AMDM-SE: Attention-based Multichannel Diffusion Model for Speech Enhancement

Diffusion models have recently achieved impressive results in reconstructing images from noisy inputs, and similar ideas have been applied to speech enhancement by treating time-frequency representations as images. With the ubiquity of multi-microphone devices, we extend state-of-the-art diffusion-based methods to exploit multichannel inputs for improved performance. Multichannel diffusion-based enhancement remains in its infancy, with prior work making limited use of advanced mechanisms such as attention for spatial modeling - a gap addressed in this paper. We propose AMDM-SE, an Attention-based Multichannel Diffusion Model for Speech Enhancement, designed specifically for noise reduction. AMDM-SE leverages spatial inter-channel information through a novel cross-channel time-frequency attention block, enabling faithful reconstruction of fine-grained signal details within a generative diffusion framework. On the CHiME-3 benchmark, AMDM-SE outperforms both a single-channel diffusion baseline and a multichannel model without attention, as well as a strong DNN-based predictive method. Simulated-data experiments further underscore the importance of the proposed multichannel attention mechanism. Overall, our results show that incorporating targeted multichannel attention into diffusion models substantially improves noise reduction. While multichannel diffusion-based speech enhancement is still an emerging field, our work contributes a new and complementary approach to the growing body of research in this direction.

Spectral or spatial? Leveraging both for speaker extraction in challenging data conditions

This paper presents a robust multi-channel speaker extraction algorithm designed to handle inaccuracies in reference information. While existing approaches often rely solely on either spatial or spectral cues to identify the target speaker, our method integrates both sources of information to enhance robustness. A key aspect of our approach is its emphasis on stability, ensuring reliable performance even when one of the features is degraded or misleading. Given a noisy mixture and two potentially unreliable cues, a dedicated network is trained to dynamically balance their contributions-or disregard the less informative one when necessary. We evaluate the system under challenging conditions by simulating inference-time errors using a simple direction of arrival (DOA) estimator and a noisy spectral enrollment process. Experimental results demonstrate that the proposed model successfully extracts the desired speaker even in the presence of substantial reference inaccuracies.

A Study of Binaural Deep Beamforming With Interpretable Beampatterns Guided by Time-Varying RTF

In this work, a deep beamforming framework for speech enhancement in dynamic acoustic environments is studied. The time-varying beamformer weights are estimated from the noisy multichannel signals by minimizing an SI-SDR loss. The estimation is guided by the continuously tracked relative transfer functions (RTFs) of the moving target speaker. The spatial behavior of the network is evaluated through both narrowband and wideband beampatterns under three settings: (i) oracle guidance using true RTFs, (ii) estimated RTFs obtained by a subspace tracking method, and (iii) without the RTF guidance. Results show that RTF-guided models produce smoother, spatially consistent beampatterns that accurately track the target's direction of arrival. In contrast, the model fails to maintain a clear spatial focus when guidance is absent. Using the estimated RTFs as guidance closely matches the oracle RTF behavior, confirming the effectiveness of the tracking scheme. The model also outputs a binaural signal to preserve the speaker's spatial cues, which promotes hearing aid and hearables applications.

(SP)$^2$-Net: A Neural Spatial Spectrum Method for DOA Estimation

We consider the problem of estimating the directions of arrival (DOAs) of multiple sources from a single snapshot of an antenna array, a task with many practical applications. In such settings, the classical Bartlett beamformer is commonly used, as maximum likelihood estimation becomes impractical when the number of sources is unknown or large, and spectral methods based on the sample covariance are not applicable due to the lack of multiple snapshots. However, the accuracy and resolution of the Bartlett beamformer are fundamentally limited by the array aperture. In this paper, we propose a deep learning technique, comprising a novel architecture and training strategy, for generating a high-resolution spatial spectrum from a single snapshot. Specifically, we train a deep neural network that takes the measurements and a hypothesis angle as input and learns to output a score consistent with the capabilities of a much wider array. At inference time, a heatmap can be produced by scanning an arbitrary set of angles. We demonstrate the advantages of our trained model, named (SP)$^2$-Net, over the Bartlett beamformer and sparsity-based DOA estimation methods.

Transient Noise Removal via Diffusion-based Speech Inpainting

In this paper, we present PGDI, a diffusion-based speech inpainting framework for restoring missing or severely corrupted speech segments. Unlike previous methods that struggle with speaker variability or long gap lengths, PGDI can accurately reconstruct gaps of up to one second in length while preserving speaker identity, prosody, and environmental factors such as reverberation. Central to this approach is classifier guidance, specifically phoneme-level guidance, which substantially improves reconstruction fidelity. PGDI operates in a speaker-independent manner and maintains robustness even when long segments are completely masked by strong transient noise, making it well-suited for real-world applications, such as fireworks, door slams, hammer strikes, and construction noise. Through extensive experiments across diverse speakers and gap lengths, we demonstrate PGDI's superior inpainting performance and its ability to handle challenging acoustic conditions. We consider both scenarios, with and without access to the transcript during inference, showing that while the availability of text further enhances performance, the model remains effective even in its absence. For audio samples, visit: https://mordehaym.github.io/PGDI/

Few-Shot Speech Deepfake Detection Adaptation with Gaussian Processes

Recent advancements in Text-to-Speech (TTS) models, particularly in voice cloning, have intensified the demand for adaptable and efficient deepfake detection methods. As TTS systems continue to evolve, detection models must be able to efficiently adapt to previously unseen generation models with minimal data. This paper introduces ADD-GP, a few-shot adaptive framework based on a Gaussian Process (GP) classifier for Audio Deepfake Detection (ADD). We show how the combination of a powerful deep embedding model with the Gaussian processes flexibility can achieve strong performance and adaptability. Additionally, we show this approach can also be used for personalized detection, with greater robustness to new TTS models and one-shot adaptability. To support our evaluation, a benchmark dataset is constructed for this task using new state-of-the-art voice cloning models.

Video Editing for Audio-Visual Dubbing

Visual dubbing, the synchronization of facial movements with new speech, is crucial for making content accessible across different languages, enabling broader global reach. However, current methods face significant limitations. Existing approaches often generate talking faces, hindering seamless integration into original scenes, or employ inpainting techniques that discard vital visual information like partial occlusions and lighting variations. This work introduces EdiDub, a novel framework that reformulates visual dubbing as a content-aware editing task. EdiDub preserves the original video context by utilizing a specialized conditioning scheme to ensure faithful and accurate modifications rather than mere copying. On multiple benchmarks, including a challenging occluded-lip dataset, EdiDub significantly improves identity preservation and synchronization. Human evaluations further confirm its superiority, achieving higher synchronization and visual naturalness scores compared to the leading methods. These results demonstrate that our content-aware editing approach outperforms traditional generation or inpainting, particularly in maintaining complex visual elements while ensuring accurate lip synchronization.

DiffusionRIR: Room Impulse Response Interpolation using Diffusion Models

Room Impulse Responses (RIRs) characterize acoustic environments and are crucial in multiple audio signal processing tasks. High-quality RIR estimates drive applications such as virtual microphones, sound source localization, augmented reality, and data augmentation. However, obtaining RIR measurements with high spatial resolution is resource-intensive, making it impractical for large spaces or when dense sampling is required. This research addresses the challenge of estimating RIRs at unmeasured locations within a room using Denoising Diffusion Probabilistic Models (DDPM). Our method leverages the analogy between RIR matrices and image inpainting, transforming RIR data into a format suitable for diffusion-based reconstruction. Using simulated RIR data based on the image method, we demonstrate our approach's effectiveness on microphone arrays of different curvatures, from linear to semi-circular. Our method successfully reconstructs missing RIRs, even in large gaps between microphones. Under these conditions, it achieves accurate reconstruction, significantly outperforming baseline Spline Cubic Interpolation in terms of Normalized Mean Square Error and Cosine Distance between actual and interpolated RIRs. This research highlights the potential of using generative models for effective RIR interpolation, paving the way for generating additional data from limited real-world measurements.

End-to-End Multi-Microphone Speaker Extraction Using Relative Transfer Functions

This paper introduces a multi-microphone method for extracting a desired speaker from a mixture involving multiple speakers and directional noise in a reverberant environment. In this work, we propose leveraging the instantaneous relative transfer function (RTF), estimated from a reference utterance recorded in the same position as the desired source. The effectiveness of the RTF-based spatial cue is compared with direction of arrival (DOA)-based spatial cue and the conventional spectral embedding. Experimental results in challenging acoustic scenarios demonstrate that using spatial cues yields better performance than the spectral-based cue and that the instantaneous RTF outperforms the DOA-based spatial cue.

Explainable DNN-based Beamformer with Postfilter

This paper introduces an explainable DNN-based beamformer with a postfilter (ExNet-BF+PF) for multichannel signal processing. Our approach combines the U-Net network with a beamformer structure to address this problem. The method involves a two-stage processing pipeline. In the first stage, time-invariant weights are applied to construct a multichannel spatial filter, namely a beamformer. In the second stage, a time-varying single-channel post-filter is applied at the beamformer output. Additionally, we incorporate an attention mechanism inspired by its successful application in noisy and reverberant environments to improve speech enhancement further. Furthermore, our study fills a gap in the existing literature by conducting a thorough spatial analysis of the network's performance. Specifically, we examine how the network utilizes spatial information during processing. This analysis yields valuable insights into the network's functionality, thereby enhancing our understanding of its overall performance. Experimental results demonstrate that our approach is not only straightforward to train but also yields superior results, obviating the necessity for prior knowledge of the speaker's activity.