cmyk neutral grayscale values ?

[Tim Lookingbill]

So the CMYK recipe calculated by photoshop is the best or there are other options?

Yes, you can work with any hungry printer with a commercial press who'll guarantee neutral CMYK grayscale with no color crossover errors and provide the instructions for you to aid in helping them reach YOUR goal. Have them sign an agreement indicating what your goals are and if they don't agree, go seek another hungry printer who will. It's not like commercial press technology is in its infancy.

These guys know how to get the job done as long as you pay them and from the list of printers available shown in Mark's linked PDF those and many others will want your business and bend over backwards to reach your goal of a CMYK neutral grayscale by any means necessary.

The "Process Color Manual" (corrected the name) result I posted above is sort of like a canary in the grayscale coal mine of printing "Standards" in that you can check the Photoshop numbers against the swatches to see how much variances you get from a random sampling off a commercial press that didn't come up with the swatches using an ICC profile.

Those swatches are pretty much like a "Raw" dump of CMYK color number combinations that present a neutral appearance comparing against a Photoshop ICC based recipe. The fact that one mid gray swatch located out of 24,000 from a Photoshop ICC profile number recipe would look neutral not knowing the variables (which you seem to be trying to compensate for unnecessarily) that generated it, shows "Standards" are pretty easy to attain by the industry that's been printing CMYK color over 50 years. You'ld think these guys would know how to produce a decent CMYK grayscale and they do, so you shouldn't be so concerned about it on the front end.

Just work with a printer who'll listen to your concerns and sign an agreement indicating what you want.

[smilem]

@MarkM
@Tim Lookingbill

I agree that commercial print shops knows what they are doing (most of them anyway), but to print correctly for them means to send correct data for me.

I checked the idealliance Proof Verifier V1.5 and they seem change the reference values for SWOP, gracol etc. That confirms you must send different data for each colorspace to get neutral gray, or lets say color if we are talking about color patches.

So there seems to be 2 ways:

1st. You must compensate your reference values for SWOP, Gracol like idealliance does by adjusting reference values since the test image is the same.

2nd. You must send different values to begin with. For every standard SWOP, Gracol you make different values that leads to same expected reference values.

I wonder why idealliance preferred the 1st. way of doing things since preparing neutral gray for a given color space as we came to this conclusion is rather easy, you use photoshop for this with CMYK working space set to the one you want to make grayscale for.

@digitaldog

I know LAB is not sullution since most commercial print shops will not accept LAB data files to begin with, but if a certain RIP or Photoshop for that matter opens LAB file it converts it CMYK AFAIK. The file is never converted to RGB.

[digitaldog]

I know LAB is not sullution since most commercial print shops will not accept LAB data files to begin with, but if a certain RIP or Photoshop for that matter opens LAB file it converts it CMYK AFAIK. The file is never converted to RGB.

Lab isn't defining an output space value(s) without the output profile. IOW, Lab is not a destination, no RIP, Photoshop or other software deciphers the output without a descriptor from Lab to output. Again, Lab doesn't help you at all.

[smilem]

Lab isn't defining an output space value(s) without the output profile. IOW, Lab is not a destination, no RIP, Photoshop or other software deciphers the output without a descriptor from Lab to output. Again, Lab doesn't help you at all.

But LAB makes me have only one file for data not 3 for every CMYK color space. I understand that the output is handled by ICC profile for that printer/ink/paper.

[Czornyj]

But LAB makes me have only one file for data not 3 for every CMYK color space. I understand that the output is handled by ICC profile for that printer/ink/paper.

So is RGB. Why bother to convert RGB to L*a*b? Unlike L*a*b, synthetic RGB is a working space, it's supported by all applications like PS, ID, Ai, easy to edit and so on.

[smilem]

Anyone knows proper custom photoshop CMYK colorspace settings to convert LAB to CMY like
ECI bvdm Gray Control Strip 2012 does ?

I tried to use Eurostandard Ink colors with K set to zero and I get L60 instead of L80 for 30% patch when entering C27 M19 Y20 K0? Any ideas? I'm looking to make 20 patch (every5%) chart with CMY inks that would be the same as K only, and the ECI has only 3 patches at 30%/50%/70%

Perhaps such chart already exists?

[Czornyj]

Here's my F39L based profile created with no K
https://dl.dropboxusercontent.com/u/19059944/FOGRA39L.icc

[smilem]

Thank you. Your profile is closer:

your profile input C27 M19 Y20 K0 --> L79 A0 B-1 should be L80 A0 B-1

L80 A0 B-1 --> C26 M18 Y19 K0 should be C27 M19 Y20
L80 A0 B0 --> C25 M19 Y20 K0 should be C27

is it possible to adjust your profile so that it would match with ECI? Seems the color error is non linear.

[Czornyj]

ECI ISO Coated v2 profile is build on the same data set (F39L), but with different profiler (Heidelberg Print Open) - my CMY profile was created with i1Profiler, so there might be a slight difference caused by different colour engines, rounding errors and so on.

[smilem]

How did you create no K profile from forgra39? MY friend has heidelberg color toolbox (don't know what version) would that work for creating better profile (I don't know anyone with printopen )?

[Czornyj]

How did you create no K profile from forgra39? MY friend has heidelberg color toolbox (don't know what version) would that work for creating better profile (I don't know anyone with printopen )?

I don't know Color Toolbox, but usually there's some option to cut off K:


I'm not sure which of PrintOpen and i1Profiler created profiles is "better", but quite sure that the difference is so small, it virtually doesn't matter

[smilem]

I decided why not email ECI to ask for fogra39L based no K profile. After all they should have one.

After emailing them at  [email protected] I got a reply that:

Technical details of permanent failure:
Google tried to deliver your message, but it was rejected by the server for the recipient domain eci.org by mxlb.ispgateway.de. [80.67.18.126].

The error that the other server returned was:
554 Sorry, no mailbox here by that name.

ECI doesn't even have mailbox configured at that address.  Why bother listing one then  

I wrote to another contact on their page: Public relations: Karl Michael Meinecke asking them:

Hello, you provide only only 3 patches 30/50/70

I would like to create my own test strip with different values. Could you send me the ICC profile (with no K ink) so that I could calculate values for other patches?

Thank you.

Is my english that bad   The reply is:

If  you go to the download area on ECI website, you can get the complete package, including documentation
http://www.eci.org/_media/downloads/eci_bvdm_gray_control_strip_2012/eci_bvdm_gray_control_strip_2012.zip

Seems like in Germany nobody understands English, even sellers from china reply better. I tried to explain myself better:

I have downloaded the color strip, but I want to make more patches in CMY that maches patches with K ink. For that I need fogra 39L based ICC profiler created with heidelberg print open (as this is what is used for fogra profile creation right?)

I tried to create a profile with i1profiler but the numbers doesn't match as it uses different engine.

If you could just send me a ICC profiler that was used to calculate CMY patches for the test strip it is all that I need. Thank you.

Reply is even more strange:

Hi,
you can Download the caracterisation data on Fogra.org and the ECI profiles on eci.org . This is all you need. The values of ECI/bvdm graycon patches are taken from the respective caracterisation data set patches.

What next email other ECI contact like Vice chairman, Chairman do they even know their employees are not communicating with each other because it looks like the response is from incompetent employee. All other companies at lest forward email to the right person, not the case with ECI.

I can't contact Board: Darrian Young, Color & Image Consulting as no email is listed 

[Czornyj]

Try converting with BPC off instead, it's almost there when you turn it off.

ECI ISO/PSO profiles were created by Heidelberg, that is a member of ECI. It's unlikely that any other ECI member has/will send you "no K" variant of such profile.

[smilem]

Well, after sending a spreadsheed to the ECI they understood what I want. They sent me instruction to just use Isocoated v2 profile with Xrite colorpicker then set zero for K and uncheck it. This way no special profiles are required.

I tried this but since colorpicker uses logo engine the results were not optimal. I got a CMY profile created by a friend. Then we made some color conversions and behold, here is a list of calculated colors !

I hope no mistakes (there could be most of the numbers were entered by hand), so feel free so post errors so I can correct it.
I also attach pixels.tif for target creation.

A also want to ask how to create CGATS? do I enter absolute data or relative colorimetric as that is what PS uses by defaut?

[smilem]

So can anyone advise me on CGTAS data ?
Do I need 3 different files each for rendering intent, Absolute, Perceptual, Relative?