By Mark Gundlach
In the past century, there has been much study and research into the measurement of color and the ability to control it so that both the seller and the buyer agree that the quality is acceptable.
Among the first people of our century to provide us with a greater understanding of our visual system was Albert Munsell. He described a “natural” color recognition system with three independent color attributes and suggested that these attributes are “built in” to the human mind and eye.
In the 3-D L*c*h° space, L* stands for “lightness” or “value,” c* for “chroma,” and h° for “hue.” Hue changes are organized around the circumference where reds shift towards oranges, oranges toward yellows, yellows towards greens. Hues are family shades identified by the “color” word, such as “greens,” “purples” and “reds.” A hue shift is essentially a change in color family. Lighter and darker colors (L*) slice through the center of Munsell’s color model. At the bottom is absolute black, and at the top, absolute white. Note that the light/dark “axis” is colorless—a series of differing grays or achromatic shades to which no color “family” can be assigned. This is known as “value,” or “lightness.”
Hues are family shades. In this illustration we see that Munsell divided the circle into 100 units, essentially starting with red and moving clockwise. Every ten units he identified a color family – “Red,” “YellowRed,” “Yellow,” etc. Today we keep pretty much the same model, except we normally divide the circle into degrees.
Chroma, or increased chromaticness, is a movement away from gray. As a color moves outward from the axis, the grayness is decreased and the strength and purity is increased. At this point, a family name, such as orange, can be assigned. Duller colors exist toward the center, and become stronger and cleaner colors as they approach the outer circumference.
L*C*h° has become the accepted color model, and Delta L*, Delta C*, and Delta H* the tolerances. These acceptance solids remain aligned to the chromatic axis of each color and, as a result, they conform more closely to the behavior of the human eye, improving the color quality assurance process by about 10 percent. Machine pass/fail judgments agreed with trained color matchers 85 percent of the time.
In 1988, CMC equations brought even more improvement. The CMC tolerancing system still describes color as hue differences moving around the circumference and chroma as a vector from the center towards the outer edge, but it creates tolerance ellipsoids in color space around the target color. These ellipsoids change shape and dimensions depending upon the position in color space. With CMC, it is not unusual to achieve machine/man decision matches that exceed 95 percent.
CMC Tolerancing System—The smaller ellipsoids, such as the grays in the middle, designate colors for which we have a small tolerance for change. The larger the ellipsoid, the great tolerance we have for change. For example, if you add 5-percent blue to light gray you will notice the color shift more than if you add a 5-percent blue to a very saturated emerald green.
Newer tolerancing schemes are constantly being introduced. In fact, the CMC equations were slightly modified in 1994 yielding the CIE1994 method. The math was altered again in 2000, where the L*C*h° point representing the center of the perception ellipsoid “floats” between the standard and the sample.
In the late 1920s, W. David Wright and John Guild conducted a series of experiments. By using three red-, green-, and blue-filtered lamps, they were able to evaluate what percentages of each color was required to create a vast assortment of colors through a two-degree field of view. The work of Wright and Guild was combined into the specification of the color space from which the CIE XYZ color space was derived.
Guild and Wright’s experiment mapped the electro-chemical, tri-stimulus response of the human eye. Their research directly showed how a mixture of red, green and blue lights might be used to create any color in the visible spectrum. Indirectly, it also revealed clues hinting at a complex, non-linear visual system wherein the observer can distinguish small differences between some shades much easier than others.
In the early 1960s the CIE updated Guild and Wright’s analysis to include a 10-degree field of view. As a general rule, the 10-degree curves are preferred when observing large objects. For very small objects, the two-degree curves are still available and are often specified. Today’s instruments and software are able to use either set of data.
The first recorded method of analyzing color came from work by Richard Hunter. His numerical tolerancing method in 1945 was the first of its kind, and was adopted by the Commission Internternationale De L’Eclairage in 1976. This is the color model known as Lab. His math was published in 1945 and was widely used with early color measurement instruments. Although his coordinate scheme only approximated the human pass/fail response with a 75-percent success rate, it was the best available at the time.
While Hunter’s Lab was quickly adopted as the de-facto model for plotting absolute color coordinates and differences between colors, it never achieved the status of a formally accepted international standard. Thirty-one years passed, and in 1976, the CIE published a new model: L*a*b* (pronounced “L star,” “a star,” “b star”), or CIE Lab (pronounced “See Lab”). With only a few small changes to Hunter’s original math, this new map became the recommended and internationally sanctioned method for reporting colorimetric values.