Digital Light & Color

As an amateur nature photographer I don't see the need to invest in a colorchecker, but occasionally it can be usefull to have more accurate colors. So I was thinking; is it possible to print a usable colorchecker with a good inkjet printer and a spectrometer? I checked the colors of the Macbeth chart and they fit almost entirely inside the gamut of my printer. Of course this printed 'reference' will not be perfectly accurate, nor will it be reliable over long time and there will be metamerisme issues. But I recon a printed reference might be better then just eyballing and relying on my memory (and it is fun to make my own colorchecker).
So I made a first attempt to print a colorchecker. The main problem with a printed colorchecker is that on an amateur colormanaged system the patches on the print won't match the original. At least on my system they don't. They look good to the naked eye, and when measured the L values are quite good, but for most patches the a en b differ several units from the original (the spectrometer output is in Lab colorspace only). I found an old thread on this forum mentioning that the values for the reference charts are internaly coded, I suppose that hasn't changed. To be usable the printed colorchecker has to be as close as possible to the original.

Is it possible to use PWP to make a 'corrected' file of the colorchecker that will result in a more accurate print? Wat would be the best approach? I suppose the first step will be to convert the spectrometer Lab values to RGB values.

You may find this paper of use:

http://www.babelcolor.com/download/RGB% ... hecker.pdf

While you might be able to print an approximation to a Color Checker, as you mention the color values are hard coded in PWP so there is no way to adjust them for use with the Match Reference transformation. A genuine Color Checker is made fade-resistant pigments with a non-shiny surface that makes it much easier to use. The miniature version is quite handy.

Hi

Does it work with a X-Rite Colorpasport?

Alain

It should work since the bottom half of the Color Passport is a Color Checker.

@tomczak: thanks for the link, that site also has an excel file with all the values of the colorchecker, very handy

I succeded in printing my own version of the colorchecker by making a first print from a file with accurate RGB values, measuring the resulting RGB values for each printed patch, calculating the difference with the original value and then creating a new file with a bitmapeditor where the RGB for each patch is corrected for the difference. I repeated this twice, and the last printed version has RGB values for the color patches that only differ about 1%-3% with the original values. At least according to my spectrometer. Most problematic is the white patch, it contains too much blue, due to the OBA's used in the paper. This might be solved by using different paper.

I made a first test shot with the printed chart under fluorescent lighting and applied the 'match reference' transformation. The result looked better on screen than the original image, but it is hard to tell wether it is actually more accurate. I will do some more tests and see if my DIY-checker actually can be used as a 'reference'. I don't expect it to work in all conditions.