Yes, the books might have been written on this but not being a technical person, I'd like to understand in short and easy to understand language - How to interpret a histogram to know whether the photo is under or over exposed?

  • \$\begingroup\$ Look at your picture instead of the histogram! A picture is over or under exposed if you can't see details in dark or light parts where you want to see them. \$\endgroup\$ Oct 21, 2014 at 20:27

3 Answers 3


Frankly, you can't. If you look at the histogram of this image contre-jour it will look wrong. But the image is neither over-, nor underexposed.

  • \$\begingroup\$ thanks about the pretty image but I request you to consider answering with reference to a general case. \$\endgroup\$ Oct 21, 2014 at 13:51
  • 3
    \$\begingroup\$ General case in art is something that escapes me ;) Seriously, the histogram you refer to is of the brightness of the final image, not of the exposure of the latent (raw) image. Another thing to consider is that exposure is an artistic tool. Yet one more is that in certain situations only the brightness of the subject matters, and the histogram needs to be of the selection. Each subject has its own brightness range, optimally, so, again, no general case. \$\endgroup\$
    – Iliah Borg
    Oct 21, 2014 at 14:02
  • 1
    \$\begingroup\$ Technically this image is over- AND underexposed, although you could argue that it was exposed for the wall. Usually though, under- and overexposure are meant in the sense of clipping blacks and whites, which is clearly the case here. It is not badly exposed, just over- and underexposed. :) \$\endgroup\$
    – J0hj0h
    Oct 21, 2014 at 19:44
  • \$\begingroup\$ Why do you say the histogram will look wrong? If the photographer understands what the histogram is telling her, and if this is the kind of image she's trying to create, the histogram will look (and will be) correct in the context of the photog's goal. \$\endgroup\$
    – Caleb
    Oct 21, 2014 at 20:36
  • \$\begingroup\$ Histogram does not show where on the image the certain level occurs. Histogram is a one-dimensional report, hiding the tonal distribution over the whole image. How much the histogram needs to climb the left and right wall is also unclear, spotmeter is a better tool because it allows to control the result instead of reshooting. For a rendered image, eyedropper is a great aid. \$\endgroup\$
    – Iliah Borg
    Oct 21, 2014 at 20:54

If the histogram is shifted very much to the right, the photo may be overexposed, and if it is shifted very much to the left, the photo may be underexposed. But as the comments and other answer have pointed out, it's not a solid rule.

If the histogram is cut off, like the first image, it means there are blown out highlights (ie unrecoverable detail in very bright things, like clouds (but again, as shown in the image in the other answer, blown-out highlights or blacks are sometimes desirable eg B&W photography)).

(Possibly) overexposed:


(Probably) underexposed:

enter image description here

  • \$\begingroup\$ you didn't show what is "balanced"? \$\endgroup\$ Oct 21, 2014 at 13:25
  • \$\begingroup\$ I think you can perfectly guess what "balanced" is \$\endgroup\$
    – Rodrigo
    Oct 21, 2014 at 13:35
  • 2
    \$\begingroup\$ I don't think it's possible for a histogram to definitively say whether a photo is over- or underexposed. It can certainly accurately show the relative distribution of tones, but it's entirely possible for a photograph to contain a lot of blown highlights or blacks yet still be correctly exposed (product photography, for example). On its own it is simply a tool, and without the context of the accompanying photograph it is only telling a part of the story. \$\endgroup\$
    – user456
    Oct 21, 2014 at 13:39

Image sensors and image files only have a limited storage space and dynamic range. If the dynamic range of the scene is bigger than the one the camera can capture and store, details will get lost. The technical term for this loss is clipping.

Since the histogram shows the amount of blacks to its far left and whites to its far right, a big spike to those ends is typically avoided. It suggests that the scene had many details in those areas, but the image simply cuts out (or clips) those details and shows a solid color (black or white) instead. A technically well exposed image would have all its colors inside the bounds of the histogram, meaning that no detail had to be left out.

As Iliah Borg pointed out, over- and/or underexposing does not necessarily mean having a bad exposure. In the picture he provided this clearly is intentional and pleasing.
Nevertheless, that case is an exception. Most of the time, you want to retain as much detail as possible, to be able to play with or pronounce later on. If you really want to keep as much details as absolutely (technically) possible, you can take a look at exposing to the right.

Do not get caught up too much in the mechanics, though. Iliah Borg made it very clear that having a pleasing overall image is more important than having a technically well exposed image. ;)


Not the answer you're looking for? Browse other questions tagged or ask your own question.