Accent conversion using discrete units with parallel data synthesized from controllable accented TTS

Open in new window