Multiple rounds of color correction color matching problems

G

Guest

Guest
Hi Walter,

Although my question was about your earlier post, I addressed it to Michael to get the view of another re-separating/re-purposing adherent. No disrespect was intended. But thank you for your positive response.

Have I understood your first paragraph above correctly?: When applied to one CMYK image, the "Optimizer" process may produce some what different values for the same CMYK pixel as the "re-separating" process, but when each is applied to a set of several different images, each process leaves the set of images not only with the same color appearance, but with neutral balances that will respond very similarly to adjustments on press when that set of images are on the same page or press sheet. I am hoping this is what you meant.

But in your first message you stated that the "CMYK -> Lab -> CMYK would negate the effects of the edits." Did you meant to say "the ill effects of the edits," and not the color appearance obtained by the edits? Otherwise, why bother? However, 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. Then this is not a work around with results similar to "Optimizer" process.

This seems contradicting. I must not be understanding this properly. Can Walter or someone else please clarify.

Thanks,

Al
 
Multiple rounds of color correction color matching problems

Hi Folks!

We are an art printer. We have been experiencing something strange. It seems when we get an Epson proof from one of our clients and their has been multiple rounds of color correction in photoshop, we are way off on color when we get on press. (As opposed to receiving the same with only say 2-4 rounds of color correction - where we do pretty well with achieving color...)

Has anyone ever seen this before?

Epson 9800
Current Mac OS
CS3 or CS2
Heidelberg Sheetfed Press
Apogee X light RIP
Film
Scans can be done on drum scanner or flat bed camera.
 
Re: Multiple rounds of color correction color matching problems

THe first thing to do is to make sure that your proofs match your presses and stop it getting to your presses.

If you are cute (mmm) make sure your proofs and your presses run to a standard (in the US G7) - and vailidate the proofs with a label (if you dont know how to do this then G7 will tell u)

So then client sends file - you proof to G7 standard with a label on saying its to standard - clients looks at label and sorts file out


Peter
 
Re: Multiple rounds of color correction color matching problems

I have seen this before. I can't explain it either. When I have an artprint that only goes through 2 or 3 rounds of correction, it prints fine. When I get up to round 4 or above in Photoshop, the proof looks perfect, but doesn't behave well on press.

The proofs and press are all calibrated, etc. etc. blah blah blah, and we are running Epson 9800's with GMG behind them.

If anybody can figure it out, I'd love to hear it also.
 
Re: Multiple rounds of color correction color matching problems

I am so glad to read someone else has this same issue. Any other folks experience this? We just can not get a handle on it and is driving us nuts here! It could be almost any color and any size file.


BTW what is GMG?

Edited by: centralnj on Feb 1, 2008 1:25 PM
 
Re: Multiple rounds of color correction color matching problems

A while back in December on the Printing Across Borders mailing list, Juergen Rosch (of GMG) wrote a post regarding standardizing separations that might speak to the issues you're referring to. Basically stating that there are multiple CMYK combinations that will result in the same visual result and will reproduce consistently on a proof, but perhaps inconsistently on press. His example referred to a solid gray patch that was converted to several different GCR/black generation settings and output side by side. On proof, they were all identical, but on press, they were visually different, and adjusting one of them toward neutral resulted in the others being casted.

A potential solution is in the workflow, either moving toward a late binding/RGB workflow (corrections done in RGB and only converted upon output), or incoporating a tool like Alwan or GMG Ink Optimzer to effectively "homogenize" the separations prior to going to press. More and more CMYK retouching is done visually rather than "by the numbers" , and this could lead to inconsistency. I couldn't tell you if this is the issue you all have experienced, but I can say that I've noticed individually CMYK retouched images that look similar on a proof appear visually different when layed up on a press form and printed. Non-homogenous separations can be problematic for a press operator who might need to adjust inking to hit color in one page of a form, while the others fall further from ideal.
 
Re: Multiple rounds of color correction color matching problems

This makes perfect sense to me. Different proofs could look the same but not the press prints.

The printing of CMYK on a press is a physical phenomena. CMYK is not a colour space. The CMY and K parameters are non linear and they are also not independent. What this means is that any attempt to control the colour outcome via curves is not sufficient to solve the problem.

I does not matter if industry Standards or calibration methods have been made based on tone curves of some kind. Curves of any kind will not solve the problem. Adjusting to some standard dot gain via curves will not solve the problem. Using a generic profile will not solve the problem either.

Of course adjusting a separation with curves helps but it can not be counted on to be highly predictable. The press prints look different because of the non linearity and non independent nature of printing CMYK dots and solids on press, while an ink jet proofer prints CMYK in a much more linear and independent way and therefore the results are more predictable.

With the increase in computing power, methods can be developed that make the process much easier and much more predictable and it would not use curves. The use of curves is an artifact of past technological approaches and needs to be eventually put aside.

In this industry, that will probably be a very long time from now.
 
Re: Multiple rounds of color correction color matching problems

Hi Centraj
The thing that worries me is you receive proofs 'from one of our clients...' What are THEY using to make the proofs and don't say 'Epson' because there is no such thing as an 'Epson' proof! It's only a proof if it's profiles and created to a TARGET. This usually means that the Epson 'proofer' is driven by advanced proofing software such as GMG, Oris, EFI Colorproof, Absolute proof, Digital Matchprint etc etc.
I would suggest the problem, whilst there definitely is a quirk in photoshop when you do multiple colour corrections, is really in the PROCESS you are using to achieve contract signed-off colour.
Sorry to say, Apogee X light on its own is not enough.
The proofs from the client...do they have a spectrophotometer-readable colour bar?
It's the old story which crops up here so often. Without profiling to a specified target and knowing the characteristics of the press the job is going to be printed on, you are leaving yourself open to colour malfunctions.
If you change you process to one where YOU supply the proof back to the client and make adjustments according to mark-up, you will not only improve your service level to the client, but extend the lives of your ertswhile pressmen and women!
Hope this helps, I realise it's radical.
 
Re: Multiple rounds of color correction color matching problems

> whilst there definitely is a quirk in photoshop when you do multiple colour corrections

Putting aside particular problems in press printing methods AND proofing methods, can you please expand on this quirk (malfunction) in Photoshop when multiple colour corrections are done, or perhaps give specific examples? How can we reliably observe instances of this quirk?

Thanks,

Al Ferrari
 
Re: Multiple rounds of color correction color matching problems

ASM,
Client uses Kodak Proofing Software for Matchprint Inkjet / Veris version 3.1.5.23 on an Epson 9800. Can you elaborate on this software and what can be done to better our situation with it?

Mr. Eddington,
Can you post a link, perhaps, of that article you wrote of? "A while back in December on the Printing Across Borders mailing list, Juergen Rosch (of GMG)"
 
Re: Multiple rounds of color correction color matching problems

I think ASM, Michael & Erik all hit on the fundamental problem. It sounds like you and your client are making proofs to some kind of standard (like SWOP), though you don't explicitly state what. Obviously, it is very important that your proofing and your client's proofing match, and that they both come pretty close to your press.

You also don't state if the rounds of color correction are occurring in RGB or CMYK, but (as Michael implied), it sounds like they are happening in CMYK. Here is my theory, which is essentially just re-stating Michael's point: When the image is initially converted to CMYK, it is probably being separated to a SWOP profile (U.S Web Coated (SWOP) v2 being the default in photoshop) and since you guys are probably proofing and printing near SWOP, this workflow "works" for a couple rounds of color. The separations that were initially created are built to work on a SWOP press and proofing scenario. However, as more rounds of correction happen, you are curving the different channels in more extreme ways, and you are moving your separations AWAY from SWOP. As Erik said, press inks behave in a non-linear fashion while inkjet proofing is relatively linear.

What is happening is that the color corrections are in response to the PROOF, and because you (or your client) are working in CMYK , the channels are getting curved to get good results on the PROOF, which doesn't have the same CMYK response as a PRESS. In other words, if you had infinite time and money, and you could color correct the image by running press runs for each round, and then making CMYK corrections based on what you saw in those press runs you would probably find that the moves you make are different from what you would make in response to the PROOF. If there are only a couple rounds of color correction, and/or the moves aren't that big overall, there will not be much difference between these two methods of color correction. However, once you get into lots of rounds, the moves become more tailored to matching the PROOF, and less related to what will happen on PRESS.

Erik stated it quite succinctly: "The printing of CMYK on a press is a physical phenomena. CMYK is not a colour space." Michael stated your solution: "...moving toward a late binding/RGB workflow (corrections done in RGB and only converted upon output)" The key is for you (or your client) to do all your rounds of color correction in RGB , and then convert to CMYK before output. If you convert to a CMYK profile that comes close to your actual press conditions, then those separations will visually match between your proof and press. By doing color corrections AFTER the conversion, you are degrading the ability for your proof and press to match.

Hope this makes some sense!

-Todd Shirley
 
Re: Multiple rounds of color correction color matching problems

I want to thank all of you folks for your input with this! It has been interesting learning some new things here and has been helpful.

May I ask one more question here? Some information written in this thread says CMYK is not a color space, which now looking into it, I can not find anywhere that it IS. We work in CMYK due to us being a printer, not RGB. If we were to keep the scan RGB but change the mode to CMYK to apply curves, USM, etc... to achieve desired color, change the mode back to RGB, then do a save would this help in any way? Hence we are making changes in a different mode, but never saving it as CMYK. We can then allow our Kodak Proofing Software for Matchprint Inkjet / Veris version 3.1.5.23 on an Epson 9800 or our RIP (Apogee X light) convert to CMYK for us.

Do you think this will help us in any way and make our proofs more press predictable, stable?
 
Re: Multiple rounds of color correction color matching problems

In short - no, don't do it.

First, CMYK is what I would call a 'generic' colorspace. But since that's misleading in this day, I'll try to explain it a little better. Let's say I have two devices. Those devices output CMYK. One device prints, and I find that it's results are different from the other device - even though I have printed the same file to both devices. The reason is that both devices print differently, and so I would need to actually create the CMYK for the intended device, and then use color management to convert the actual numbers in the file when sent to the second device, to make the results look the same between the two. A press and an inkjet are two such devices. You would build for the press using the correct press profile (which describes the paper/ink combo used on that press and contains the visual information of how the device prints with this ink and paper). Then you would convert that press CMYK to inkjet printer/proofer CMYK and get totally different CMYK numbers - but when those different numbers are printed on the inkjet, it shows how the press sheet will look. Now the press CMYK is not just CMYK colorspace, it's "GRACoL2006_Coated1v2" colorspace. And the proofer CMYK is not just CMYK colorspace, it's "your_custom_proofer_profile" CMYK (or CMYKcmk, etc.). CMYK is not generic. It should be made for the device it prints on.

Now to InDesign. If keeping as RGB, it's for the purpose of converting to CMYK to print. You can even keep the RGB in the PDF you export from InDesign if doing a PDF/X-3 or PDF/X-4 workflow. The RGB stays RGB when placed within InDesign. The CMYK is done at the lastest it can be in the workflow. This is late-binding workflow, which means the files get prepared for final output late in the workflow (like at the rip).

The ONLY reason you'd want to output ALL RGB from this InDesign doc (which probably contains at least some CMYK elements) is if going to the web, in which case you'd *TEMPORARILY* change your Blending Space in the InDesign doc to RGB, export a standard PDF for the web, and change the Blending Space (under Edit menu) back to CMYK. Of course, with a late binding workflow, this could be also done in the rip, so would not be necessary from InDesign.

The goal = One PDF/X-4 that contains fully embedded fonts, live transparency and blend modes, all profiles included for images and document, and of course full images (masked if needed) included in the PDF/X-4. Then the workflow would need a PDF editor (like Neo), and with a workflow or via PDF editor, this one PDF could be repurposed for output to many different devices (even sRGB IEC61966-2.1 for the internet).

Don
 
Re: Multiple rounds of color correction color matching problems

Hi All,
Sorry for late reply - I'm in a far away timezone. Al - the Photoshop 'Quirk' I think was eloquently theorised by Todd Shirley. I can't pinpoint is from a software code-writer's viewpoint but all I can add is that my designers all say 'the more colour corrections you do in PhotoShop, the less reliable the proof match to the printed result.'
At the beginning, our friend here said he is an art printer. This is mission-critical colour and I do believe a case for fingerprinting the press. Someone mentioned the luxury of running sheets on the press at every stage of colour correction and re-proofing. This is essentially what you do to fingerprint a press so you'll need to have it at your disposal for half a day in my experience. Make sure you let the sheets dry before measuring colour.
Once fingerprinted, so long as you don't change anything major like ink brand or proofing profile/s - you should not need to keep doing this. One of the best press fingerprinting methods I have seen is Mellow Colour from the UK and (I think) handled by Chromaticity in the USA. Mellow Colour uses ISO 12647 as the standard and this is gaining global recognition. www.mellowcolour.com (colour with the 'u')
The proofing software your client is using - MatchPrint/Veris - should be capable of working in with a press fingerprint, but the two ends have to talk to each other.
This is a great opportunity to do what so many printers avoid - getting the creatives into the presshall and working with the pressmen to build a closed-loop colour managed workflow that eleiminates all the variables.
By the way if it is SWOP space your proofs are made to, as an art printer you need to change. SWOP is Standard for Web Offset Printing - aimed at maximising the limited gamut of heatset web presses. Your gamut map space (hope you've mapped it) must be much wider than SWOP.
 
Re: Multiple rounds of color correction color matching problems

>I can't pinpoint is from a software code-writer's viewpoint but all I can add is that my designers all say 'the more colour corrections you do in PhotoShop, the less reliable the proof match to the printed result.'

Thanks for responding to my post ASM, but does that mean that if the same number of color corrections are done using another image editing application the problem does not occur?

Al
 
Re: Multiple rounds of color correction color matching problems

Hi Al - Yes it does. If the Photoshop/PDF/CS2 file is parsed in a high-end proofing app such as GMG's ColorProof or Oris ColorTuner, you can edit to your heart's content. It's because these 2 apps actually enable editing of the ICC profile I believe. They have some unique CIE L*ab*-based capabilities in there in the CM kernel and I heard that Adobe would just love to get their hands on them!
 
Re: Multiple rounds of color correction color matching problems

In your experiences, what happens if the CMYK file with multiple edits is "re-separated"? Does the CMYK -> Lab -> CMYK negate the effects of the edits?

In my opion it would. If I am right, then this could be a workaround for the jobs with many rounds of proof corrections. After the 2nd or 3rd correction, take and reseparate the image. My suggestion would be to generate another profile with a different name from the same dataset as your working space. Now in Photoshop, assign this temporary profile to the image. Nothing will change. The colour will not shift. Then "convert to" the original working space. Photoshop will pass the colour values through the Lab colour space, and in my opinion, bring all similar shades into numeric CMYK balance again.

Walter
 
Re: Multiple rounds of color correction color matching problems

ASM,

Although it is correct that in any case the best profile to use for separation and color correction is the one that that shows how the press prints with that particular paper type and ink...

... the gamut of the SWOP separation, when printed on #1 paper, is about the same as GRACoL2006_Coated1v2 when printed on #1 paper. Since separations are done using Relative Colorimetric Intent, then the SWOP paper is not taken into account.

Do a test to prove it:
Convert the same RGB to both U.S. Web Coated (SWOP) v2 and GRACoL2006_Coated1v2. Print or proof both on GRACoL2006_Coated1v2 and compare. Are they much different? No. The blue might be a little darker, but not much.

Since the gamut is made mostly by the paper, when NOT printing a SWOP separation on SWOP paper, the gamut is not SWOP.

Do another test for uncoated:
Convert the same RGB to both U.S. Web Coated (SWOP) v2 and ISOuncoated. Print or proof both on ISOuncoated and compare. Are they much different? No. The only difference that stands out is the blue going to purple issue when using the ISOuncoated profile (which is why I wouldn't use ISOuncoated to separate even if printing on uncoated).

Don
 
Re: Multiple rounds of color correction color matching problems

Walter,

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.

IMO repurposing after doing color correction is only adding confusion. It may work, yes (if the customer used the original CMYK profile to edit with, and they did not use their press profile to soft-proof with while editing). If one has a workflow where they repurpose, that would need to be done before the proof is printed. So if they need to come back to do color correction, they'll be doing it on a file in the correct colorspace (the press profile's colorspace). Either that or just leave the CMYK as is, and soft-proof with the correct output profile if color correction needs done after the first proof.

Can anyone see that when soft-proofing the CMYK (whether SWOP, GRACoL2006_Coated1v2, or custom) with the correct output profile, that what the person sees on-screen should be what the press prints, so it really shouldn't matter what CMYK that it is, as long as using the correct output profile to soft-proof the CMYK when doing color correction?

Don
 

PressWise

A 30-day Fix for Managed Chaos

As any print professional knows, printing can be managed chaos. Software that solves multiple problems and provides measurable and monetizable value has a direct impact on the bottom-line.

“We reduced order entry costs by about 40%.” Significant savings in a shop that turns about 500 jobs a month.


Learn how…….

   
Back
Top