Using Categorical Observers to Minimize Metameric Failures on Wide Color Gamut Displays

It is common industry practice to calibrate displays using measurement data based on the CIE-1931 2° Standard Observer color-matching function (CMF) to represent the human visual system response. Anecdotal observations, with some experimental evidence, have shown that adjusting two displays with different spectral power distributions to the same CIE-1931 tristimulus values (TSVs) produces a calibration that is not a visual match for some, if not all, human observers. In other words, displays adjusted so that they should produce metameric matches do not do so. — Research shows that this is due to deviations in CMFs between observers that also differ from the CIE-1931 2° Standard Observer; and that populations of observers can be clustered into categories, each of which have largely similar CMFs. It may be that selection of an appropriate alternate set of CMFs for any given color grader, cinematographer, color editor, or others making critical color judgments or adjustments across spectrally dissimilar displays may reduce metameric failures and increase productivity throughout a motion-picture workflow. — This paper presents results from a psychophysical study to determine if there is a “best-match” CMF for a real human observer from Asano and Fairchild's 10 categorical observers that reduces metameric failures between 2 displays that have significantly different spectral power distributions. The results are further analyzed to support a discussion about the practical implementation of categorical observer transforms in a motion-picture workflow.