How to Make an Accurate Voice Model

How much of your voice do they actually need to be able to then use? And presumably this stuff has been advanced with AI and the other programs that they can use. That if I have a certain length of sample of your voice, that I can then use it to generate whole sentences, maybe even conversations. The media exaggerate this a bit and they come out with this number like three seconds. But realistically, 45 seconds to three minutes will make a far more accurate voice model. And it's the same with your face. So right now we've just, you know, I could have this call recorded and I can see your face right now,. For everyone who doesn't have a video