What cropping the world out does
Every other size gives the audience a choice about where to look. A close up removes it. There is one subject, it fills the frame, and every flicker on it is now the whole event. That is why the size is spent on turning points: the moment a character decides, realises, lies, or gives in.
The cost is symmetrical. The information a wide shot was carrying, where people are and what surrounds them, is gone. Audiences will tolerate that for a while and then start feeling untethered. A close up borrows against the geography that an earlier wide shot paid for, and if that earlier shot never happened, the debt shows.
Scale is also strange at this size. A face filling a screen is many times life size, which means small things become large: a swallow, a shift of focus in the eyes, a tightening at the jaw. Performers work smaller in a close up for this reason, and the same principle explains why a big theatrical gesture looks false when the frame is tight.
Sizes within the family
- Medium close up. Chest up. Still shows shoulders and posture, the standard interview size.
- Close up. Head and shoulders. A whole face with a little headroom.
- Big close up. Chin to forehead. The face is cropped by the frame, which raises intensity.
- Extreme close up. A part rather than a whole: eyes, mouth, hands, a keyhole, a wedding ring.
- Insert. A tight shot of an object or an action cut into a wider scene, such as a hand turning a key.
Objects are worth treating differently from faces. A close up of a face is about interiority. A close up of an object is usually about plot, so it works best when the audience already wants to know what the thing is. Inserting a tight shot of a prop nobody is curious about reads as the film pointing at itself.
Where it stops working
The commonest problem is inflation. Once a sequence is running on tight shots, the next turn needs something tighter, and eventually the size has nowhere to escalate. Keeping a scene mostly in mediums means one close up will do the work of five.
The second problem is a close up shot with nothing to hold. This size needs either performance or texture. A neutral face doing nothing at this scale simply looks like a photograph, and cutting to it drops the scene's pulse rather than raising it.
Prompting one
This is the strongest size in image generation, because portrait data is everywhere. That strength is also the trap: without direction you get a posed studio portrait.
- Reliable:
close up shot, head and shoulders, 85mm, shallow depth of field - For a part rather than a whole:
extreme close up of the eyes, filling the frame - Weak on its own:
closeup, which some models treat as a generic beauty shot
Three things make results feel like film rather than headshots. Give the gaze a direction, since looking off frame left immediately implies a scene. Name one physical detail the size is there to show, such as sweat at the temple or a chipped nail on the trigger, because that tells the model why the frame is tight. And name the lens, because a wide focal length at this distance produces a bulging nose that no amount of prompt polish will fix afterwards.