The lighting and camera language that consistently produces stronger images.
Image models respond especially well to two kinds of language: how a photographer would describe the light, and how a camera would describe the shot. Leaning on both is a shortcut to results that look intentional rather than generic.
Borrow a photographer's vocabulary
Terms like "golden hour," "soft diffused light," "backlit," or "hard studio light with a single key" carry an enormous amount of visual information in a few words - far more precise than "nice lighting." The same goes for camera language: "shallow depth of field," "wide-angle establishing shot," "shot at eye level" all steer composition directly.
Keyword-stuffing works against you
Long strings of disconnected adjectives read as noise, not direction. A prompt that names five contradictory moods, or repeats the same idea three ways, forces the model to average them out into something muddier than any single clear choice would have produced.
- 1Name one lighting condition, precisely.
- 2Name one framing or camera angle.
- 3Name one dominant mood, not three.
- 4Read the prompt back - if two phrases contradict each other, cut one.
If a result looks generic, the fix is usually not more words - it is more specific ones. Swap "nice lighting" for "late afternoon sun through a sheer curtain" and watch the difference.
The same discipline - one clear idea instead of five vague ones - is even more important once things start moving. That is the subject of the next lesson.