Metamerism Index

Consider a pair of colours which match to a given observer in daylight. They have different spectral reflectance curves and so will be metameric. The observere now moves to tungsten light and sees a difference of say five times a 'just noticeable difference'. He can with some justification say, that for him, the degree of metamerism on that pair which match in daylight is 5 JND in tungsten light.

This is a very useful concept to describe the extent of the mismatch and could have considerable commercial value. However, it suffers from the usual deficiencies of subjective colour assessment. It assumes that all this observers estimations of mismatch differences in tungsten light will be the same and it assumes that other observerswill estimate similar observed differences in the same way. Neither of these assumptions holds true. But much worse is the fact that it completely ignores observer metamerism. The differing relative spectral sensitivities of different observers retinal cones can cause such enormous differences of opinion as to render the whole concept futile. Nevertheless an index of metamerism such as this is gaining in importance due to the rapid spread of colour measurement and colour prediction.

Instead of our human observer consider the same case using the CIE standard observer colour matching functions and illuminants such as D65 and A as our daylight and tungsten light. For describing the colour difference we could use equations such as CMC, BFD or CIE94. This would appear to be a neat objective way of describing the difference in tungsten light. Unfortunately it ignores the fact that all the colour difference equations were established for daylight only. They may not work for observed differences in illuminant 'A' as the eye completely changes its sensitivity. The success in practice of the 'metamerism index' is due to the work that has been done in attempting to define mathematically this change in sensitivity as shown in the following section.