Improving the Prosody of RNN-Based English Text-To-Speech Synthesis by Incorporating a BERT Model.

Tom Kenter, Manish Sharma, Rob Clark

Browse the full Interspeech paper archive.