Possibility to convert csv-file with only L*A*B* values into a ArgyllCMS ti3 file

yes indeed, likewise my i1Pro1 and i1Pro2 white tile also is bound with a seria number to the device and cannot be exchanged, because each tile/spectro combo is factory calibrated to match.

Here is a collar and 300mm rule for 10mm square patches.
Rule and collar on your 247 patch target scaled by forced printing at 518 dpi.

Align the rule to the edge of the patches with notches central. Press the button and slide to the next notch. I can't see how that isn't faster and less error prone than manually locating each larger patch and it is half the ink, half the paper, or twice as many patches.

STL attached. You may need to tweak printer calibration to make the collar a snug fit.

@Epatcola : that is a fantastic design you have made....making it possible to scan miniature patches as well. That will greatly increased the amount of patche on a sheet of A4/Letter.

I am struggling with the larger 247 patch target. There is not the same error as previous, but running colprof.exe and the system tends to hang without any error message and does not proceed:



@x64 : maybe you can have a look at, not sure what is wrong with it and can you produce a profile with the data ?

EDIT: the target set is now corrected to the right patch distribution and should be OK now. For those who already downloaded the previous 247-target.zip: please discard this incorrect version and use the new "247-target for CR30 with space bars.zip", which is correct one and should not have this problem.

Running almost 3 hours: the process is still hanging:



Not sure what is causing it: it does not give an error that some data is wrong or corrupted, just hanging. The previous 228-target works flawlessly. But I tried a slightly larger patch count of 247 (one extra row of 19 patches added), but it does not seems to like it.

Comparing that ti2 and csv to the files we used to design the converter I notice some obvious problems/differences:
-You used random order
-Your 247 csv has a lot of empty data, see the fields populated with 0 across the file:
-;M0;D50/2°;42,23;63,57;-3,94;0;0;0;0;16,44;18,91;19,08;18,54;17,56;16,31;14,63;12,34;9,68;7,2;5,39;4,19;3,47;3,22;3,16;2,74;2,08;2,54;6,58;15,43;26,62;38;47,39;54,27;58,92;62,72;65,06;66,42;66,88;67,22;67,56;0;0;0;0;0;0;0;0

OK, now I understand the underlying problem. I will regenerate the the ti2 not the use random order and remove the extra data: these are values of the spectral data at 360-390 nm and from 710 to 780 nm. Thanks for the update.

I'm not sure if the random order is causing it, because with our interpretation it is not 'random' (e.g. index 1 on ti2 = index 1 on csv = index 1 on ti3). I'd try without the empty data first, then try the non random order. If the order is causing it, I'm sure we can find a workaround so you can use random order, but I currently don't have time for it.

@x64 : You were right the randomisation was the cause of the hanging of the colprof process. I now have changed the patch layout to be non-randomized, but in doing so the grey ramp is sometimes difficult to determine where the next grey patch starts. To visually overcome this I have introduced a tiny separation line between patches so it is now easily visible where the next patch is located.

For those who alreaady downloaded the previous 247-target: please discard the earlier version and download this latest version:

This is the target layout on A4 (to be printed in the borderless setting, NO Expansion, retain size in order to maximize the printing surface without running of the paper):

NOTE: it is very important to start reading from the first patch starting from the patch closest to the letter, so in the below image the white patch on the upper left white patch (A1)....up to the last patch on the bottom right purplish blue patch (M19). It took me about 10 minutes to scan and save the data with the ColorQC2 program. I have activated the registration of the spectral data in the measurement setting from 400-700 nm (setting M0 and D50/2 degrees as the colprof command will use the 1931_2 viewing angle (2 degrees)), so these necessary data is included in the exported .xlsx file. Afterwards use OpenOffice/Excel to export the data again into a .csv file (comma separated file) to be feeded to the python scrip build_profile.py after having modified the profile_config.ini file to your taste and file names used and desired.

OK, I just generated a larger patch set target for the Colorimeter to be printed on 2 sheets of A4.

It has twice amount of patches compared to the previous 247 patch target.