Seedance 2.0 Lip Sync: Talking Heads That Don’t Mush the Mouth

Seedance 2.0 lipsync only looks easy when the still is framed for speech. I animated the same portrait three ways on Seedance 2.0 and kept the winner: tight crop, calm camera, clean VO.

TL;DR

Lever	What worked
Source crop	Medium close-up, mouth visible
Camera	No orbit / no zoom
Audio	VO first, match length
Motion	Subtle head + blinks only

Motion prompt

Talking head, subject speaks naturally to camera, subtle head motion, accurate mouth movement, soft blinks, stable shoulders, softbox light, no zoom

Soft CTA

Stills on Generate, speech motion on Seedance. More control: first/last frame.

FAQ

Singing? Harder than speech—start with VO.
Need a talking video reference? Strong still + clean audio covers many social uses.