Boosting Self-Supervised Embeddings for Speech Enhancement

TL;DR

This paper enhances speech enhancement by integrating SSL embeddings with spectrograms, improving performance and noise robustness analysis, and demonstrating superior results over existing SSL-based methods.

Contribution

It introduces a cross-domain feature combining SSL and spectrograms, analyzes noise robustness via CN distance, and shows improved speech enhancement performance without complex architectures.

Findings

SSL embeddings with lower noise robustness are more important for SE.

Fine-tuning SSL with SE models outperforms state-of-the-art SSL-based SE methods.

CN distance increases after fine-tuning, indicating improved robustness.

Abstract

Self-supervised learning (SSL) representation for speech has achieved state-of-the-art (SOTA) performance on several downstream tasks. However, there remains room for improvement in speech enhancement (SE) tasks. In this study, we used a cross-domain feature to solve the problem that SSL embeddings may lack fine-grained information to regenerate speech signals. By integrating the SSL representation and spectrogram, the result can be significantly boosted. We further study the relationship between the noise robustness of SSL representation via clean-noisy distance (CN distance) and the layer importance for SE. Consequently, we found that SSL representations with lower noise robustness are more important. Furthermore, our experiments on the VCTK-DEMAND dataset demonstrated that fine-tuning an SSL representation with an SE model can outperform the SOTA SSL-based SE methods in PESQ, CSIG and COVL without invoking complicated network architectures. In later experiments, the CN distance in SSL embeddings was observed to increase after fine-tuning. These results verify our expectations and may help design SE-related SSL training in the future.