Lip sync, and why it is a within-shot problem
Lip sync is the alignment of a character's mouth with the words being heard. It is a different problem from cross-episode continuity: it lives inside a single generation, it is the failure a viewer notices first, and it is solved at a different scale from a cast. As of 2026-09-12.
| The scale of the problem | What has to match | Why it is separate from a cast |
|---|---|---|
| Inside one shot | Mouth against audio, frame by frame | No stored object is involved |
| Across a scene | Which character is speaking | A cast holds appearance, not attribution |
| Across a language | Timing against a dubbed track | The words change and the picture does not |
Inclusion rule. Three scales at which mouth and audio can disagree, ordered from the narrowest. The table describes the problem rather than any product's handling of it. Order. From the narrowest scale to the widest.
1A page bundling lip sync with episode continuity is answering two questions at once
One page on this file syncs dialogue to lip movement, locks faces across every episode and assembles everything on a timeline in a single sentence. Those are three operations at three scales, and a reader cannot tell which of them the sentence is strongest about.
This file grades the cells separately for that reason. Within-shot accuracy says nothing about whether a face survives to the ninth instalment.
2Dubbing turns lip sync into a new problem
A cut dubbed into a second language has a picture composed for the original timing. Where a page states that timing is preserved, that is a claim about the audio rather than about the mouths, and the two are not the same guarantee.
A line generated in the target language while the shot renders avoids the problem altogether, and is a different pipeline with a different schedule.
3Attribution is the failure that looks like lip sync and is not
When two characters are in frame and the wrong one appears to speak, nothing is out of sync: the words are attached to the wrong mouth. That is a storyboard problem rather than an audio one.
It is also the commonest reason a two-hander has to be regenerated, which is why a dialogue-heavy segment is the right thing to put in a trial.
4Where these figures come from
A definition kept for this register: the term, and what vendors state about it, each figure read on 2026-09-12. The tools themselves are on the register. Related: Season, Ingredient.
5One pipeline's continuity
Serial work like “Lip sync” depends on continuity; SceneMixer's guides describe how it keeps it:
A producer working in German can make a series whose cast speaks Indonesian: the interface and the working documents follow the interface language, the spoken dialogue follows the project setting.
SceneMixer, AI Short Drama in 15 Languages (2026): Interface, Scripts and the Dialogue Your Cast Actually Speaks — “Can the interface be in one language and the dialogue in another?” (read 2026-10-13)
A SceneMixer segment comes out with its sound already in it: each line spoken in that character’s voice, the speaker’s lips moving with the words, and the ambience of the place — rain, traffic, a clock.
SceneMixer, Dialogue and Sound in One Generation (read 2026-10-13)