CIE publishes CIE TC 1-90 Technical Report CIE 224:2017 and CIE 2017 Colour Fidelity Index for accurate scientific use

The CIE Colour Rendering Index (CRI), defined in CIE 13.3-1995, in particular the general colour rendering index, Ra, is widely adopted and used by the lighting industry, in regulatory documents and in international and regional standards and specifications. However, limitations of the CRI have been recently addressed, especially for solid-state light sources, whereby the Ra values do not always correlate well with visual evaluation by general users. This mismatch arises, first, from inaccuracies of the CRI in its intended role as a colour fidelity index; and second, from perception-related colour quality effects beyond colour fidelity. It was determined by the CIE that, for both aspects, better colour quality characterization methods are needed to measure and specify white-light sources, and the work was divided into two corresponding tasks: (1) to develop a scientifically accurate colour fidelity index, assigned to TC 1-90, and (2) to develop one or more perception-related colour quality measures beyond fidelity, assigned to TC 1-91 for initial work.

This Technical Report, developed by TC 1-90, is a research report describing a general colour fidelity index, Rf, as a scientifically accurate measure of colour fidelity with respect to a reference illuminant, although there still remain some technical issues for further research. This colour fidelity index, based on the fidelity index of the Illuminating Engineering Society of North America, defined in TM-30-15, addresses aspects for only the first part of the limitations of the CRI – it does not address the need for perception-related colour quality measure(s) beyond fidelity. However, it does address several previously reported inaccuracies of the CRI as a colour fidelity measure. The important improvements of this measure, relative to the CRI, are the update of the colour difference calculation, in particular the object colour space, and the incorporation of 99 test-colour samples which provide a more uniform distribution of slope and curvature values as a function of wavelength and which have colour appearance values that are more widely and uniformly distributed in the three dimensions of a uniform colour space.