Post Doc – Unifying controlled text-to-speech and voice conversion

By instance, self-supervised representations such as Wav2vec 2.0 [3], that achieve SOTA results on ASR [4] and speaker recognition tasks [5], replace usual TTS…