Demystifying Histograms
"What does this little chart thingie on the back of my camera, and in Photoshop?
What does it mean and how do I use it?"
Histograms are arguably the most important visual clue that you can get from your camera as to proper exposure and what the image will actually look like when printed.
Most digital cameras from, advanced professional to simple point-and-shoot have the ability to display a histogram with the image preview on the back of the camera. The histogram is a graphic representation of the tonal information that makes up the image, a distribution chart of all the light values from white (on the right) to the black (on the left).
The histogram is the light meter of the digital age. It will let you know if your image is overexposed (the information in the diagram will be substantially stacked up on the right side), or underexposed (information stacked against the left side). Trying to judge from the image that shows up on the tiny LCD screen is not a good way to check exposure. If it were the manufacturer would not give options for varying the brightness of the display.
sample images and their histograms
As you can see, the image on the left has a histogram that indicates a strong presence of mid tones and falls off gradually towards the light and dark tones. There is a tiny peak on the highlight end as a result of the reflection of the window sill in the top center. The histogram of the image on the right shows strong shadows and highlights with less information in the midtone range. The shadow and highlight ends of the diagram fall just short of peaking, known as clipping, to absolute black or white.
There are times, when we can try our hardest, but cannot get a perfect exposure. Keep in mind that there are no good or bad histograms, there just are. They exist as a visual graphic representation of the tonal structure of the image.
The image on the left is dull, flat and uninteresting, and the histogram reflects that.
There are several methods to improve the image, I am only displaying one option. That is to open the levels dialogue box (control/command - L) and move the right and left sliders until them are sitting right under the edges of the graphic mountain in the dialogue box. What you would end up with is the image on the right.
This is a much more pleasing image and would make a great print as long as it was no larger than 5 x 7.
Why, you ask? The histogram again is the key to the problem. Remember, it is a graphical representation of the structure of the image, and this one is full of holes, or lines, whatever, there is a lot of missing information indicated by the white lines running vertically through the graph. Each white line is indicative of a tonal level that does not exist. This can result in an image that has a lot of noise, banding or is pixilated or posterized when enlarged. The way to fix this problem is to convert your image to 16bit prior to the levels adjustment, then convert it back to 8bit afterwards. The histogram for the same image looks like this:
Much nicer and structurally friendly. The process sounds more complicated than it is. A simple action to convert, open and tweek the levels and then reconvert the files makes it all a one button process. In the actions area it is called, levels adj.
It is important when processing images and subsequently sending them to the lab for printing, that each image submitted is processed and optimized for color and levels as part of your workflow. Of course, if your really do incorporate image optimization into your workflow, it will give us a lot less material with which to quietly ridicule and make sarcastic comments, but on the upside, it will also make printing easier and more efficient at the lab, save you time and energy on any future file transfer and printing.