Check whether the light change is intentional
short-drama-edit identifies usable footage before matching a scene. Do not equalize every full-frame average after noticing a jump. Framing, clothing area and background change those statistics.
Original material setup: at a night service counter, a clerk hands over a key, followed by a close view of the visitor’s hand. The light state is unchanged, but one shot is yellow and the other washed out. Compare the desktop and sleeve. A later lamp switch intentionally darkens the space and should remain.
These are fictional diagnostic descriptions, not measurements. Play actual footage and record the first defect’s time and region. Gradual brightening within a shot differs from a jump at a cut; two stills alone are insufficient.
Compare equivalent regions rather than whole-frame averages
| Check | Example | Avoid |
|---|---|---|
| Scene and light | Same counter, lamp still on | Grouping a new scene or lamp-off moment |
| Desktop | Same visible material | Using a reflective key as neutral reference |
| Sleeve | Same garment, unclipped area | Mistaking tighter framing for changed color |
| Character | Skin and identity both correct | Expecting grading to repair a face change |
Built-in matching uses scene-level statistics for approximate correction. It cannot infer intentional lighting. Exclude deliberately changed segments and inspect equivalent regions and actual cuts afterward. There is no universal correct brightness for every model.
Correct modestly and recheck the story information
Compare untreated footage with a modest correction. If the table matches but the face turns gray or key detail disappears, revise the affected segment rather than adding a strong filter to the episode. Tool temperature values are not kelvin, and supported ranges are not quality standards.
Record correction or exclusion decisions so effects do not stack unintentionally. Use a grading tool where needed and document actual treatment. Writing consistent color does not modify footage.
For an overexposed tail, use a complete information interval before the defect if available. If an essential action occurs in the defective region, request repair rather than deleting the handoff to avoid the flaw.
Inspect both joins and the complete scene
Scrub the desktop, garment and face around each join, then play the scene normally to see whether the change distracts from its purpose. Frames or measurements help locate it; they do not establish final readability alone.
Record corrected differences, retained intentional changes and upstream defects separately. Automatic matching success is not acceptance of the whole film.
After regenerating a segment, recheck it and both neighboring joins; old values may no longer fit. Preserve correct order, dialogue and results while repairing observed differences. Planning light before generation is a separate task.
FAQ
Will equal brightness make every cut match?
No. Framing, materials, clothing area and intentional light changes affect averages. Compare equivalent regions under the same light state, preserve story changes and judge actual playback.
Do it with the skill
Use short-drama-edit to inspect brightness and color across these shots. Confirm scene and light state, compare regions, preserve intentional changes, and review modest corrections without masking identity or action defects.
Read the method: Usable intervals and shot matching