Multiple rounds of color correction color matching problems

[meddington]

Re: Multiple rounds of color correction color matching problems

Hi Don.

>Why do you think repurposing will get rid of the problems? I wouldn't think so. Why? Let's say a person has done color correction on a SWOP image while soft-proofing for their press (using their press profile)? It should still look on press like the soft-proof did.

Think of it this way...a number of images start out with similar black channel and GCR properties. After several round of color corrections, each image now has different black channel and GCR properties, and when printed together on a press form, they can ll move in different directions when inking is changed. If the images are repurposed to "homogenize" the separations, the issues on press are reduced. It doesn't have to be "repurposing", but could be the result of a late binding/RGB workflow. Also, if the corrections are done by a single operator with a consistent approach, editing by the numbers, this could be less of an issue, but its really more work to go that route.

I was convinced of the importance of a common approach to black generation/ CMYK editing a long while back while observing a fashion catalog printing involving lots of dark clothing. All black clothing looked similar on the proof, but because they were edited somewhat extensively in CMYK by a number of operators, they all printed differently, and it was impossible for the press operator to adjust for one while maintaining the others. He may have been able to hit any one of the images fairly closely, but not all of them together on the form. Thus the products and backgrounds printed inconsistently. Re-purposing these images would have minimized this effect.

 

[cnjprinter]

Re: Multiple rounds of color correction color matching problems

O.K. I am kind of following along here, some of this stuff is deep.

A couple folks used the words "repurposed" and "re-separated". How is this done? I looked through PS help and found nothing on this.

BTW... the client sends us files with from their flat bed camera, scanner, or their client supplied files. Scans as RGB they converts to CMYK and applies profile "US Sheetfed ctd v2 (8bpc)" { 350% total percent ink coverage} after this then they produce their first Epson proof. From there it stays CMYK, gets placed into InDesign for us to run film. (Yes - we still run film)

 

[ASM]

Re: Multiple rounds of color correction color matching problems

Don,
That's a very interesting point about SWOP and relative colorimetric intent and even without trying it, I can see what you mean. I hark back to analogue Cromalin/Waterproof days and we used to have a swatch of several bases on which to proof - all of which changed the gamut. CP1, CP2, etc. Thanks for making this point.

 

[Guest]

Re: Multiple rounds of color correction color matching problems

Hi Michael,

Can the procedure outlined in the post by Walter Zacharias be considered a manual implementation of what the GMG software does automatically? Is it a lossless process? Aside from the creation of the "temporary" profile, could it be automated with a Photoshop action?

I am thinking that if it works as well as suggested, without any negative effects, it could be adopted as the finishing step for CMYK corrections. In addition it could be used by output service providers as a preventive measure with incoming image files for which the separation profile is available.

Al

 

[walterz]

Re: Multiple rounds of color correction color matching problems

Al asks:
"Can the procedure outlined in the post by Walter Zacharias be considered a manual implementation of what the GMG software does automatically?"
Yes and No. Colour servers like CMYK Optimizer and I assume GMG, can do this automatically but also can do way more than this. When I use the work "re-separate", I am referring to the profile-to-profile conversion of a CMYK image using basic tools like Photoshop. When you do this transform in Photoshop, the CMYK data gets converted behind the scenes to Lab, then back to CMYK. I call this re-separating as the data is being "separated" from a 3-colour space to a 4-colour space again. (Colour servers like CMYK Optimizer can do the CMYK-CMYK conversion without first going to a 3-colour space. This process would not produce the same effect - it will not necessarily restore the same balance we are considering in this thread.)

"Is it a lossless process?" Yes.

"Aside from the creation of the "temporary" profile, could it be automated with a Photoshop action?" Yes

As you say, so far this is fairly theoretical. I have been trying to think of how I could test and "prove" or verify this theory. If I exaggerate it becomes quite obvious. (Take a neutral gray of 38, 31, 30, 14. If you would edit the image so drastically that you have edited all the black out of it, the values could be 50, 40, 40, 0. Taking this and re-separating it back to the original GRACoL profile would restore the initial CMYK values of 38, 31, 30, 14.) So this seems to make sense. But the unaswered question still remains, at what point will the press begin to struggle? How much of a change is too much? How sensitive is this balance?

Walter

 

[walterz]

Re: Multiple rounds of color correction color matching problems

Al
"Have I understood your first paragraph above correctly?" Yes

"... in your first message you stated that the "CMYK -> Lab -> CMYK would negate the effects of the edits." Did you mean to say "the ill effects of the edits," and not the color appearance obtained by the edits?" Yes

"...toward the end of your second message you say that if a neutral gray of 38, 31, 30, 14 was edited in CMYK to 50, 40, 40, 0, then re-separating it back to the original GRACoL profile would restore the initial CMYK values of 38, 31, 30, 14. But then the desired color appearance from the edits is lost."

The effects of the colour edits is not lost, it is the balance that has been restored. If you have an image that would have a cement driveway in the image with a gray value as stated above. Then you perform some extreme editing which would cause the CMYK values to be well out of balance, like the 50, 40, 40, 0. The balance can be reestablished by the CMYK-Lab-CMYK transform without changing the colour. Both 38, 31, 30, 14 and 50, 40, 40, 0 values relative to the GRACoL profile produce the same appearance L=60, a=0, b=0. That is why I chose to use these values. I cannot practically see how someone would or could edit the colour so significantly that this would ever happen. But like I prefaced the discussion, the theory seems plausable if one considers extreme examples.

Walter

 

[Guest]

Re: Multiple rounds of color correction color matching problems

Walter,

Thanks for clarifying that.

>(Colour servers like CMYK Optimizer can do the CMYK-CMYK conversion without first going to a 3-colour space. This process would not produce the same effect - it will not necessarily restore the same balance we are considering in this thread.)

So Optimizer restores some other neutral balance than the re-separating method. Is that a concern? As I understood the Juergen Roesch 12/21/07 post in the PAB list, and Michael's 2/6 post above, what is important is that the neutrals for various images on the same press sheet be "harmonized" so that they respond similarly to on press adjustments.

Returning to the re-separatig method, in your first post in this thread, you suggested generating another profile with a different name from the same dataset as the current working space to be assigned temporarily to the image, followed by a profile to profile conversion to the original working space in order to cause a CMYK to LAB to CMYK transformation that would bring all similar shades into numeric CMYK balance again.

Can't this temporary profile be more easily created by renaming the profile of the working space with the ColorSync Utility on the Mac, or ColorThink from Chromix for the Mac or the PC?

Thanks,

Al

Edited by: Al Ferrari on Feb 8, 2008 11:25 PM

 

[dremaley]

Re: Multiple rounds of color correction color matching problems

Color correction and change depends on where it is in the color scale. Example, browns, purples and grays are problem colors at press - but why? It's because their C-M-Y values are near the midtone values (40-50-60%).
For many years I managed a high end color trade house, we made 3-4% changes all the time, however, the final files were converted to high amounts of GCR so the printer could not change the color.
Using GCR pushes the C-M-Y values away from the midtone area towards the saturated and de-saturated areas, and what do we know about saturated and de-saturated areas on press? They don't change, midtone changes and thats in the black printer now.
Send me a email and I'll send a pdf of my PC Reference brochure that's used in my training program, please share. . . see attached item on gray balance.

Dan Remaley

 

[michaelejahn]

Re: Multiple rounds of color correction color matching problems

@ Don Isbell

Wow. It is wonderful to see someone advocating a PDF/X-4 workflow. I got a little confused by your explanation of CMYK, as you (at the begining say);

"CMYK is what I would call a 'generic' colorspace"

then later say;

"CMYK is not generic"

Without getting into a protracted semantics based argument on what CMYK "is" - lets agree that as all digital proofing devices, all monitors and nearly any color output device will NOT be using the same exact imaging approach or colorant set as is used on a CMKY printing press.

So, that being said, the important point here is that either;

1. You talk to your print service provider, and ask THEM what THEY would like to assure that you give them what you need to get what you are expecting.

Most often, this would mean PDFX-1a - or a CMKY PDF with an output intenet and all objects that use transparency "flattened"

This is Standard Operating Procedure with Magazines where they require that you 'follow the rules' so they can accept color ads and not be involved with conversion processes that happen outside of their control - I bring this up as color is complex to exchange, and without conversion BEFORE exchange, many things can go wrong.

here is a typical web site where you can see this is fact;

http://direct2time.timeinc.com/

click over one of the magazine titles - Sports illustrated for example - and you will see that besides insiting on CMYK, there are a few other important things they require;

PDF/X1a SPECS FILE FORMAT

PDF/X1a format is the preferred file format for file submission. Careful attention must be paid to the proper creation of PDF/X1a files to ensure that they will reproduce correctly.
• PDF/X1a file saved with only one ad per file.
• All High-resolution images and fonts must be included when the PDF/X1a file is saved
• Only use Type 1 Fonts - No TrueType fonts or Font Substitutions
• Include all images, which must be SWOP TR 005 and at minimum 200 dpi TIFF or EPS format files.
• Total area density should not exceed SWOP 300% TAC
• Images should be CMYK or grayscale only. Images. Do not nest EPS files into other EPS files. Do not embed ICC profiles within images
• All required image trapping must be included in the file(s)
• Right Reading, Portrait Mode, 100% Size, No Rotations
• Standard Trim, Bleed and Center Marks in all separations, 1/2" outside Trim (No marks included in the "live" image area)
• Maximum dimensions, including all Marks, Color Bars, and Agency info:
Single page = 11" wide by 17" high
Spread page = 22" wide by 17" high
• Digital proof (Color or B&W) which represents the final digital file
• For color guidance on press—Digital, Press, or Off-press proofs which meet SWOP specifications (call magazine production department for number required)
• Include a 6 mm 5%, 25%, 50%, 75%, 100% C, M, Y, K patch strip for quality control - You can download an EPS file containing this strip from our web site

So, I will ALWAYS advocate that you work with your printer to find out what they need. If they do not know, or you are working through some third party agency and do not have direct contact with the service provider, demand to know what "Printing condition" is to be used, because as Don Isabell explained "not all CMYK is equal"

From the GHent PDF Workgroup;

"Whenever a document that will be printed is created, the designer should have a specific printing condition in mind so that images and specific tints such as those in logos can be correctly prepared. The idea of a printing condition encapsulates both the print technology to be used (sheet or web offset, flexo, gravure, etc), and the media that it will be printed on. It therefore includes data about tone value increase (TVI, or dot gain) and total area coverage (TAC), as well as the actual color that will be printed for any combination of the process inks. Common printing conditions include SWOP (CGATS TR001), Fogra27, etc.


If that document is submitted for printing as a PDF/X file the intended printing condition must be recorded in the file as an “output intent” structure."

So, while I appreciate the flexibilities that a highly color managed workflow offers - I tend to think it is best to learn your print condition, set up your files AND applications to enable you to SIMULATE they print condition, and you will NEVER be surprised.

if this is not followed, you will continue to live in the chaos you are currently experiencing.

 

[disbellj]

Re: Multiple rounds of color correction color matching problems

I agree with Michael Jahn.

Having said that, when I get PDF/X-1a in, I still am not going to print it on SWOP paper. It's going to print on #1, #2, or #4 paper.
When I get native files instead of PDF, I get as many images without a profile embedded as I do get images with a profile embedded. Which one? SWOP. No, it's not the right profile to use, but until Adobe changes it or my boss decides to tell customers to use a different profile, then it is what it is.

I say PDF/X-4 because of edit-ability, which is taken away mostly when I receive a PDF/X-1a or PDF/X-3 - because transparency has been flattened. Do I care if the CMYK is color managed? Not at this point. If it is, it is. If it's not, it's not. As long as it's done correctly if it is done as all, then so be it. For here, I would test PDF/X-4 color conversions in my rip before I let it automatically do conversions to incoming CMYK to production files.

But since I'm working at entering another career, and my boss shows no signs of changing anything, and we're supposedly on the brink of recession, I'll likely be retired from prepress by the time I'd have to do anything different than I am now anyways.

Don