How to read a histogram: a reliable exposure check

The rear screen lies. A frame that looked fine at night often turns out underexposed the next morning, because screen brightness, ambient light and the panel itself all skew what you see. The histogram just counts pixels, which makes it the one exposure check on the camera you can trust.

A monitor showing an image editing application

1. What the two axes mean

A histogram counts pixels. It is not a thumbnail of the frame.

One histogram with the x-axis labelled brightness from black to white, the y-axis labelled pixel count, and clipped-black / clipped-white zones marked at both ends — x is brightness, y is how many pixels sit there. The hump can be anywhere; touching a wall is the problem.

Horizontal is brightness, from pure black at the left to pure white at the right. It has nothing to do with left and right in the picture.
Vertical is how many pixels sit at that brightness. Taller bar, more pixels.

So a night scene piles up on the left, a snow scene on the right, and both are fine.

2. Clipped blacks and whites

Only the ends matter. When the graph slams into an edge and stacks up there, those pixels have run out of range:

The same frame under-, correctly and over-exposed, each with its histogram beneath: one piled left, one centred with headroom, one slammed into the right wall — Three exposures, three shapes — each histogram is computed from the frame above it.

Where	Name	Consequence
Hard right	Blown highlights	The data is gone; RAW cannot recover it
Hard left	Crushed shadows	Detail lost, but RAW can often bring some back

The two ends are not equal, so err towards slight underexposure. Colour negative film is the opposite and handles too much light better than too little.

Turn on the highlight warning. Most cameras can blink blown areas on playback, and many show zebras in live view. Whatever flashes is already pure white.

Try one in the browser

Mark adds the settings strip right in your browser. Free, no account, nothing uploaded.

Open Mark — try one for free

3. There is no correct shape

You will read that a good histogram is a hill in the middle. That only holds for mid-toned scenes.

High key (snow, white walls, backlit portraits) belongs on the right. Force a hill and the snow turns grey.
Low key (night, dark still life) belongs on the left. Lift it and you lift the noise.
High contrast, like a room with a bright window, shows bars at both ends and a dip between. One frame cannot hold it: choose what to keep, or bracket and blend.

The one rule that always applies: neither end clipped. The subject decides the shape.

4. Using it while shooting

Shoot a frame and check the histogram on playback, or overlay it in live view if your body can.
Right end clipped: dial in −1/3 to −1 stop of exposure compensation and reshoot.
Left end crushed and you want the shadows: add compensation, or use a tripod and a slower shutter.
Nothing clipped: carry on, whatever the shape.

For timing the best light, see golden and blue hour; for trading the settings against each other, see the exposure triangle.

5. FAQ

Is the camera histogram the same as the one in my editor?

Not quite. The camera shows the histogram of a JPEG preview built with your picture style, and the RAW holds a little more range. Just touching the right edge on a RAW is often recoverable, but there is not much headroom.

What is the RGB histogram for?

It shows red, green and blue separately. On red flowers the red channel often clips first while the luminance histogram looks fine, and the petal detail is gone. Use it for saturated subjects.

What does "expose to the right" mean?

Push the histogram as far right as it goes without clipping, then pull brightness down in editing. More light means cleaner shadows, and it works on digital, but go too far and the highlights are gone for good. Starting out, just avoid clipping.