Fast, low-artifact speech synthesis considering global variance.

Matt Shannon, William Byrne

Browse the full ICASSP paper archive.