How to use the Image Histogram
- Drop, paste or choose an image.
- Switch between the luminance, RGB overlay and individual channel views.
- Read the statistics to spot clipped highlights or shadows.
Image Histogram features
- Luminance histogram, RGB overlay and individual red, green and blue channels
- Log scale to reveal small pixel counts
- Shadow, midtone and highlight zones
- Mean, median, standard deviation and tonal range
- Highlight and shadow clipping warnings in plain language
- Runs locally; your last image is kept in your browser
Read the exposure of any photo
An image histogram is a graph of how many pixels sit at each brightness level, from pure black on the left to pure white on the right. Photographers use it to judge exposure far more reliably than by looking at a screen, which can be too bright, too dim or badly calibrated. This online histogram shows the luminance histogram and the red, green and blue channels of any image, with statistics and plain-language notes about clipping and tonal range.
Drop in a photo straight from your camera or phone, a scan or an edited image. It is analysed in your browser and never uploaded.
How to read a histogram
The horizontal axis runs from 0 (black) to 255 (white). The height of the graph at any point shows how many pixels have that brightness. The three zones marked under the graph help you read it:
| Zone | Levels | Contains |
|---|---|---|
| Shadows | 0–85 | Dark areas: deep shade, hair, night skies |
| Midtones | 86–170 | Most skin, foliage, mid-gray and typical subjects |
| Highlights | 171–255 | Bright areas: sky, white clothes, reflections, light sources |
- Bunched to the left: the image is dark. That is right for a night scene or a moody low-key portrait, but may mean underexposure otherwise.
- Bunched to the right: the image is bright. Fine for snow, beaches and high-key shots; otherwise it may be overexposed.
- A narrow hump in the middle: a narrow tonal range, so the image may look flat or hazy. Adding contrast spreads it out.
- A tall spike at either edge: clipping. Detail has been lost to pure black or pure white.
There is no single "correct" shape. What matters is that the histogram matches the scene and that important detail is not clipped.
Clipping warnings
The statistics include the percentage of pixels in the darkest levels (0–2) and the brightest levels (253–255). When more than 2% of the image is clipped, you get a clear warning, because detail in those areas cannot be recovered by editing. Small amounts of clipping in light sources, specular reflections or deep blacks are normal; clipped skies, faces or white dresses are usually not.
Luminance vs RGB histograms
The luminance view shows perceived brightness, using the Rec. 709 weights for sRGB (0.2126 R + 0.7152 G + 0.0722 B), so green counts most and blue least, just as in human vision. The RGB view overlays the three color channels, and you can also view red, green or blue alone. Channel views reveal problems the luminance histogram hides: a bright red flower can clip the red channel while the overall brightness looks fine, losing texture in the petals. Shifts between the channels can also point to a color cast.
Turn on log scale to make small amounts of pixels visible. It is useful for spotting a thin line of clipped highlights that would otherwise be too short to see.
Statistics explained
- Mean and median: the average and middle brightness. Around 110–140 is typical for an evenly exposed scene.
- Standard deviation: how spread out the tones are. Low values mean low contrast.
- Darkest / brightest: the actual tonal range used. A range much narrower than 0–255 often means a flat image.
- Pixels counted: the number of pixels analysed. Mostly transparent pixels are left out so they do not skew the graph.
Fixing what the histogram shows
For quick corrections, use Image Filters to adjust brightness and contrast and watch the effect. For future shots, use exposure compensation on your camera or phone: if highlights are clipping, expose a little darker; if the histogram is crammed to the left with no clipping on the right, you can usually expose brighter. Many photographers "expose to the right" and darken later, because shadows recover with more noise than midtones.
Check the camera settings that produced a photo with the Image Metadata Viewer, or find its overall color with Average Color of an Image.
Image Histogram: frequently asked questions
What does a good histogram look like?
There is no single right shape. A low-key night photo will lean left and a bright beach photo will lean right. What matters is avoiding unintended clipping at the edges.
How is luminance calculated?
With the Rec. 709 weights used for sRGB: 0.2126 R + 0.7152 G + 0.0722 B.
What does clipping mean in a histogram?
Pixels piled up at pure black (0) or pure white (255). Detail in those areas is lost and cannot be recovered by editing.
How do I know if my photo is overexposed?
Look for a histogram bunched to the right with a spike at the right edge. The tool warns you when more than 2% of pixels are clipped to white.
Why use the RGB histogram?
A single color channel can clip while overall brightness looks fine, for example in red flowers or blue skies. The channel views reveal this and help spot color casts.