search

DigitalPhonetics / IMS-Toucan

Multilingual and Controllable Text-to-Speech Toolkit of the Speech and Language Technologies Group at the University of Stuttgart.
Apache License 2.0
1.17k stars 135 forks source link

Emotional Voice #140

Closed Simbaprince closed 2 weeks ago

Simbaprince commented 1 year ago

How to get correct value of parameters of IMS_toucan for several emotional voices such as angry, sad, excited, cheerful, shouting, whispering, terrified, friendly, unfriendly, hopeful, normal.

Flux9665 commented 1 year ago

Some of the things you mention are not necessarily emotions, but you can achieve different speaking styles by providing a reference audio of that style and speaker to the set_utterance_embedding method of the inference interface.

Flux9665 commented 1 year ago

Versions are cumulative, so the most recent version has the best quality yet.