Refining Color Profiles

Hello all!

First, I would like to apologize if a lot of this information has been covered in other posts, and trust me I looked around, however it seems that there is not a lot of public information concerning my interests.

I would like to also say that although I am going to focus my questions towards scanner calibration, a lot of information I am trying to seek deals with color profiles in general. I am going to start in small steps here and see if I can get a discussion going.

The fact is I am not happy with the ICC profiles I am generating from IT8 targets. I have a nice scanner, it has a great D-max rating, the specs on its color gamut are impressive, and I have been able to run tests that verify that the manufacturers specifications are indeed correct. I am not going to name the scanner because I want to talk about theory and I dont want specifics to cloud things.

Now, I also have a number of spectrocolorimeters (SCMs), and quite a bit of software ranging from the low-cost consumer space to the high-cost pro market. I am no longer interested in purchasing any more equipment or any more targets, but I am willing to spend more on software if it does what I need. I have a number of IT8 targets, and the correct descriptions. I can take my SCMs and verify that the IT8 descriptions match the Lab values on the actual targets. For the most part, X-Rites profile maker gives me the most accurate profile for my scanner, but it is off in a couple of areas, usually in the yellows. What I would like to do is print / develop my own IT8 target that consists of nothing but shades of yellow, measure the target, and generate my own description file. I would then like to use this yellow-IT8 set to create a second ICC profile and average it with the profile created off the stock IT8 target.

Does anyone have experience building their own IT8 targets, and are there are good tools for generating description files? So far it looks like this is going to be a laborious process of manually editing fields in a text file and converting each Lab value to XYZ, Lch, and then work out the statistical variance by hand.

Does anyone know of any decent software that combines information from multiple profiles? Will this work and will it help me achieve greater accuracy?

Does anyone have a better idea on how to accomplish what I am looking to do? I just really feel that a standard IT8 doesnt have enough data to get the level of accuracy I am looking for, and I have enough printing and measuring equipment that I should be able to come up with a custom solution to improve things.

Thank you for your time!


Hi Martin,

It sounds like you are wanting to make custom profiling targets, esp. for scanners, and therefore need to be able to measure the targets and get accurate reference information. What you are calling a description file is usually referred to as a “reference” file. This is the file that the measurement file is compared to in order to create a profile.

If you have a dongle-version of MeasureTool, it has a Testchart Generator that you can use to easily create custom profiling targets of any size for a number of different spectrophotometers. Colorport by Xrite will also allow you to create profiling targets. X-Rite also has a sort of “in house” software that is not supported called Colorlab which does many of these same functions. All of this software is available from X-Rite. Colorport and Colorlab are free. All of these will generate a reference file without you having to manually edit text files.

Also a spreadsheet program is very handy here. Excel can allow you to copy and paste whole rows or columns, and fill in columns using the same numbering pattern. Excel can import and export in text format also.

Once you have a printed target image, you would of course have to measure the patches in order to create the reference file which you would then use with the scanned image of your target to make a scanner profile. So, keep in mind your ability to read the patches when you layout your target (patch size, etc.)

Keep in mind that this printed target would be limited to the gamut of your printer - not necessarily well represented of the gamut of your scanner. Scanner profiling software generally tends to go beyond the limits of the measured target in creating a scanner profile. It sort of makes some educated guesses as to the more saturated portions of the profile, so that the profile will have some idea of what to do with colors that are outside of the gamut of the profiling target.