EA has patented the ability to generate in-game voice acting using a player’s own voice.
As spotted by VeryAli Gaming, the patent – which is named ‘Generating speech in the voice of a player of a video game’ – was filed in 2020 but was published last month.
The patent describes a process in which players “input data that represents speech content”, which is then passed through a synthesiser module, a voice convertor and a vocoder to generate new dialogue based on the sample.
The idea is that players will be able to put their own voice in the game, with the most obvious example being the ability to make the protagonist sound like the player.
The patent explains that previous attempts at generating speech in games often involve text-to-speech systems, where the game receives text and uses a synthesiser to generate audio speech.
However, the patent says that “such approaches usually require many speech samples (e.g. hours of speech samples) in order for the synthesiser to accurately capture the voice in the synthesised speech audio”, and that if a different voice is needed, the synthesiser would need to be trained again with hours of new samples.
The methods described in the patent, however, “enable speech audio to be generated in the voice of a player of a video games, using speech audio provided by the player (e.g. a small amount of speech audio such as minutes of speech audio from the player), and a voice convertor.
The process sounds similar to the numerous examples of AI voice generator software currently found online.
These can receive voice recordings of a person talking then use AI to turn these into a text-to-speech tool which can say anything in that person’s voice, in a manner that sounds convincing enough to fool some people.
As ever, the filing of a patent is no guarantee that such technology will ever make it into a video game, but it could shed light into how future narrative-based EA games may evolve character customisation.