*** From the Archives ***

This article is from May 9, 2011, and is no longer current.

Histograms Are the Best Photography Tool You Already Own: Part 1

Digital photography has many advantages over film, but perhaps the most significant is the in-camera histogram. This simple tool lets you immediately spot potential exposure problems as you shoot. Once you’re done shooting, the histogram in your image-editing software shows you what’s wrong with an image and how to adjust tone and color.
The histogram scares a lot of people because it looks like it has something to do with math. While there are a lot of numbers behind it, the histogram is actually easy to understand.
Why Should I Bother?
The LCD screen on the back of your camera is good for only one thing: judging composition. It’s too small to accurately report focus—even if you zoom in—and its brightness and saturation are always cranked up so you can see the image on the LCD in bright light. If you rely on a camera’s LCD screen to determine exposure, you’ll often get it wrong.
When you know how to use your camera’s histogram, you can immediately tell if an image is over-exposed, if the shadows have gone to complete black, if you have a contrast problem, and whether you’ve captured the maximum possible amount of image data. Capturing enough data will ensure you can perform all of the edits that you want without seeing ugly artifacts.
Histogram Basics
The histogram is simply a bar chart that graphs the distribution of tones in your image, from black on the left to white on the right. The graphy-looking thing in the upper-right corner of Figure 1 is an example of a histogram from a Canon camera.
Figure 1. A typical in-camera histogram display.

Figure 2 is a histogram generated by Photoshop.
Figure 2. Photoshop can also display histograms, which are an essential part of the image editing process.

The histogram data graphs black on the left, and white on the right. From the histogram in Figure 2, you can see that there’s no actual black in the image because the data doesn’t go all the way to the left side (A). There’s a tremendous amount of dark gray, and then a decreasing amount of light gray, until it hits white on the right.
Notice, too, that first spike toward the left end of the histogram (B). It represents all the dark gray tones that make up the shadows in the rocks, and the dark bits of the flamingo’s leg.
The really huge hump (C) covers the area around 65 to 70% gray. This is the tone that comprises the bulk of the image: the baby flamingo itself, the dull shade of the ground, the dull gray of the rocks. This image is predominantly a single tone, and so that particular part of the histogram is piled high with data.
Let’s try a quick thought experiment: What happens to the histogram if you flip the image horizontally (Figure 3)?
Figure 3. Do you think the histogram for this image looks different now that it’s flipped?

If you answered “Nothing,” you’re exactly right. There’s no correspondence between the histogram and any geographic part of the image. It’s simply a graph of how much of each tone there is in the picture. Flipping the image doesn’t change any tonal quantities, just their locations within the image.
Figure 4 shows an image with an underexposure problem.
Figure 4. This image is underexposed.

That big spike on the far left of the histogram indicates that the image is underexposed, and that some pixels have gone to complete black.
Photographers usually don’t worry about underexposure because a shadow that’s gone to complete black simply looks like a very dark shadow. But if you’re hoping to preserve details in the shadowy parts of your scene, you may not want underexposure. A quick check of the left side of the histogram of your image will tell you if you need to brighten your exposure.
Overexposed highlights, on the other hand, are almost always a problem. When a highlight overexposes, it loses all detail and turns completely white. Figure 5 is an image that’s been overexposed.
Figure 5. The highlights in this image are overexposed.

When you’re shooting in a situation that you’re afraid might overexpose, check the histogram after you shoot. Is there a big spike on the right side? If so, you’ll need to underexpose.
You can also use the histogram to determine if you’re capturing as much contrast as you want. When all of the tones are grouped in the center of the histogram, as in Figure 6, you can be sure that your image will lack contrast.
Figure 6. This rather blah image is rather blah because it lacks contrast, as revealed by the histogram.

I shot that image in the middle of the day, when the light was flat and boring, and so I suspected that I might have a contrast problem. I dialed in some over-exposure, shot again, and saw this histogram:
Figure 7. A simple overexposure gives the image more contrast and therefore more punch.

The histogram in Figure 7 reveals a much greater distance between the darkest and lightest tones, indicating more contrast. This ability to determine how much contrast you’ve captured is especially useful if you’re planning on converting the image to black and white. Because you tend to edit black and white images a lot, you want as much data as you can get, so that you can push your edits farther before seeing visible artifacts in your image.
It’s important to note that there’s no correct shape for a histogram. You don’t aim for a bell curve, or a camel shape, or something that looks like topiary.
Your Camera’s Histogram
Most digital cameras can display a histogram of any image you’ve shot. Typically, you play back the image on your camera, and then change the display to show a histogram alongside an image thumbnail. When you display the histogram, many cameras will flash any pixels in the thumbnail that are overexposed.
Your camera might also display a three-channel histogram like the one shown in Figure 8. All colors in an image are made up of a combination of red, green, and blue channels (the primary colors of light). A three-channel histogram shows a separate histogram for each of these color components. If one component is heavily out of registration with the others, then the odds are good that your camera has a white balance issue or other color cast.
Figure 8. A typical three-channel histogram display on a camera.

Some cameras can also display a live histogram that appears when you’re shooting and updates in real-time (Figure 9). I don’t find these particularly useful, as they clutter the screen and are in constant motion.
Figure 9. A typical live histogram, which updates in real time as you frame your shot.

Applied Histograms
Now that you know how to interpret a histogram, you’re ready to use that knowledge to fix the problems it shows. In the next part of this series, I’ll look at how to use your camera’s exposure compensation control, in conjunction with the histogram, to get better exposures.
 

  • Anonymous says:

    Very quick and clear explanation.

  • Anonymous says:

    Nice explanation. Thanks!

    Jason

  • Anonymous says:

    looking forward to part 2 !

  • Anonymous says:

    I look forward to the next installment – Great stuff – never did understand what I was doing when i did it.

  • smlphoto says:

    Great info, well written… I want to read more of yours… looking forward to part 2…

  • mslori says:

    I am embarrassed to say I have not noticed the histogram on my camera, so I will look for it. Although I have seen and used it in Photoshop before.

  • Anonymous says:

    By “ove-rexposure,” I presume the author means he increased the exposure relative to what the meter considered normal. He means that he added exposure to the meter’s recommendation to produce the effect he intended. In the old days, of course, there were no meters and an over-exposed image was one lighter than intended. In those days, over-exposure was a mistake. Now that “over-exposure” has been redefined to mean “lighter,” what word or phrase do we use for an image that is mistakenly too light?

  • Anonymous says:

    Now that “over-exposure” has been redefined to mean “lighter,” what word or phrase do we use for an image that is mistakenly too light?

    CRAP!!!

  • >