Gamma in the popular press


A major contributor to confusion about nonlinear coding, and gamma correction, is misinformation and lack of information in the academic literature and in the "popular" technical literature. Another major contributor is careless use of terminology by video engineers who are unfamiliar with the principles of color science, and consequent publication in video textbooks of ambiguous information.

In the domain of computer graphics, consider Computer Graphics: Principles and Practice, Second Edition, by James D. Foley et al. (Reading, Mass.: Addison-Wesley, 1990). 300,000 copies of various editions and adaptations of that book have been printed. Page 589 makes this unequivocal statement (with the emphasis in the original):

The YIQ [color] model is used in U.S. commercial television broadcasting ... The component of YIQ is not yellow but luminance, and is defined to be the same as the CIE Y primary.

This is flat-out, 100% wrong. The video Y' component - often carelessly called "luminance", and properly called luma - is computed as a weighted sum of nonlinear (gamma-corrected) R'G'B' components. True, CIE luminance is expressed as a weighted sum of linear (not gamma-corrected) RGB components.Foley et al. fail to understand that video is nonlinearly coded.

A definitive reference to the image processing field is William K. Pratt's Digital Image Processing, Second edition (New York: Wiley, 1991). The only entry in the index for gamma concerns film. There is no discussion of gamma in video, the power function of a CRT, or nonlinear coding. There is an encyclopædic collection of 3 x 3 color transform matrices on facing pages 66 and 67, but there is no reference to which matrices are intended to apply to intensity data, and which are intended to apply to nonlinear (gamma-corrected) data.

Gregory A. Baxes does marginally better in Digital Image Processing: Principles and Applications (New York: Wiley, 1994). At least he mentions the nonlinearity of a CRT. But he also fails to discuss nonlinear coding.

If you can't trust information from the definitive sources, could the popular press be any better? Let's go to Borders, and start in the Computer Graphics section at the letter A.

Consider Adobe Web Design and Publishing Unleashed, Dave Brown, Aandi Inston, Curt Akin, et al., (Sams.net Publishing, 1997). The back cover copy advertises that "Everything users need to know about the latest design, publishing techniques, and authoring with Adobe's entire product family is included in this comprehensive tome!" Three index entries relate to gamma: "Gamma control panel (Photoshop)"; "gamma level, movies"; and "gamma rays (Premiere)". The first referenced passage comprises a short paragraph that concludes:

Through several tests and disasters of perfectly-planned graphics, I have found that graphics viewed on a PC monitor tend to be darker than on a Mac monitor.

The second passage isn't topical. The third passage doesn't actually discuss gamma rays, that was just a cutesy section title. It says,

A nicely lit movie on a Macintosh will play rather darkly on a Windows machine. A gamma setting of 0.7 or 0.8 is usually adequate to compensate for cross-platform gamma differences.

"Unleashed" indeed - more like running rough-shod. This book has 1086 pages, and it's branded "Adobe." I don't want "tests and disasters," I want facts! Where's the science?

How about Digital Camera Companion, by Ben Sawyer and Ron Pronk (Scottsdale, Ariz.: Coriolis Group, 1997). According to the back cover copy, this contains "everything you need to know about buying a digital camera and publishing your photos on the web." Sorry: There's no index entry for gamma. No mention of gamma in film, no mention of Mac vs. PC, no mention of cross-platform issues.

Consider Digital Cinematography by Ben de Leeuw (Boston: Academic Press, 1997). Back cover copy commences, "With the advent of movies like Toy Story ... ." Mmm, this is promising. Sounds high-end. Cinematography has to achieve good tone reproduction; let's see if gamma is mentioned. Sorry, nope. There are only two index entries under G: games, and gunfire. (The book turns out not to be about cinematography at all, it's about computer animation.)

There's the inevitable Digital Photography for Dummies, by Julie Adair (Foster City, Calif.: IDG Books, 1997). I'll bite my tongue here, and say just that there's no index entry for gamma, and I halted the experiment at "D."


See also: Gamma FAQ - Linear and nonlinear coding

Charles Poynton
Copyright © 1998-03-12