Reading Digital Histograms

understanding digital histograms

Whether you are shooting with a digital camera or are editing an image on the computer, an understanding of digital histograms is so useful as to be essentially mandatory. Histograms give you essential feedback on the overall exposure and the contrast range of an image.

When shooting digitally, checking the histogram of what you just shot tells you if you nailed the exposure; if the exposure is poor, then you have a chance to re-shoot after correcting the exposure.

This is something you can’t do with the film. With film, you have to wait to get your prints or slides back, and by then it’s too late to re-shoot the scene. Score one for digital cameras! (Not all digital cameras have a histogram feature, a feature I strongly recommend you require when shopping for a new camera)

When editing an image on the computer or when setting up a film scan, examining the image histogram is vital for getting optimal results.The general idea of histograms:

So . . . just what is this terrific histogram thing? In general, a histogram is a chart of the frequency of the occurrence of something. In photography, a histogram is a chart of the frequency of pixel color values in your image. If this sounds confusing, read on for specific examples to help clarify things.

To help explain what a histogram is, I’ve made up an example of a histogram of the salaries of the hypothetical Mega-Mart corporation:


The “something” that is being plotted in this case is the salary for every employee in the company. The more employees that have a particular salary, the higher the bar is for that salary. Where the histogram is high, there are a lot of employees with the salary indicated at the bottom of the chart; where the histogram is low there are few employees with the indicated salary.

This histogram is a precise and compact way of understanding the company’s salary structure. You can see at a glance that the peons are paid in the vicinity of $40,000, the middle managers get around $90,000, and the head honchos get around $180,000.

Note that the top executives are paid over $200,000 and appear in an “overflow channel” ­ all entries that would appear in channels over $200,000 if they had been plotted, are added up, and bunched into this one channel.

Photo histograms and exposure:

Now let’s see how histograms are used in digital imaging. The first plot below shows a histogram of tonal values in an image that has a full range of colors from black to white. Pure black is at the left end of the histogram, pure white is at the right.

The second plot represents an image of a scene that is rather dark in tone so that most of the pixels’ values are on the dark end of the plot.

The third plot is for an image of a scene that is rather light in tone, so most pixels fall into the light end of the plot. Each of these histograms represents good image exposure for the corresponding scene.

histogram-averageimage histogram of a scene with a full tonal range
histogram-darkimage histogram of dark scenehistogram-lightimage histogram of light scene


Now let’s see how underexposure and overexposure look in a histogram. The key is to note whether the histogram seems to be “piling up” at one side or the other. If you can check the overflow channel (rightmost) and underflow channel (leftmost), so much the better: if these channels contain any significant amount of pixels, you have over or underexposure.

In the first histogram below, the image is underexposed, with lots of dark tones collapsed into pure black. The second shows an overexposed image, with lots of light tones pushed into pure white. If the histogram is piling up at the dark end, if it looks like it wants to continue off the left edge of the plot, you have an underexposed image.

Similarly, if the histogram is piling up at the light end and looks like it wants to continue off the right edge of the plot, you have an overexposed image.

histogram-underhistogram of underexposed imagehistogram-overhistogram of the overexposed image


The same principle applies when scanning film into a digital computer file. If the histogram of your scan shows that you underexposed or overexposed in the scanning, you can re-scan to get it right. In fact, most scanning software lets you set the dark and bright limits of the output histogram before the scan is performed, so you are guaranteed to get a suitable scan the first time.

Histograms are also vital when manipulating a digital image on the computer. The histogram feature of your image editing program will reveal if the print you are creating will be too dark or too light, and you can adjust the image accordingly.

Photo histograms and contrast:

So far I’ve only spoken of image histograms in terms of how light or dark the overall image is, i.e. overall exposure. They are even more useful than that, for they also give you information on the overall contrast of an image. The overall contrast is the range of tones in the image.

An image with a wide tonal range – very dark darks and very bright lights — has high contrast. An image with a limited tonal range — the darks aren’t much darker than the lights — is a low contrast picture. Depending on the subject, sometimes high contrast looks better, and sometimes low contrast looks better.

The more spread out a histogram is – that is, the farther apart the dark end of pixel values is from the light end — the higher the contrast of the image. When shooting with a digital camera, you are pretty much stuck with the tonal range of the scene you are photographing.

But if you are scanning film or editing images on the computer, you get to adjust the histogram in the software to achieve the overall contrast you want for that image. As an example, below are two versions of the same image along with their histograms. One has low contrast and one has a high contrast.

The histograms reveal a broader range of tones for the higher contrast image. In this case, the high contrast image looks better. It has more snap, crackle, and pop.

high-contrast high contrast imagelow-contrastlower contrast image
histogram-high-contrasthistogram of the high contrast imagehistogram-low-contrasthistogram of the lower contrast image