A synthesis method based on concatenation of demisyllables and a residual excited vocal tract model.

Steve Pearson, Nick Kibre, Nancy Niedzielski

Browse the full Interspeech paper archive.