3 ms·

I say this as someone who's tried all the state-of-the-art voice generation systems: so far, nobody has trained an AI to "chew the scenery" nearly as effectivel

by btown 2y ago

I say this as someone who's tried all the state-of-the-art voice generation systems: so far, nobody has trained an AI to "chew the scenery" nearly as effectively and effusively as these voice actors having the time of their lives playing southern belles and bean company owners who are blisteringly envious that a dead mermaid is stealing their spotlight by dying from a megalodon attack.

withinboredom 2y ago
Haha, the problem is: we still haven't gotten AI to generate voices! "State of the art" just uses regular TTS engines and then adds extra inflection as "special sauce" or stretches it out, etc. At least, that is how it works for these AI generators that need to do it "at scale." When you can spend more on classic speech models, you can go well beyond that (see Siri, Google, Alexa, etc).