Perceive and predict: self-supervised speech representation based loss functions for speech enhancement
This work addresses speech enhancement for applications requiring high-quality audio, but it is incremental as it builds on existing self-supervised representation methods.
The authors tackled the problem of improving speech enhancement by proposing a loss function based on distances between self-supervised speech representations of clean and noisy speech, which correlates with quality and intelligibility measures. They demonstrated improved performance over baseline loss functions, achieving gains in objective metrics like PESQ and STOI.
Recent work in the domain of speech enhancement has explored the use of self-supervised speech representations to aid in the training of neural speech enhancement models. However, much of this work focuses on using the deepest or final outputs of self supervised speech representation models, rather than the earlier feature encodings. The use of self supervised representations in such a way is often not fully motivated. In this work it is shown that the distance between the feature encodings of clean and noisy speech correlate strongly with psychoacoustically motivated measures of speech quality and intelligibility, as well as with human Mean Opinion Score (MOS) ratings. Experiments using this distance as a loss function are performed and improved performance over the use of STFT spectrogram distance based loss as well as other common loss functions from speech enhancement literature is demonstrated using objective measures such as perceptual evaluation of speech quality (PESQ) and short-time objective intelligibility (STOI).