Perceive and Predict: Self-Supervised Speech Representation Based Loss Functions for Speech Enhancement.

George Close, William Ravenscroft, Thomas Hain, Stefan Goetze

Browse the full ICASSP paper archive.