ArgyllCMS profiles: Black Point Compensation gives composite black when checked

[Ink stained Fingers]

I think we are at a point where we think as it should be or could be but it is not really solving the open question , I propose a test which should deliver some numbers to confirm this or that assumption or not . There is a difference between the profiles - ColorMunki and ArgyllCMS , the color mgmt module active with this or that profile delivers different black/gray values to the printer, and this should be directly visible in comparison of the .tiff files before and after application of the resp. profile - with the additional options of BPC and rendering options.

 

[Ink stained Fingers]

I made an attempt to work out differences between an X-Rite/ColorMunki and an ArgyllCMS profile. I exchanged profiles several months ago with @pharmacist as I was looking for a way to assess the accuracy of a profile.
I'm using here some of the data from that activity, he provided me a set of profiles - in this case for his Epson P800 with several different inks - I'm picking the inkset Octobestjet_3 for this evaluation - with a profile generated with the ColorMunki and with ArgyllCMS. The paper was the Netbit/Aldi glossy 300gr paper for these tests.

What did I do following the actions as described above , I'm using Qimage to print, I installed the P800 driver and printed the patch sheet with the dark gray patches - with both profiles - ColorMunki and ArgyllCMS - to a .tiff file which Qimage allows me to do - I'm using the premium glossy paper selection in the P800 driver with the quality level 4 and use the rendering intent rel. colorimetric in Qimage - with and w/o BPC - black point compensation and with color adjustments turned off in the driver since Qimage via the profile controls the colors.

This approach shows me the gray level data as they are sent to the printer - after the application of the resp. profiles to the data - tlhis is the state after profile color adjustments and before print.
The profiles in use have these black point values - the ColorMunki profile has a value of Lab 8,78 -0,07 -2,37 stored which equals to RGB = 24 25 28 , the ArgyllCMS profile has a blackpoint stored in the profile of Lab 11,19 0,01 -2,20 which equals RGB = 29 30 33

( I'm using this Lab - RGB converter
https://www.nixsensor.com/free-color-converter/ scroll down to the middle of the page )

I'm getting these numbers for the ColorMunki

and these for the ArgyllCMS profile

There are quite significant differences between these profiles at the dark end - again - that's data before print.
I don't have a P800 to look for the data after print, but that's not really necessary since this ws already done with the profile generation - the link between the colors before and after print, the profiling math creates tables in both directions . So I can directly display that data with the old Gamutvision software - with the display of the Black and White response.

This is the graph - an overlay of 2 graphs - with and w/o BPC for the ColorMunki profile

It behaves as expected - a straight horizontal line from L=0 to the paper black point of L=8.78 and then raising upwards, the corresponding line for BPC is directly bending upwards .

This is the graph for the ArgyllCMS profile

There are 2 differences - the line from L=0 to the paper black point is already raising upwards slightly and the graph with BPC on starts at the higher gray level than the curve w/o BPC. This is the status for 2 profiles only up to here, this does not imply similar differences for other profiles - for other papers and inks and driver settings. I would consider this - or a similar approach - just as a starting point for further evaluation.

 

[nertog]

Thanks for the interesting comparison, @Ink stained Fingers. 2 interesting issues:

1. Enabling BPC raises the actual black point of that specific profiled printer-ink-paper combination. Shouldn´t happen, IMHO
2. There is a significant difference in black point between X-rite and Argyll.

 

[Ink stained Fingers]

Yes - @nertog - there are several questions open which you address. I think it would be necessary at this point
to do several things

- redoing the current profiles with exactly the same parameters to see how large - or small - process - scanning instrument variations are and influence the results

- doing the same test with other papers - and/or another printer and inks - to see if the effects - your points 1 and 2 - remain - are gone or are getting wider - I just don't know it at this point, I'm not an ArgyllCMS user and cannot really proceed from here. And I don't know if there are any optional parameters for the profile calculation which may have an impact on the observed issues.

 

[nertog]

Yes - @nertog - there are several questions open which you address. I think it would be necessary at this point
to do several things

- redoing the current profiles with exactly the same parameters to see how large - or small - process - scanning instrument variations are and influence the results

- doing the same test with other papers - and/or another printer and inks - to see if the effects - your points 1 and 2 - remain - are gone or are getting wider - I just don't know it at this point, I'm not an ArgyllCMS user and cannot really proceed from here. And I don't know if there are any optional parameters for the profile calculation which may have an impact on the observed issues.

I can help with the Argyll stuff, but my hardware here is different (printers and spectro). As far as I can see, Argyll offers the option to focus more on neutral or darker values and apply smoothing to the measurement data. None of those makes any difference to the black point.

 

[Ink stained Fingers]

The best would be if a printerknowledge member with both - X-Rite7Colormunki/i1Studio and ArgyllCMS - on the computer could do all testing together, but there is one test you can do quite easily

1. Enabling BPC raises the actual black point of that specific profiled printer-ink-paper combination. Shouldn´t happen, IMHO

Pick a few of your Argyll profiles and check wtih the Gamutvision B/W display function if you see that effect as well in your profiles; if you are not familiar with Gamutvision please send me 2 - 3 of those Argyll profiles and I'll try to get it done.

 

[Ink stained Fingers]

I screened some archive data to find some ArgyllCMS profiles which I got from other forum members longer time ago incl. @palombian .

This is a view of the B/W reponse of a PCK3HD_RR_CanonProPlatinum, it very much matches my expectations and is similar to the Colormunki profile above

The profile Netbit_Glossy_300_gsm looks somewhat different at the black end, the difference between the lines with BPC on or off is much smaller than above.

And here the diagram for a PC65SE_CanonProPlatinum

There is a visible difference to the plots in the previous posting - there is no increase of the black point value for the curves with BPC on , I don't have an explanation for this variance.

Some explanation for the B/W density graphs can be found here

http://www.gamutvision.com/docs/blackpoint.html

Please see the note at the bottom that the plots with the perceptual rendering intent are with BPC on - aways, that's why I use the colorimetric setting instead.

 

[marco565]

The best would be if a printerknowledge member with both - X-Rite7Colormunki/i1Studio and ArgyllCMS - on the computer could do all testing together, but there is one test you can do quite easily

Pick a few of your Argyll profiles and check wtih the Gamutvision B/W display function if you see that effect as well in your profiles; if you are not familiar with Gamutvision please send me 2 - 3 of those Argyll profiles and I'll try to get it done.

I can do test I have the Color munki with both argycms and i1 studio on my mac.

EDIT: In attachment you will find a profile done with colormuki with Xrite and with ArgyCMS

 

[marco565]

That's why I recommend creating a perceptual table for your source space (or better, a device link) if you want a nice color reproduction.

Could you elaborate more on this point I don’t really understand. I tried to look at the documentation but I could not figure it out.
If I want to get the best perspective what source profile I should use? Most of the examples that I have seen uses adobeRGB. For example what would be the best for a canon pro 10?

 

[Ink stained Fingers]

@marco565 - this is a quick view of the 2 profiles you provided - they are very similar overall - across the complete luminance range - they just differ by a bit here and there. I'll come back later to the B/W response .
Yellow is for the ArgyllCMS profile, and orange for the Xrite profile, the cross-section is at L=50