BUDDy: Single-Channel Blind Unsupervised Dereverberation with Diffusion Models

TL;DR

This paper introduces BUDDy, an unsupervised single-channel dereverberation method using diffusion models that estimates room impulse responses and enhances speech without prior acoustic knowledge.

Contribution

The paper proposes a novel unsupervised approach leveraging diffusion models for blind dereverberation and room impulse response estimation, outperforming prior methods.

Findings

Significantly outperforms previous blind unsupervised baselines.

Demonstrates robustness to unseen acoustic conditions.

Operates without prior room or paired data.

Abstract

In this paper, we present an unsupervised single-channel method for joint blind dereverberation and room impulse response estimation, based on posterior sampling with diffusion models. We parameterize the reverberation operator using a filter with exponential decay for each frequency subband, and iteratively estimate the corresponding parameters as the speech utterance gets refined along the reverse diffusion trajectory. A measurement consistency criterion enforces the fidelity of the generated speech with the reverberant measurement, while an unconditional diffusion model implements a strong prior for clean speech generation. Without any knowledge of the room impulse response nor any coupled reverberant-anechoic data, we can successfully perform dereverberation in various acoustic scenarios. Our method significantly outperforms previous blind unsupervised baselines, and we demonstrate its increased robustness to unseen acoustic conditions in comparison to blind supervised methods. Audio samples and code are available online.