Learning with Learned Loss Function: Speech Enhancement with Quality-Net to Improve Perceptual Evaluation of Speech Quality
This work addresses the challenge of improving perceptual speech quality for applications like communication systems, but it is incremental as it builds on existing methods for approximating non-differentiable metrics.
The paper tackled the problem of training speech enhancement models with a human-perception-related objective function, specifically PESQ, by proposing a differentiable approximation learned from data, resulting in a 0.18-point increase in PESQ score compared to using MSE loss while maintaining speech intelligibility.
Utilizing a human-perception-related objective function to train a speech enhancement model has become a popular topic recently. The main reason is that the conventional mean squared error (MSE) loss cannot represent auditory perception well. One of the typical hu-man-perception-related metrics, which is the perceptual evaluation of speech quality (PESQ), has been proven to provide a high correlation to the quality scores rated by humans. Owing to its complex and non-differentiable properties, however, the PESQ function may not be used to optimize speech enhancement models directly. In this study, we propose optimizing the enhancement model with an approximated PESQ function, which is differentiable and learned from the training data. The experimental results show that the learned surrogate function can guide the enhancement model to further boost the PESQ score (in-crease of 0.18 points compared to the results trained with MSE loss) and maintain the speech intelligibility.