Speech Enhancement Using Non-Negative Spectrogram Models with Mel-Generalized Cepstral Regularization.

Li Li, Hirokazu Kameoka, Tomoki Toda, Shoji Makino

Browse the full Interspeech paper archive.