Perceptual Contrast Stretching on Target Feature for Speech Enhancement

TL;DR

This paper introduces a perceptual contrast stretching method based on critical band importance to enhance speech enhancement models, improving performance while maintaining efficiency and compatibility across architectures.

Contribution

The proposed PCS method is a novel perceptual feature modification technique that enhances speech enhancement performance without affecting model training or causality.

Findings

Achieves state-of-the-art PESQ scores on VoiceBank-DEMAND dataset.

Compatible with various SE architectures and training criteria.

Reduces online computation compared to post-processing methods.

Abstract

Speech enhancement (SE) performance has improved considerably owing to the use of deep learning models as a base function. Herein, we propose a perceptual contrast stretching (PCS) approach to further improve SE performance. The PCS is derived based on the critical band importance function and is applied to modify the targets of the SE model. Specifically, the contrast of target features is stretched based on perceptual importance, thereby improving the overall SE performance. Compared with post-processing-based implementations, incorporating PCS into the training phase preserves performance and reduces online computation. Notably, PCS can be combined with different SE model architectures and training criteria. Furthermore, PCS does not affect the causality or convergence of SE model training. Experimental results on the VoiceBank-DEMAND dataset show that the proposed method can achieve state-of-the-art performance on both causal (PESQ score = 3.07) and noncausal (PESQ score = 3.35) SE tasks.