Special Index of Metamerism Across Illuminants

This is the additive correction of the special index, which subtracts the reference mismatch from the test difference; CIE 15 also describes a multiplicative correction, and the two agree closely while the reference mismatch is small and diverge once it is not. Where the D65 difference is a substantial fraction of the tolerance, correct the match first and re-measure rather than trusting either correction to see through it. The index is a property of the PAIR, not of the sample: a batch can be metameric against a dyed standard and acceptable against a pigment-printed one, or vice versa, because metamerism arises from the difference in spectral curves rather than from either curve alone. That is also why the remedy is a change of dye combination - typically replacing one component with a dye whose absorption band sits where the one in the standard sits - and why it must be settled at lab-dip stage, since no production correction reaches it. Illuminant A and F11 are the conventional pair because they bracket the practical failure modes, tungsten warmth and narrow-band fluorescent spikes, but a buyer with LED store lighting is testing under something this pair does not represent well, and an LED spectrum should be added where that is the selling environment.