Skip to content

Comment on Show HN: VoiceClonr attempts to reconstruct human voicesparent

Comments

That means that you can't simply reverse engineer voice from, say, a sample text read by an voice actor? I mean, down to the tiny bits of audio waveform? I mean, how hard could it be? :)

If you can have a speaker read through a specific list of items, a useful singing model can be constructed. That's how Vocaloid works.

What hasn't been done well yet is extracting a model from existing uncontrolled voice samples. That's what this is trying to do. Once this works well, software clones of dead singers will be popular. The RIAA is going to hate this.

That's exactly how unit selection systems like Festival work. The trouble starts when you hit a previously unencountered phoneme sequence and you have to interpolate. Sometimes good, sometimes bad.

Edit: Text to pronunciation is a whole other problem.

Given a text and waveform, how do you know how to match them up exactly?

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.