Local spectral attention for full-band speech enhancement

TL;DR

This paper introduces local spectral attention to full-band speech enhancement, limiting frequency span to improve performance and reduce residual noise, demonstrating effectiveness over state-of-the-art models.

Contribution

It proposes a novel local spectral attention mechanism that limits frequency span in full-band speech enhancement models, improving performance.

Findings

Improved objective scores on VoiceBank+DEMAND dataset

Local spectral attention enhances overall speech enhancement quality

Effective in reducing residual noise

Abstract

Attention mechanism has been widely utilized in speech enhancement (SE) because theoretically it can effectively model the inherent connection of signal both in time domain and spectrum domain. Usually, the span of attention is limited in time domain while the attention in frequency domain spans the whole frequency range. In this paper, we notice that the attention over the whole frequency range hampers the inference for full-band SE and possibly leads to excessive residual noise. To alleviate this problem, we introduce local spectral attention (LSA) into full-band SE model by limiting the span of attention. The ablation test on the state-of-the-art (SOTA) full-band SE model reveals that the local frequency attention can effectively improve overall performance. The improved model achieves the best objective score on the full-band VoiceBank+DEMAND set.