Simultaneous Denoising and Dereverberation Using Deep Embedding Features

TL;DR

This paper introduces a joint deep learning approach for simultaneous speech denoising and dereverberation, leveraging deep embedding features and a two-stage process to improve speech quality in noisy, reverberant environments.

Contribution

It proposes a novel joint training method using deep embedding features for combined denoising and dereverberation, outperforming traditional methods especially at low SNRs.

Findings

Outperforms WPE and BLSTM baselines in low SNR conditions

Uses deep clustering features for effective speech separation

Employs a two-stage neural network approach for denoising and dereverberation

Abstract

Monaural speech dereverberation is a very challenging task because no spatial cues can be used. When the additive noises exist, this task becomes more challenging. In this paper, we propose a joint training method for simultaneous speech denoising and dereverberation using deep embedding features, which is based on the deep clustering (DC). DC is a state-of-the-art method for speech separation that includes embedding learning and K-means clustering. As for our proposed method, it contains two stages: denoising and dereverberation. At the denoising stage, the DC network is leveraged to extract noise-free deep embedding features. These embedding features are generated from the anechoic speech and residual reverberation signals. They can represent the inferred spectral masking patterns of the desired signals, which are discriminative features. At the dereverberation stage, instead of using the unsupervised K-means clustering algorithm, another supervised neural network is utilized to estimate the anechoic speech from these deep embedding features. Finally, the denoising stage and dereverberation stage are optimized by the joint training method. Experimental results show that the proposed method outperforms the WPE and BLSTM baselines, especially in the low SNR condition.