Blind speech dereverberation using convolutive nonnegative matrix factorization with mixed penalization

RUBEN SPIES; FRANCISCO IBARROLA; LEANDRO EZEQUIEL DI PERSIA

Comodoro Rivadavia

Congreso; VI Congreso de Matemática Aplicada, Computacional e Industrial; 2017

Asociación Argentina de Matemática Aplicada, Computacional e Industrial

When a signal is recorded in an enclosed room, it typically gets affected by reverberation. This degradationrepresents a problemwhen dealingwith audio signals, particularly in the field of speech applications, such as automatic speech recognition. Although there are some approaches to deal with this issue that are quite satisfactory under certain conditions, constructing a method that works well in a general context still poses a significant challenge. As aneffort in this direction, we propose a method based on convolutive nonnegative matrix factorization that mixes two penalizers in order to impose certain characteristics over the time-frequency components of the restored signal and the reverberant components. An algorithm for finding such a solution is described and tested. The results show a significant improvement on the quality of the restored signals.