Pages: [1]   Go Down

Author Topic: Argyll with i1 Pro 2 - profiles good size and smooth, but colours off.  (Read 723 times)

kamma1

  • Newbie
  • *
  • Offline Offline
  • Posts: 31

I recently got a 2015 i1 Pro 2 with very little use, and have been trying unsuccessfully to get usable profiles with several matte papers. 

For comparison, I have profiles from the OEM and custom profiles done with an iSis and i1 Profiler.   I also made some profiles with an i1 Studio about a year ago, and though I no longer have those, all three of these sources produces very similar profiles in gamut, smoothness of transitions, etc. 

The i1 Pro 2 matches the gamut size, and is perhaps better in transitions, but the colours are off:  in particular the saturated blues are shifted toward violet, and the very saturated yellows and oranges lose detail.   These characteristics are consistent with every paper I have tried, and whatever patch count. 

I'm using a Macbook, so I unfortunately can't use Gamutvision for comparisons, but using Copra a comparison of the linear graph shows a dogleg in the blues, again consistent in every profile, and unlike the OEM/Custom profiles.   And a comparison of the gamuts shows that the gamut has a very similar shape to the custom iSis profiles, but the whole thing seems to be shifted away from the blues and the darks, and toward the yellows.   

At this point I have run through a lot of paper, and run out of ideas.   I wonder if I am missing something basic in Argyll, or whether the i1 Pro 2 measurements are off.  I've tried a lot of variations (including specifying the default values for illuminant, etc), but I am mostly using these commands for targen and colprof:

targen -v -d2 -G -e8 -B8 -g128  -f2025   outputFile
colprof -v -qh  -sAdobeRGB1998.icc -cmt -dpp inputFile

Does anyone have any suggestions, or ideas about what is going on?

I'm attaching images of the line graph from Copra, and the gamut comparison - the wire frame is the iSis profile (again, this is virtually identical to the OEM profile) and the solid colours are the i1Pro2. 

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 596
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: Argyll with i1 Pro 2 - profiles good size and smooth, but colours off.
« Reply #1 on: December 25, 2021, 06:50:17 pm »

You may be better off posting to the ArgyllCMS Mailing list for issues related to the software, because it's easier to communicate and other with knowledge can add their 2c. 

To be able to help at all some pertinent information is needed - gamut surfaces aren't very useful in communicating what issues you have.
Most important is providing a reference to the .ti3 data, so that your problem can be reproduced. Next is an example of an image that demonstrates the problem, and a description of what specifically is the problem in reference to the image. Finally the workflow needs to be clear - what the source color space is (AdobeRGB1998 ?) and what intent you are using. You've already provided the colprof details, so that's good.

You should try updating to the (just released) ArgyllCMS V2.3.0 to see if that helps with the blue, since there are changes in that area.
Logged

kamma1

  • Newbie
  • *
  • Offline Offline
  • Posts: 31
Re: Argyll with i1 Pro 2 - profiles good size and smooth, but colours off.
« Reply #2 on: December 27, 2021, 10:37:54 am »

Hi Graeme,

Thanks very much for the response, I'll put together a post for the mailing list, with the information you suggest. 

But to clarify a bit here:  both the images I posted are of the absolute rendering intent.   They differences in the gamuts exist in all rendering intents, while the dogleg in the blues in the line graph is visible in all intents except perceptual. 

I'm not sure what you mean by "source color space" - in colprof I'm specifying AdobeRGB1998 for the perceptual intent. 

And I'm not sure quite what additional information I should provide for workflow.  I posted the targen and colprof commands and parameters, the chart is printed without any color management via Adobe Printer Utility, and chartread is simply:

chartread -v -H -T0.4

I was using version 2.2.1, but have now installed 2.3.0 and rerun colprof, and didn't see any significant difference.  Should I redo the process from the start to see if there's a difference? 

I didn't mention that magenta doesn't seem to quite reach maximum saturation (in either Argyll version). 

Thanks again.

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 596
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: Argyll with i1 Pro 2 - profiles good size and smooth, but colours off.
« Reply #3 on: January 05, 2022, 03:49:07 am »

But to clarify a bit here:  both the images I posted are of the absolute rendering intent.   They differences in the gamuts exist in all rendering intents, while the dogleg in the blues in the line graph is visible in all intents except perceptual.
As I already indicated, these aspects are not very helpful in understanding or diagnosing problems. Neither are involved in the construction or use of profiles. To understand and be able to reproduce problems, the necessary files are needed, as well as the workflow details.
Quote
I'm not sure what you mean by "source color space" - in colprof I'm specifying AdobeRGB1998 for the perceptual intent.
Yes you are, but the colorspace of the images you are printing has not been stated, and a disconnect between this and the way you are previewing and printing the images can result in unexpected color reproduction.
Quote
And I'm not sure quite what additional information I should provide for workflow.
Source image colorspace. Source ICC profile. Identity of CMM/application being used for printing and any setup details. CMM/application intents (in and out, if these are not the same).
Quote
I was using version 2.2.1, but have now installed 2.3.0 and rerun colprof, and didn't see any significant difference.  Should I redo the process from the start to see if there's a difference? 
No.

Without the .ti3 file, source image examples and workflow details, then there is no prospect of reproducing or understanding what you are seeing.
Logged
Pages: [1]   Go Up