Keep only the lead singer: vocals -> lead + backing, and a lead-activity curve that drives the mouth (so it does not open on backing lines). — ComfyUI node…

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

vocals

share

floor_db

Keep only the lead singer: vocals -> lead + backing, and a lead-activity curve that drives the mouth (so it does not open on backing lines).

Name	Type	Default	Description
vocals	AUDIO		The vocal stem (from LSDub Vocal Split).
share	FLOAT	0.350.05–1	How much of the vocal energy must be the lead for a frame to count as lead. Lower keeps more; higher drops more backing.
floor_db	FLOAT	-40-80–-10	Quieter than this (dB) counts as silence.

Name	Type	Description
lead	AUDIO	The lead vocal.
backing	AUDIO	Everything else in the vocal stem.
lead_activity_csv	STRING	CSV time_s,lead_active,f0_hz per analysis frame: the mouth drive.

LSDub Lead Vocal

Inputs (3)

Outputs (3)