Audio2Face Character R&D
I was contracted to prototype a fully rigged game character based on a reference image to be used with NVIDIA Omniverse Audio2Face lipsync in Unreal within a week.
Retargeted Idle Anim test + Omniverse Live Link. Disregard the latency due to low system specs
DIY MetaBody Mocap to Unreal
Animation Retargeting Setup for said mocap
Made variations of character face
Character (custom metahuman blueprint) controlled by NVIDIA Audio2Face ARKit stream
Upscaling & repainting textures, revision from client notes
In-editor screenshots
In-editor screenshots (studio lighting)
Character (custom low poly character - non metahuman) controlled by NVIDIA Audio2Face ARKit stream
Character hand painted texturing to match the style of "Fortnite"
NVIDIA Omniverse Live Link with custom non metahuman character
UE5 rig
Control Rig
46 / 52 blendshape setup for Omniverse / ARKit
USD generation of blendshapes
Omniverse setup
UVing character
Modeling character
Hair uses curves, later retopologized into single mesh and painted.