ArgyllCMS profiles: Black Point Compensation gives composite black when checked

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,555
Reaction score
1,253
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
I discovered that ArgyllCMS created printer profiles tend to use composite black when BPC is checked. This is most visible on pigment printers like the Epson SC-P800. When BPC is unchecked the black area is much more darker and intenser when BPC is disabled, thus improving contrast of the picture printed significantly. Using Colormunki software BPC does not affect the blackness at all and 100% is always printed with the dedicated photo/matte black unlike ArgyllCMS created printer profiles. Anyone knows how to change this during the profile creating process ?
 

Ink stained Fingers

Printer VIP
Platinum Printer Member
Joined
Dec 27, 2014
Messages
5,838
Reaction score
6,965
Points
363
Location
Germany
Printer Model
L805, WF2010, ET8550
You may use the old Gamutvision software - it's freeware - which has a display mode 'Black and White density response' and you directly above can select 'Black Pt. Comp' on/off per display window, and you can select various rendering intents. The software regrettably does not display the actual black point value, that's only possible with ColorThink by Chromix.
Gamutvision shows you the relation of L* out vs. L* in

Gamutvision.png
 

nertog

Print Addict
Joined
Jun 30, 2009
Messages
156
Reaction score
82
Points
161
Location
Regensburg, Germany
Printer Model
Epson R3000 - PX720 - P50
This is a valid concern that I have noticed as well if you print using relative colorimetric rendering. Looking at the output RGB curves for an X-rite profile, you´ll see that the RGB curves reach 0 at or before the input values reach 0 (= pure black input). An Argyll profile often has substantial output values when the input is absolute black. Maybe the aim is to neutralize black? No idea, but it gives less than ideal black values as you said.

Now, the main culprit seems to be the actual inkset, as I have only seen this when using non-linearized printer - ink - paper combinations...but as we normally cannot linearize our printers and many of us use non-standard inks, this is an issue.

My solution was to modify the .ti3 file and neutralize the measured black patches (keep the L value as measured and put a=b=0). I don´t like it, but it looks ok and solves the raised black levels for most cases. If the difference is large, you might need to do a manual "fade-to neutral" adjustment in measurements. Good luck :)

Let´s see what @Graeme Gill thinks about this!
 

mikling

Printer VIP
Platinum Printer Member
Joined
Jul 2, 2006
Messages
3,239
Reaction score
1,471
Points
313
Location
Toronto, Canada
Just an opinion.
Though I sell aftermarket ink, color management cannot and does not create magic. Review the role of color management in your workflow. Are the inks up to the task? IMO put poor tires on a great print engine, no amount of traction control, vector steering, awd and steering and wheel ALIGNMENTs will compensate for the lack of initial grip of the tires. When you make the suggested changes are you setting up the print engine to run on an oval by inserting misalignments ? just so you can run in circles.

As a parallel in audio, Digital Signal Processing (DSP) cannot make poor performing drivers/speakers perform like drivers that are up the task. In each case DSP like Color Management which is trying to achieve the same thing for a different set of senses) can improve each system but the end results will differ.
 
Last edited:

nertog

Print Addict
Joined
Jun 30, 2009
Messages
156
Reaction score
82
Points
161
Location
Regensburg, Germany
Printer Model
Epson R3000 - PX720 - P50
Just an opinion.
Though I sell aftermarket ink, color management cannot and does not create magic. Review the role of color management in your workflow. Are the inks up to the task? IMO put poor tires on a great print engine, no amount of traction control, vector steering, awd and steering and wheel ALIGNMENTs will compensate for the lack of initial grip of the tires. When you make the suggested changes are you setting up the print engine to run on an oval by inserting misalignments ? just so you can run in circles.

As a parallel in audio, Digital Signal Processing (DSP) cannot make poor performing drivers/speakers perform like drivers that are up the task. In each case DSP like Color Management which is trying to achieve the same thing for a different set of senses) can improve each system but the end results will differ.
Sure, I agree 100%. That being said, a good DSP can make a tremendous difference...and the same holds true for a good ICC profile. The problem here is a secondary effect that limits the full performance of the system. OP and I want to know why this happens and how to work around/solve it using the means we have at hand.
 

Ink stained Fingers

Printer VIP
Platinum Printer Member
Joined
Dec 27, 2014
Messages
5,838
Reaction score
6,965
Points
363
Location
Germany
Printer Model
L805, WF2010, ET8550
I don't think at this time that the observed effect is ink related , I did a test with @pharmacist recently as reported in the ArgyllCMS thread that we were exchanging profiles - I printed his patch sheet and mailed the print to him, and he created an ArgyllCMS profile with it; I printed at the same time - same driver settings - same inks - etc a profile with i1profiler based on 283 patches, the gamuts very much match as reported earlier.

The first Gamutvision print above shows the B/W response for the ArgyllCMS profile - rel. col. intent - no BPC - and this screenprint shows the i1profiler B/W response - with the same parameters .

When you look to the dark end - to the values of L* below 10 the plots indeed differ , I would consider the i1profiler response more consistent - the L* out stays flat up to the point where L* in reaches the black level of the ink/paper combination, and then raises up pretty much linearly. I don't see here that an ink can have any influence - it's the same for both profiles.
 

Attachments

  • Gamutvision-1.png
    Gamutvision-1.png
    172 KB · Views: 135
Last edited:

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,555
Reaction score
1,253
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
Just an opinion.
Though I sell aftermarket ink, color management cannot and does not create magic. Review the role of color management in your workflow. Are the inks up to the task? IMO put poor tires on a great print engine, no amount of traction control, vector steering, awd and steering and wheel ALIGNMENTs will compensate for the lack of initial grip of the tires. When you make the suggested changes are you setting up the print engine to run on an oval by inserting misalignments ? just so you can run in circles.

As a parallel in audio, Digital Signal Processing (DSP) cannot make poor performing drivers/speakers perform like drivers that are up the task. In each case DSP like Color Management which is trying to achieve the same thing for a different set of senses) can improve each system but the end results will differ.
Actually it could be the car itself causing havoc to the tires.....Many years ago I had a Spyderprint scanner and it made terrible profiles because dark blues tends to become printed purplish. I had to tweak the target of the blue patch to change to hue to purple to compensate for the error but it was less than ideal. I sold my Spyderprint scanner and got my current Colormunki: it produces deep blues without any purplish hue as in the case with the Spyderprint. I think it is software problem with ArgyllCMS, because the problem does not occur when using the Colormunki proprietary software.
 

nertog

Print Addict
Joined
Jun 30, 2009
Messages
156
Reaction score
82
Points
161
Location
Regensburg, Germany
Printer Model
Epson R3000 - PX720 - P50
I don't think at this time that the observed effect is ink related , I did a test with @pharmacist recently as reported in the ArgyllCMS thread that we were exchanging profiles
This compares different printers + different profile engines. Also, I don´t see the actual issue mentioned by @pharmacist. The RGB curves nicely converge at the black point. Here is a problematic one that results in raised blacks when selecting BPC when printing. The question is: why does Argyll need to keep the >0 for an input level of RGB=0. And no, it does not result in a deeper black.

Ncorrected.jpg

Uncorrected measurement with relatively blueish PK ink. Note Lmin = 4.9 and the red channel is still at 26 (10% of max)!

corrected.jpg

Same .ti3 file (XYZ measurement data for Argyll) where I manually modified the pure black patches to neutral. Note Lmin is now 3.5 and the RGB output curves reach 0.
 

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,555
Reaction score
1,253
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
I suggest to print with peceptual intent with and without BPC with an ArgyllCMS created printer profile and see for yourself. Relative colorimetric intent I did not tested before, maybe with relative colorimetric intent the problem is not there: I will do a print test this evening and see what happens.
 
Top