While deep fake videos may be the most talked about application of using AI to replicate people, the Stable Voices YouTube channel has been doing it convincingly with just voices for some time now, with the latest character to be imitated being Severus Snape.
“Alan Rickman’s voice was synthesized using an AI model trained on his speech patterns,” The video description reads, “While the voice may be computer generated, the script is written by something vaguely resembling a human.”
We assume that last part is a bit of self-deprecating humour on the part of whoever wrote the script for the video (embedded below), but the writing here is pretty funny for a short demo of the software at work.
Stable Voices is most famous for its AI voice recreation of Donald Trump in its series of videos called “A Stable Genius Explains”. Here the president of the United States tackles serious topics such as bone spurs, impeachment and walls.
It’s just a bit of fun and the backing track – called Splashing Around – sounds like background music from SpongeBob SquarePants. It’s surprisingly fitting.
The full A Stable Genius Explains series is available in this YouTube playlist that you can get through in around ten minutes due to how short each video is.
As you may have guessed by Severus Snape being the topic of a new video, the channel has since branched out to other people and characters.
Some are simply the voice reading a funny script, while others involve some kind of skit. We’ll let you watch them for yourself, but Obama does sing The Fresh Prince of Bel-Air theme song.
The voices are of varying levels of quality in terms of being accurate to the real people, but they’re consistently good and we highly recommend subscribing to the channel.