Varianceflow: High-Quality and Controllable Text-to-Speech using Variance Information via Normalizing Flow.

Yoonhyung Lee, Jinhyeok Yang, Kyomin Jung

Browse the full ICASSP paper archive.