Towards High-Quality Neural TTS for Low-Resource Languages by Learning Compact Speech Representations

Open in new window