A basic guide (see post #1) to setting up ARGYLL CMS profiling on your computer

Ink stained Fingers

Printer VIP
Platinum Printer Member
Joined
Dec 27, 2014
Messages
5,837
Reaction score
6,964
Points
363
Location
Germany
Printer Model
L805, WF2010, ET8550
It is quite difficult to compare profiles this way particularly when you use the perceptive rendering mode. Profile programs allow you to introduce adjustments - intentionally - they call it differently - about one option focusses onto the rendering of colors close to the gamut limit, how an increasing color saturation is mapped to the output colors - I could adjust that between neutral to colorful, Datacolor has other similar sliders, I'm not familiar with ArgyllCMS. And there is the gray axis, you can have the choice that the gray axis runs from paper white to paper/ink black which would give lighter grays a similar color tone as the paper color itself, or the gray axis creates neutral grays regardless which type of paper you use except at the top and bottom where you only can have the paper white and ink black tones which both are not exactly neutral. And you typically can adjust a profile to the color temp of the viewing light - in most cases D50 but you can change that. And some software let you adjust the profile as well to the brightness of the viewing light, probably doing some Gamma adjustment. Gamutvision offers some method to check profiles for their integrity, but if you find a flaw it is not really clear how to go from there to correct that.
 

martin0reg

Printer Master
Joined
May 9, 2010
Messages
1,058
Reaction score
746
Points
273
Location
Germany Ruhrgebiet
I'm learning a lot, what I didn't know and what I forgot, from your comments and by making mistakes...
After having noticed that the argyll profile from CAP480 seems to be "relative colorimetric", and setting the correct RI, the prints are very different, stronger reds in particular.
(therefor I just deleted my "first impression" in previous post..)
And when I open the profile in gamutvision (thanks for the hint, emulator), there is an option for RI which confirms what I see in the prints using different RI: the displayed gamut becomes bigger(!) with RI=rc than with RI=perceptual.
 

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,553
Reaction score
1,252
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
Hi Martin,

This is the package to be used to produce my profiles based on that magical 480 patch target with optimizations around the neutral axes to get really neutral B&W prints and good color fidelity.

It produces better profiles compared to the 2 step Colormunki method, even when you use 2 optimizations on top of the produced profiles to increase accuracy.

I tested this many times and each time the profiles were way better using the ArgyllCMS method.

It is a compressed TIF file for the target (no data loss, as the compression is LZH losless). Print it on A4 borderless setting with border expansion set to none (so the printing resolution is about 360 dpi).

The 480target.ti3 is calculated after you read the target with your Colormunki or other spectrophotometer.
 

Attachments

  • 480target.zip
    2.8 MB · Views: 501

Ink stained Fingers

Printer VIP
Platinum Printer Member
Joined
Dec 27, 2014
Messages
5,837
Reaction score
6,964
Points
363
Location
Germany
Printer Model
L805, WF2010, ET8550
It's more a matter of interest - why do you use a 16 bit color depth for a file which should provide 480 colors ? Did you do any resizing of that file - the color count of unique colors gives me a number of 46 000, and even if Iwould reduce the color depth to 8/24 bit it still counts 16 000 of them . The edges between the color patches and separation lines appear to be resampled . Just magnify some area by 10x. These additional colors are along the edges and most likely not in the middle within the reading circle of the spectro and would not change the results, but I'm wondering about the patch file generation process.
Patch magnified.JPG
 

Attachments

  • 30_BPlan_28_T2_Ab1_fortschreibung.pdf
    1.5 MB · Views: 461

martin0reg

Printer Master
Joined
May 9, 2010
Messages
1,058
Reaction score
746
Points
273
Location
Germany Ruhrgebiet
Thank you for the target, Pharmacist!
Some time ago you have posted a "ruler set" including a different A4 target (660 patches) and a "makeprofil.bat".
- Is it possible to use this batch file together with the now posted 480target?
- Using a profile, do I have to print with a specific RI, like relative colorimetric? BPC?

BTW after you mentioned to understand german quite well:
I'm "beta testing" a new version of the simple GUI for argyll print, called "printercal", here is the latest, possibly unstable version:
http://www.dslr-forum.de/showthread.php?p=14496191#post14496191
...not for the argyll experts around here, but for lazy people who won't learn command lines..
 

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,553
Reaction score
1,252
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
Hi Martin,

yes you can without problems. The makeprofile.bat is universal and can be used for every ArgyllCMS created target.

For printing: ArgyllCMS tends to be printed best with relative colorimetric intent and BPC enabled.

Actually you can use that ruler without problems. Just give it a try.
 

pharmacist

Printer VIP
Platinum Printer Member
Joined
May 29, 2007
Messages
2,553
Reaction score
1,252
Points
313
Location
Ghent, Belgium
Printer Model
Epson SC-P800,WF-7840,XP-15000
It's more a matter of interest - why do you use a 16 bit color depth for a file which should provide 480 colors ? Did you do any resizing of that file - the color count of unique colors gives me a number of 46 000, and even if Iwould reduce the color depth to 8/24 bit it still counts 16 000 of them . The edges between the color patches and separation lines appear to be resampled . Just magnify some area by 10x. These additional colors are along the edges and most likely not in the middle within the reading circle of the spectro and would not change the results, but I'm wondering about the patch file generation process.
View attachment 5178

Not sure why I did use 16 bit color depth, but I presume it would increase accuracy. I did resize the file using Photo Shop to elongate the patches so more samples are taken from each color patch when reading the rows with my Colormunki. Note: everything is done using tiff without data loss and compressing the final file is done with LZH compression (no data is lost). I think the elongating process seems to have introduced some artifacts near the color borders, but as you have mentioned: it doesn't change the results at all.
 

Emulator

Printer Master
Platinum Printer Member
Joined
Jan 4, 2012
Messages
1,675
Reaction score
1,308
Points
277
Location
UK
Printer Model
Canon Pro9000 II
CAP957 is the one I prefer to use of late. It takes more care in the reading process, (it needs a good mechanical guide, as illustrated elsewhere in this thread (Page 17, #166) and care to avoid ambient light creeping in) but gives me a profile that seems to work with all the limited range of the papers I use, it's lazyness I know.

Incidentally I am still using Argyll 1.6.3

Anyway how are you getting on Martin?
 
Last edited:

martin0reg

Printer Master
Joined
May 9, 2010
Messages
1,058
Reaction score
746
Points
273
Location
Germany Ruhrgebiet
Hi Emulator,
I'm now trying 3 different methods using argyll,

- the most convenient for users who don't want to learn anything regarding DOS command lines:
"PrinterCAL", a simple GUI for argyll, latest version (all beta!) 0.5.0:
http://www.dslr-forum.de/showthread.php?t=1606257&page=24 (download link in post 232)
The default option I tried was "medium quality": a 110patch precondition target on one A4 sheet and a 220patch target on two sheets, makes 3 A4 target prints, which can be read without ruler for the colormunki.
First prints are fine - but not visibly better than profiled with CM original software.

- CAP480.bat, first prints turned out fine... but not better than the 220target + precondition.
And chartreading with a ruler could be fiddly. As you just said, a good mechanical guide is needed, mine is a paper cut out..

- Pharmacist's special 480target: I wasn't able to run the "makeprofile.bat", the dos-window disappears when I hit enter.

Questions
@pharmacist: How should I handle the filenames in "makeprofile.bat"?
- Only the name or with full path?
- the filename with or without ending (.ti1 .ti2 .ti3)

Argyll is installed and "PATH" is modified... cap480.bat works fine, but it generates a different target. So can I use step 3 and 4 of cap480 and using pharmacist's target for chartread and the corresponding files - and skip step 1 and 2?

My conclusion for now: I haven't found a profiling method, which would be the killer app, compared to the original colormunki software.
But for special purpose (like profiling remote printers with only one target) it may be worth the additional effort of using argyll
 
Top