A wedding ceremony is rarely lit consistently — a church goes from window-side daylight at the entrance to candlelit altar to harsh exit-door backlight in 40 minutes. Trying to colour-match the whole ceremony to one reference frame averages out and ends up flattering nothing. The right move is to cluster the footage by light condition, build one reference per cluster, and let AI Color Match Quick Mode handle the bulk of each cluster.
How to do it
Cluster the ceremony footage into 3-4 light groups: aisle entrance, altar wide, ring exchange / vows close-up, recessional. Each cluster is a separate AI Color Match job.
Pick one well-exposed hero frame per cluster as the reference. For LOG footage this still works — AI Color Match supports LOG directly without needing a LUT pre-conversion.
Run AI Color Match in Quick Mode for the majority of each cluster. Click reference → Upload for Color Matching (or Recommended sample if no hero exists yet) → preview → Apply Effects to render the look.
Switch to Control Mode for outliers in the cluster — frames where Quick produced too-strong or off-direction matches. Control Mode gives you the fine-tune you need without leaving the AI Color Match flow.
Use Reference Management to organize the 3-4 references by colour or style group so the next-day editor can re-apply the same references to additional clips without re-uploading.
Try it
On your next wedding ceremony edit, cluster the footage into 3-4 light groups before running any AI Color Match. Build one reference per cluster, run Quick Mode across the cluster, then Control Mode on the 2-3 outliers per cluster. Compare against your previous one-reference workflow on the same ceremony — the look should hold across light changes without the global averaging that flattens the look.
Watch out for
Trying to match the candlelit altar to the daylight aisle. They're different colour worlds; the goal is consistency within a cluster, not identical across the whole ceremony.
Forgetting the Apply Effects render step. AI Color Match preview shows you the result, but until you click Apply Effects the look isn't rendered into the clip and won't playback consistently.
Treating every clip as its own AI Color Match job. The clustering step is what saves time; per-clip matching defeats the purpose.
Source: Evoto Video AI Color Adjustments