Using continuous lexical embeddings to improve symbolic-prosody prediction in a text-to-speech front-end.

Asaf Rendel, Raul Fernandez, Ron Hoory, Bhuvana Ramabhadran

Browse the full ICASSP paper archive.