Everyone has probably noticed from time to time that a weird term comes up on their computer monitor, or in a discussion with some film or broadcast professional called Gamma. Often there are these very complex and technical explanations of what it is, that has to do with mathematics. However, the easiest way to understand it, is to see it.

If look carefully at both images, which one looks like it is evenly going from black to white?

The one with gamma correction, right?

But, here’s the trick. the values in the gradient from the image without gamma correction is in fact going from 0 to 100% evenly. That’s right, the human eye sees 18% gray as the middle on the road, due to the way our eyes perceive light. So, as an enhanced feature for us humans, monitors and all sorts of devices come preset with the brightness of the screen adjusted with a curve that leaves the blacks and whites at the same place, but pushes the middle grays out to the center of the spectrum.

This curve is called a logarithmic curve in math. The gamma value that you usually see is 2.2 that is common in video and most monitors and operating systems. This refers to the logarthmic value of the curve.

So, we can push and pull the curve to decide where we want the exact “middle gray” to be.

Gamma on OS X and Final Cut Pro may not be the same as your television set, since they have put the gamma value at 1.8 (if you have a version of OSX prior to snow leopard). So, Final Cut Pro may be making adjustments on the fly for example. But, the standard for HDTV and sRGB is a gamma value of 2.2.

This is great for viewing images on a monitor, however, it can be problematic if you are doing picture manipulation. For example, if you take a middle gray value at 18%, and add it to another middle gray value of 18%, what should the gray value be? Well, if it is not gamma corrected, then 18% + 18% will give you 36%. However, if it is gamma corrected, then 18 + 18% will give you equivalent of 100% because 18% is in the middle and the middle + middle takes you right to the edge!

So, gamma correction is best when we’re viewing the picture, but not necessarily when we’re doing the math to adjust colors or compositing multiple pictures together over each other.