Image Histogram

How to use the image histogram

1. Choose an image.
2. Switch between brightness and the colour channels.
3. Check the clipping figures.

About this tool

A histogram counts how many pixels fall at each brightness level, from black on the left to white on the right. It is the one honest way to judge exposure, because a screen's own brightness makes an image look darker or lighter than it is while the histogram does not change.

Brightness here uses the perceptual weighting — green counts for roughly ten times as much as blue — rather than averaging the channels. Averaging treats pure green and pure blue as equally bright, which is wrong by about a factor of ten to the human eye and makes the resulting histogram useless for judging exposure.

The clipping figures are the number worth reading. A pixel at 0 or 255 has no detail in it at all, and no amount of editing recovers what was never recorded. A little clipping is normal — a specular highlight on water should be pure white — but a large proportion means the exposure was wrong when the shutter opened.

There is no correct histogram shape. A photograph of snow should lean right and a night scene should lean left. What matters is whether pixels are stacked against an edge, which means lost detail, rather than merely gathered near one.

The bars are drawn on a square-root scale, since a single dominant tone would otherwise flatten everything else into an invisible line. Large images are sampled down before measuring, which changes the shape not at all.

Common uses

• Judging whether a photo is over or underexposed.
• Finding out how much highlight detail was lost.
• Checking a colour cast by comparing channels.

Frequently asked questions

What does a good histogram look like?

There is no single right shape. Snow should lean right, a night scene left. What matters is whether pixels pile up against an edge, which means lost detail.

What is clipping?

Pixels at pure black or pure white, which hold no detail. It cannot be recovered by editing, because it was never recorded.

Why is green weighted so heavily?

The eye is far more sensitive to green than to blue. Averaging the channels equally would make a green image and a blue one look equally bright, which they are not.