Pages: 1 [2]   Go Down

Author Topic: Comparing Argyll and I1Profiler printer profiles  (Read 8340 times)

Ernst Dinkla

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 4005
Re: ArgyllCMS + best B&W workflow ?
« Reply #20 on: January 02, 2019, 09:01:39 am »

Hi Ernst,

When you create your ArgyllCMS profiles how many additional gray step patches do you add?  I do a step of 50 for my Epson 3880 if I am going to use the normal print driver (otherwise I use QTR with the B/W option).

Alan

Alan, 

So far I did not add extra gray patches and used the default 1728 target of the HP Color Center targets.  Before I go in target enhancement I want to know whether that Rel.Col. route is an option for B&W. The point is if I have to make perceptual profiles for more color spaces and that for several media choices there will be a huge list of profiles I have to plow through every time, not just the effort to create them. For most work I can do it with Rel. Col. given the saturation in the images including reproduction work.

Met vriendelijke groet, Ernst

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

RobWignell

  • Newbie
  • *
  • Offline Offline
  • Posts: 33
Re: Comparing Argyll and I1Profiler printer profiles
« Reply #21 on: January 10, 2019, 06:00:28 am »

Thanks, Doug Gray, for your comparative analysis. 

I have come in on this conversation quite late in the piece.  I use Argyll with an X-Rite DTP70 for scanning targets but I also own an i1Pro V2 so I have an option to use this for generating profiles too.
 
My experience through observation rather than measurement is that profiles produced with ArgyllCMS are noticeably more neutral in their grey scale production that those I have produced using the i1Pro, however, in my study I have not been able to sort out what number is being quoted when DE2000 figures are quoted.  For instance, in your initial post you have quoted low DE figures of significantly less than 1 but the Argyll command, "profcheck" outputs three figures: errors(CIEDE2000) max, avg and RMS.  In addition, the Argyll command "colprof" produces three more figures that are usually higher: peak error, average and RMS.
 
I presume that you are quoting the errors(CIEDE2000) avg figure generated by "profcheck" but would be grateful for a confirmation.

On a related note I picked up that the X-Rite profiles use a black point compensation in the A2B0, Intent-0 table and ArgyllCMS doesn't.  Does this have the effect of increasing the gamut volume and make it appear that gamut volumes generated by i1Profiler tare larger than gamut volumes generated by ArgyllCMS when using comparable targets, on the same paper, printer and ink combination?

Best wishes

Rob Wignell
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing Argyll and I1Profiler printer profiles
« Reply #22 on: January 10, 2019, 11:51:11 am »

Thanks, Doug Gray, for your comparative analysis. 

I have come in on this conversation quite late in the piece.  I use Argyll with an X-Rite DTP70 for scanning targets but I also own an i1Pro V2 so I have an option to use this for generating profiles too.
 
My experience through observation rather than measurement is that profiles produced with ArgyllCMS are noticeably more neutral in their grey scale production that those I have produced using the i1Pro, however, in my study I have not been able to sort out what number is being quoted when DE2000 figures are quoted.  For instance, in your initial post you have quoted low DE figures of significantly less than 1 but the Argyll command, "profcheck" outputs three figures: errors(CIEDE2000) max, avg and RMS.  In addition, the Argyll command "colprof" produces three more figures that are usually higher: peak error, average and RMS.
 
I presume that you are quoting the errors(CIEDE2000) avg figure generated by "profcheck" but would be grateful for a confirmation.

I don't use profcheck. Rather I use Matlab functions that calculate the DE2000. I don't generally bother testing the target colors against the generated profile preferring to compare a distinct set of printed colors that don't correlate to the ones used to generate a profile. I believe profcheck can be used against an independent color set too. I prefer the Matlab functions since it easily generates histograms of errors as well as other ways of looking at printer accuracy.
Quote


On a related note I picked up that the X-Rite profiles use a black point compensation in the A2B0, Intent-0 table and ArgyllCMS doesn't.  Does this have the effect of increasing the gamut volume and make it appear that gamut volumes generated by i1Profiler tare larger than gamut volumes generated by ArgyllCMS when using comparable targets, on the same paper, printer and ink combination?


One shouldn't use the AtoB0 tables for gamut volume calculations as they are undefined colorimetrically. Stick with the AtoB1 tables which are. As an aside, I find gamut volume metrics greatly overstate differences between printers. I rarely bother with them. Here's an interesting example of how gamut volume can be misinterpreted:

Make a print of some random image that is largely in-gamut on both matte and luster using Rel. Col. Now view them side by side in a room where the lighting is completely diffuse so you see no reflections from the luster print and viewed at a distance such that one cannot see the luster texture.

The matte print will have more saturated colors. Yet the matte print's gamut will be far smaller. This is because profiles are made from light at 45 degrees which eliminates the reflections from the luster (or gloss) surface. This produces markedly different profiles than if the profiling instrument measured with a diffuse illuminant thus included specular reflections. It makes a big difference with glossy and luster papers but almost no difference at all with matte papers.
Logged

RobWignell

  • Newbie
  • *
  • Offline Offline
  • Posts: 33
Re: Comparing Argyll and I1Profiler printer profiles
« Reply #23 on: January 10, 2019, 06:53:51 pm »

Doug

Thanks for the clarification that you use Matlab for DE2000 calculations and not the Argyll profcheck command.  I am not familiar with Matlab but will look into it. 

For gamut volume calculations, I don't know what tables Argyll or i1Profiler use but, considering the differences, I guess that Argyll uses the AtoB1 table and i1Publisher uses AtoB0 .

Thanks also for the comment about gamut volumes and the example of printing matte and lustre images for comparison. 

I have not been able to confirm that my X-Rite DTP70 conforms to the standard of using 45 degree illumination - there are very few references to this machine on the web and it has been abandoned by X-Rite.  My attraction to using it is the convenience and speed of operation and it seems to have a reasonable degree of consistency.  With multiple patch readings I have not noticed a variation before the third significant figure.

Cheers

Rob Wignell
Logged

GWGill

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 608
  • Author of ArgyllCMS & ArgyllPRO ColorMeter
    • ArgyllCMS
Re: Comparing Argyll and I1Profiler printer profiles
« Reply #24 on: January 11, 2019, 12:16:47 am »

I have not been able to confirm that my X-Rite DTP70 conforms to the standard of using 45 degree illumination
Almost certainly yes. 45/0 is a good way of approximating how we view things, and is what graphic arts instruments use. Almost all glossy surfaces spectacularly reflect the color of the illuminant, not the underlying surface, and we tend to regard such specularity as a distinct quality ("shininess") rather than an attribute of surface color.

Industrial color measurement often uses diffuse (0/d or 8/d) measurement geometry which includes the specular component. This is very useful when you want to know the colorant content of a substance, independent of its surface finish.

« Last Edit: January 13, 2019, 01:17:00 am by GWGill »
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 2197
Re: Comparing Argyll and I1Profiler printer profiles
« Reply #25 on: January 11, 2019, 12:40:26 am »

Almost certainly yes. 45/0 is a good way of approximating how we view things, and is what graphic arts instruments use. Almost all glossy surfaces reflect the color of the illuminant, not the underlying surface, and we tend to regard such specularity as a distinct quality ("shininess") rather than an attribute of surface color.

Industrial color measurement often uses diffuse (0/d or 8/d) measurement geometry which includes the specular component. This is very useful when you want to know the colorant content of a substance, independent of its surface finish.

And there is something somewhat appealing about glossy print specular reflections. Luster is an interesting animal. Rather than just reflecting a portion of light hitting it like a mirror, it reflects the same but scatters it over about a 20 to 30 degree cone angle. It has some of the characteristics of glossy but spreads out reflections so the specular reflections seem attenuated. In a situation where viewing a print with entirely indirect, diffuse light, one doesn't see much of the specular reflection. It remains, of course, but is just distributed so has the effect of reducing color saturation and raising the black point in such diffuse, indirect lighting.

OTOH, both glossy and luster prints have much larger gamuts and can produce those more saturated colors when viewed with directed lighting at an angle such that the specular reflections become negligible. This somewhat models the way 45/0 spectros measure color.
Logged
Pages: 1 [2]   Go Up