Why negative film forgives one direction and not the other.
An in-camera meter assumes the world averages to middle grey. Most of the time it is close enough. The trouble starts in high contrast, where the meter splits the difference and the shadows land below the point where the film records anything at all.
Underexposed shadows are gone. There is no density on the negative to recover, and no amount of care in the darkroom or the scanner brings back what was never there. Highlights behave differently: they compress, they get dense and stubborn, but they usually hold something.
So the working rule is asymmetric. Meter the darkest area where you still want visible texture and place it two stops below middle grey.
| Placement | Result on the negative |
|---|---|
| 3 stops below | Near-black, texture mostly lost |
| 2 stops below | Dark with detail — the target |
| 1 stop below | Dark grey, safe but flat |
| Middle grey | Reference point |
Many people rate a 400-speed film at 200 and leave it there. This is not mysticism. It buys a stop of shadow margin at the cost of slightly denser negatives, which most films handle without complaint.
Do it deliberately, though, and keep it consistent. Changing the rating and the development time and the developer at once teaches you nothing.
If the light falling on the subject is the same light falling on you, an incident reading removes the guesswork entirely. It ignores how reflective the subject is, which is exactly the variable that confuses a reflected meter in snow or against a dark wall.
Expose for the shadows, develop for the highlights. Old advice, still the most useful sentence in the craft.