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

[Epatcola]

OK, so for the moment only non-randomized patches will work. This is not a big deal, but when using several grey steps, because the hue increased with each following patch with a little but and sometimes without spacebars it is difficult to determine the end/beginning of the next patch. Therefore I have uploaded a modified i1Pro based target with tiny space bars between the patches making it much easier to determine where the next patch is starting.

I am using 8mm patches and the rule I designed so you only need to see a couple of patch 'joints' to align the rule.

I have never had any ti3 files from chartread to look at. It appears the colprof needs the input RGB values (from the ti2 file) followed by measured values on each data line. I don't see that colprof needs SAMPLE_LOCs for anything. What it does need is the RGB values for a line followed by the measurements from the patch associated with the line. I think the simplest way to achieve that with randomized targets would be to sort the data lines of the ti2 file on SAMPLE_LOC. I will try that.


Also I don't see anything wrong with the csv files ColorQC writes. Comma separated with decimal points in numbers. Did you have it mess up using commas instead of decimal points? Probably some locale induced crap.

 

[Epatcola]

Yesterday I created an A5 target with 384 randomized 8mm patches. I printed it on half an A4 sheet of Epson Premium Glossy paper on an ET-7750.

I measured the target this morning which took nearly 20 minutes. I am sure with practice and a bit more care that could be reduced. Aligning the rule with the paper sliding around took time and the USB lead was shorter than ideal.

I don't know python at all to be able to modify x64's script. I do know awk enough to write a quick and dirty script to create ti3 files for randomized targets. I generated a profile and apart from a small kink on the orange edge it looks good to me. Maybe I misread a patch or two.

I attach a zip with all the files and the Epson provided profile for comparison.

You can get awk (gawk) for windows here.

This is the first profile I have created with argyll. Any advice on optimum parameters particularly for colprof would be welcome.

 

[Epatcola]

And a few batch files named paper_sizexpatch_countxpatch_size to generate various targets on one or two sheets.

 

[pharmacist]

with te latest version of ArgyllCMS I get an errorthat the ink limit is set to 290% by standard using colprof after I have done the the chartread command to read my targets. The older generated targets with ti2 do not have that problem. Not sure what is happening. However: it did not effect the resulting profile.

Anyhow I tried the trick again today but this time using a 300-patch target that is still readable without any jig/ruler to scan with the CR30 colorimeter. I Always look at my screen if the scanned color is more or less the same to check I have read the right patch color. This target has 32 grey steps.

For those having both spectro's (i1Pro1/2) and this el cheapo Colorimeter can make profile side by side and see for yourself how good the CR30 is. I just printed the same matrixlarge.jpg image with 2 different profiles: one generated with the CR30 and the other with the i1Pro2 and the magic is still there: visually exactly the same (gamut volume are more or less the same).

 

[pharmacist]

There is also a bug in the gui.py (@x64, @micronica ): despite you can give an external icc name, the internal icc name (that is displayed in the profile setting of PS or Qimage) is just "profile". I think this should be corrected. I had to use small ancient tool called changedescription to change the internal name after the profile has been created. I had to find on date because I could not find my profile in the profile setting in Qimage.....seems the internal name was "profile".

 

[pharmacist]

Today I printed the CR30 collar (thanks @Epatcola ), but I discovered the hole is a bit too tight and I had to increase the hole of the collar a bit. I redesigned it and also the separate ring.