In cylindrical color spaces, such as represented by the (for many colors near the spectral limit) the CIELAB equation predicts about 3 times the hue difference be-Munsell System, colors of equal hue are by definition arranged on radial planes in the cylinder. By definition, tween adjacent Munsell Hues compared to chroma 6; there should be 7.5 Munsell 100 Hues between two hues remain constant on these radial planes, regardless of the chroma and lightness of the color. Visual examina-Munsell 40 Hues at chroma 18, but only 2.5 hues at chroma 6. This would require about 300 hues at chroma tion of the Munsell Book of Colors confirms that this is closely so. The visual hue difference between two neigh-18, clearly much more than can actually be discerned. Conversely, if 100 hues are discernible at chroma 18, boring Munsell Hues is the same regardless of chroma. However, calculated hue differences, using a formula then at chroma 6 only 33 should be discernible. There is no problem discerning the 40 Munsell Hues represented such as CIELAB and those derived from it, increase with increasing chroma. Thus, for example, the CIELAB hue by chips at chroma 6 and, in most cases, it is easy to imagine at least one hue step in between. Figure 1 (Right) difference between 7.5PB4/6 and 2.5P4/6 is 9.0 units. At 4/12 for the same two hues, the hue difference in-illustrates schematically the situation described above.
Figure 1 (Left) illustrates schematically the correspond-creases to 17.2 units. The implication of this result is that at 4/6 a similar hue difference is represented by ing perceptual situation. Below about chroma 4-6, there appears to be a less-neighboring samples 5PB4/6 and 5P4/6 (CIELAB hue difference 18.1 units). When examining these sample ening of discernible hues as the hue component begins to be overwhelmed by the achromatic component. How-pairs, this is clearly not the case. Sample 5PB4/6 is noticeably greener than 7.5PB4/12 and 5P4/6 is clearly ever, at chroma 2, there are still 40 hue steps visible. A visual examination indicates an increase from about 40 redder than 2.5P4/12. This example is not unique, but can be documented around the hue circle.
discernible hues at chroma 2 to about 100 at chroma 4-6, from whereon the number remains the same until the Another implication of the CIELAB type of calculation is that new hues would appear continuously between two spectral limit.
To calculate hue differences as the remainder after Munsell Hues as chroma increases or, conversely, hues disappear continuously as one moves from higher chroma the chroma difference is subtracted from the chromatic to lower chroma. It is not obvious that this is so. If one difference appears incorrect. If the above observations are interpolates the number of discernible spectral hues from valid, they imply that, for supra-threshold differences, the wavelength discrimination functions and adds an estimate CIELAB formula calculates a chroma-dependent error for for the nonspectral hues, one arrives at a number of about hue differences of up to a factor of 3 or more. Hue differ-100-110. That is about three times the number of the 40 ences that are in better agreement with visual judgments hues of the Munsell Book of Colors. Since at chroma 18 can be based on hue angle differences with some adjustment at low chromas. Alternatively, they can be calculated conventionally with a chroma-dependent adjustment