On the Number of Discernible On the Number of Discernible Colors Colours
I was surprised that, in their study of the number of dis-The authors are grateful to Cal McCamy for his timely cernible colors, Pointer and Attridge 1 missed the early response to their original article. Neither they, nor the work and the classic derivation by Nickerson and others they contacted, were aware of the relevance of Newhall. In 1896, Titchener estimated about 33,000 colthe contributions cited by Mr. McCamy. That the lower ors. In 1939, Boring upped that to 300,000. In the same number derived by Nickerson and Newhall is not too year, Judd took the quantum leap to 10,000,000. Nickerdissimilar from the authors' total of 2.28 million colours son and Newhall were intimately familiar with the is very encouraging. That a reference can be cited for the Munsell color-order system, in the ''hands on'' or, if you original quoted figure of 10 million is satisfying, even if please, ''eyes on'' sense. They derived a psychological it is still a little clouded in mystery. As I read the paper color solid. Knowing what they did about color judgby Judd and Kelly I find reference to ''10 million surface ments, they were able to conclude that, under usual condicolours distinguishable in daylight to the trained human tions, a normal observer could distinguish 1,875,000 coleye,'' but not a derivation of that number. ors and, under the best viewing conditions, 7,500,000 At a presentation of the work, given to the Colour colors. Based on their derivation, I consider the latter the Group (Great Britain) in October 1997, David Chambertotal number of colors distinguishable by the human vilin remarked that R. K. Schofield, an early collaborator sual system. They made it clear that it all depends on with The Tintometer Company, manufacturers of subwhat is meant by the number. These estimates are very tractive visual colorimeters, had always taken a figure of sensitive to the limen chosen, because it is cubed in the approximately 220 discernible steps in the Lovibond Red computation. Their lower estimate is remarkably close Scale, 220 in the Yellow Scale, and 185 in the Blue Scale. to the estimate by Pointer and Attridge, considering the These figures combine to give 8,954,000 discernible steps different derivations. It relates visual limits in the Munsell in the Lovibond colour solid. Formal reference to this is system to the significance of the unit of CIELAB color anecdotal but an early item of Company literature, written difference in industrial color applications.
by G. J. Chamberlin, contains the statement: When the Society for Imaging Science and Technology named me Guest Lecturer for 1985, and asked me to ''Here is a picture of the glasses you have seen made speak to the various chapters, I organized an after-dinner in Salisbury, which can be used singly or in pairs, or slide show billed as ''The Joy of Color.'' It was also the three glasses together to make nine million different first Franc Grum Memorial Lecture, sponsored by the colours and with which the colours of any object can Council for Optical Radiation Measurements, in 1986.
be measured and the results described in figures.'' My opening gambit was skylines of several large U. S. cities, with total population of 10 million. I asked the That this number is considerably higher than the 2.28 audience to imagine all those people, showed people million quoted above is probably due to the use of a wearing colored ribbons, and said all 10 million could dark surround in the Tintometer visual colour measuring don perceptibly different colors. Today I might leave out instruments. This will serve to increase discriminability, Dallas and Denver, but it would still be an eye-opener.
especially for dark, saturated colours. The figure is, in 1. M. R. Pointer and G. G. Attridge, The number of discernible colours, fact, closer to that quoted by McCamy from Nickerson