Pages: [1]   Go Down

Author Topic: Possible problem with printer profiles created with i1Profiler  (Read 1615 times)

Louie

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Possible problem with printer profiles created with i1Profiler
« on: February 20, 2018, 07:55:12 pm »

I recently resumed making my own print profiles after purchasing the X-Rite i1 Photo Pro package (i1Pro 2, i1Profiler with everything except CMYK). Basically everything works in that I can create and print patch sets, read the targets with the i1Pro 2 and create a profile. To keep things simple to start with I am not trying to do anything with UV. So it's just the M0 measurement data that I am using.

To validate the new profiles I am using a couple of color reference images from DigitalDog.net.  Using the new profile I printed both of these images from Lightroom with both Perceptual and Relative rendering intents. The results of this was that I am seeing much less banding especially with saturated red/magenta and green/cyan areas of the test prints. So far so good especially since I do a lot of landscape with saturated sunsets and green trees.

Now to the part where I think I found a problem with profiles I made, with a profile from Canson for Platine Rag on the P600 (and maybe all profiles made by i1Profiler?). Before I continue I want to say I am not a CMS expert. I am simply a reasonably competent computer guy/photographer that is comfortable running complex software with sufficient guidance/documentation.

This is what is describe below and it produced an anomaly which I am completely at a loss to explain. So I decided to being it to the "colorati" here to see if you guy have any insights about the effects I am seeing and to what extent they do or do not represent a problem with the profile or if it is really a problem with the test itself.

So to continue, in my quest to learn more I discovered an interesting YouTube video produced by Pat Herold over at Chromix titled  Viewing a Rendered Gamut. This is basically a torture test of a profile using ColorThink Pro to plot out in 3D how well the output profile maps out of gamut colors back to the gamut of the output device. So this sounds cool and I tried it with my newly minted profiles. 

When examining the resulting plot of the colors mapped to the output space there are big holes in the saturated reds and saturated yellow. For some strange reason these seem to be plotted way into the interior of the gamut.
I will attempt to attach an image one of these plots. If I am unsuccessful you also see a good example from a blog post by Pat titled ColorThink to the rescue!.

I will add that this anomaly is worst when you assign a wide gamut space to the color list and completely disappears when you assign Adobe RGB to the color list.

So I have several questions

What is going on?

Is this anomaly indicative of other problems with the profile?

Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Possible problem with printer profiles created with i1Profiler
« Reply #1 on: February 20, 2018, 09:41:39 pm »

Is this anomaly indicative of other problems with the profile?
Yes, to some degree. I say some degree because it's very difficult to see what you've done in CTP. A vector plot between reference values and what the profile produces? If so, don't expect to see the vectors dead on or even close to each other. And differing colors in color space will, as I think I see in your screen capture, have 'larger' or smaller vector plot differences.


Best thing to do is print some images (reference images of course*) and examine how the various rendering intents appear on the print itself.

* http://www.digitaldog.net/files/Gamut_Test_File_Flat.tif

Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Possible problem with printer profiles created with i1Profiler
« Reply #2 on: February 20, 2018, 11:45:20 pm »

You might find these I1P profile anomalies interesting.

It's a different way to look at how OOG conversions map. I'm fond of fixed L* slices with a 2D a*,b* plane.

http://forum.luminous-landscape.com/index.php?topic=114653.msg944584#msg944584
Logged

Louie

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Possible problem with printer profiles created with i1Profiler
« Reply #3 on: February 21, 2018, 10:06:00 am »

Yes, to some degree. I say some degree because it's very difficult to see what you've done in CTP. A vector plot between reference values and what the profile produces? If so, don't expect to see the vectors dead on or even close to each other. And differing colors in color space will, as I think I see in your screen capture, have 'larger' or smaller vector plot differences.
Hi Andrew,

Here is a summary of the workflow as described in Pat's video. I used a color list file provided by Pat and available for download from the YouTube page and loaded it into a CTP worksheet. Then I applied a color space to that list in this case ProPhoto RGB. Then I added the device profile to the worksheet plotted the result in 3D. In order to make the holes clearly obvious. I choose "cube" as the point shape so what you are seeing in the attached plot are really large points.

What is not clear from the attached image is that all those points that you would expect to be filling those gaps are plotted way inside the gamut volume. This is hard to see without displaying the 3D plot in CTP. I can't really see them until I start to rotate the plot.

I have attached a screen shot of Color Worksheet that illustrates what I am describing above.


Quote
Best thing to do is print some images (reference images of course*) and examine how the various rendering intents appear on the print itself.

* http://www.digitaldog.net/files/Gamut_Test_File_Flat.tif

Yes I am using this reference image. Using it in fact started me down this latest exercise as profiles I was using were showing some pretty noticeable  banding in red/magenta and green/cyan. The profiles generated by i1Profiler are quite a bit better in this regard.

tks, louie
Logged

Louie

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
Re: Possible problem with printer profiles created with i1Profiler
« Reply #4 on: February 21, 2018, 10:17:16 am »

You might find these I1P profile anomalies interesting.

It's a different way to look at how OOG conversions map. I'm fond of fixed L* slices with a 2D a*,b* plane.

http://forum.luminous-landscape.com/index.php?topic=114653.msg944584#msg944584

Yes that is interesting could possibly be related. As Graeme Gill commented it does look like some kind of numerical calculation problem.

Did you contact X-Rite and did they ever respond?

-louie
Logged

digitaldog

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 20630
  • Andrew Rodney
    • http://www.digitaldog.net/
Re: Possible problem with printer profiles created with i1Profiler
« Reply #5 on: February 21, 2018, 11:53:03 am »

Here is a summary of the workflow as described in Pat's video. I used a color list file provided by Pat and available for download from the YouTube page and loaded it into a CTP worksheet. Then I applied a color space to that list in this case ProPhoto RGB. Then I added the device profile to the worksheet plotted the result in 3D. In order to make the holes clearly obvious. I choose "cube" as the point shape so what you are seeing in the attached plot are really large points.
Not sure what's special about Pat's ref file but that's only part of the process. I'm not sure why you applied (assigned) ProPhoto RGB and that too could be an issue as that color space contains device values that are NOT colors and cannot be printed. In such a case, you'll see much larger dE vectors.
What you want to do IMHO is compare what the profile predicts to what the profile produces. Compare the reference values to the measured output values and run a report on that set of data. Again, the color list values and their assigned profile will make a difference here.
Logged
http://www.digitaldog.net/
Author "Color Management for Photographers".
Pages: [1]   Go Up