Your camera's rear screen brightness changes with ambient light and its own settings, so judging exposure by looking at the picture is unreliable — a photo that looks perfect on a dim screen at night can be badly underexposed. The histogram is the objective measurement.
What the Graph Shows
The horizontal axis is brightness: pure black on the left, pure white on the right, midtones in the middle. The vertical axis is how many pixels fall at each brightness level. That's all it is — a count of tones, not a quality score.
There Is No "Correct" Shape
This is the most common misunderstanding. A histogram describes the scene, and different scenes legitimately look completely different:
- A snow scene should be bunched to the right. Forcing it to the middle makes grey snow.
- A night scene should be bunched to the left. Centring it makes a muddy, noisy mess.
- A high-contrast scene has peaks at both ends and a dip in the middle.
- A foggy scene is a narrow hump in the middle with nothing at the edges — correct, and flattening it out would be wrong.
Ignore anyone who says the graph should be a nice hill in the centre.
What Actually Matters: Clipping
The one thing to watch for is data piling up hard against either edge:
- Right edge (highlight clipping) — those pixels are pure white with no detail recorded. This is usually unrecoverable. Once a highlight is blown, there's nothing in the file to bring back.
- Left edge (shadow clipping) — those pixels are pure black. Somewhat more recoverable in RAW, though lifting them reveals noise.
A tall spike touching an edge means clipped areas. A gentle slope reaching the edge is usually fine — a few pure white specular highlights (sun glints, light bulbs) are natural and shouldn't be corrected.
Using It in Practice
- Enable the highlight warning ("blinkies") alongside the histogram. It shows *where* in the frame the clipping is, which the graph can't. Blown highlights on a plain sky may not matter; blown highlights on a bride's white dress do.
- Expose to the right (ETTR): for RAW shooting, push exposure as far right as you can *without* clipping. Digital sensors record far more information in the bright end, so a brighter-then-darkened file has less shadow noise. Don't do this for JPEG, where the tone curve is already baked in.
- Check the RGB histogram, not just luminance. A single channel can clip while the overall brightness looks fine — very common with saturated reds and with skin under warm light.
The Screen Lies, the Histogram Doesn't
Two practical habits:
- Never judge exposure by the preview image, especially outdoors in bright sun or indoors at night.
- Trust the histogram, but read it as a description of the scene rather than a target to hit.
When shooting a high-contrast scene where you genuinely can't hold both ends, decide which one matters: protect highlights for landscapes and weddings, protect shadows when the subject is dark and the background is expendable.