Duplicate shot detection

What it means

Action footage is full of repetition. You hit the same kicker five times, two cameras record the same wave, a drone circles the same peak twice. Duplicate shot detection finds these near-identical clips so that an edit can keep one strong version and drop the rest.

Exact duplicates, the same file copied twice, are easy to catch by comparing file hashes. Near-duplicates are harder, because the shots differ slightly in timing, framing and light. Systems compare video embeddings or visual fingerprints and treat clips that sit very close together as one group. The step works hand in hand with shot scoring, which decides which member of each group survives.

How it works in practice

On a multi-camera trip, duplicates come from angles as well as attempts. A helmet camera and a friend on the shore filming the same jump are not duplicates in the strict sense; they are two views of one moment, which an editor might use back to back like a replay.

Repeated attempts are where detection saves most time. If you filmed ten laps of a bike park line, the tool can pick the cleanest lap and maybe one crash, instead of showing all ten.

Copying the same card twice by mistake is common, especially when a card is offloaded on two laptops. Detection silently fixes that too, but tidy folders still help.

What to watch out for

Aggressive thresholds can drop meaningful variation. Two runs down the same line can look alike, yet one is the run where everything finally worked. If a moment matters, make sure its attempt is clearly different, for example a different angle or a celebration at the end.

Very loose thresholds do the opposite and keep too many similar shots, making the edit feel repetitive no matter how good each clip is.

Duplicates across chaptered files are not duplicates at all: they are continuous parts of one recording. Good tools treat chapters as one clip before comparing.