Pages: [1]   Go Down

Author Topic: ArgyllCMS - very poor results with 1000+ patch target  (Read 1287 times)

podll

  • Newbie
  • *
  • Offline Offline
  • Posts: 6
ArgyllCMS - very poor results with 1000+ patch target
« on: October 04, 2014, 01:27:09 pm »

It was suggested I try this forum instead of the Printing one on my other thread, so I apologize for posting this twice. Here's my issue: I have been trying to make watercolor reproductions on traditional cold press watercolor paper, unfortunately with very mixed results using existing profiles. Luckily I got my hands on a Colormunki to help fix this, and using the Colormunki software with only 3 targets (130 patches total) I was able to make the best matched print I've ever made.

Enter ArgyllCMS. I've read great things about it. I thought I'd take it to the next level and produce a great profile, so I generated a target with 1,050 patches, printed it out using the ColorSync app with the setting "Print as color target" (this turns off color management based on what I have read), and read it in with the ColorMunki with the following commands:

targen -v -d2 -f1050 -R -G 1050_Arches

printtarg -v2 -iCM -h -t360 -a.6 -m0.0 -M0.0 -P -p190x320 1050_Arches

chartread -v -H -B -T0.4 1050_Arches

After getting all the rows, I grabbed my local AdobeRGB1998.icc from my ColorSync folder (as I've read you're supposed to do), copied it to my working directory and used this command to generate the profile:

colprof -v -qh -S AdobeRGB1998.icc -dpp -D "Arches 300gsm Canon Pro 100" 1050_Arches

And here are the results (I apologize for picture quality) with the original on top, the Colormunki software profile on the bottom left, and Argyll's profile on the bottom right. This was printed in Photoshop with relative colormetric selected on my Canon Pro 100:



Notice the blacks are completely missing in the face and the clipping on the blue wash on the side. It is difficult to tell from this picture, but there is a green cast as well. Also, here's a 3D view from the ColorSync app comparing the Colormunki generated profile and ArgyllCMS, with the smaller ArgyllCMS profile in color and the Colormunki shown in a white outline. It's obvious that there's quite a bit less coverage in Argyll's profile. To be honest, I'm baffled that the gamut is so much smaller considering I used more than 8x the patches with Argyll:



Perhaps I've missed a valuable step in generating my target or profile - I sure hope so. I'm a complete novice at profiling (I'm just a guy trying to help his wife), but I was really hoping that 1,050 patches would produce a superior profile to the 130 patch Colormunki profile. If anyone has any ideas as to what might have gone wrong here, or what else I need to do to improve this, I would greatly appreciate any help. I'm comfortable on a command line, but I will admit I simply used parameters/settings that I found from others on forums.
Logged

Robert Ardill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 658
    • Images of Ireland
Re: ArgyllCMS - very poor results with 1000+ patch target
« Reply #1 on: October 04, 2014, 02:28:38 pm »

It was suggested I try this forum instead of the Printing one on my other thread, so I apologize for posting this twice. Here's my issue: I have been trying to make watercolor reproductions on traditional cold press watercolor paper, unfortunately with very mixed results using existing profiles. Luckily I got my hands on a Colormunki to help fix this, and using the Colormunki software with only 3 targets (130 patches total) I was able to make the best matched print I've ever made.

Enter ArgyllCMS. I've read great things about it. I thought I'd take it to the next level and produce a great profile, so I generated a target with 1,050 patches, printed it out using the ColorSync app with the setting "Print as color target" (this turns off color management based on what I have read), and read it in with the ColorMunki with the following commands:

targen -v -d2 -f1050 -R -G 1050_Arches

printtarg -v2 -iCM -h -t360 -a.6 -m0.0 -M0.0 -P -p190x320 1050_Arches

chartread -v -H -B -T0.4 1050_Arches

After getting all the rows, I grabbed my local AdobeRGB1998.icc from my ColorSync folder (as I've read you're supposed to do), copied it to my working directory and used this command to generate the profile:

colprof -v -qh -S AdobeRGB1998.icc -dpp -D "Arches 300gsm Canon Pro 100" 1050_Arches

And here are the results (I apologize for picture quality) with the original on top, the Colormunki software profile on the bottom left, and Argyll's profile on the bottom right. This was printed in Photoshop with relative colormetric selected on my Canon Pro 100:



Notice the blacks are completely missing in the face and the clipping on the blue wash on the side. It is difficult to tell from this picture, but there is a green cast as well. Also, here's a 3D view from the ColorSync app comparing the Colormunki generated profile and ArgyllCMS, with the smaller ArgyllCMS profile in color and the Colormunki shown in a white outline. It's obvious that there's quite a bit less coverage in Argyll's profile. To be honest, I'm baffled that the gamut is so much smaller considering I used more than 8x the patches with Argyll:



Perhaps I've missed a valuable step in generating my target or profile - I sure hope so. I'm a complete novice at profiling (I'm just a guy trying to help his wife), but I was really hoping that 1,050 patches would produce a superior profile to the 130 patch Colormunki profile. If anyone has any ideas as to what might have gone wrong here, or what else I need to do to improve this, I would greatly appreciate any help. I'm comfortable on a command line, but I will admit I simply used parameters/settings that I found from others on forums.

Hi,

In general you seem to be doing the right thing.  However you shouldn't have the -H (high res) argument in chartread, or the -T argument (read tolerance).  Only use the -T if you are getting misreads, and then you should make it >1, not <1.

The Argyll-generated gamut seems quite wrong, but then again so does the Colormunki-generated one.  There are some weird nobs and dips that really shouldn't be there.

My guess is either that the print is wrong, or the scanning was wrong.  Try rescanning your targets without the -H and -T arguments in chartread.  If that still gives a very poor profile then the most likely problem is with the prints.

I've never used the Colormunki (I use an i1Pro2) so there could be issues there that I'm not aware of (could be that with that instrument you would be better using fewer spot colors).  If you are still having problems the best place to post your problem is to send an email to argyllcms@freelists.org ... just checked and I see you've just done that :).  You will get help there, it's a very supportive and useful forum, and Graham Gill is absolutely the guru when it comes to Argyll as he is the author.

Robert
Logged
Those who cannot remember the past are condemned to repeat it. - George Santayana
Pages: [1]   Go Up