How to Interpret Histograms

How to Interpret Histograms.
Much has been written about how useful it is to see the Histogram of your image. However not every one understands exactly what he or she is seeing.

The Histograms horizontal axis ranges from dark valued pixels (O=black) at the left to the bright valued pixels (255=white) at the right. The horizontal axis shows the luminance channel and does not tell you anything about the individual Red, Blue and Green channels.

The vertical axis is the number of pixels in the image with a particular luminance value. This axis scales with the data, and is not particularly important (other than to identify what’s happening with a particular tone vis-à-vis others).

So what does a well-exposed image look like? It’s actually easier to define what constitutes a poorly exposed image.

Here are some things to look for:

• 
  Most pixels skewed to the right of the histogram. If a significant number of pixel values exist at the very right, it’s likely the shot is overexposed. Histograms that are “right heavy” make it difficult to control highlight detail. Check out the Highlights display to see if you have blown out any highlight detail.
  
• 
  Most pixels skewed to the left of the Histogram. If a significant number of pixel values exist at the very left edge, it’s likely the shot is underexposed. Histograms that are “left heavy” tend to have troublesome shadow details. If there is little or no exposure shown to the right of the Histogram, you need to add more exposure to the shot. Note that underexposed images are easier to recover detail from than overexposed images.
  
• 
  Pixels are evenly scatted over the entire width of the Histogram. The overall image is likely to be high in overall contrast. While a broadly scattered pattern in the Histogram is OK, you might not be satisfied with colour saturation or contrast of the final image.
  
• 
  Pixels are in a very narrow band in the Histogram. The image is very likely low in contrast (or it could be monochromatic).
  
• 
  Any spike at the right edge means lost highlight detail. This is probably the worst thing you can see in a Histogram. The higher the line crawls up the right edge of the Histogram frame, the more blown-out pixels you have in your image. What makes this bad is that your eyes immediately go to the brightest area of a photo when we view it, and all those pixels stacked up at the right of the Histogram will eventually print as paper.
  
• 
  Any spike at the left edge means lost shadow detail. Or it could simply mean you have some totally black areas in your shot. Our eyes aren’t bothered as much by dark areas in a picture (unless, I suppose, that is your subject).
  

In general, you are looking for a moderately wide distribution of the pixel values, with the largest peaks for the important portions of your scene somewhere in the middle three-quarters of the range. If you are working in a scene that has many bright values (e.g., snow), the peaks may be to the right of the Histogram. Likewise, if you are working in a scene that contains many dark values (e.g., unlit, shadow areas), the peaks may be to the left of the Histogram. Either case is usually OK, as long as you have a wide distribution of pixels and neither extreme run off the edge of the Histogram.

Note. Most users find it easier to “fix” dark images (e.g., increase shadow detail) than to fix bright images (e.g., “pull back” highlight detail).