Every photograph is mostly the part you cut off

The frame is the first claim

A photograph is defined as much by its edges as by its contents. Everything outside the rectangle has been excluded by somebody, and the viewer has no way of knowing what it was.

This is why the familiar defence (we did not alter the image) is weaker than it sounds. You do not need to alter an image to make it say something that is not true. You need only to choose where the edges go, and choosing where the edges go is not optional; there is no photograph without it.

A crowd photographed tightly around a confrontation is a riot. The same crowd photographed wide is forty people and an argument. Neither image has been manipulated. Both are accurate. They make opposite claims.

Four decisions the camera makes for you

1. Height. A camera below eye level looks up at the subject, which is the geometry of a smaller person looking at a larger one, and reads as imposing or authoritative. Above eye level reverses it and reads as diminished or observed. Level is the neutral one, and neutral is also a choice.
2. Distance and lens. A wide lens used close up exaggerates what is nearest (a nose, a hand, a machine) because the distance ratio between near and far parts of the subject is extreme. It reads as aggressive, comic or distorted. A long lens from far away compresses depth, flattens features, and reads as observational or surveillant. The face has not changed; the relationship between the camera and the face has.
3. The moment. Speech makes unflattering shapes. Any politician can be photographed mid-vowel looking deranged, and any calm person can be caught blinking. A fifteen-frame difference is the difference between competent and shifty, and no editing is involved.
4. What is in focus. Throwing a background out of focus removes information from the frame as effectively as cropping it. The context is present, illegible, and deniable.

Where the industry draws its line, and why the line is odd

News agencies have explicit rules. Reuters, the Associated Press and their peers forbid adding or removing content (no cloning out a distracting object, no compositing) while allowing cropping, and allowing tonal adjustment within what a darkroom could have done. Photographers have been dismissed for cloning out a shadow.

The rule is defensible as a bright line, because “did you add or remove pixels” is checkable and “is this fair” is not. But it is worth seeing clearly what it permits: a crop can mislead far more comprehensively than a cloned-out shadow, and it is entirely within the rules. The line is drawn where it is because it can be enforced, not because it tracks the harm.

This disagreement is live. Documentary photographers, news editors and academics argue about it continuously, and there is no settled resolution to hand you.

The working test for commercial and documentary work alike

You will rarely be deciding whether to clone something out. You will constantly be deciding how tight to crop, which frame to pick, and where to stand.

The question that does useful work: would the person in the photograph, or the person who was standing there, recognise this as what happened? Not would they like it (nobody likes their own photograph) but would they say this is a fair account of the moment.

It is not a rule and it does not decide hard cases. It does catch the ordinary ones, which are the ones you will actually face: the crop that makes a small workshop look derelict, the low angle that makes a friendly person look severe, the frame chosen because it was dramatic rather than because it was representative.

Generated and heavily edited images

An image that is wholly or partly generated carries a further problem: it depicts something that never occurred, in a register the viewer reads as photographic.

The duties here differ by country and keep changing: India’s rules for synthetic media, China’s labelling requirements and the European transparency provisions all point towards visible marking, on different timetables and with different scope, and none of this is legal advice. Provenance standards such as C2PA attach a signed record of how an image was made, and are being adopted unevenly; metadata also does not survive a screenshot.

The design consequence is the durable part: a visible label in the frame is the only disclosure that survives being copied. Everything attached to the file is lost the first time somebody screenshots it.

What this does not mean

None of this makes photography dishonest or unusable. Every representation selects: a sentence selects, a diagram selects, a film selects. The point is not to achieve neutrality, which is unavailable, but to stop believing you have it by default.

A photographer who knows the frame is an argument can make the argument deliberately, defend it when asked, and notice when they are making one they did not intend. A photographer who believes the camera records is making arguments constantly and cannot see any of them.