- Edited
Silly-Sausage I get what you're saying and would have thought the same a few years ago. I've seen how some AI voice models have grown to perfectly match someone's voice and this is seeming quite plausible. Have you seen AI generated voice for music covers where multiple singers AI voices are stitched for a single song.
And this won't be for the entire scene, just whenever she mentions the viewer by name. So hopefully a lil controlled. This will only get better as AI's capabilities improve.
And do you remember VRBangers having a studio where they faceswapped the performer(one was a Hime Marie) scene. I ain't asking for something that drastic. Just that performer's lips should follow what the AI voice generates, and in short bursts.
Anyways it's on SLR how they implement it so that AI voice doesn't stick out and is stitched seamlessly. But keeping this optional should make it fine I guess. Just have a regular vid and in case you want it personalized with ya name, flip a toggle.