Problems between MAC & Win - Color Management

[Doug Gray]

Hi Abdo, 1728 patches is actually OK. One can make good profiles with even less. Ethan Hansen published a very interesting post some time back dealing with appropriate patch numbers - more is not necessarily better - and identified the various patch configurations he has tested that make very good profiles. Worthwhile having a look at that. With an iSis it doesn't matter much because it's so automated, but anyhow, just saying the patch numbers probably does not explain issues you had with the canned profiles. That said, with good equipment and if you are doing it correctly a custom profile should anyhow be preferable. The comparison between routes 1, 3 and 4 should be interesting.

I would suggest trying all three with the default-Isis target. It has 957 patches and fits on a single US letter size page. Make 2 prints with each of the three which will allow you to compare print to print variation. Then use I1Profilers chart comparison tool to get metrics on how much variation occurs between them.

[Abdo]

Good Morning !

Here's a file, 2 charts

Canson Rag 310g Paper
Chart: 809 patches
2 Charts - I1 profiler and Print Studio Pro
2 Archives XMF
1 TXF Chart

Who can say thank you.

I'll see if I can still do the same on Windows today.

Thanks

[Abdo]

attach Win - generate I1profiler + color management off.

[Mark D Segal]

Good Morning !

Here's a file, 2 charts

Canson Rag 310g Paper
Chart: 809 patches
2 Charts - I1 profiler and Print Studio Pro
2 Archives XMF
1 TXF Chart

Who can say thank you.

I'll see if I can still do the same on Windows today.

Thanks

What am I supposed to do with this and the attachment in the last post above?

[Abdo]

Mark,

I did 3 profiles.

Win - Made with I1Profiler with management off
Mac - made with i1profiler and other with Canon Print Studio Pro (color management off)

Unfortunately I can not put 1 zip on account of size.

[Mark D Segal]

You sent one MXF file, which is not a profile, it's a set of i1Profiler measurements and you sent one ICM which is a profile and only takes 2 MB, so with another post and another attachment you could send the other profile whether ICM or ICC, doesn't matter. But none of this is properly labelled to know what is what, so it's essentially unusable. Anyhow, what are we supposed to do with these files? What have you done with them to verify what differences they are revealing? If you are going to let us help make progress in this thread, the messaging will need a lot more clarity and precision.

[Abdo]

Attached file of the configuration used.

Is that what you need to be able to evaluate?

Thanks

[Mark D Segal]

Abdo - sorry, but totally useless and I don't have time to do all your work for you. You need to print the targets, construct the profiles and test them comparatively, then let us see what you come up with.

[Abdo]

Abdo - sorry, but totally useless and I don't have time to do all your work for you. You need to print the targets, construct the profiles and test them comparatively, then let us see what you come up with.

Sorry, Mark, I did not mean to waste your time on this.

I just put the files step by step how I'm running, if that does not have to analyze, I apologize.

I'll still print with these new profiles to see the differences in impressions.

Thanks for your help.

[Mark D Segal]

OK Abdo fair enough. Don't get me wrong - another participant in this discussion may find those useful, I don't know, I'm just responding on my own behalf. My basic suggestion to you is just to take the options you've identified and run through the whole cycle of printing the targets, making the profiles, printing known printer evaluation images with those profiles and then it will be useful and minimally time-consuming to see what works better than what. Whether we'll know why is another question, but one thing at a time.

[Doug Gray]

Abdo,

Here's a quick view of the gamut boundaries of the 3 profiles and sRGB at L*=50.

The one from print studio is quite stunted so there are some issues with the settings running it. No idea what. I don't have Macs or run Print Studio.

Note that the Rag 809 I1 and Rag 809 Win are very close and overlap. I'll run some more tests on this to see how close and will update this post in a bit but wanted to give a quick view now. As an aside, the two matching profiles are reasonably close, but slightly larger in shape to the profile generated for my Canon 9500 II (not shown) using the same paper.


Edited: Additional Info:

I've compared the mxf files and the Mac one "Rag Sec Cor I1 Profiler" and Windows one "Rag 809 Windos (Color Man. OFF)" match exactly like I would expect two target prints printed consecutively on the same system and setup. Further the gamut's have the appropriate shape and are reasonably similar to that on my Canon 9500 II using the same paper. I've added the results of the comparison to this post. You can do this yourself by using an "A-B" comparison available in I1Profiler under the "Analysis" tab on the left that is available when you use "advanced mode."

So a profile made from either of the two matching processes can be used with both Mac and Windows. What I find particularly interesting is that the files match so closely that I can safely say the Mac and Windows, and the device level, (color unmanaged, max gamut) printing algorithms match.

Added:
I added a comparison with the Mac Print Studio mxf file. Delta E 1976's are shown as dE2000 becomes quite unreliable for large errors.

[Stan Prevost]

ACPU has a known (to Adobe) problem of incorrectly resizing some outputs, which prevents scanning by iSIS.  Lower in priority to them than working on Photoshop print pipeline, even though it disables the functionality it was created to have to work around the removal of the option to turn off color management.

Turning off color management can cause Windows (or Photoshop, not sure) to apply a conversion to sRGB to data going to the printer driver.  An example of how this could occur is setting Photoshop (and maybe other apps) to "Printer Manages Color" and then in the printer driver setting "Color Management Off", in an attempt to work around the inability to apply "No Color Management" in Photoshop.

Latest version of iProfiler won't print.  XRite says it is a known problem to be fixed in next release. Have to print TIFF to file and use ACPU.  Interesting that XRite produces a product for which actually printing the test chart is a central feature doesn't place top priority on a bug that prevents it from doing so. the TIFF/ACPU workaround works but you still have to beware of the sRGB problem in Windows.

https://www.xrite.com/service-support/rgbprintertestchartwontprintini1profiler171onwindows
"RGB Printer Test Chart Won't Print in i1Profiler 1.7.1 on Windows
The RGB Printer test chart will not print in i1Profiler 1.7.1...this is a known bug reported to engineering,.The RGB Printer profiling “Print” button does not work in the Test Chart workflow step. This has been verified and logged as a software bug to fix in the next update of i1Profiler. In the meantime, you have a few options:
  1.Some users have reported that if they uninstall one of their printers (preferably one that you no longer use) and then reboot the PC, the “Print” button worked.
2.You can save the test chart out as a TIFF and then print it with all color management turned off.
3.You can uninstall i1Profiler 1.7.1 and X-Rite Device Services Manager in Control Panel > Programs & Features, reboot the PC, and then reinstall i1Profiler 1.6.7 using the following link
http://www.xritephoto.com/ph_product_overview.aspx?ID=1397&Action=Support&SoftwareID=1672 "



So with the ACPU resizing problem and the sRGB conversion issue and the nonprinting defect in iProfiler, and  none of these things afforded much publicity, it is no wonder that people are having problems.

Says Dave Polaschek of the Adobe Photoshop printing team (2015):

"Picking "Printer Manages Colors" on Windows guarantees that your color data will be converted to sRGB. We could adopt XPS printing to get around that, but there are many new and exciting bugs waiting to be found by the first application to head down that path.

On Mac, Picking "Vendor Color Management" and "Adobe RGB" does not mean the color data will automatically be converted to Adobe RGB. It means that untagged color data will be treated as Adobe RGB. When I last discussed this with the Apple printing engineers, I got a long explanation of when color conversions will happen, and the printer driver gets a chance to say to the OS "Yes, I can use color data in that color space" or "No, please convert that color data to XXXXX profile for me" (where XXXXX is the default profile for the printer, which you can set in ColorSync Utility.app).

When using "Printer Manages Colors" on Mac OS X, with a printer driver that says "Yes, send the color data as-is" when asked about what the application is offering to print, that data will be passed to the driver without conversion. I've done this with CIELAB data. I don't believe Epson's drivers will accept CIELAB, but when last I investigated (when the x900 printers shipped), they would accept ProPhoto RGB when using "Printer Manages Colors." There was only one conversion of the color data to the printer's final profile, and that was done by the driver.

All else being equal, fewer conversions between color spaces is preferable."

Dave Polaschek said (2015): "This always happens. When using "Printer Manages Colors" on Windows, Windows ICM 2.0 assumes that any color data passed to it is in sRGB, so Photoshop converts the colors to sRGB (so users don't get wrong colors when printing). We can't know whether the user has picked ABW or not (the driver-specific print settings are a black box to us at this point), and we must provide sRGB to the OS in this case."
 Dave's comments from:  http://forum.luminous-landscape.com/index.php?topic=122927.40
 This thread also mentions some printing differences between PS and LR.
 

 Adobe instructions for ACPU:  https://helpx.adobe.com/photoshop/kb/no-color-management-option-missing.html
 


John Nack of Adobe re ACPU resizing: 
http://blogs.adobe.com/jnack/2010/12/a-no-color-management-print-utility-for-photoshop.html
 [Thanks for bringing this to our attention. Apparently it’s a known issue, and we’d like to issue an update, but I don’t have an ETA for that happening (as enhancing Photoshop’s print engine needs to take priority). –J.]
 
 ACPU does no color conversions at all."
 
 Dave Polaschek — 10:24 AM on December 03, 2010 (Same link as John Nack above):
   
 "On Windows, the main difference between using the mode that Michael describes and this application is that with Printer Manages Colors, the image data is converted to sRGB before sending it to the Windows print system (because that’s a requirement of the print system). So if your target is tagged with a color profile (it really shouldn’t be, but some are), there will be a color conversion done by Photoshop.


-------
Stan

 

[Doug Gray]

ACPU has a known (to Adobe) problem of incorrectly resizing some outputs, which prevents scanning by iSIS.  Lower in priority to them than working on Photoshop print pipeline, even though it disables the functionality it was created to have to work around the removal of the option to turn off color management.

Yes, ACPU shrinks the print about 3% putting it right on the edge where ISIS may, or may not, be able to scan it

Turning off color management can cause Windows (or Photoshop, not sure) to apply a conversion to sRGB to data going to the printer driver.  An example of how this could occur is setting Photoshop (and maybe other apps) to "Printer Manages Color" and then in the printer driver setting "Color Management Off", in an attempt to work around the inability to apply "No Color Management" in Photoshop.

It depends on the settings. Windows does not always convert to sRGB when selecting printer manages color.

Latest version of iProfiler won't print.  XRite says it is a known problem to be fixed in next release. Have to print TIFF to file and use ACPU.  Interesting that XRite produces a product for which actually printing the test chart is a central feature doesn't place top priority on a bug that prevents it from doing so. the TIFF/ACPU workaround works but you still have to beware of the sRGB problem in Windows.

https://www.xrite.com/service-support/rgbprintertestchartwontprintini1profiler171onwindows
"RGB Printer Test Chart Won't Print in i1Profiler 1.7.1 on Windows
The RGB Printer test chart will not print in i1Profiler 1.7.1...this is a known bug reported to engineering,.The RGB Printer profiling “Print” button does not work in the Test Chart workflow step. This has been verified and logged as a software bug to fix in the next update of i1Profiler. In the meantime, you have a few options:
  1.Some users have reported that if they uninstall one of their printers (preferably one that you no longer use) and then reboot the PC, the “Print” button worked.
2.You can save the test chart out as a TIFF and then print it with all color management turned off.
3.You can uninstall i1Profiler 1.7.1 and X-Rite Device Services Manager in Control Panel > Programs & Features, reboot the PC, and then reinstall i1Profiler 1.6.7 using the following link
http://www.xritephoto.com/ph_product_overview.aspx?ID=1397&Action=Support&SoftwareID=1672 "

The latest version of I1Profiler fixes this problem. However, simply installing the latest version may not work. I had this problem on Windows 10 x64. It resolved after completely uninstalling it then re-installing.

So with the ACPU resizing problem and the sRGB conversion issue and the nonprinting defect in iProfiler, and  none of these things afforded much publicity, it is no wonder that people are having problems.

Says Dave Polaschek of the Adobe Photoshop printing team (2015):

"Picking "Printer Manages Colors" on Windows guarantees that your color data will be converted to sRGB. We could adopt XPS printing to get around that, but there are many new and exciting bugs waiting to be found by the first application to head down that path.

This is quite simply an incorrect statement. I and others here have noted that it is totally possible to print using "Printer Manages" an image in ProPhoto RGB and have it rendered selecting the appropriate profile in any of the 4 standard Intents using Windows. It has been brought to Dave's attention but he seems to have higher priorities. Understandable. Photoshop is a complex beast.

On Mac, Picking "Vendor Color Management" and "Adobe RGB" does not mean the color data will automatically be converted to Adobe RGB. It means that untagged color data will be treated as Adobe RGB. When I last discussed this with the Apple printing engineers, I got a long explanation of when color conversions will happen, and the printer driver gets a chance to say to the OS "Yes, I can use color data in that color space" or "No, please convert that color data to XXXXX profile for me" (where XXXXX is the default profile for the printer, which you can set in ColorSync Utility.app).

When using "Printer Manages Colors" on Mac OS X, with a printer driver that says "Yes, send the color data as-is" when asked about what the application is offering to print, that data will be passed to the driver without conversion. I've done this with CIELAB data. I don't believe Epson's drivers will accept CIELAB, but when last I investigated (when the x900 printers shipped), they would accept ProPhoto RGB when using "Printer Manages Colors." There was only one conversion of the color data to the printer's final profile, and that was done by the driver.

All else being equal, fewer conversions between color spaces is preferable."

Dave Polaschek said (2015): "This always happens. When using "Printer Manages Colors" on Windows, Windows ICM 2.0 assumes that any color data passed to it is in sRGB, so Photoshop converts the colors to sRGB (so users don't get wrong colors when printing).

The problem with a known expert is that false beliefs are more easily propagated. Quite unfortunate.

We can't know whether the user has picked ABW or not (the driver-specific print settings are a black box to us at this point), and we must provide sRGB to the OS in this case."
 Dave's comments from:  http://forum.luminous-landscape.com/index.php?topic=122927.40
 This thread also mentions some printing differences between PS and LR.
 

 Adobe instructions for ACPU:  https://helpx.adobe.com/photoshop/kb/no-color-management-option-missing.html
 


John Nack of Adobe re ACPU resizing: 
http://blogs.adobe.com/jnack/2010/12/a-no-color-management-print-utility-for-photoshop.html
 [Thanks for bringing this to our attention. Apparently it’s a known issue, and we’d like to issue an update, but I don’t have an ETA for that happening (as enhancing Photoshop’s print engine needs to take priority). –J.]
 
 ACPU does no color conversions at all."
 
 Dave Polaschek — 10:24 AM on December 03, 2010 (Same link as John Nack above):
   
 "On Windows, the main difference between using the mode that Michael describes and this application is that with Printer Manages Colors, the image data is converted to sRGB before sending it to the Windows print system (because that’s a requirement of the print system). So if your target is tagged with a color profile (it really shouldn’t be, but some are), there will be a color conversion done by Photoshop.


-------
Stan

http://forum.luminous-landscape.com/index.php?topic=120585.0
Sheesh

[Abdo]

Abdo,

Here's a quick view of the gamut boundaries of the 3 profiles and sRGB at L*=50.

The one from print studio is quite stunted so there are some issues with the settings running it. No idea what. I don't have Macs or run Print Studio.

Note that the Rag 809 I1 and Rag 809 Win are very close and overlap. I'll run some more tests on this to see how close and will update this post in a bit but wanted to give a quick view now. As an aside, the two matching profiles are reasonably close, but slightly larger in shape to the profile generated for my Canon 9500 II (not shown) using the same paper.


Edited: Additional Info:

I've compared the mxf files and the Mac one "Rag Sec Cor I1 Profiler" and Windows one "Rag 809 Windos (Color Man. OFF)" match exactly like I would expect two target prints printed consecutively on the same system and setup. Further the gamut's have the appropriate shape and are reasonably similar to that on my Canon 9500 II using the same paper. I've added the results of the comparison to this post. You can do this yourself by using an "A-B" comparison available in I1Profiler under the "Analysis" tab on the left that is available when you use "advanced mode."

So a profile made from either of the two matching processes can be used with both Mac and Windows. What I find particularly interesting is that the files match so closely that I can safely say the Mac and Windows, and the device level, (color unmanaged, max gamut) printing algorithms match.

Added:
I added a comparison with the Mac Print Studio mxf file. Delta E 1976's are shown as dE2000 becomes quite unreliable for large errors.

Doug,

Thank you very much for the analysis, I'm glad that the path I chose is the best one.

I already followed the standard printed in Windows, but I had problems in the past with the MAC, it seems that now they fit.

Actually Canon's Print Studio, I already tried several targets and the result is always the same, just looking at the printout to notice something erased and dead in the colors.

I just did not test Print Studio on Windows to see if it repeats the same as on the Mac.

When I play the ICC to analyze the graphic, it is very clear that it loses its result with Print Studio Pro.

Thanks for the study, one more time.

[Ernst Dinkla]

ACPU has a known (to Adobe) problem of incorrectly resizing some outputs, which prevents scanning by iSIS.  Lower in priority to them than working on Photoshop print pipeline, even though it disables the functionality it was created to have to work around the removal of the option to turn off color management.

snipped because of length
   
 "On Windows, the main difference between using the mode that Michael describes and this application is that with Printer Manages Colors, the image data is converted to sRGB before sending it to the Windows print system (because that’s a requirement of the print system). So if your target is tagged with a color profile (it really shouldn’t be, but some are), there will be a color conversion done by Photoshop.


-------
Stan

Windows running Qimage Ultimate, I wrote the following in 2010 and I do not think anything has changed since of before:

>>Qimage has 3 main choices on color management:

1/ CM-Off = Qimage doesn't convert the file, sends the plain RGB data to
the driver so with the color space profile stripped from the file.

In that mode it will also not assign a colorspace to an untagged file,
for example a target file for profile creation. And Windows will not
assign a profile either, it leaves that task to applications.

2/ Let printer manage color = Qimage does the above but the color space
profile isn't stripped from the file so the printer driver CM can do
something sensible with the file.

3/ Qimage CM on, Qimage does the conversion and sends the converted data
to the printer driver, expecting that the last's CM is set to Let
application do CM, so in fact with the driver's CM off..

In both 2 and 3 Qimage could assign a colorspace to an untagged file
based on EXIF data, camera model as an educated guess or without any
hint available it will assign a color space, the default sRGB or a
preferred choice of the user, say AdobeRGB or nothing if the user
doesn't like guesswork.

The printer driver has usually two choices: let driver do CM and the
other one let application do CM.  With application CM set the driver will
do no conversion at all and expects that the application does the work.
That is also the setting you will use to transfer targets through with
Qimage on CM off. When the driver does CM it will expect files with
certain colorspaces, my Z3200 can be set on AdobeRGB  or sRGB, the PS
driver has more spaces including CMYK versions.<<

So the paths I can take for target printing are through HP's Print Utility for the Z3200 or Qimage Ultimate settings + the Z3200 driver.  I do not think that QU is the only application in Windows that can act like that. For more printer brands.

That Adobe with 15000 employees still creates confusion for its customers on profile target printing is a shame. 10 years? The scaling of the target -3% with ACPU not solved either in time .......... Where both Apple and Adobe are to blame for the Mac confusion it is Adobe that should have taken responsibility for the Windows issues.

Met vriendelijke groet, Ernst

http://www.pigment-print.com/spectralplots/spectrumviz_1.htm
March 2017 update, 750+ inkjet media white spectral plots

[Stan Prevost]

The latest version of I1Profiler fixes this problem. However, simply installing the latest version may not work. I had this problem on Windows 10 x64. It resolved after completely uninstalling it then re-installing.

The current version of i1Profiler is 1.7.1.  Note that the XRite cut-and-paste quote in my post references the problem as pertaining to 1.7.1.  I had just downloaded and installed 1.7.1 prior to a profiling session, and the product UI showed the installed version as 1.7.1.  When I experienced the no-print problem, I began searching for applicable information and found the XRite post, and since the product developer/vendor described my exact problem as existing in the exact version I was using, I did not proceed with troubleshooting.  I just started working on a workaround, like ACPU (in which you still have to set up the printer driver correctly).  After reading your post, I have uninstalled, redownloaded, reinstalled and restarted, and it seems to work fine.  Thanks.  It is also possible that I did not restart my computer after the previous install, I just made sure the product UI screen showed the correct new version. Obviously insufficient.

Stan

[Doug Gray]

The current version of i1Profiler is 1.7.1.  Note that the XRite cut-and-paste quote in my post references the problem as pertaining to 1.7.1.  I had just downloaded and installed 1.7.1 prior to a profiling session, and the product UI showed the installed version as 1.7.1.  When I experienced the no-print problem, I began searching for applicable information and found the XRite post, and since the product developer/vendor described my exact problem as existing in the exact version I was using, I did not proceed with troubleshooting.  I just started working on a workaround, like ACPU (in which you still have to set up the printer driver correctly).  After reading your post, I have uninstalled, redownloaded, reinstalled and restarted, and it seems to work fine.  Thanks.  It is also possible that I did not restart my computer after the previous install, I just made sure the product UI screen showed the correct new version. Obviously insufficient.

Stan

Xrite still seems confused: In your link they state the problem and possible workarounds but in the download link for the windows version 1.7.1 they state it fixes this problem. It didn't, until after I did the I1Profiler uninstall then reinstall.


https://www.xrite.com/service-support/downloads/i/i1profiler-i1publish_v1_7_1

Notes: This update to i1Profiler is a bug-fix release to fix the following issues:

•The print button was not working when trying to make RGB printer profiles.