Guest Posted May 3, 2020 Posted May 3, 2020 Anyone here tried to use neural networks to generate voice files? I'm gathering information about it, and want to know where to start. Where can I download a neural network program? How can I use it to generate voices for Skyrim? Yesterday I visited the site for the Wavenet project: https://deepmind.com/blog/article/wavenet-generative-model-raw-audio But they are still developing it. Is there a parametric solution I could use for mods?
bobbokins Posted May 5, 2020 Posted May 5, 2020 In terms of full publicly available implementations corentin is probably your best bet. https://github.com/CorentinJ/Real-Time-Voice-Cloning there is more work being done in the subject to improve performance to human levels and add control-ability but most talent goes to venture firms and big tech company's with something to gain from having the best voice synthesis. https://google.github.io/tacotron/publications/prosody_prior/ https://google.github.io/tacotron/publications/hierarchical_prosody/ you could probably make a pretty great fallout 4 synth companion with it at the moment but it won't be until starfield is out till full speech synthesis is viable for modders.
Recommended Posts
Archived
This topic is now archived and is closed to further replies.