You are reading a single comment by @deleted and its replies.
Click here to read the full conversation.
-
We did something like this at work a few years back... The software gets you to speak X number of words and then synthesises a complete set of phonetics from that... Type away and the software talks back at you in that exact voice/accent... It was pretty amazing... Software got pulled a couple of months after release, I assume one of the big tech companies snapped it up...
deleted
stevo_com
Better/worse than deepfake. Deepfake needs an existing video to adapt, as far as I know. This just needs a short snippet of audio to then be able to construct something totally false from scratch. I'm sure their examples are carefully chosen and it is not yet capable of spoofing any combination of syllables, but it is still pretty amazing/terrifying.
The optimist in me goes down the route of voice synthesizers for people who have lost the ability to speak. If there was a short recording of them when they could still speak, then they could have more a more recognisable sound from their device, giving them back a sense of identity that they had lost.