Building a sequence out of an action
Shots do not add, they multiply
In the 1920s Lev Kuleshov cut the same neutral shot of an actor's face against three different images — reportedly a bowl of soup, a child in a coffin, a woman on a divan — and audiences described the actor as hungry, grieving and desirous. The face was identical every time.
The experiment's details are disputed and the original footage is lost, so treat the anecdote as a demonstration rather than an evidence base. The principle behind it is not disputed and you can test it yourself in twenty minutes: the meaning of a shot is produced by the shot next to it. Editing is not selection; it is juxtaposition.
This is the whole craft in one sentence. It also means you can build things you never filmed.
The sequence, as a unit
A sequence is a group of shots that together perform one action or make one point. A person makes tea. Someone arrives at a building and goes inside. A process is explained. A sequence has a beginning, a middle and an end even if the film around it does not.
The classic construction, and it works for almost any physical action:
- Wide — establishes where we are and what is about to happen.
- Medium — the person doing the thing.
- Close on the hands — the actual detail of the action.
- Insert — the object itself, very close.
- Reaction — the face, or somebody else watching.
- Wide again, or a new wide — the state of the world after.
Six shots, and the viewer believes they watched a continuous event. They did not: they watched six fragments, probably filmed in a different order, possibly over an hour.
The shooting list this implies
If you also shoot, this is the most valuable paragraph in the block. For any action you intend to include, film it at least three times: once wide, once medium, once close on the hands. Repeat the whole action each time rather than filming fragments — repeated whole actions cut together; fragments do not, because the match points do not exist.
The thirty seconds this costs on set buys you the ability to compress the action to any length you like, hide any mistake, and cut away at any moment.
Cause and effect, and the word "therefore"
A sequence of shots joined by "and then" is a list. A sequence joined by "but" or "therefore" is a story.
This is a test you can apply to your own assembly. Read your shot order aloud, inserting the joining word between each pair. "She picks up the kettle, and then she fills it, and then she puts it down." Nothing is happening. "She picks up the kettle, but it is empty, therefore she goes to the tap." Now there is a small engine running.
In unscripted material you find these relationships rather than write them, but they are almost always there. Somebody tried something, it did not work, so they did something else. The edit's job is to make that visible.
The insert that does the work
The single most useful shot in any sequence is the close insert of the object. It is useful for three separate reasons at once:
- It tells the viewer what matters in the scene.
- It is neutral enough to be placed almost anywhere, so it hides a cut you could not otherwise make.
- It compresses time invisibly. Cut to hands for two seconds and cut back, and the viewer accepts that any amount of time has passed.
If you shoot nothing else extra, shoot inserts.
Compressing time without saying so
Real actions are boring at their real length. A person walking from a door to a desk takes eight seconds; a sequence can do it in two.
The mechanism is that the viewer reconstructs continuity from partial information. Shot of the door opening; cut to feet on the floor; cut to the person sitting down. Nobody asks what happened to the intervening six seconds, because the shots establish direction, and direction implies the journey.
The failure mode: compress too far and the viewer loses the geography. The test is whether they could draw the room. If not, add one wide.
Montage and its two meanings
The word means two different things and both are useful.
Montage as compression — the passage-of-time sequence, usually over music. Days of work in thirty seconds. Works because each shot is an instance of a repeated activity rather than a step in a chain.
Montage as argument — cutting two unrelated things together to produce a third idea. A politician's promise, cut to a closed factory. This is powerful and it is also the technique most easily used to say something untrue, because it asserts a relationship without stating it. If the juxtaposition would be a lie in words, it is a lie in pictures, and the fact that you never said it out loud is not a defence.
Today
Film somebody making tea, three times, from three distances, performing the whole action each time. Cut it into a six-shot sequence lasting fifteen seconds. It will take twenty minutes and it teaches the whole block.
The one thing to keep
A sequence is built out of fragments the viewer stitches into a continuous action, so shoot every action whole from three distances and join the shots with 'but' and 'therefore' rather than 'and then'.
Before you move on
An editor needs to compress a two-minute process into twenty seconds but only has one continuous wide shot of it. Why does adding a close insert of the object, filmed separately afterwards, solve the problem that speeding up the wide does not?
Pick the one you would defend. Nobody sees your answer.