Adobe GRACoL profile doesn't match GRACoL profile

[disbellj]

I plugged in the Lab numbers for Adobe's GRACoL profile (CoatedGRACoL2006.icc, named Coated GRACoL 2006 (ISO 12647-2:2004) in Color Settings) into a calculator I made a few years ago that I made to fix plate curves between printing conditions.

When I saw the differences, I saw that Adobe's GRACoL profile was about 3% lower than GRACoL2006_Coated1v2 in the 10% area on Magenta. This was the largest difference between the two. Anyone can do the same tests using Bruce Lindbloom's instructions and his LabDotGainCalculator2 files and instructions from his website Welcome to Bruce Lindbloom's Web Site, or they can just use my calculator I made (attached).

Maybe this will not be any problem at all, but it makes me wonder why the difference if Adobe used the same characterization data to make the ICC profile as GRACoL provided. Maybe Adobe made their own press run and used their own characterization data? And it makes me wonder if colors converted using Adobe's GRACoL, if that separation, when printed on GRACoL, will look that much darker than it should be (like 6% darker than it should be).

Zipped Excel file attached.

Anyone know about this and can shed some light?

Thank you.

Don

 

[eran]

I've been really curious as to the difference here as well... Unfortunately I've got precisely no input- I just want to be subscribed.

 

[meddington]

Hey there Don,

I think the issue could be the result of rounding when obtaining the Lab values (presumably out of Photoshop, yes?). So the 10% Mag source is 91, 6, -3, and destination is 90, 6, -3.

If you render the Lab values from ColorThink pro using a P2P25 target and these profiles, the values are 89.96, 6.63, -3.29 & 89.95, 6.67, -3.35 respectively. Those values plot closer to each other in your spreadsheet.

Looking at the comparisons in ColorThink, the data used to create the two profiles appears to be identical.

 

[tmiller_iluvprinting]

The data may be identical, but the profiles are not. Black start, TAC, and I believe a different amount of GCR were used to create the Adobe profile.
Best regards,
Todd

 

[meddington]

The data may be identical, but the profiles are not. Black start, TAC, and I believe a different amount of GCR were used to create the Adobe profile.
Best regards,
Todd

Totally agree. Even if all those were identical you could still get different results (from negligible to significant) based on the different gamut mapping algorythms of the profiles...and who's to say which one would be right?

 

[disbellj]

Michael,

EXACTLY! And this is why we all need to be using one profile, and not have different profiles out there. A customer's RGB Blue (or any out-of-gamut color of the printing condition that has to get squeezed or clipped into the printing condition) might look different from one printer to another even if all printers are printing to GRACoL2006_Coated1v2. Of course, theorhetically we should all be converting from RGB or Lab into CMYK or Grayscale using Absolute Colorimetric, but we know that's not practical (unless the new sRGB profile or something keeps CMYK 0,0,0,0 paper from becoming simulated paper, not CMYK 0,0,0,0 but dots where there should be no dots, or maybe there is better than Relative Colorimetric Intent, since I haven't looked at this stuff in a couple years).

Thanks for pointing this crucial fact out.

Regards,

Don

Totally agree. Even if all those were identical you could still get different results (from negligible to significant) based on the different gamut mapping algorythms of the profiles...and who's to say which one would be right?

 

[rich apollo]

The data may be identical, but the profiles are not. Black start, TAC, and I believe a different amount of GCR were used to create the Adobe profile.

There are positives and drawbacks to each of the profiles.

I don't think it's necessary, or even desirable, for everyone to be using one ICC profile. Each users' needs may differ radically.

 

[TerryWyse]

And this is why we all need to be using one profile, and not have different profiles out there.

And which profile should that be Don? I don't see any reason to have a single profile describing the a particular printing condition such as GRACoL Coated1. I feel they should all be made from the same *data set* but there are a host of reasons that we shouldn't be restricted to a single profile.

Regards,
Terry

 

[disbellj]

Terry,

I guess because when a designer sends a blue to be printed on the same paper type at two different printers, it would be nice for them to have the output to look the same.

Both printers using the same ICC profile is what must happen for the blues to match.

If the designer is told to design in RGB, then they softproof for CMYK but don't convert to CMYK.

The gamut mapping is in the ICC profile-making software, and different ICC profiles map the out-of-gamut colors of the printing condition differently.

It all has to do with software used to make the profile, because the different programs create different ICC profiles with different gamut mapping, even when making a profile for the same paper type/characterization data.

So the short answer is: To make out-of-gamut colors look as intended (or at least the same between printers).

ECI has profiles that all users use in the production chain, so they don't have this problem, but the U.S. is lagging behind even though I mentioned this problem over a couple years ago when I used to post here. Heck, the U.S. doesn't even have a standard profile to use for #4 (uncoated) paper. It doesn't matter, right? Users are blissfully ignorant anyways. SWOP profile is good enough to convert to (which is true especially for commercial printing), and no one really needs to do soft-proofing (most don't), and who really cares if a blue or green, etc. doesn't match between printers (for those that don't convert to CMYK before sending to prepress or printer)? Guess just me, since I am an artist at heart and see it from their point of view.

Regards,

Don

 

[meddington]

ECI has profiles that all users use in the production chain, so they don't have this problem

I think Europe does tend to use the same profile often as you say, but I'm sure many use custom profiles for their own black generation settings. And the ISO profiles we're relativily recently updated for for gamut mapping anomolies (purple blues). Also photoshop does have its pen profile based on fogra39 as well (CS4).

 

[disbellj]

Michael,

Maybe they've gotten better, but I can tell you that at one time (a couple or few years ago when I studied all this), all GRACoL ICC profiles based on the same characterization data, that I saw from the different packages, all mapped out-of-gamut colors differently. So really unless all programs that make ICC profiles map the out-of-gamut colors the same, then there will still be differences. I'm just thinking of the customer here. If I was designing, I'd want my pictures to look the same all places I sent them to print (notwithstanding the different appearance on different paper types, since I'm talking about sending a job to get printed on the same paper type printing at different printers.

Oh well, according to a Huffington Post article dated today, guess which industry will lose another 16% jobs in the next ten years? #5 Printing and related support activities. We'll eventually get automated out either way. Never thought I would peak in my career in prepress about the same time I finished my journeyman (about 7-8 years ago now - about half my career back) until it happened. Things are so much better than they used to be, I just have too much time on my hands, hence the post on something nobody really cares that much about (or is an old argument that maybe has been fixed by now).

Thanks Michael.

Regards,

Don

 

[meddington]

Michael,

Maybe they've gotten better, but I can tell you that at one time (a couple or few years ago when I studied all this), all GRACoL ICC profiles based on the same characterization data, that I saw from the different packages, all mapped out-of-gamut colors differently. So really unless all programs that make ICC profiles map the out-of-gamut colors the same, then there will still be differences. I

Yep, I agree. Not necessarily apples and oranges, but certainly some differences. Which gamut mapping algorithm would be correct though, and in choosing one as a defacto standard, would this not alienate the other competitors? For this reason I don't think you'll see standard/specifications mandating a particular profile creation package, lest they feel the wrath of the vendors' lawyers.