0
\$\begingroup\$

Thank you for your potential interest, I hope someone can please help me. Apologies in advance for the long post.

TL;DR : why are source colors that are known to be printable excluded from the generated gamut of a custom RGB ICC printer profile based on those colors?

Background: I’m just getting started with an Epson P5000 (10 inks). I’ve been working with a commercial service who provides custom ICC profiles via i1Publish. They recommend printing the i1 profile targets (3155 patches total) using Adobe Color Print Utility. I couldn’t get this to work on my end, details omitted for brevity.

Workaround: against their advice, I used Photoshop CC 2022 to print the 5 profile target images. I double-checked each step of the process, making sure the images were untagged and that color management was off, in both the Photoshop and Epson print dialogs.

Evaluation: when I got the custom ICC profile back, I opened it in ColorThink Pro for a sanity check, along with one of the ICC print target images (in theory containing 631 patches). Quite a lot of the source color patches are significantly out-of-gamut vs. the new custom profile. (see snip).

3D Lab-space : profiling target image and generated custom ICC printer profile

My concerns:

  • Many of the out-of-gamut colors look like they could be fitted to various smooth surfaces. Shouldn’t the custom profile’s gamut boundary have interpolated and incorporated those surfaces (and thereby the “outer” colors)?

  • I loaded into ColorThink Epson’s stock P5000 profile for the same paper (not shown here). The stock profile gamut boundary generally falls just inside my custom profile’s boundary, this seems normal. Consequently, all the same “outer” print target color patches are similarly out-of-gamut with respect to Epson’s stock P5000 profile. This suggests my results should be reproducible, and that either I didn’t do anything wrong or else both Epson and I made the same mistake(s).

  • If no mistakes were made, (why) is it correct and maybe even expected to get so many out-of-gamut colors from the test image and its generated printer profile?

Thank you for reading this and for any explanation you may be able to offer.

\$\endgroup\$
3
  • 2
    \$\begingroup\$ As you cannot turn off colour management without the Adobe Color Print Utility you're testing a false premise. You're going to have to ask Adobe &/or i1. \$\endgroup\$
    – Tetsujin
    Jun 13, 2022 at 17:15
  • \$\begingroup\$ ^ It's absolutely critical that color management be disabled. ACPU can do that and there are techniques using Photoshop that do as well (Null Transform). Untagged isn't sufficient in Photoshop. If you are using Windows the Null Transform works great and I've used it for over 10 years because ACPU slightly shrinks the target in Windows. \$\endgroup\$
    – doug
    Jun 18, 2022 at 0:18
  • \$\begingroup\$ The Null Transform technique requires: 1. Assign any printer profile to the image. 2. Select no color management in printer setup->printer settings in the print dialog. See the printer's manual for how to set this up using their supplied ICC profiles. 3. Then select Photoshop Manages Color. Then select exactly the same profile you assigned to the image in step 1. You will get a warning. Cancel it and print the target. \$\endgroup\$
    – doug
    Jun 18, 2022 at 0:36

1 Answer 1

1
\$\begingroup\$

TL;DR : why are source colors that are known to be printable excluded from the generated gamut of a custom RGB ICC printer profile based on those colors?

The stock profile gamut boundary generally falls just inside my custom profile’s boundary...or else both Epson and I made the same mistake(s).

Clearly the source colors were not printable and your assumption of "known to be printable" is in error. Maybe the P5000 could reproduce more of those colors on a different paper with different color management applied...

If you mean "known to be printable by some other printer;" well maybe, but I doubt it. You can't define the capabilities without finding the limits... if all of the test colors were reproduced accurately then the gamut of the custom ICC profile would be that of the test target, not that of the printer.

\$\endgroup\$
2
  • \$\begingroup\$ Hi Steven, Thank you for your reply and comments. I see what you mean, I was definitely confused. Yes, clearly the only “known” color values are the device-independent Lab values of the color patches in the target images. Corresponding Lab values of the device-dependent, printed color patches are only discovered empirically, as they are measured by the spectrophotometer. As you said, the presence of all those out-of-gamut colors in the 3D plot simply means my printer cannot accurately reproduce those hues, on that paper. \$\endgroup\$ Jun 14, 2022 at 23:32
  • \$\begingroup\$ The profile targets are composed of a distributed set of RGB values. They cover the RGB space completely and are all printable and in-gamut by definition. They are not defined colors in any sense. They are printed as RGB values. They are then measured by a spectro and those colors are associated with the RGB values. From that a profile can be made either by i1profiler or Argyll (free open source). \$\endgroup\$
    – doug
    Jun 18, 2022 at 0:22

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.