#Piper model for whispering
1 messages · Page 1 of 1 (latest)
Think the closest thing that can do this is kokoro, but it'd need a Wyoming wrapper. But i think it supports whispering effects
Well i'm ok with completely different model, that is only whispering. I'm just curious if there is one... But Kokoro is great tool!
Seems someone has already started working in this
Reading into it more, i don't really see how you'd make this whisper. Might be a particular voice model that does it..
Ok so if you go to https://kokorottsai.com/ and use the Nicole voice, it whispers sort of
Kokoro TTS - Advanced AI text-to-speech model with only 82M parameters, delivers HQ and efficient speech synthesis.Turn text into natural, lifelike voices.
Oh crap, you did a lot of investigation for me. Thanks!
Damn she could seduce me with that ASMR :)))
Spun up kokoro, but can't hear anything. Will dig.
So it doesn't work from satellite, but works from "Try voice" in pipeline editor...
Clear tts cache maybe?
so I spun up the docker, it's a bit buggy. When I went to try the af_heeart voice for instance it breaks with an error in the logs
but I tried the af_bella and af_nicole voices and they both worked for me on my Voice PE