The main purpose of an image reproduction system is to deliver picture that are judged acceptable - even pleasurable! - by human viewers. It has been known for many, many decades that this goal is not achieved by correct physical reproduction of the intensities in the original scene. Rather, contrast must be raised somewhat, and saturation must be increased somewhat. If an image reproduction system does not do this, its images will be judged by viewers as being wrong.
"Gamma" was introduced to film long before image digitization or computers were invented. Read chapter 1 of any photography textbook to learn about tone scale and gamma. Gamma, and the S-shaped response of film, is not an intrinsic aspect of the physics or chemistry - these characteristics are put there by the film designers, to compensate the subjective effects! (Video systems need the same sort of end-to-end alteration of mathematical correctness, to achieve subjectively acceptable images.)
"Gamma" is now more complicated, because the term was introduced in the video domain, and is used to describe the power law exponent of a CRT. In computing, "Gamma" describes the imposition of a power law at image scanning or encoding time. "Gamma" is even more complicated still in digital photography, because the CRT and computer aspects of the term apply, but the original film intent of "gamma" must be considered as well, because this is photography! Because "gamma" is so complicated and technical, the Gamma FAQ is complicated and technical.
Timo believes that image reproduction should be linear, that is, intensity in, intensity out.
See also: Gamma FAQ - Linear and nonlinear coding
Charles Poynton
Copyright © 1998-07-27