If I were to use, say, five or so minutes of clips in order to form a particular voice, would it be suitable? I don't know much about how this ai works, but I'm planning to set it up when i get back to my desktop so I'd like to know how much training material the AI is going to need if I'm to get those uncanny humanity was born to flee this thing ones like the particularly true to form one I was being shown some time ago. When you know someone, and then they procede to swap their voice over to the one you know and it's so uncannily similar you spend the whole time attemptng to connect voice to human it makes me fear for where this stuff can go next

5 mints is good for a model

you can pm me if you need any thing about that thing