Here are results from color matching tests conducted by a customer using Colibri ColorMatch. Ink recipes predicted for 10 target colors were mixed and produced by drawdowns on Leneta card. The color of each drawdown was compared with the corresponding target color by measuring on spectrophotometer. The ΔE values in the chart below show the difference between drawdown and target color. The average ΔE is 0.94 for the 10 colors.
The results show the accuracy and predictive capability of Colibri’s multi-flux optical model. Both the calibration sets and test sets were made on Leneta cards. This represents the best results that can be expected under ideal conditions.

A second set of colors were tested on a paper with higher ink absorbency. Colibri translates ink calibration data from low-absorbency Leneta cards to higher absorbency papers. The need to create separate ink calibration sets on different materials is eliminated. The results below represent the results from this exercise. The average ΔE is 1.7 for the 7 colors.

There is some loss of accuracy in translating data from non-absorbent to absorbent substrates. But 80% of the colors are typically acceptable at first proof. These results are more typical of what is encountered in practice.

In cases where the first drawdown proof does not produce an acceptable match, a color correction can be made using Colibri to predict an adjusted ink recipe.
For this test, first predictions were purposely made to have a high ΔE by NOT applying the absorbency compensation to the calibration data. This lets us see the performance of color correction more clearly.
The hashed area in the bar chart shows the improvement from first proof to the corrected second proof. Correction improves the color on average by 2.2 ΔE. About two-thirds of the deviation can be removed on average by a single correction.