How to choose which clips belong in your compilation

Most people overestimate how many clips a good compilation needs. A 60-to-90-second video, even one made of short scenes, rarely holds more than 8 to 15 clips comfortably. If you've imported 40 clips from an event or trip, the real editing work is choosing which dozen or so actually make the cut — everything after that is arranging and trimming. This guide is specifically about that selection decision.

Start with quality scores, but don't stop there

Every imported clip gets an automatic quality score used for ordering suggestions. It's a reasonable first filter — a genuinely shaky, dark, or out-of-focus clip usually scores low for a real reason. But quality scoring can't tell you which clip captures the moment that actually matters to your compilation's purpose. A slightly soft-focus clip of the winning goal is more valuable than a technically sharp clip of the crowd milling around beforehand. Use the score to flag candidates for removal, not to make the final call by itself.

A selection framework

1. Watch every imported clip once at full attention using the inspect view, before making any decisions.
2. Mark clips that are technically unusable (shake, focus, dark exposure) using the quality score as a first pass.
3. Of the remaining clips, identify the one or two strongest candidates for hook and peak roles.
4. Group the rest loosely by what role they'd play — build, escalation, payoff — rather than by chronology.
5. Cut any clip that doesn't clearly serve one of those roles, even if it's individually decent.
6. Check remaining clips against the duplicate flag for near-identical takes you might want to consolidate to one.

Handling near-duplicate clips

Exact duplicate files get flagged automatically, but near-duplicates — three takes of the same trick, several angles of the same toast — are a judgment call the system won't make for you. Generally, keep the single best take rather than multiple similar ones; two very similar clips back to back reads as repetitive rather than reinforcing, even if each one individually looks fine. The exception is when the near-duplicates genuinely show progression (attempt one fails, attempt two lands), in which case both earn a place because together they tell a story a single clip couldn't.

Choosing clips against your length target

Work backward from your target length. At 60 seconds with short scenes, you might realistically use 8-12 clips; at 90 seconds, maybe 12-18. If you've selected noticeably more candidates than that, the fix isn't to make each scene shorter until they all fit — that flattens pacing — it's to cut further. Generate a plan at your intended length early, in draft, so you're choosing against a real constraint rather than in the abstract.

Clips that look good alone but don't belong

• A visually striking clip that doesn't match the energy or subject of the rest of the set.
• A long establishing or context clip that doesn't have a strong moment inside it once trimmed to compilation length.
• A clip that only makes sense with narration or backstory you can't include, since there's no voiceover or subtitle feature to carry that context.
• A technically excellent clip that duplicates a beat another clip already fills better.

Mistakes in the selection process

• Including a clip because it took effort to film, not because it earns its place in the final cut.
• Relying entirely on the quality score without watching flagged-low clips yourself — some low scores are wrong for your purpose.
• Keeping three similar takes of the same shot instead of picking the strongest one.
• Selecting clips first and only checking your target length afterward, forcing an awkward final trim.
• Treating chronological order as automatically the right order for selection and sequencing both.

When you don't have enough clips

If your honest selection leaves you with only 4-5 usable clips, that's fine for a 60-second compilation — don't pad it with weak filler to hit a higher clip count. A short, honest set of strong clips at 60 seconds outperforms a stretched 90-second cut carrying material that doesn't hold up. If you consistently don't have enough usable footage, that's a filming problem to solve before your next shoot, not something editing choices can fix after the fact.