Text-to-Speech
English

Suggestion: training unique voices

#128
by meyomon - opened

Since Kokoro has the ability to "blend" voices together, I imagine that a good strategy for achieving a large arsenal of voices would be to train "unique" voices since we can then "blend" these to achieve a greater number of differently-sounding voices.
A bit how we can blend "unique" colors to get any color in-between (think how RGB-displays work)

If you think this is a good idea, please upvote this post to increase visiblity

image.png

interesting idea.

Sign up or log in to comment