I've started a course on image processing for beginners, and am trying to understand Gamma and Gamma Correction. Unfortunately, I seem to have come across conflicting explanations across multiple websites. I'm not an artist or computer scientist (just a very humble beginner) but I wanted to pitch what I understand so far so you could correct what I find conusing?

So far, I seem to be finding two different explanations, that go kinda like this:

Fist Explanation: Compensating for the Eye

The human eye does not interpret changes in light intensity accurately; if I double the intensity of light outputted by my sources in a seen, by definition there is double the intensity of light in the scene. When though you have a very dark room and increase the true light intensity say 10%, our eyes & brain interpret a much higher percentage increase. And correspondingly, in bright situations, if you increase the true light intensity by Y%, our eyes/brain interpret a much lower increase. Thus we are more sensitive to darker scenarios, and less so to brighter ones. The curve for the non-linear way our eyes perceive different light intensities is below.

enter image description here

Cameras however capture the 'true' intensities of light; if you double the light in a scene, the camera detects and records 'double' the light on its sensor (too bright a scenario though forces the camera to normalise its data so the max intensity is 255 though).

So, from what I understand from half the explanations, this is where Gamma comes in: light rays in a scene hit the camera each with a 'true' intensity, call it x. Cameras modify the raw data they capture by raising the value for the intensity x of each pixel to the power of 1/2.2, so that the values of intensity for each pixel that are passed onto your computer now have the intensities a human eye would detect in your scene.

Monitors then apply an inverse function to these intensities, raising all the pixels' intensities to the power of 2.2, so thus restoring the intensities to their original, 'true' x value, as they actually were in the scene when they hit the camera.

These 'true' intensities then are what is displayed by the monitor, and when those emitted light rays hit our eyes, our eyes/brain apply their non-linear interpretation of intensities, so the image we look at on the screen looks as if we were at the scene instead of the camera.

Now - this is the interpretation as I've managed to understand it from half the sources I've found. But here's what I find confusing; if the intent of Gamma correction on a monitor is to modify pixel intensities back to what a camera would detect, why does the camera apply its 1/2.2 gamma correction in the first place? Why not let the camera just capture the true raw intensities and display them on the monitor so that our eyes would interpret the scene as if we were actually there?

Which brings me to the 2nd set of explanations I've found:

My above suggestion of letting cameras just capture the raw, true intensities of a scene's colours, displaying those straight to a monitor with no gamma modification, so that these 'true' scene intensities reach and are interpreted by our eyes (as if they were actually in the scene) - is what camera/screen makers would prefer to do. But, they don't simply because of legacy reasons: whereby digital cameras had to apply a gamma function after capturing an image because older technology did it that way for some different reason, and so computer screens have to apply an inverse 'Gamma Correction' to the data they receive, so that it reaches the eye with 'true intensities'.

To me, the 2nd explanation seems a bit strange - i.e. that the industry would do all this gamma stuff just to comply with old legacy tech standards - but I can't seem to get past the problem with explanation 1? So I'm unsure which is correct.

The other very likely circumstance however is that I've totally misunderstood the ideas behind Gamma. If so, it would be really helpful if anyone could point out where I went wrong/lay out what I theorize is actually a deceptively simple subject?

Thanks very much for your help.


The way my professor explained it is that gamma correction is a transformation that was mainly used for old computer screens because the old computer screens had a transfer function and did not output the correct colors. Gamma correction was so that the output portrays the true image. You can also see the input output map of a gamma transform. It shifts all the brighter frequencies to 1 (very bright) and all the darker ones to 0 (very dark) depending on the gamma (gamma has a different effect if it is more than 1 and less than one). That is what i know about it. Good luck with your course! I would recommend Digital Image Processing Book by Rafael C. Gonzalez and Richard Eugene Woods as a reference for with your course.


Your first interpretation is wrong, and second one is partial and for a better understanding at least please read the following related question Why does gamma correction of image use two different definitions of gamma?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.