#Training Piper model on "unnatural" sounding voice?

1 messages · Page 1 of 1 (latest)

teal gust
#

Hey guys, I'm trying to train a Piper TTS model on the AI character EDI from the Mass Effect series. If you aren't familiar, her voice sounds like this: https://www.youtube.com/watch?v=4MebtEtRQDM

My friend told me the voice isn't going to sound right because voice models are trained on natural, human voices, so it's just going to filter EDI's distortion out.

Can anyone weigh in on this? Is it even worth chopping that video up and seeing what comes out of training?

undone kite
#

36 minutes is definitely not enough for training... It will be all over the place.
Check NetworkChuck's video https://www.youtube.com/watch?v=3fg7Ht0DSnE&t=4s he used a lot of his own videos (with free speech, which is better than just phrases) to train voice model, and it's still not that good.

Stop storing your secrets and API keys in your code!! Try Keeper, a password manager you can use in the terminal: (built for devs/admins): https://www.keeper.io/networkchuck

I did it…..after days of frustration, blood, sweat and coffee..I finally figured out a way to clone a voice to use with my fully local, AI voice assistant!!!! This isn...

▶ Play video
teal gust
undone kite
#

Well, maybe it will be enough for such a "robotic" voice 🙂 GlaDOS also doesn't have much of samples, but its model is great.

teal gust