[argyllcms] Re: Fluorescence

Gerhard Fuernkranz wrote:
Many thanks for the interesting link. I must admit, so far I have not concerned myself with the issue "accuracy of observer CMFs". Afer looking at this article, I am somewhat distressed. Are the various observer CMFs really unable to explain/predict a metameric match better than with an accuracy of circa 4 dE_a*b* average? This sounds nearly hopeless. How should we be able to achieve an accurate metameric color reproduction under this circumstances, by profiling devices via measurements? We care much about things like profile fitting accuracy, FWA, etc., try to eliminate each thenth dE if possible in this areas - and actually we do have a significant error source directly "at the roots", i.e. at the observer CMFs, and their variability? Hmmm ... 

It appears that bad, yes. That's just the reality, nothing you
can do about it when stuck with tristimulus reproduction. To overcome
the problem we would need to move more towards spectral reproduction.

Speaking to some people at the CIC about CMFs, I got the comment that
although there are suspicions that currently accepted CMFs may not be
as good as they could be (ie. supported properly by widely researched
measurements of observer populations), there seemed little interest
amongst those who control funding, in doing much work in this area.
The original 1931 CIE 2 degree Color observer is actually from a very small
sample population, and wasn't done as well as it could be using
modern equipment.  The 10 degree observer is from a much larger
sample population.

Graeme Gill.


Other related posts: