Most gallery photos fail for the same reason. The photographer points the camera at the art, gets a flat rectangle of someone else’s work, and goes home with a bad reproduction of a good painting. The room is the more interesting picture. Walls, spacing, wall text, and one person deciding where to stand. This frame works because the subject is turned away from the big canvas and reading the panel instead.

The stall had everything a photographer thinks they want. Whole crabs on crushed ice, whelks in their shells, three grades of oyster, lemon wedges cut that morning, parsley banked up like hedging. I shot the counter twice, wide, both frames from standing height. Both were mush. Too much of everything and no reason for the eye to stop anywhere. Then I bought one oyster, the vendor handed it over in a paper tray, and the picture was already made.

An open kitchen is one of the least cooperative environments you can point a camera at. The light source is whatever the restaurant installed for cooks, not photographers, and it rarely agrees with itself from one fixture to the next. In this frame the copper hood throws a warm, almost orange glow across the top third, while the stainless prep line below sits under a cooler, harder white. Shooting straight into that split without a plan produces a mess of clashing color temperatures.

Evaluative metering is correct most of the time. It analyzes the scene, weights the meter reading toward the focus point, and produces an exposure that protects the most information across the widest area of the frame. In scenes with even illumination and a subject that occupies a significant portion of the frame, it works without adjustment. In scenes where the subject is lit differently from the background — a face in open shade against a bright street, a person near a window with an unlit room behind them — evaluative metering averages toward a wrong answer.