Skip to main content

Week 3: Prompting as direction

Treating prompt-writing as a directing discipline — shot type, blocking and coverage as textual grammar, and where that grammar runs into the medium's actual limits

8 March 2027

Text-to-video generation systems increasingly document their own prompting guidance in the vocabulary of cinematography, not of software: shot scale (wide, medium, close-up), camera move (dolly, pan, tilt, tracking shot, static), and framing (over-the-shoulder, low angle) are surfaced as controllable parameters because the training footage was captioned by people who already used that vocabulary to describe what a camera was doing. A prompt that says “a woman walks into a kitchen” gives the model almost nothing to work with. A prompt that says “medium wide, static camera, a woman enters frame right and crosses to the counter” is doing what a shooting script does: specifying not just what happens, but how it is seen happening. That specification is directing, done in prose instead of on a set.

What carries over from the set, and what doesn’t

Three concepts do real work once you start writing prompts this way:

Shot type and camera move control scale and motion the way they always have — a close-up reads as intimate or threatening regardless of who or what generated it, because the grammar comes from a hundred years of cinema the model was trained on captions of, not from anything specific to synthetic video.

Blocking — where a body is in frame and how it moves through it — has to be written explicitly, because the model has no set to physically place someone on. “Enters from the left” is now a sentence doing the job a floor mark used to do.

Coverage — shooting a scene from enough angles that an editor has material to cut with — is the concept that breaks. A human crew shoots coverage of the same take: the actor’s jacket, the room’s light, the prop positions are physically continuous across the wide, the two-shot and the close-up, because it’s the same afternoon on the same set. A prompt sequence generates each shot from a fresh prior with no memory of the others. You can specify the jacket’s colour in every single prompt and still get three different jackets, because “coverage” in the traditional sense assumes a continuity the medium doesn’t structurally have yet.

Case study: Bright Aperture’s six-shot method

A small production outfit that appears, in one form or another, on a specimen shelf most semesters: they generate every scene as exactly six fixed prompts — establishing wide, two singles, two close-ups, one insert — with every continuity detail (wardrobe, prop, lighting direction, time of day) copy-pasted verbatim across all six rather than described once and trusted to persist. Their footage still drifts — a ring appears on the wrong hand in the insert shot in roughly one take in five, by their own account — but far less than a crew that varies the wording shot to shot.

The actual claim

Prompting-as-direction is a real and useful frame, not a metaphor stretched past its use — the vocabulary transfers, the judgment calls transfer, the eye for composition transfers. What doesn’t transfer is the assumption, baked into a century of film-set practice, that continuity is free once you’ve built the set and dressed the actor. In this medium, continuity is the thing you pay for, in redundant words, every single shot.

Teaching team