DNSMOS: A Non-Intrusive Perceptual Objective Speech Quality metric to evaluate Noise Suppressors
This addresses the need for a reliable no-reference speech quality metric for noise suppressors, which is incremental as it builds on existing approaches but improves correlation with human perception.
The paper tackles the problem of evaluating noise suppression algorithms without needing a reference clean speech signal, introducing DNSMOS, a multi-stage self-teaching based perceptual objective metric that achieves high correlation with human ratings in challenging test conditions.
Human subjective evaluation is the gold standard to evaluate speech quality optimized for human perception. Perceptual objective metrics serve as a proxy for subjective scores. The conventional and widely used metrics require a reference clean speech signal, which is unavailable in real recordings. The no-reference approaches correlate poorly with human ratings and are not widely adopted in the research community. One of the biggest use cases of these perceptual objective metrics is to evaluate noise suppression algorithms. This paper introduces a multi-stage self-teaching based perceptual objective metric that is designed to evaluate noise suppressors. The proposed method generalizes well in challenging test conditions with a high correlation to human ratings.