The Solver Results Analyzed
Looking at the shift in de da/db and da/dL diagram the Solver was successful i.e the green lines are much closer to the origin that the initial red lines.
Also, the combined color difference under three light sources is reduced from 1.25 to 0.46 by adding 1.2 Blue, no Yellow1, 6.2 White, 0,5 Yellow2, and 0.39 Red. Well, this seems a rather big beneficial step. However, the additions are substantial, i.e. 8.3. Can we achieve similar success with less addition? Yes, we can!
Let’s first look at the colorants. Perhaps they are working against each other. From Graph 1, it can be concluded that some colorants partly work against each other. There is, however, an alternative way to detect this. Using the Data/Data Analysis/Correlation option on the tinter table, we obtain the following table:
It is easy to see that Blue and Yellow 2 are negatively correlated, which means that if one of them are set to zero, the other will be less also. We run the Solver again and set the value of Yellow 2 as constraint to 0. Hitting Solve produces the following table:
And indeed, the additions go from 8.3 to a total of 2.2 and the combined dE hardly increased. Using these options, you can play around with different constraints to see what happens with the color of your batch.
One of the aspects that is important for some businesses is the metameric index (MI) and the combination of MI with color difference. For instance, a specification on dE might require it to be smaller than 0.5 AND the metameric index to be smaller than 0.5 as well. In our example, this effect would however be rather minimal.