NVIDIA has made Audio2Face open source
, by Quentin
NVIDIA has released its Audio2Face tool under an MIT license. This technology generates real-time facial animations and lip sync from a simple audio stream, with integrations already available in Maya and Unreal Engine 5.
The SDK and pre-trained models are available, allowing studios to adapt the solution to their needs. Devs such as The Farm 51, Codemasters, and GSC Game World are already using it to speed up their animation pipelines.
Source: developer.nvidia.com