3 ms·

This is from 2018 – does anyone know if there are pretrained models and code for this? I found https://github.com/CorentinJ/Real-Time-Voice-Cloning https://gith

by moyix 7y ago

This is from 2018 – does anyone know if there are pretrained models and code for this? I found https://github.com/CorentinJ/Real-Time-Voice-Cloning https://github.com/CorentinJ/Real-Time-Voice-Cloning , but the generated audio quality was much worse than the samples here.

JaRail 7y ago
The biggest missing piece is WaveNet, which is Google's proprietary voice-synthesizer. With only the models trained for this paper, the best you could build is a voice-recognizer. As far as I know, Google only allows people to do TTS with one of their provided voices. I don't expect them to open it up until other companies/academics have achieved similar results. It's too much of a competitive advantage right now. Alexa, Siri, etc all sound like robots compared to WaveNet (google assistant).