Q: Anyone knows if there are any service that is equivalent of 'This Voice Does Not Exists'?
Google Text-to-Speech has good Wavnet based voices[0], but the focus seems to be generating human like voice intonation, accents but not generating infinite (or large) set of different voice characters.
Then there are services like Lyrebird[1] where focus is on cloning a real person's voice (ex. your own).
But there doesn't seem to be any which creates fictitious voice characters which is totally new every time (synthesizing pitch, tone, accents, intonations); same as perhaps what GAN does when generating 'deepfake faces'.
Comments
Q: Anyone knows if there are any service that is equivalent of 'This Voice Does Not Exists'?
Google Text-to-Speech has good Wavnet based voices[0], but the focus seems to be generating human like voice intonation, accents but not generating infinite (or large) set of different voice characters.
Then there are services like Lyrebird[1] where focus is on cloning a real person's voice (ex. your own).
But there doesn't seem to be any which creates fictitious voice characters which is totally new every time (synthesizing pitch, tone, accents, intonations); same as perhaps what GAN does when generating 'deepfake faces'.
[0] https://cloud.google.com/text-to-speech
[1] https://www.descript.com/lyrebird-ai