Produce different voices for the same sentence

TTS (Text-to-Speech)
#1

Hi,
I wanted to know if there is a way to generate different voice samples for the same sentence, using any of the pre-trained models ?

Thanks

#2

different speakers you mean ? If so, no there is not.

#3

Out of interest, for the same speaker, would the same text always produce the exact same audio output?

Is there a way to produce slight variations, in the same way that people don’t say a phrase in precisely the same way each time?

#4

you can enable the prenet dropout at inference

2 Likes