Light metering and the histogram

What the meter is trying to do

The light meter in your camera measures how much light comes back from the scene and picks settings that make it look average. Average here means mid-grey, a tone roughly halfway between black and white (often described as about 18% reflectance, an approximate convention).

This works well for ordinary scenes with a mix of light and dark. It goes wrong when the scene really is mostly bright or mostly dark. A snowfield gets pulled down to dull grey; a black dog on dark ground gets pushed up to washed-out grey.

The fix is exposure compensation. For snow or a bright beach, try about +1 to +2 stops. For a dark subject filling the frame, try about −1 stop. These are starting points; check the result.

Metering modes and when to use them

Evaluative (also called matrix or multi-zone): the camera divides the frame into many zones and weighs them, often giving extra weight to the focus point. Use it as your everyday default.

Centre-weighted: reads the whole frame but counts the middle most. Useful for a subject in the centre with a bright or dark background, and predictable because it does not try to be clever.

Spot: reads only a small circle, typically a few percent of the frame. Use it for a lit face on a dark stage, the moon, or a backlit subject, where only one small area matters.

Whatever the mode, the meter still aims for mid-grey in the area it reads. Spot-metering a white wedding dress will still make it grey unless you add compensation.

How to read a histogram

A histogram is a bar graph of brightness. The left edge is pure black, the right edge is pure white, and everything in between is a shade of grey from dark to light.

The height at any point shows how many pixels have that brightness. A tall peak on the left means lots of shadow; a tall peak on the right means lots of bright area.

On this page, a mountain landscape sits next to its histogram: the dark forest forms a hump on the left, the grey rock fills the middle, and the bright sky and snow sit towards the right.

The histogram is more reliable than the rear screen, whose brightness changes with screen settings and sunlight.

Clipping and blinkies

If the graph is pressed hard against the right edge, some highlights are pure white with no detail left. This is called clipping. Against the left edge, shadows have become solid black.

Many cameras can flash clipped highlights on the preview image, often called blinkies or highlight warning. Turn it on: it shows exactly where detail is lost.

Small clipped areas are fine for the sun, street lamps or specular glints. Clipping on a face, a white shirt or cloud detail you care about is worth fixing, usually by reducing exposure by a third to a full stop.

Expose to the right, simply put

Digital sensors record more clean detail in bright tones than in dark ones. Brightening a dark photo later also brightens its noise.

So, when shooting raw, a useful habit is to make the exposure as bright as you can without clipping highlights you need, pushing the histogram towards the right edge. You then darken it slightly when editing, with less noise in the shadows.

Do not overdo it: once highlights clip, they cannot be recovered. With JPEG, or when the scene is fast-moving, a normal exposure is the safer choice.

What a good histogram looks like

There is no ideal shape. A good histogram is one that matches the scene and avoids unwanted clipping.

Snow scene or bright beach: weighted to the right. That is correct, not overexposed.

Night city or low-key portrait: weighted to the left, with a few spikes on the right for lights.

Foggy morning: a narrow hump in the middle, with nothing at either edge.

Sunset into the sun: often both edges touched. You may need to accept some clipping, or use bracketing to combine exposures.