deep, data-driven modeling of room acoustics: literature review and research perspectives

Our everyday auditory experience is shaped by the acoustics of the indoor environments in which we live. Room acoustics modeling is aimed at establishing mathematical representations of acoustic wave propagation in such environments. These representations are relevant to a variety of problems ranging from echo-aided auditory indoor navigation to restoring speech understanding in cocktail party scenarios. Many disciplines in science and engineering have recently witnessed a paradigm shift powered by deep learning (DL), and room acoustics research is no exception. The majority of deep, data-driven room acoustics models are inspired by DL-based speech and image processing, and hence lack the intrinsic space-time structure of acoustic wave propagation. More recently, DL-based models for room acoustics that include either geometric or wave-based information have delivered promising results, primarily for the problem of sound field reconstruction. In this review paper, we will provide an extensive and structured literature review on deep, data-driven modeling in room acoustics. Moreover, we position these models in a framework that allows for a conceptual comparison with traditional physical and data-driven models. Finally, we identify strengths and shortcomings of deep, data-driven room acoustics models and outline the main challenges for further research.

The trajectoRIR Database: Room Acoustic Recordings Along a Trajectory of Moving Microphones

Data availability is essential to develop acoustic signal processing algorithms, especially when it comes to data-driven approaches that demand large and diverse training datasets. For this reason, an increasing number of databases have been published in recent years, including either room impulse responses (RIRs) or recordings of moving audio. In this paper we introduce the trajectoRIR database, an extensive, multi-array collection of both dynamic and stationary acoustic recordings along a controlled trajectory in a room. Specifically, the database features recordings using moving microphones and stationary RIRs spatially sampling the room acoustics along an L-shaped, 3.74-meter-long trajectory. This combination makes trajectoRIR unique and applicable in various tasks ranging from sound source localization and tracking to spatially dynamic sound field reconstruction and system identification. The recording room has a reverberation time of 0.5 seconds, and the three different microphone configurations employed include a dummy head, with additional reference microphones located next to the ears, 3 first-order Ambisonics microphones, two circular arrays of 16 and 4 channels, and a 12-channel linear array. The motion of the microphones was achieved using a robotic cart traversing a rail at three speeds: [0.2,0.4,0.8] m/s. Audio signals were reproduced using two stationary loudspeakers. The collected database features 8648 stationary RIRs, as well as perfect sweeps, speech, music, and stationary noise recorded during motion. MATLAB and Python scripts are included to access the recorded audio as well as to retrieve geometrical information.

A Comparative Analysis of Generalised Echo and Interference Cancelling and Extended Multichannel Wiener Filtering for Combined Noise Reduction and Acoustic Echo Cancellation

Two algorithms for combined acoustic echo cancellation (AEC) and noise reduction (NR) are analysed, namely the generalised echo and interference canceller (GEIC) and the extended multichannel Wiener filter (MWFext). Previously, these algorithms have been examined for linear echo paths, and assuming access to voice activity detectors (VADs) that separately detect desired speech and echo activity. However, algorithms implementing VADs may introduce detection errors. Therefore, in this paper, the previous analyses are extended by 1) modelling general nonlinear echo paths by means of the generalised Bussgang decomposition, and 2) modelling VAD error effects in each specific algorithm, thereby also allowing to model specific VAD assumptions. It is found and verified with simulations that, generally, the MWFext achieves a higher NR performance, while the GEIC achieves a more robust AEC performance.

A Zero-Shot Physics-Informed Dictionary Learning Approach for Sound Field Reconstruction

Sound field reconstruction aims to estimate pressure fields in areas lacking direct measurements. Existing techniques often rely on strong assumptions or face challenges related to data availability or the explicit modeling of physical properties. To bridge these gaps, this study introduces a zero-shot, physics-informed dictionary learning approach to perform sound field reconstruction. Our method relies only on a few sparse measurements to learn a dictionary, without the need for additional training data. Moreover, by enforcing the Helmholtz equation during the optimization process, the proposed approach ensures that the reconstructed sound field is represented as a linear combination of a few physically meaningful atoms. Evaluations on real-world data show that our approach achieves comparable performance to state-of-the-art dictionary learning techniques, with the advantage of requiring only a few observations of the sound field and no training on a dataset.

Integrated Minimum Mean Squared Error Algorithms for Combined Acoustic Echo Cancellation and Noise Reduction

In many speech recording applications, noise and acoustic echo corrupt the desired speech. Consequently, combined noise reduction (NR) and acoustic echo cancellation (AEC) is required. Generally, a cascade approach is followed, i.e., the AEC and NR are designed in isolation by selecting a separate signal model, formulating a separate cost function, and using a separate solution strategy. The AEC and NR are then cascaded one after the other, not accounting for their interaction. In this paper, however, an integrated approach is proposed to consider this interaction in a general multi-microphone/multi-loudspeaker setup. Therefore, a single signal model of either the microphone signal vector or the extended signal vector, obtained by stacking microphone and loudspeaker signals, is selected, a single mean squared error cost function is formulated, and a common solution strategy is used. Using this microphone signal model, a multi channel Wiener filter (MWF) is derived. Using the extended signal model, an extended MWF (MWFext) is derived, and several equivalent expressions are found, which nevertheless are interpretable as cascade algorithms. Specifically, the MWFext is shown to be equivalent to algorithms where the AEC precedes the NR (AEC NR), the NR precedes the AEC (NR-AEC), and the extended NR (NRext) precedes the AEC and post-filter (PF) (NRext-AECPF). Under rank-deficiency conditions the MWFext is non-unique, such that this equivalence amounts to the expressions being specific, not necessarily minimum-norm solutions for this MWFext. The practical performances nonetheless differ due to non-stationarities and imperfect correlation matrix estimation, resulting in the AEC-NR and NRext-AEC-PF attaining best overall performance.

State-Space Estimation of Spatially Dynamic Room Impulse Responses using a Room Acoustic Model-based Prior

The estimation of room impulse responses (RIRs) between static loudspeaker and microphone locations can be done 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 point 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.

Reference Microphone Selection for the Weighted Prediction Error Algorithm using the Normalized L-p Norm

Reverberation may severely degrade the quality of speech signals recorded using microphones in a room. For compact microphone arrays, the choice of the reference microphone for multi-microphone dereverberation typically does not have a large influence on the dereverberation performance. In contrast, when the microphones are spatially distributed, the choice of the reference microphone may significantly contribute to the dereverberation performance. In this paper, we propose to perform reference microphone selection for the weighted prediction error (WPE) dereverberation algorithm based on the normalized $\ell_p$-norm of the dereverberated output signal. Experimental results for different source positions in a reverberant laboratory show that the proposed method yields a better dereverberation performance than reference microphone selection based on the early-to-late reverberation ratio or signal power.

Frequency Tracking Features for Data-Efficient Deep Siren Identification

The identification of siren sounds in urban soundscapes is a crucial safety aspect for smart vehicles and has been widely addressed by means of neural networks that ensure robustness to both the diversity of siren signals and the strong and unstructured background noise characterizing traffic. Convolutional neural networks analyzing spectrogram features of incoming signals achieve state-of-the-art performance when enough training data capturing the diversity of the target acoustic scenes is available. In practice, data is usually limited and algorithms should be robust to adapt to unseen acoustic conditions without requiring extensive datasets for re-training. In this work, given the harmonic nature of siren signals, characterized by a periodically evolving fundamental frequency, we propose a low-complexity feature extraction method based on frequency tracking using a single-parameter adaptive notch filter. The features are then used to design a small-scale convolutional network suitable for training with limited data. The evaluation results indicate that the proposed model consistently outperforms the traditional spectrogram-based model when limited training data is available, achieves better cross-domain generalization and has a smaller size.

One-Shot Distributed Node-Specific Signal Estimation with Non-Overlapping Latent Subspaces in Acoustic Sensor Networks

A one-shot algorithm called iterationless DANSE (iDANSE) is introduced to perform distributed adaptive node-specific signal estimation (DANSE) in a fully connected wireless acoustic sensor network (WASN) deployed in an environment with non-overlapping latent signal subspaces. The iDANSE algorithm matches the performance of a centralized algorithm in a single processing cycle while devices exchange fused versions of their multichannel local microphone signals. Key advantages of iDANSE over currently available solutions are its iterationless nature, which favors deployment in real-time applications, and the fact that devices can exchange fewer fused signals than the number of latent sources in the environment. The proposed method is validated in numerical simulations including a speech enhancement scenario.

Topology-Independent GEVD-Based Distributed Adaptive Node-Specific Signal Estimation in Ad-Hoc Wireless Acoustic Sensor Networks

A low-rank approximation-based version of the topology-independent distributed adaptive node-specific signal estimation (TI-DANSE) algorithm is introduced, using a generalized eigenvalue decomposition (GEVD) for application in ad-hoc wireless acoustic sensor networks. This TI-GEVD-DANSE algorithm as well as the original TI-DANSE algorithm exhibit a non-strict convergence, which can lead to numerical instability over time, particularly in scenarios where the estimation of accurate spatial covariance matrices is challenging. An adaptive filter coefficient normalization strategy is proposed to mitigate this issue and enable the stable performance of TI-(GEVD-)DANSE. The method is validated in numerical simulations including dynamic acoustic scenarios, demonstrating the importance of the additional normalization.