Description
Making a portrait photo talk usually yields mismatched lips or a frozen head. SadTalker learns realistic 3D motion coefficients from audio to animate a single image into a natural talking-head video.
Lips, expressions and head pose follow the voice across styles, with a web UI and a Stable Diffusion WebUI extension; CVPR 2023.
Single image:One photo.
Lip sync:Audio-driven.
Head motion:3D coefficients.
WebUI:SD extension.
Lips, expressions and head pose follow the voice across styles, with a web UI and a Stable Diffusion WebUI extension; CVPR 2023.
Features
Single image:One photo.
Lip sync:Audio-driven.
Head motion:3D coefficients.
WebUI:SD extension.
