Rolf G. Kuehni

Color space and its divisions

A synthesis of the authors recent work on color-order systems and color-difference evaluation is provided in context of current knowledge and practices. The development of a colorimetric model is demonstrated using Munsell “Celtic crosses” as a model of perceptual space. Issues surrounding color-matching functions, unique hues, the Helmholtz–Kohlrausch effect, and lightness and chroma crispening are addressed, as is the difficulty of reconciling a difference-based hue, chroma, lightness model with an Euclidean model. A new lightness scale and treatment of lightness crispening is proposed. The results indicate that, despite problems, relatively simple modified opponent-color models provide good accuracy in predicting color-order system and supra-threshold small color-difference data.

Determination of unique hues using Munsell color chips

Forty observers determined their unique hues from arrays of Munsell chips in a standard surround under artificial daylight. There was some discrepancy in the results of males and females. Essentially no variability due to age of observer was found. The standard deviation around the total mean was less than 1 Munsell 40 hue step. Simple linear opponent-color a and b values were calculated. The ranges were found to straddle in all but the red color the system axes for the CIE 2°observer, but not for the 10°observer. NCS unique hues determined at similar chroma and lightness values fell in all cases within the ranges. The shifts in cross-over wavelengths of the color-matching functions necessary to match the extreme range values were determined to be between 6–11 nm. The results provide support for an opponent-color system based on subtractions of color-matching functions. They also point to significant variation in color normal observers.

Unique-hue stimulus selection using Munsell color chips

Presented are intra- and inter-observer variability data comparing the unique-hue (UH) selections of sets of males and females, using two different visual experimental procedures incorporating Munsell color chips of varying hue but identical chroma and value. Although 34 of the 40 Munsell hue chips were selected by at least one observer as a UH, selections were generally repeatable. In addition, intra-observer variability represented approximately 15% of inter-observer variability. Also, when only three consecutive Munsell chips were viewed at a time, females showed significantly larger intra-observer variability than males, especially when making unique green selections. However, variability in UH selections was statistically insignificant between males and females when all Munsell chips were viewed simultaneously. No correlation was found between UH selections or intra-observer variability and hue ordering ability.

Focal Color Variability and Unique Hue Stimulus Variability

The degree to which physiology and culture have affected the formation of primitive color categories continues to be a matter of discussion. In this paper the degree of agreement between the ranges of individual color term foci for the four hue-based color categories yellow, green, blue, and red and individual choices of Munsell samples representing for the observers Herings four unique hues is investigated. The color term focus range data are extracted from the survey results of the 110 unwritten languages of the World Color Survey, also in terms of the Munsell color order system. Agreement of approximately 90% between the two has been found, indicating the likelihood of a strong color vision system related physiological component in the formation of these four primitive hue categories.

HUE UNIFORMITY AND THE CIELAB SPACE AND COLOR DIFFERENCE FORMULA

The hue uniformity of the CIELAB system is investigated using a hue circle of Munsell colors at value 6 and chroma 14 and experimentally determined hue coefficient data. CIELAB hue differences for equal Munsell hue increments are found to vary up to nearly a factor 4, and hue coefficients differ from the experimentally determined ones by up to 40% at certain wavelengths. Dominant wavelengths assigned by the CIELAB system to individual Munsell hues are found to vary up to 35 nm from those of the Munsell Renotations. Four other color space systems are compared with widely differing but comparable results. The CIE 2° color-matching functions are adapted to result in a set of opponent-color functions accurately representing the Munsell Hue and Chroma data. A call is made for the experimental determination of the “standard hue observer” as a step toward an improved color space/color-difference formula.

Perceptual prominence of Hering's chromatic primaries

Reported are results of an experiment involving perceptual assessment of very large color differences using samples representing approximate mean Hering opponent generic unique hues (guHs) based on subject selections, intermediate hues (iHs) using Munsell samples intermediate between guHs, and pairings of both guHs and iHs with a neutral gray. Sample pairs were assessed by 28 color normal subjects twice, with a gap of at least 24 hours between assessments. Results were calculated for individual subjects and the entire group. The hypothesis was that perceived chromatic differences of Herings guHs are larger than those of iHs, and this was found to be statistically valid at the 99% confidence level based on a t-test. In addition, gray as a percept was found to have prominence comparable to that of generic unique hues.

Analysis of Five Sets of Color Difference Data

In a systematic optimization process five sets of recent color difference data have been analyzed for commonalities. Adjustment of the X tristimulus values and application of a systematic, surround dependent SL function was found to be beneficial in all cases. Other modifications of the CIE94 color-difference formula were found to bring improvements only in some cases and may be spurious. Application of what seem to be nonsystematic scale factors in a range of 0.78–1.38 improve correlation between calculated and visual color differences in all cases. After optimization, calculated color difference values explain between 80–90% of the variation in visual color differences. Some of the datasets are shown not to be well suited for formula optimization. Optimization in all cases by set, for three sets of data by quadrant in the a*b* diagram, and for one set by subset did not reveal any additional systematic trends for improvement. It appears that the basic structure of CIE94, with the recommended modifications, is a good approximation as a model for color-difference evaluation in the range from 0.5–10 units of difference. The model is surround dependent. A number of issues remain to be resolved.

Threshold color differences compared to supra-threshold color differences

Threshold color difference ellipses, as represented by the MacAdam ellipses have been compared to unit contours implicit in the CIE94 color difference formula, found to represent small supra-threshold color differences well, both in the CIE chromaticity diagram as well as the a*, b* diagram. The results indicate that the two types of contours are significantly different. While threshold differences are well described by incremental changes in cone sensitivity supra-threshold differences are well-described by increments in the opponent color parameters a* and b*. A comparison of the hue angles in the a*, b* diagram of the MacAdam ellipses and the CIE94 contours for the same 25 colors indicates no correlation (correlation coefficient −0.1). The conclusion is that threshold and supra-threshold color difference perception operates at different levels in the color vision system.

Uniform color space modeled with cone responses

An opponent-color model based on simple subtractions of color-matching functions and found useful for modeling the uniform color space is expressed in terms of cone activity functions approximating those of Smith and Pokorny. Taking into account neuro-physiological information, it is found that two stages are necessary to model the yellowness-blueness function and three steps for the redness-greenness function. The third step is similar to one proposed by Muller and Judd. It provides significantly increased hue discrimination in the spectral region of 470–500 nm. The model is shown to be superior for representing uniform color space compared to those proposed by Guth et al. and DeValois and DeValois.

Focal Colors and Unique Hues

Unique hues, as recently determined by 40 observers using Munsell color chips, have been compared to published information on focal hues. While there is good agreement in the case of three hues, there is a significant disagreement (nearly 4 Munsell hue steps) in the case of green. An experiment was performed using mostly identical observers, identical color chips, and identical viewing conditions to have 40 observers determine their focal green hue. While their average unique green hue is located at Munsell Hue 2.75 BG, their average focal green is located at 2.5G, thus being in agreement with the average World Color Survey focal green. The result points to different concepts of the observers for unique green and focal green.