Meta has recently unveiled two innovative AI tools: a video face generator called the MetaHuman Animator and Voicebox audio voice generator. These tools can generate convincingly realistic material based on a person’s face or voice samples.
The MetaHuman Animator captures an actor’s performance using an iPhone or a head-mounted camera system and then applies it as high-fidelity facial animation on any MetaHuman character. This technology is straightforward to use, requires no manual intervention, and produces results in minutes1.
On the other hand, Voicebox is a non-autoregressive flow-matching model trained on over 50,000 hours of speech. It outperforms existing models in tasks like text-to-speech synthesis, noise removal, content editing, style conversion, and diverse sample generation2.
While these tools open up new possibilities in content creation, they also raise concerns about the misuse of synthetic media. To address this, Meta has partnered with Microsoft to develop strategies to manage the risks associated with these new technologies.
Moreover, Meta plans to make more of its AI technology available as open source. This move not only promotes competition with tech giants like Microsoft and OpenAI, but also allows companies to leverage these advanced tools to build their own applications.
Key References:
