Left is black. Right is white. Height is how many pixels are that bright. That's the entire reading.
A spike hard against the right edge means clipped highlights — those pixels are pure white and no editing brings them back. A spike against the left means crushed blacks. Everything between is distribution, and distribution is a creative decision, not a mistake.
What the axes mean
The horizontal axis is brightness, from pure black at the far left to pure white at the far right. The vertical axis is how many pixels in the image sit at that brightness.
That's it. A histogram is a bar chart of your photo's tones. It contains no information about where in the frame those tones are — a photo and its mirror image produce identical histograms.
The thing to internalise: the height means quantity, not importance. A huge spike in the shadows might be a black background you don't care about. A tiny sliver at the right edge might be your subject's face blowing out. Height and importance are unrelated.
There is no correct shape
The most common bad advice in photography is that a good histogram is a nice bell curve in the middle. It isn't. The histogram describes your scene, and scenes differ:
| Scene | Correct histogram |
|---|---|
| Snow, beach, white studio | Piled against the right. Anything else is grey snow. |
| Night scene, low-key portrait | Piled against the left, with small highlights. |
| Fog, mist, overcast | Bunched in the middle, nothing at either end. |
| Backlit subject, sunset | Two humps — dark subject, bright sky. |
| Average daylight scene | Spread fairly evenly. The textbook one. |
If your camera meters a snow scene to a middle-of-the-histogram average, the meter is doing exactly what it was designed to do — assume the world averages to mid-grey — and it's wrong, because snow isn't grey. That's the situation exposure compensation exists for.
Clipping, and when it's fine
Clipping means pixels have hit the maximum or minimum value and stopped recording. There is no detail there, because no detail was captured — not because it's hiding.
Highlight clipping is usually the expensive one. Blown highlights render as flat white paper, and the eye finds it jarring. Crushed shadows read as normal darkness far more readily.
But clipping is not automatically a failure:
- Fine: the sun in the frame, a bare bulb, chrome reflections, a light source. These hold no useful detail, and forcing them down makes the whole picture flat and grey.
- A problem: skin, clouds with structure, a white shirt, snow with texture, anything where the detail mattered.
The practical rule: look at what is clipped, not how much. A frame that's 3% clipped because the sun is in it is fine. A frame that's 0.4% clipped on a forehead is not.
You can check any photo's exact clipped percentage — per channel — with the histogram analyzer.
The comb pattern
Sometimes a histogram shows regular vertical gaps, like a comb. It means the image has been stretched in editing and is running out of levels to describe the tones with.
An 8-bit JPEG has 256 brightness levels per channel. Increase contrast, and you spread those levels further apart — the gaps are levels that no longer have any pixels. Push far enough and smooth gradients start to band: the stepped rings you see in skies.
Combing is a warning that you've asked too much of the file. It's the single clearest argument for editing from raw, which carries 12 to 14 bits — thousands of levels instead of 256 — and can absorb the same edit without gaps.
RGB versus luminance
Most cameras show one of two things, and it's worth knowing which:
- Luminance — a single curve of perceived brightness. Simpler, and what most people mean by "the histogram."
- RGB — three overlaid curves, one per colour channel. Uglier, and more informative.
RGB matters because channels clip independently. A red rose in bright sun routinely blows the red channel while green and blue are nowhere near the top. Luminance shows a comfortable histogram; the red channel has already lost all its detail, and the rose renders as a flat red shape.
If your camera offers RGB, use it. Saturated reds and deep blues are the usual offenders, and they're invisible on a luminance-only display.
Why your camera lies about RAW
This one surprises people, and it changes how you shoot.
When you review a shot on the back of the camera — or on a phone screen — the histogram is almost never computed from the raw data. It's computed from the embedded JPEG preview: the raw file with the camera's contrast curve, picture profile, saturation and tone mapping already applied.
That preview clips earlier than the raw file does. Highlights that look completely gone on the camera screen frequently pull back in editing — typically between half a stop and a full stop of extra headroom, sometimes more.
Learn your camera's offset. Deliberately clip a bright scene, then see how much you can recover from the raw. Once you know it's roughly two thirds of a stop, you can expose closer to the right — capturing more signal and less noise — while the on-camera histogram still shows a warning you now know to discount.
Zebra stripes: the spatial version
The histogram's blind spot is location. It tells you 2% of the frame is clipped; it can't tell you whether that's the sky or your subject's face.
Zebra stripes solve exactly that. Borrowed from broadcast video monitors, they draw diagonal stripes over any area brighter than a threshold you choose, directly on the live image. You see instantly which part is going.
Thresholds are given in IRE, a video brightness scale where 0 is black and 100 is nominal white:
- 100 IRE — marks genuinely clipped pixels. Use it as a hard warning.
- 95 IRE — warns a fraction of a stop before clipping. The safest general setting.
- 90 IRE — the classic setting for placing lit skin tones. If a face is striping at 90, it's about to be too bright.
Histogram for the overall picture, zebra for the specific problem. Together they're most of what a light meter used to be for.
Common questions
How do you read a histogram?
Left is black, right is white, height is how many pixels are that bright. Spikes at either edge mean clipped tones. Everything else is distribution, which is a creative decision rather than a right answer.
What does a good histogram look like?
Like the scene. Snow piles right, night piles left, fog bunches in the middle. The only genuine error is clipping you didn't intend.
Should I "expose to the right"?
For raw, pushing exposure as far right as you can without clipping captures more signal and less noise. It's a real technique with a real cost: you must actually check the highlights, and you must edit afterwards, because the file looks too bright straight out of camera.
Does the histogram work for video?
Yes, and it matters more, because video has far less latitude than raw stills. Most operators run a histogram plus zebra continuously while shooting rather than checking afterwards, since you can't re-expose a take that already happened.
Can I see a histogram on an iPhone?
Not in the stock Camera app. Third-party camera apps can display one live in the viewfinder, which is far more useful than checking afterwards — you can fix the exposure while the shot is still in front of you.
Keep going
A live histogram, before the shot.
Lucid renders a real-time RGB histogram and selectable zebra stripes right in your iPhone's viewfinder — where they can still change the photograph.