What I'd really like is a way to generate synthetic voices, not clone an existing voice. Something like Stable Diffusion's Dreambooth, ControlNet, or LoRa/A for voices.
Extra points for adding the ability to adjust individual performances with tags or "style matching", like Img2Img.
I'd be willing to spend money to get that ability, possibly by the minute (say, $1/minute) or a monthly fee or whatever.
Extra points for adding the ability to adjust individual performances with tags or "style matching", like Img2Img.
I'd be willing to spend money to get that ability, possibly by the minute (say, $1/minute) or a monthly fee or whatever.