Framing and Crop: Telling the Model How Close to Stand

Framing is the single decision that changes an image most and gets specified least. It is not about the shape of the canvas — that is a separate setting — but about how much of your subject the frame contains, and where its edges fall on the body. Left unstated, image models default to a comfortable middle distance: the whole subject, centred, a little air around it. That default is safe and boring, and it is responsible for a great many prompts that “look generic” for reasons the writer cannot name.

The Words That Actually Move the Camera

The reliable framing vocabulary comes from film, and it works because each term names a specific edge on the body rather than a feeling. Extreme close-up means the frame cannot hold a whole face. Close-up is head and a little shoulder. Medium close-up cuts around the chest, medium shot around the waist, medium long at the knees, full shot includes the feet with a margin, and wide shot puts the figure small inside its surroundings.

These land far more consistently than adjectives. “Intimate” and “epic” describe your intention; “close-up” and “wide shot” describe a distance. If you want the figure to occupy most of the frame, say the crop, not the mood.

The same applies to objects. A macro shot of a wristwatch and a product shot of a wristwatch are different pictures, and the words are doing the work of a lens change without you having to name one.

Cropping Is a Positive Instruction

The awkward part of framing is that models are reluctant to cut things off. Ask for a close-up and you will often still get the full head with space above it, because a complete subject reads as a safer answer than a severed one. The fix is to state the crop as something present rather than something missing: “framed from the eyes to the chin, top of the head cropped by the frame edge” works, while “don’t show the whole head” tends to be ignored or inverted.

Headroom deserves the same treatment. The default leaves a generous band of empty space above the subject, which flattens portraits. Saying “tight headroom, top of the frame just above the hair” removes it, and saying it in that direction — describing where the edge sits — is more dependable than asking for less of anything.

Framing Fights With Everything Else in the Prompt

A prompt that asks for a close-up and then spends thirty words on the room is contradicting itself, and the model splits the difference: you get a mid shot with a visible room. If the crop matters, the rest of the description has to shrink to match. A close-up prompt should describe a face and the quality of light on it, and nothing about the architecture behind it, because none of that is in the picture anyway.

That is the useful discipline in framing. Choosing a crop is choosing what you are not going to talk about, and prompts get shorter and sharper as soon as you commit.