From now on, every AI voiceover I make uses Eleven v4, the model your voice is trained on. The dubbing tool uses it too.
The Glance short's dubs, as shipped in September (v3) and re-rendered today on v4. Same translations, timing and background music. Only the voice model changed. First 14 seconds of each.
This is the disaster simulation center in China. Couples come here to trauma bond. You get trapped in a sinking car, swim through a storm, and get shot with hail. Sounds fun, right? It's not. It's actually terrifying. The whole idea is you learn how to survive real natural disasters before one actually happens to you. So today we're trying every single one, all the way to the final flash flood. I'm already regretting this. And if I don't make it out, well, the video ends early. Let's go.