A Supervised Speech enhancement Approach with Residual Noise Control for Voice Communication

TL;DR

This paper introduces a generalized loss function for supervised speech enhancement that balances speech distortion and noise reduction, incorporating residual noise control to improve practical voice communication quality.

Contribution

It derives a flexible generalized loss function that includes existing loss functions as special cases, enabling better residual noise control in supervised speech enhancement.

Findings

The generalized loss function improves the trade-off between speech quality and noise suppression.

Residual noise control significantly enhances both objective and subjective speech enhancement results.

The approach outperforms traditional methods in practical noise reduction scenarios.

Abstract

For voice communication, it is important to extract the speech from its noisy version without introducing unnaturally artificial noise. By studying the subband mean-squared error (MSE) of the speech for unsupervised speech enhancement approaches and revealing its relationship with the existing loss function for supervised approaches, this paper derives a generalized loss function, when taking the residual noise control into account, for supervised approaches. Our generalized loss function contains the well-known MSE loss function and many other often-used loss functions as special cases. Compared with traditional loss functions, our generalized loss function is more flexible to make a good trade-off between speech distortion and noise reduction. This is because a group of well-studied noise shaping schemes can be introduced to control residual noise for practical applications. Objective and subjective test results verify the importance of residual noise control for the supervised speech enhancement approach.