Copyright © 1999 Massachusetts Institute of Technology
Color vision is the ability to detect and analyze changes in the wavelength composition of light. As we admire a rainbow, we perceive different colors because the light varies in wavelength across the width of the bow.
An important goal of color science is to develop PSYCHOLOGICAL LAWS that allow prediction of color appearance from a physical description of stimuli. General laws have been elusive because the color appearance of a stimulus is strongly affected by the context in which it is seen. One well-known example is simultaneous color contrast. Color plates illustrating the color contrast can be found in most perception textbooks (e.g., Wandell 1995; Goldstein 1996; see also Evans 1948; Albers 1975; Wyszecki 1986).
In everyday use, we describe color appearance using simple color names, such as "red," "green," and "blue." There is good agreement across observers and cultures about the appropriate color name for most stimuli (see COLOR CATEGORIZATION). Technical and scientific use requires more precise terms. The purpose of a color order system is to connect physical descriptions of color stimuli with their appearance (see Derefeldt 1991; Wyszecki and Stiles 1982). Each color order system specifies a lexicon for color. Observers then scale (see PSYCHOPHYSICS) a large number of calibrated color stimuli using this lexicon, and from such data the relation between stimuli and names is determined. Examples of color order systems include the Munsell Book of Color and the Swedish Natural Color System (NCS). Note that color order systems do not address the problem of how context affects color appearance. Rather, each system specifies a particular configuration in which stimuli should be viewed if the system is to apply.
People are sensitive to wavelengths between 400 nano-meters and 700 nanometers (nm); hence this region is called the visible spectrum. The chromatic properties of light are specified by how much energy the light contains at every wavelength in the visible spectrum. This specification is called the light's spectral power distribution. Color vision is mediated by three classes of cone photoreceptor. Each class of cones is characterized by its spectral sensitivity, which specifies how strongly that class responds to light energy at different wavelengths; the three classes of cones have their peak sensitivities in different regions of the visible spectrum, roughly at long (L), middle (M), and short (S) wavelengths. How the cones encode information about the spectral power distribution of light is discussed in RETINA and COLOR, NEUROPHYSIOLOGY OF (see also Wandell 1995).
Physically different lights can produce identical responses in all three classes of cones. Such lights, called metamers, are indistinguishable to the visual system (see Wyszecki and Stiles 1982; Brainard 1995). It is possible to construct metamers by choosing three primary lights and allowing an observer to mix them in various proportions. If the primaries are well chosen, essentially any other light can be matched through their mixture. This fact is known as the trichromacy of human color vision (see Wyszecki and Stiles 1982; Wandell 1995; Kaiser and Boynton 1996). Trichromacy facilitates most color reproduction technologies. Color television, for example, produces colors by mixing light emitted by just three phosphors, and color printers use only a small number of inks (see Hunt 1987).
Some people are color blind, usually because they lack one or more classes of cone. Individuals missing one class of cone are called dichromats. Consider a pair of lights which produce the same responses in the M- and S-cones but a different response in the L-cones. A normal trichromat will have no difficulty distinguishing the lights because of their different L-cone response. To a dichromat with no L-cones, however, the two lights will appear identical. Thus dichromats confuse lights that trichromats distinguish. In the most common forms of dichromacy, either the L- or M-cones are missing. This is often called red-green color blindness because of consequent confusions between reds and greens. Red-green dichromacy occurs more frequently in males (about 2% of Caucasian males) than females (about 0.03% of Caucasian females). The rate in females is lower because the genes that code the L- and M-cone photopigments are on the X-chromosome. There are other forms of color blindness and anomalous color vision (see Pokorny et al. 1979).
Different species code color differently. Many mammals are dichromats with correspondingly less acute color vision than most humans. In addition, for two species with the same number of cones, color vision can differ because the cones of each species have different spectral sensitivities. Bees, for example, are trichromatic but have cones sensitive to ultraviolet light (Menzel and Backhaus 1991). Thus bees are sensitive to differences in spectral power distributions that humans cannot perceive, and vice versa. Note that the color categories perceived by humans are unlikely to match those of other species. Behavioral studies indicate that humans and pigeons group regions of the visible spectrum quite differently (see Jacobs 1981).
A key idea about postreceptoral color vision is the idea that signals from the separate cone classes are compared in color opponent channels, one signaling redness and greenness and a second signaling blueness and yellowness (Wandell 1995; Kaiser and Boynton 1996). An informal observation that supports the idea of opponency is that we rarely experience a single stimulus as being simultaneously red and green or simultaneously blue and yellow. Quantitative behavioral and physiological evidence also supports the idea of color opponency.
Why is color vision useful? With some notable exceptions (e.g. rainbows and signal lights), we rarely use color to describe the properties of lights per se. Rather, color vision informs us about objects in the environment. First, color helps us distinguish objects from clutter: ripe fruit is easy to find because of color contrast between the fruit and leaves. Second, color tells us about the properties of objects: we avoid eating green bananas because their color provides a cue to their ripeness. Finally, color helps us identify objects: we find our car in a crowded parking lot because we know its color (for more extended discussions, see Jacobs 1981; Mollon 1989).
The spectral power distribution of the color signal reflected from an object to an observer depends on two physical factors. The first factor is the spectral power distribution of the illuminant incident on the object. This varies considerably over the course of a day, with weather conditions, and between natural and artificial illumination. The second factor is the object's surface reflectance function, which specifies, at each wavelength, what fraction of the incident illuminant is reflected. For color to be a reliable indicator about object properties and identity, the visual system must separate the influence of the illuminant on the color signal from the influence of the surface reflectance. To the extent that the human visual system does so, we say that it is color constant.
The problem of color constancy is analogous to the problem of lightness constancy (see LIGHTNESS PERCEPTION). More generally, color constancy embodies an ambiguity that is at the core of many perceptual problems: multiple physical configurations can produce the same image (see COMPUTATIONAL VISION). Human vision has long been believed to exhibit approximate color constancy (e.g., Boring 1942; Evans 1948). Developing a quantitative account of human color constancy and understanding the computations that underlie it is an area of active current research (see Wandell 1995; Kaiser and Boynton 1996).
Albers, J. (1975). Interaction of Color. New Haven: Yale University Press.
Boring, E. G. (1942). Sensation and Perception in the History of Experimental Psychology. New York: D. Appleton Century.
Brainard, D. H. (1995). Colorimetry. In M. Bass, E. Van Stryland, and D. Williams, Eds., Handbook of Optics: vol. 1. Fundamentals, Techniques, and Design, Second edition. New York: McGraw-Hill, pp. 26.1-26.54.
Derefeldt, G. (1991). Colour appearance systems. In P. Gouras, Ed., Vision and Visual Dysfunction, vol. 6: The Perception of Colour. London, Macmillan, pp. 218-261.
Evans, R. M. (1948). An Introduction to Color. New York: Wiley.
Goldstein, E. B. (1996). Sensation and Perception. Pacific Grove, CA: Brooks/Cole Publishing Company.
Hunt, R. W. G. (1987). The Reproduction
of Colour. Tolworth, England: Fountain Press.
Jacobs, G. H. (1981). Comparative Color Vision. New
York: Academic Press.
Kaiser, P. K., and R. M. Boynton. (1996). Human Color Vision. Washington, DC: Optical Society of America.
Menzel, R., and W. Backhaus. (1991). Colour vision in insects. In P. Gouras, Ed., Vision and Visual Dysfunction, vol. 6: The Perception of Color. London: Macmillan, pp. 262-293.
Mollon, J. D. (1989). Tho' she kneel'd
in that place where they grew. The uses and origins of primate color
vision. Journal of Experimental Biology 146:21-38.
Pokorny, J., V. C. Smith, G. Verriest, and A. J. L. G.
Pinckers. (1979). Congenital and Acquired Color Vision Defects. New
York: Grune and Stratton.
Wandell, B. A. (1995). Foundations of Vision. Sunderland, MA: Sinauer.
Wyszecki, G. (1986). Color appearance. In Handbook of Perception and Human Performance. New York, Wiley, 9.1-9.56.
Wyszecki, G., and W. S. Stiles. (1982). Color Science -- Concepts and Methods, Quantitative Data and Formulae. New York: Wiley.
Copyright © 1999 Massachusetts Institute of Technology