PerfX Color Management Pro

Hello everyone,

A few months ago I was contacted by Louis Dery of TGLC (http://www.tglc.com). TGLC is a profiling s/w company from Canada and they provide an ICC color management solution that “uses artificial intelligence for the color calculation”. I was in the middle of testing the IT8.7/4 profiling targets and it appeared as a good opportunity to perform the same tests with profiles made from TGLC’s software.

As I understand TGLC’s solution, it is a profiling software that uses an advanced (intelligent) statistical approach for the construction of the LUT, allowing for more accuracy in the prediction of the color behavior of the profiled device. Additionally, there is a different approach regarding the generation of the black (K) channel. TGLC’s software –and the statistics applied therein- seems to cope better with the nonlinearities of color modeling.

ICC provides the framework for the application of color management technologies, allowing each profiling software developer to come up with a better solution. From a scientific point of view, it is fortunate that there are companies that take advantage of this freedom and implement technologies that are traditionally used in other fields of science.

The graph below displays the cumulative relative frequency (CRF) curve of ∆E*ab from the software I used for the main testing. The further to the left the curve, the least the color difference. The grey curve displays the spatial variation of KPG Approval (which is very good).



As you can see, PerfX software has a very good performance. The green and blue curves are made from the profiling s/w that was used for the rest of the experiment. The blue curve displays the results of the profile made with the profiling target with 6mm and the green with the one constricted at 4mm. Both have large color differences after the 80th percentile. Closer inspection showed that these color differences occur at the dark neutral region (four color overprints). PerfX profiles manage to minimize the color difference at the dark neutrals, and this is the reason why they display less color difference after the 80th percentile. The results were nevertheless almost identical with the performance of the randomized target, which, in general, had better results from the visual one.

Taking these into account, a wish to provide the space of my blog to increase the awareness of the market for TGLC’s software. Any s/w that has something new to offer is welcome, and the choice relies on the end user. I consider increasing awareness to be beneficial for competition, and therefore for the entire market.