The mouth controls for the dub. Defaults are the proven recipe; change one thing at a time. — ComfyUI node from LSDub. Run it in the cloud, no install required.

Lip-sync for music videos in ComfyUI: dub the singers' mouths to the song, shot by shot, with QA. Every input is a slot (bring a clip and song, or generate them with LTX-2.5, Qwen-Image-Edit and ACE-Step); vocal split, lyric alignment, lip kits, sync QA.

LSDub

lead

word_close

face_target

snap

auto

continuity

settings

jaw_csv

The mouth controls for the dub. Defaults are the proven recipe; change one thing at a time.

LSDubTweaks

tweaks

LSDub Mouth Tweaks

Name	Type	Default	Description
lead	FLOAT	0.060–0.3	How far the mouth leads the sound, s. Lower = later mouth.
word_close	FLOAT	0.060–0.3	Closure between words, s. Higher = crisper word edges.
face_target	FLOAT	512256–1024	Face height the shot is upscaled to before warping, px.
snap	BOOLEAN	true	Snappy mouth: 2x frame rate warp and sharper attacks.
auto	BOOLEAN	true	Also try the full-frame warp and keep whichever measures better in sync.
continuity	BOOLEAN	true	After the dub, repair single-frame pops against the source.
settings	STRING		Advanced: extra mouth-warp flags, passed through as is (e.g. --warp-sigma 0.8).
jaw_csvopt	STRING		Optional: a jaw-curve CSV for the whole song (Audio2Face-3D blendshape export, columns frame,time_s,...,jawOpen; song time). Drives the jaw in the silences more cleanly. Empty = the vocal envelope only.

LSDub Mouth Tweaks

Inputs (8)

Outputs (1)