Matching screen with proofs (FOGRA52)

[thomas911]

Hi guys,

We are using the following workflow: Calibrated screen Eizo CG277 (5000K, 80 cd/m2, gamma 2,2) -> exporting pdfx4 from InDesign with FOGRA52 > Printing on an Epson SureColor P500 via GMG ColorProof that is calibrated for FOGRA52 (paper GMG OBA Matte 150). The print is measured and passed the FOGRA52 standard. The problem is that the print has quite a bit more red than the screen...

Any ideas why they don't match better? The client is using the pictures for both print AND web, so if we compensate to make them look less red on the proof, they will look too green on the web... (unless we make two versions)
Should we trust the calibrated screen or the calibrated proof printer?

 

[Stephen Marsh]

The great monitor and calibration are not enough in itself. Keep in mind that this is a better monitor than most “web viewers” will be using and most of them will not have calibrated or profiled monitors, nor be in suitable viewing conditions etc.

How is the PDF being viewed (what software) and has the viewing software been correctly configured for softproofing?

A good introductory guide for softproofing can be found here:

https://www.fogra.org/en/downloads-for-softproof.html

The hardcopy that has been measured and passed verification would be the better “yardstick” IMHO. This is not to say that you can’t do this via a softproof, just that it takes more work.


Stephen Marsh

 

[thomas911]

Thanks for answering Stephen,
And thanks for the link, I will take a look. For softproof, we mainly just use PhotoShop with proof setup > fogra52. But if I view a pdf in Acrobat (with fogra52 preview) that's been exported with fogra52 in InDesign, it's identical to the image in Photoshop from what I can see.
We also make physical proofs using fogra47, and those are close to what we see on screen. But the fogra52s are redder.., but only on the physical proofs, and not on the softproof. Shouldn't a physical proof of f47 and f52 look more or less the same in terms of color? Both measured and passed for the respective standard.

 

[cementary]

thomas911 , for correct 47 proof one should use no-oba paper. For correct 52 proof - oba paper. So question number 1 is your paper correct? And question number 2 what are the viewing conditions?

 

[gordo]

thomas911 , for correct 47 proof one should use no-oba paper. For correct 52 proof - oba paper. So question number 1 is your paper correct? And question number 2 what are the viewing conditions?

And measured M1?

 

[thomas911]

thomas911 , for correct 47 proof one should use no-oba paper. For correct 52 proof - oba paper. So question number 1 is your paper correct? And question number 2 what are the viewing conditions?

Thanks. Yes we use no-oba paper for 47 proof and oba paper for the 52 proof. We use a ligh booth measured to 5000K and desktop lamps also at 5000K.

 

[cementary]

Thanks. Yes we use no-oba paper for 47 proof and oba paper for the 52 proof. We use a ligh booth measured to 5000K and desktop lamps also at 5000K.

What about uv content in your lamps? do they (lamps) conform 3664 ISO? And also gordo 's question — M1 for both or M0 for 47, and M1 for 52?

 

[Puch]

Your hardcopy proofs are the right color if they pass the certification. FOGRA 51 and 52 are unfortunately not that easy to softproof than their predecessors due to their OBA-aware white point definition. After reading a long argument on the Apple ColorSync mailing list (which I would suggest subscribing to, if you're into colormanagement) about a similar topic, I would say that we have to wait for some kind of official 'tweaking" or guidance from FOGRA to successfully employ these profiles in a practical prepress environment.

 

[Puch]

A follow up. Apparently the ICC conversion engine in Photoshop is not up to the job. I did a small test run using the ECI supplied device link profiles and Colorlogic ZePrA. I selected two relevant images from Roman16, converted them to FOGRA 52 and FOGRA 47 in ZePrA, then converted the resulting tagged CMYK images to sRGB (for display purposes) using Photoshop (Absolute Colorimetric). There are differences, but the images are very similar, and I think the FOGRA 52 versions, printed on today's OBA-stuffed uncoated papers would produce excellent results. I can supply the high-res images for anybody interested, just PM me, please.

 

[cementary]

A follow up. Apparently the ICC conversion engine in Photoshop is not up to the job. I did a small test run using the ECI supplied device link profiles and Colorlogic ZePrA. I selected two relevant images from Roman16, converted them to FOGRA 52 and FOGRA 47 in ZePrA, then converted the resulting tagged CMYK images to sRGB (for display purposes) using Photoshop (Absolute Colorimetric). There are differences, but the images are very similar, and I think the FOGRA 52 versions, printed on today's OBA-stuffed uncoated papers would produce excellent results. I can supply the high-res images for anybody interested, just PM me, please.

I've read your post several times. And I still can't get it. What were you expecting? Similar images? How so? WP of 52 and 47 are not the same.
Also why so many conversion steps? Why not just take eciRGB of Roman16 and convert to needed profile right in photoshop?

 

[Puch]

I've read your post several times. And I still can't get it. What were you expecting? Similar images? How so? WP of 52 and 47 are not the same.
Also why so many conversion steps? Why not just take eciRGB of Roman16 and convert to needed profile right in photoshop?

OK, I was in a hurry. I'm trying to elaborate.

The whole point is to prove that the color conversion engine in Photoshop – in it's current state – can't make a proper conversion to the new FOGRA 52 profile. It's producing way too much magenta in the separation, that's the phenomenon the original poster was writing about.

Everybody expected – including me – that after the very successful and useable FOGRA 39, 45, 46 and 47 the new profiles will be a 'drop in' replacement. The offset printing technology didn't change too much, so why would the profiles describing the process change? According to FOGRA, the big difference is in the definition of the white point; the new profiles reflect today's OBA-rich papers, enabling a better color reproduction through the process. Eg. FOGRA 39 isn't accurate anymore since the white point representation in a softproofing scenario (or on a hardcopy proof) is very warm in comparison to a real coated stock. Enter FOGRA 51 & 52 with a definitely more blueish white point.

People try to use these new profiles like the old ones: to convert images directly from RGB to CMYK, using Photoshop. And they get inferior results. My experiment showed that – currently – we need to use some other tools and an intermediary color space to perform the color conversion. Using FOGRA 39 or the new eciCMYK as an intermediate color space, then converting to FOGRA 51 or 52 (by using device links) produce better results than going directly from RGB to CMYK in Photoshop.

I hope Adobe will address this problem in the near future, and we will be able to directly convert images with the PSO Coated v3 and PSO Uncoated v3 profiles in Photoshop.

 

[cementary]

I hope Adobe will address this problem in the near future, and we will be able to directly convert images with the PSO Coated v3 and PSO Uncoated v3 profiles in Photoshop.

IMHO, Adobe color conversion engine works great.
I think that maybe you do not fully understand how color management work.
* You take one RGB image, convert it to ISO_coated_v2 profile — you get some CMYK values.
* You take same RGB image, convert it to PSO_coated_v3 profile — you get some other CMYK values.
Next, you convert those images in RGB via Absolute colorimetric, which takes into account WP of source profiles.
The results would be different as they supposed to be.

That's how ICC profiles work.

P.S. I honestly believe that "new" eciCMYK is pure evil just because people do not understand how all this stuff works
I can agree only on one of your points — device-links from eci are not good enough. Want best DVL — create them by yourself