Hi Marc, Thank you for taking the time to post here. I failed to mention that these were i1Profiler custom output profiles created out of IT8.7/4R measurements. The two IT8s were measured a few months apart, I confess, the 7890 in early June and the 7890 yesterday, both using the same Minolta FD-9 in M0 mode -- I'd be surprised the instrument would have drifted that much over such a short period of time? The media was the same for both printers yet there could be some slight variations coming off the media, contributing to the differences I see? Marc, I did exactly what you suggested. To assess the resultant conversion, at the end of my work, to do "quality control", I always run an IT8 on the printer through the workflow to test the quality of the simulation. The difference between the two printers "simulating" GRACoL2006_C1 is 0.66 dE2000 with a max of 2.37, 95% of patches are below 1.45, see the following screen capture from PatchTool Compare function: https://1drv.ms/u/s!AkD78CVR1NBqkaJgOTwFNP1Mpd1O7Q?e=P4lSsR I feel I should be "resting my case" but I'm curious to dig into the possible spectral differences across both ink sets? To that end, I assembled a summary analysis of both proofs "simulations" here, in a PDF: https://1drv.ms/b/s!AkD78CVR1NBqkaJhE5Sy0r9fFZ2UQg?e=DhA18K I am pleased to find similar spectra across both printer simulations, the largest difference, spectrally, coming of the CMYKRGB set is on 100M, as William noted. Yet, in terms of color difference, that difference in spectral shape only translate into 0.56 dE2000? Coming back to the visual difference between the prints, the largest and most questionable visual color difference (so far) comes from CMYK values around 66c 72m 62y 60k or CIE Lab 21 7 1, some kind of "dark gray violet", you see? 21 7 1 is the Lab value coming out the 7890 IT874. While on the 7800, according to Photoshop color picker, the CIE Lab value for this build is 19 8 1, a 1.74 quite "reasonable" dE2000 color difference. Yet, visually, across both prints, under a decent ISO-3664:2009 viewing booth, the color difference in that area of the image seems way bigger than 1.74? It's not a difference that immediately jumps to the eye when glancing at the two prints, side by side, overall, but upon closer inspection, these two areas don't seem as close as the rest of the imagery are, suggesting some possible difficult selective color edits. At a packaging printer I sometimes consult with, in their ink room, I've seen color differences in some dark blues samples, if I recall, that, colorimetrically, were rather small, well under a typical 2 dE2000 tolerance yet visually, appeared quite different. (Next time I'll go, I'll be sure to grab some samples to measure and document...) In the IT8.7/4R measurements, I found a patch that's closest in term of CMYK build to the offending "dark gray violet", patch ID 1238, 70c 70m 70y 80k. It is interesting to note that on the 7800, this is CIE Lab 16 1 1 while it is CIE Lab 15.2 1.9 1.38 on the 7890, a dE2000 difference of only 0.96. Of course, both proof pass G7 tolerances with "flying colors". I think I'll leave my analysis at that and move on to some other project. Best / Roger -----Original Message----- From: colorsync-users <colorsync-users-bounces+graxx=videotron.ca@lists.apple.com> On Behalf Of Marc Levine via colorsync-users Sent: Thursday, September 26, 2019 3:32 PM To: colorsync-users@lists.apple.com Subject: Re: Slight color differences Hi Roger, It's been so long since I posted here. I hope this is the right place. ; )
From a scientific perspective, you might run a small test. You could transform an IT8 using the workflow you described, and then compare the resulting Labeez of the 2 converted images. That would tell how ho far off you were before you even printed.
It was unclear from your post whether the profiles you mentioned are "canned profiles" or custom profiles that you made. In the case of canned profiles, there is a whole host of things that could be different and unknown between the 2 profiles. If there are both custom, following the same exact process, I would expect the electronic result to be very close and for the printed result to present a greater-but-reasonable difference compared to the electronic files. It sounds like the differences you are see are pushing the boundary of "reasonable". Here I would fall back to what others are saying about the output colorspace and the differences in the inks (assuming that both printers are also moving the material at the same speed, laying ink down at the same rate, and that the ink is drying at the same rate). Different inks will produce different behaviors in the colorspace that the profiling engine needs to accommodate (which is why proofing software like GMG and CGS tools iterate). These different behaviors might be handled slightly different but the profiling engine and would likely introduce some of the effect you are seeing. If you are feeling adventurous, you might try to add some "training data" to your profiles. That p rocess would look like: -Prepare a "special target" (an Lab file with measurable colors that are important to you). Grays are always good. -Make your regular-ol profiles -Transform your special chart into the colorspace of the printer (gives you a set of device values) -Bolt those device values together with the original IT chart values to make a new super-special profiling target -Print that target and build a new profile -Rinse and repeat for "printer #2" -Compare -Enjoy ; ) -Marc _______________________________________________ Do not post admin requests to the list. They will be ignored. colorsync-users mailing list (colorsync-users@lists.apple.com) Help/Unsubscribe/Update your Subscription: https://lists.apple.com/mailman/options/colorsync-users/graxx%40videotron.ca This email sent to graxx@videotron.ca