News for hackers. Stuff that matters.

A comment on WhisperSpeech – An open source text-to-speech system built by inverting Whisper.

Re:WhisperSpeech – An open source text-to-speech system buil... by londons_explore on Thursday January 18, 2024 @10:54AM (#39043283)
The first demo on that page was trained from a 32kbps crappy sound quality clip of winston churchill...?

garbage in, garbage out?

[ Parent | on HN ]
Re:WhisperSpeech – An open source text-to-speech system buil... by stavros on Thursday January 18, 2024 @11:32AM (#39043820)
Must have been, it sounds very much like the quality of the "we shall fight on the beaches" speech.

A bit unfortunate choice for a demo, sadly.

[ Parent | on HN ]
Re:WhisperSpeech – An open source text-to-speech system buil... by jpcl on Thursday January 18, 2024 @01:16PM (#39045343)
Good point, thanks. And I was thinking it will show that the model can really synthesize very varied samples... ;)
[ Parent | on HN ]
Re:WhisperSpeech – An open source text-to-speech system buil... by stavros on Thursday January 18, 2024 @01:20PM (#39045416)
It does, but maybe put it last!
[ Parent | on HN ]
Re:WhisperSpeech – An open source text-to-speech system buil... by jpcl on Thursday January 18, 2024 @01:23PM (#39045480)
Did exactly that, thanks for spotting that. :)

https://github.com/collabora/WhisperSpeech/commit/398b889060...

[ Parent | on HN ]