Speech dereverberation constrained on room impulse response characteristics

TL;DR

This paper introduces a physically interpretable deep learning approach for single-channel speech dereverberation that regularizes the model with a novel room impulse response coherence loss, ensuring the dereverberated speech aligns with room acoustics.

Contribution

It proposes a new physical coherence loss to regularize dereverberation models, making them more interpretable and physically consistent with room acoustics.

Findings

Enhanced physical coherence of dereverberated signals

Preservation of original dereverberation quality

Improved interpretability of deep learning models

Abstract

Single-channel speech dereverberation aims at extracting a dry speech signal from a recording affected by the acoustic reflections in a room. However, most current deep learning-based approaches for speech dereverberation are not interpretable for room acoustics, and can be considered as black-box systems in that regard. In this work, we address this problem by regularizing the training loss using a novel physical coherence loss which encourages the room impulse response (RIR) induced by the dereverberated output of the model to match the acoustic properties of the room in which the signal was recorded. Our investigation demonstrates the preservation of the original dereverberated signal alongside the provision of a more physically coherent RIR.