> The question is are those spectral data directly usable to calculate > CIE tristimulus values or should I somehow normalize > those spectral values. Clearly, you cannot directly use the 4nm of a PR-650 in tristimulus calculations because, to my knowledge, the CIE has not made Color Matching Functions tables available in 4nm increments. If this case, you have no choice but to interpolate the 4nm data to 5nm or 10nm. I gather that CIE Publication Number 15.2 recommends that "interpolation be carried out using a third-degree polynomial from neighboring data within twice the measurement interval". I'm quoting from Computational Colour Science using Matlab, Chapter 4, Computing CIE Tristimulus Values. ISBN 0 470 84562 7 Once you have the data in 5nm or 10nm form, then it is straigthforward to compute tristimulus values. Regards, Roger Breton | Laval, Canada | graxx@xxxxxxxxxxxx http://pages.infinit.net/graxx