ICC Profile Creation and Modeling Errors


Document ID: 
William Li, Martin Habekost


Digital, Non-Member: 
Photo, Member: 
Photo, Non-Member: 


ICC profiles are a well-established part of color management. There is always interest in how large the color gamut is for a given combination of ink and paper.This color gamut affects practical factors such as how many of the Pantone colors can be simulated using CMYKcombinations for a given ink and substrate combination. When converting color imagery in the RGB color space for rendition on a given CMYK device, clipping often occurs. The result of this clipping is that only a certain percentage of the original colors can be colorimetrically reproduced using CMYK combinations. It is also known that the conversion of the resulting CMYK device inking through the profile connection space (PCS) back to original RGB device space can unexpectedly result in a different color than the one that the conversion was started with. This round trip error from the A2B through L*a*b* to the B2A table and back is known, but it is of interest to know how big this round trip error can be. Furthermore, it is interesting to separate out the error due to the inherent modeling technique from any differences caused solely due to gamut mapping. A method of doing so is to run the round trip at least two times; to ensure that all one
is measuring is the level of error caused by modeling.

These modeling errors will be demonstrated on well-known ICC-profiles. Three different methods are being evaluated and the results of these methods are shown in this paper.

When profiles are evaluated for their round-trip analysis it is important to use profiles that were made with similar conditions in regards to ink-limiting and the amount of GCR that is being applied. The gamut tag has an obvious influence on the average round-trip DE*ab error. The i1Pro profile does very well on the round-trip analysis in
the far interior of the gamut, not quite as well near the surface, and likely not as well in the bottom of the gamut. The A2B -B2A -A2B round-trip analysis is heavily weighted towards sampling in the bottom of the gamut due to where the CMYK grid points map to. The ColorFlow (CE) profile does fairly well in most of the round-trip analysis.

Return To Search Results

Search Again