Preserving naturalness in synthetic voices while minimizing variation in formant frequencies and bandwidths.

Niels-Jrn Dyhr, Marianne Elmlund, Carsten Henriksen

Browse the full Interspeech paper archive.