When to cut, and why most cuts fail

A cut is an instant jump from one place to another. In life this never happens, and yet nobody notices it in a film. Understanding why is the difference between editing and arranging.

A cut is invisible when something carries across it

The audience does not see the cut if their attention is already travelling. A movement that starts in the first shot and finishes in the second. A question asked before the cut and answered after. A sound that runs through both.

A cut is visible when nothing carries. Two static shots of unrelated things, joined, and the audience has to stop and work out what happened. That pause is what people mean when they say an edit feels choppy, and it has nothing to do with how fast the cutting is.

So the question at every edit point is not where to cut. It is what the audience is following and whether it survives the cut.

Motivation

Cut when there is a reason. Somebody looks off screen, so you show what they are looking at. Somebody says a name, so you show that person. A sound arrives from somewhere unseen, so you go there.

Cutting on a beat because it has been four seconds is a rhythm, and rhythm works for a montage. In anything with an argument in it, cutting without motivation reads as restlessness.

Continuity and the things that break it

Continuity is things staying consistent across a cut. A glass that is full then empty. A subject whose head jumps six inches. Light that changes because two takes were an hour apart.

The audience does not consciously catalogue any of this. They feel a small wrongness and their attention breaks for a second, and a second of attention is expensive.

The 180 degree rule is the big one. Keep the camera on one side of the line running between two people. Cross it and on the cut they appear to swap places, which reads as them turning away from each other. In an interview, the equivalent is eyeline: the subject should be looking off the same side of the lens in every setup.

Jump cuts

Cut two nearly identical shots together and the subject jumps, because the framing did not change enough for the brain to accept it as a new view. That is a jump cut.

Three ways out. Change the shot size by a clear step or the angle by at least 30 degrees, and the cut becomes legible. Cover the join with B roll. Or commit to jump cuts as a style and do them often enough that the audience reads them as deliberate, which is what vlogging did and why it works there and nowhere else.

B roll does three jobs at once

It shows what is being described, so the audience understands rather than imagines. It hides edits in the interview, so you can remove a minute of rambling and nobody sees the seam. And it gives the eye somewhere to go, which matters because a face talking for ninety seconds is a long time to look at one thing.

Shoot three times more than you think you need. Every editor who has ever run short of B roll has said this and nobody believes it until it happens to them.

Lay it over the interview audio using an L cut, so the voice continues and the picture changes. Cutting picture and sound at the same moment makes both feel abrupt.

Pacing

The most common fault in a first edit is uniform shot length. Every clip around four seconds, because that felt about right each time. It reads as mechanical even to an audience who could not tell you why.

Vary it hard. Hold the important shot longer than is comfortable, and let the audience sit with it. Cut away from the unimportant shot sooner than feels polite. A sequence that runs two seconds, two seconds, eleven seconds, one second has a shape. One that runs four, four, four, four has none.

The cure for a slow edit is almost never faster cutting. It is cutting out the parts that were not earning their place.

Judging your own work

You lose the ability to see your own edit after about the fourth pass, because you know what is coming and your brain supplies the timing. Three things get it back.

Leave it overnight. Nothing else works as well and nothing else is as unpopular.

Watch it at double speed. Structural problems survive the speed change. If a section is boring at 2x it is fatal at 1x.

Watch it with the sound off, then listen to it with the screen off. Each reveals a different half of the problems, and a piece that works in both usually works.

Things worth doing every time

• Leave handles, a second beyond each in and out point, so you can adjust a cut later without going back to the source.
• Scan the finished timeline for gaps. Black frames and silent frames are usually accidents.
• Keep the first five seconds ruthless. On every platform that is where people leave.
• Pick the thumbnail deliberately. It does more work than the opening shot.