How 'prime colors' determine sight

Seeing light and color is one of mankind's more transcendent capabilities. It is of course one of the principal avenues by which we take in beauty as well as vast quantities of useful information. But when the sun goes down we have a problem.

Cavemen had it, and spacemen have it: how to illuminate human activities without wasting energy.

Man has his own peculiar requirements for seeing. A butterfly and a frog surely differ in their views of the world. The butterfly may see well in the violet light of dawn, and the frog hardly at all. The frog must see well in the dim light of his underwater haunts, while the butterfly is high above him in the brilliant sunshine. Each creature responds differently to the various colors of light. Each has a visual system that may respond more efficiently to blue than to red light, or to green than to yellow.

Man likes to light up his surroundings after dark. What kind of light should he use, if he must see well for the least expenditure of electrical energy? Efficiency of illumination requires a knowledge of the spectral response of human seeing. How does sunlight feed into this spectral response? Sunlight is a mixture of about a 150 distinguishable, brilliant colored lights we call the spectral colors, each of a different wavelength, together covering the visible spectrum, from deep violet through blue, green, yellow, orange, and red, to deep red. Human seeing must respond best to one -- or at most a few -- of these brilliant colored lights. If we knew which few, we might mix them to form white light of maximum visual efficiency. Since we have to generate white light from energy resources at hand, it helps if that white light is used with maximum efficiency by those whose activities it illuminates.

Scientist George Palmer, Thomas Young, and James Clerk Maxwell, representing the period from 1780 to 1850, sensed that there is a unique set of three very specific colored lights that relate closely to human vision. In recent years, this particular set of three has been called the "prime colors," to distinguish them from the term "primary colors," of which there are many useful but nonunique sets of three. It was recognized that there is an innate three-ness about human color vision, and that this implies three sensors, three spectral responses, three colored lights that characterize human vision. Nevertheless, around 1900 we took a wrong turn.

It has long been of interest to compare the brightness of two lights, even if they are different colors. With Edison and others bringing electric light rapidly into common use, this need to measure comparative brightness became acute. Somehow, a method of presenting the two lights flickering alternately on the same area of a screen presented itself. Picture two slide projectors, one with a green filter and the other with a red filter, shining on the same area of the same screen. Now blink the projectors, several times a second, so that first the green beam and then the red beam falls on the screen. If the blinking rate is increased to perhaps 10 blinks per second or a little more, one can no longer see a succession of green and red blinks, but only a flickering yellowish light. Adjusting the intensity of the green or the red beam with a dimmer will minimize the apparent flicker. At this point, for reasons hard to understand today, the green light and the red light were taken to be "equally bright." When pairs of lights differing in color, and including all the color combinations in the spectrum, were adjusted to minimum flicker, it was found that yellow-green light required the least intensity, and all the other colors greater intensity. A series of questionable presumptions followed, and the dashed curve of the illustration became regarded as "visual response per watt." A being with such a response could not, however, see color. Because of the presumptions mentioned, he would probably not even agree with a human observer watching the flickering lights and judging brightness.

Because of the great need from that time to this for some mechanical means (i.e., a meter) for measuring brightness, the dashed curve gradually became the basis for the units of light intensity called the lumen and the foot-candle, and for the "light meter" to measure them. Because the curve, and the units it defines, gradually became accepted as characteristic of brightness as the human observer sees it, even the design of commercial lamps was for 50 years unduly influenced.

In the last 15 years we have been reawaking to the more reasonable, and actually inarguable, three-response viewpoint of Palmer, Young, and Maxwell. The illustration shows what some of us now believe to be a fairly accurate picture of the three independent spectral responses of the human visual system. They are narrower and more sharply defined than the traditional Subresponse. They center on three brilliant colored lights or spectral colors, which, by name and wavelength, are blue-violet near 450 nm, pure green near 540 nm, and orange-red near 610 nm (an "nm," or nanometer, is one-billionth of a meter). These are the long-sought "prime colors" of human vision.

What tipped us off to their identity? The first discovery has to do with the color rendering of white light, which can range from beautiful to disastrous. We found in 1966 that if white light is made from the three prime colors, it renders colors in an illuminated scene very much like sunlight, even though all but three of the spectral colors in sunlight are missing. It now is clear that the prime colors are the most important constituents of sunlight. This is simply because the visual system samples incoming lights at these wavelengths and needs little else present in the illumination (as indicated by the three peaks in the visual response graph in the illustration). Another curious discovery, still not understood, is that prime-color white light colors the things it illuminates in a most attractive way -- considerably more attractive than the coloration in sunlight. More understandably, we find that much less lighting power is required in prime-color illumination, whether that illumination is used for general seeing or for performing difficult office tasks; the lighting power is fed directly into the peak sensitivities of the visual system.

Artifical lighting is not the only field affected by the unfolding of the prime colors of vision. In color photography, film response now may be matched to that of the visual system and preferred coloration will result. Some of the complexities of color television may be bypassed. In color printing and dyeing, shaping reflectances to the prime colors will help hold coloration at its most colorful, whatever the illumination.

In lighting, it is a happy circumstance that tailoring the composition of white light to save energy brought with it the unexpected benefits of clarity and beautiful coloration.