Use camera tracks your face with the front camera, in the browser. In the iPhone app, ARKit and the TrueDepth camera drive it.
Measured depth for hair, shoulders and background. Pushes the pins, sets dispersal layers.
Labels every point as eye, lip, jaw, brow or nose. jawOpen gates the cymatics.
Separates what the face mesh can't see: hair, clothing and background.
Speech level and bands drive the mouth ripples and the pin waves.
The face is MediaPipe's canonical face model with a modelled head. Without a camera it sways and speaks on its own. In the browser, Use camera runs MediaPipe face tracking on the front camera and the microphone. In the Facefield iPhone app, ARKit's face mesh, blendshapes and the TrueDepth depth map drive the same shaders: the pin-screen shows measured depth, hair and shoulders included.