EA, definitely one of the globe’s companies that you would absolutely want to offer up large chunks of your voice data, has reportedly filed a new patent as shared by Insider Gaming for a game tech solution asking players to do just that – all in the name of making your fully customised in-game hero more you than ever.
The overarching idea is that players would record their own voice, which is then put through a series of synthesisers and converters to crate a vocoder capable of replicating not only the sound of a voice but non-verbal components like tone and emphasis.
“A computer-implemented method of generating speech audio in a video game is provided. The method includes inputting, into a synthesizer module, input data that represents speech content,” reads a portion of the patent (which you can view in full here).
“Source acoustic features for the speech content in the voice of a source speaker are generated and are input, along with a speaker embedding associated with a player of the video game into an acoustic feature encoder of a voice convertor. The target acoustic features are processed with one or more modules, to generate speech audio in the voice of the player.”
Which is all a very technical way of saying, EA will process audio of you speaking into a voice module for your in-game character, who will then read out voice lines in your voice – in some cases even allowing players to write out their own dialogue for their character to then replicate in-game.
If you’ve seen any number of those voice filters that use AI to replicate famous voices to a scarily-accurate degree, or even to scam call folks by masquerading as loved ones using their own voice, you’ll know the kind of territory this is crossing over into, not to mention the implications to the work of voice actors. That said, if it works as well as EA is describing with as little data as possible and is used safely, it could be a very cool way of completing the idea of role playing as your own hero in video games.