News for hackers. Stuff that matters.

A comment on Silero V3: fast high-quality text-to-speech in 20 languages with 173 voices. Parent comment

Re:Silero V3: fast high-quality text-to-speech in 20 languag... by grumbel on Monday June 20, 2022 @06:54AM (#31808751)
> what do I type in my shell to convert some_txt.txt to some_txt.wav

Doesn't look like they have any command line tool included, but they have a simple Python example in the README. I used that, added some argparse around it and build a Nix package out of it. In case you are running Nix with flakes enabled, you can just type:

   nix run github:grumbel/silero-test -- -t "Hello World" --speaker en_0 -o /tmp/out.wav
Source is pretty much "embarrassingly simple":

   https://github.com/Grumbel/silero-test/blob/master/silero-test
It's currently hard codes the English language model, other models can be found at https://models.silero.ai/models/tts/

PS: This is probably reinventing the wheel, haven't looked around if there is anything ready to use already.

[ Parent | on HN ]
Re:Silero V3: fast high-quality text-to-speech in 20 languag... by ur-whale on Monday June 20, 2022 @08:22AM (#31809461)
Super, this is very useful, thank you!
[ Parent | on HN ]