CHARACTERIZING COLOURFULNESS AND GAMUT AREA OF WHITE LIGHT SOURCES, IN ADDITION TO COLOUR FIDELITY (PO08, Pages 236-246)

This paper explores the impact of chromatic adaptation, colour space and test-colour samples on fidelity values and obtained colour differences. The CAT02 transform and CIELUV colour space do not result in remarkably different Ra values compared to current CIE13.3 computation process. Colour differences computed with CIELAB, CIEDE2000 and CAM02-UCS are highly correlated, but different from those computed with CIELUV. Uniform distribution of test-colour samples, under the reference illuminant, depends on the CCT of the light source. It has been demonstrated that highly saturated samples result in larger average colour differences, but smaller changes in relative gamut area index values. Combining samples with different saturation levels into one average fidelity or relative gamut area index value masks valuable information on actual hue-specific colour changes. The number of samples should be between 18 and 30 to cover the whole hue-angle circle without significant information loss compared to using 90 test-colour samples.