Our paintings reflect our love of color, and they contain an infinite variety of hues and shades, but this variety also condemns us to be forever vigilant to its appearance in different conditions.

Color is an experience enjoyed by almost all the human race from a very early age. As we grow we learn to recognize and often name specific colors such as sky blue and grass green and yet we can never be sure that all persons derive precisely the same sensation from a given stimulus.

When we look at the physical causes of color it is apparent that they may be easily defined and therefore any personal differences of sensation are a problem of perception. Paint pigments are much more complex today than in past history. They are mixed with other materials or coated to give visual shift or other active color effects. In order to gain an appreciation of color theory and the problems of color matching, it is essential to consider the physics of sight in some detail. Before continuing, though, some background knowledge is required.

The sensation of color is achieved when electromagnetic waves between the limits of approximately 380 and 760 nanometers are incident upon the eye. These particular electromagnetic waves are commonly referred to as light waves. The shorter wavelengths produce the sensation of violet and as the wavelength increases the sensation changes to blue, green, yellow, orange and red successively. Thus whenever a visual sensation is received which is red, for example, it means that the light incident upon the eye has a predominance of long waves.

The Nature of Color

Having said that light that appears red contains a predominance of long waves begs a question. In the real world the objects around us appear to have a variety of colors and yet the sources of light energy, such as daylight, frequently appear to have little or no color. What is it then that gives rise to the appearance of color in these objects? The answer lies in the demonstrations of Newton that "white" light contains all the wavelengths within the visible spectrum and the fact that when light is incident upon the objects in the real world it is attenuated by them.

For an observer to perceive an object it is necessary for them to receive a complex pattern of light rays that have emanated from that object and contain the information required. Yet the majority of objects do not in themselves emit light, as can be readily shown by removing any sources of light energy. It is quite apparent, therefore, that the light necessary for perception generally comes from an extraneous source of energy and it is the way in which this is attenuated by the object that defines the perception.

When electromagnetic waves are incident upon any object three possibilities arise, they may be transmitted, reflected or absorbed, or any combination of the three. An object that appears red when viewed in daylight has generally absorbed a predominance of the shorter wavelengths of the visible spectrum and reflected or transmitted the remaining light into the eye of the observer. The absorbed light energy is then usually re-emitted as heat. Thus it is apparent that perception of color in objects is totally dependent upon the way in which that object attenuates the light waves incident upon it. The relative absorption of the various wavelengths determine its color and the total absorption the lightness of that color The manner in which the light is reflected or transmitted is important in determining gloss, texture and translucency.

Definitions

It is necessary to define some of the terminology used in color science in order more readily to discuss the perception of color and its measurement. It is generally agreed that one can define an isolated color stimulus by three attributes. Hue is the phenomenon that has given rise to color names and specifies whether a color appears red, orange or blue, etc. Saturation, is a phenomenon that denotes to what degree the color is pure or is mixed with white, gray or black and hence all other wavelengths of light. Lightness is a phenomenon that specifies the extent to which the color appears to emit light, and will range from black to white (or colorless for transmitting samples) since black denotes no light emission and a perfect white reflects all light incident upon it. (For self-luminous objects, such as light sources, the terms luminosity or brightness are used.)

The three attributes hue, saturation and lightness (or luminosity) are subjective and cannot be readily determined by measurement. Furthermore they change with the state of adaptation of the observer as will be discussed later. Since the measurement of color is an important observer independent control procedure it is necessary to have equivalent terms which are objective, these are dominant wavelength, purity and luminance respectively. For many purposes the difference between the subjective and objective attributes is not of great importance and since the subjective terms are by far the more common and widely understood (although perhaps inaccurately) these will be used in this article except where necessity demands.

One other term worthy of mention at this time is chromaticity. This is again an objective term and relates to the effect of hue and saturation in combination.

Perception

The subject of color perception is an area of science that is still a long way from being completely understood. Apart from the lack of complete understanding of the mechanics of the visual process there are also the problems of adaptation and contrast effects. For many purposes the mechanism of the visual process or the influence of adaptation effects may be ignored and then color measurement will offer a perfect definition of a color but as we shall see this is all too frequently impossible.

For very many years the theory of color vision most widely accepted by color scientists was that first postulated by Young (1807) and later developed by Helmholtz (1852, 1866). It was suggested that in the retina of the eye there exist three receptors, one principally responsive to short wavelengths of the visible electromagnetic spectrum, one to middle wavelengths and one to long wavelengths. These receptors, known as cones, are connected to the visual cortex of the brain by a series of neural networks and together these produce the color perception.

This particular theory received strong supporting evidence over many years, particularly since it shows that all colors may be matched with suitable mixtures of three widely separated monochromatic radiations. There is little doubt today that it is the accepted hypothesis of the mechanism of, the response of the eye to color despite the fact that no cone pigments have as yet been definitely isolated. Nevertheless it does have certain shortcomings as a complete theory of color vision, partly in the classification of those with defective color vision, but perhaps more interesting is the peculiarity of yellow.

Suitable mixtures of red and green light may produce yellow light; at the same time it is a color that has no apparent redness or greenness within it. This type of property is shared by only three other colors red, green and blue. These four colors together are known as the psychological primaries and have unitary hues; that is, each of them is perceived as being totally independent of the other three primaries. Based on this aspect of color vision Hering (1878) developed an opponent response theory. He postulated the existence of three pairs of response processes; white-black, yellow-blue and red-green which take place in the visual mechanism, suggesting the existence of the psychological primaries.

As a complete theory of color vision the Hering hypothesis is not valid, because of the color matching evidence described earlier. Nevertheless, it can be developed to a form that fully explains the phenomenon of defective color vision. It is likely that any hypothesis of color vision that is finally proved correct is likely to contain within it both the Young-Helmholtz and Hering concepts and define some complex neural connections bridging the gap between them. One such theory, postulated by Muller (1930), has attracted much interest.

Having discussed the mechanics of color vision, lets look at the influences of extraneous factors on the perception of color and in order to do this it is necessary to describe one of the most important single phenomena occurring in color perception, that of metamerism.

Metamerism

Metamerism may be defined as the property of the eye and brain to receive the same color sensation from two objects with different spectral energy distributions. In other words despite the fact that at each wavelength in the visible spectrum the visible light energy emitted by the two stimuli may be different both appear the same to the observer. Two objects matching in this way are known as metamers. The reason this occurs is that the eye has three receptors that are color sensitive. Extending the thought, it means that the sole requirement for two colors to match is that the total light energy, with respect to the sensitivity of the receptors, should be the same for both objects.

Metamerism may be important in a variety of ways. There are three conditions to consider:

  • Observer metamerism;
  • Illuminant metamerism; and
  • Object metamerism.

Observer metamerism occurs because of individual differences in sensitivity to color. We know that minorities of the population have defective color vision (primarily men) and even among the remainder differences in sensitivity occur. Thus, if two colors are a metameric match for one person it is likely that they may not match exactly for anyone else. Alternatively, they may match for two individuals with quite different sensitivities but not for a third. However, while individual sensitivity differences may be quite large, unless the degree of metamerism is very high we can ignore it.

With illuminant metamerism we consider the effects of varying the spectral power distribution of the illumination in a special way. Imagine that we have a standard observer and two objects with the same spectral reflectance. If, under illumination with different spectral power distributions, the objects match, we have an example of illuminant metamerism. If the spectral power distribution of either illuminant is then changed, but such that the two are still not equal, it is likely that the match will break down despite the equivalence of the spectral reflectance of the objects. Once again for most practical situations this aspect of metamerism is not of crucial importance. While we are concerned with changes in illumination we cannot use the illumination to obtain a color match. For our purposes object metamerism is the most important.

Object metamerism is concerned with differences between the spectral reflectance of the objects. Consider an example where we have two objects that have different spectral reflectances and yet for a given observer under controlled illumination they match for color. If we now change either the observer or the illumination, it is likely that the match will break down.

Here it is apparent that the biggest difference between them is the spectral reflectance at the long wavelengths. Later we shall see the relative light output at the longer wavelengths of tungsten light is very high compared to that of daylight. Furthermore, the sensitivity of the eye at long wavelengths is fairly low and the combination of these effects is sufficient to ensure an approximate match under daylight, despite the large difference in spectral reflectance. As an aside, I own a red car that I often lose at night in parking garages, since the bright red of the car in sunlight transforms to a muted brown under sodium lighting. I must literally examine the contents of the car to be convinced that it is mine.

A great deal of emphasis has been placed upon metamerism and the reason for this is its fundamental importance to color science, matching and reproduction. It is the foundation stone of colorimetry, arises very frequently when undertaking a color match, and without it photographic, printing and television color reproduction would be impossible. For these reasons it is essential that the concept of metamerism be understood. The importance of ensuring that standard viewing conditions are used for both color matching and the assessment of applied coatings becomes obvious, as do the major pitfalls to look for in color matching.

Viewing Colors in the Real World

Lets briefly look at the problems of color perception in the real world. To a large extent the discussion thus far has been for single isolated stimuli, or in the case of metameric colors a pair of isolated stimuli which match for color. Generally, however, we do not see color in this way but surrounded by other stimuli each of which influences our perception of that color. This is shown in the following illustrations. They may not work as well on a backlit monitor screen (depending upon its calibration) as they do on a printed page, but the metameric effects are still visible.

Adaptation is an important phenomenon to consider in any discussion on the perception of color. Light and dark adaptation we are all familiar with; a car headlamp which can be totally dazzling at night appears quite dim during daylight suggesting that the eye has the capability of increasing and reducing its sensitivity to suit the ambient conditions. The fact that a bright light can be dazzling for quite a few minutes at night when the eye is dark-adapted demonstrates that these adaptation effects are not instantaneous.

Brightness adaptation is only one aspect of such sensitivity changes, however, color adaptation also occurs and can be as easily seen. A yellow patch seen with an eye adapted to green light will appear to be red; it seems that the adaptation to green light has decreased the sensitivity of the eye to green thereby allowing the red content of yellow to predominate. Such effects may be total, in that the whole eye is adapted, or local as seen in illustration 1. When you looked away, you saw a temporary reduction in sensitivity of only part of the retina causing an after-image of a complementary color to that of the area viewed.

Color adaptation produces one significant visual phenomenon known as color constancy with which we are all familiar, though perhaps without realizing it. Objects seen in a room illuminated by daylight do not dramatically change their color appearance at night when illuminated by artificial light. Nevertheless the spectral emission from those objects will be quite different in both situations and only rarely could the effect be explained by illuminant metamerism. Under special conditions it is possible to demonstrate that this color constancy can be complete in so far that neutral colors can be made to appear neutral under a wide range of colored lights but under normal conditions the adaptation is rarely complete. It is further complicated by fluorescent sources because most have strong line spectra and can make some of the color distortions appear very large. When such a situation exists, a lamp is said to have poor color rendering.

Simultaneous contrast, is another significant phenomenon in that the appearance of a color is highly dependent upon any colors surrounding it. A color which is moderately bright when seen against a gray background can be made to appear gray if viewed against a background of color of the same hue but very much brighter. This is very extreme, but under many practical conditions quite large changes in appearance can be observed by moving a color from one part of a room to another in order to change the surround conditions. Unlike adaptation, contrast effects appear to be instantaneous.

If we now return to the single isolated color stimulus, with no other object in the field of view, it is apparent that the effects of simultaneous contrast and successive contrast when viewing this are not significant, at least if a short period of time has elapsed. Pre-adaptation effects may be important for some minutes, but after that the eye can be assumed to be totally adapted to the stimulus itself. It is under these conditions that all color matching data used for the majority of color measuring systems has been obtained. It has been shown that a metameric color match produced under one set of conditions does not break down when the conditions of adaptation change. The implication of this is that color measurement cannot specify the change in appearance of that color due to adaptation effects, it can only specify the physical stimulus itself. It is for this reason that there is not complete correlation between-hue, saturation and lightness and dominant wavelength, purity and luminance factor.

With the light effects of pigments commonly used in coatings becoming more complex, utilizing pigments that change color with changing light wave incidence, metamerism becomes very important. It is hoped that this brief insight into a few aspects of perception will help to emphasize the importance of being very careful when describing the appearance characteristics of colors.