Can we generate emotional speech with different tags like pauses silence laughing etc.

[dharmesh8b]

[YaoyaoChang]

Explicit speech tagging is not supported at the moment.

[YaoyaoChang]

Here is an interesting way to control speech and emotion by voice prompt, details can be found via HF discussion12.

[Arcitec]

Thanks @YaoyaoChang for sharing the great idea.

To sum it up:

• Use the 4 speaker "slots" and put the same voice in all of them, but with different voice clips containing different emotions like shouting etc.
• Then write your script with "Speaker 0: Hello! Speaker 1: What's up!" syntax, to trigger the different emotions. You have 4 slots, so you can mark 4 different emotions this way.