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

[pharmacist]

I tried to find the "testprofile v2.csv" in the profile_config.ini but could not find it to remove the wrong file name. I will give it a try tomorrow. Meanwhile you might want to have a look at my files above and see if you can generate the profile and compare it to the one generated with the i1Pro1. The 3D model should be different and larger than the previous one (which was made with pigment ink and on matte paper, so a significantly smaller gamut than this new one on glossy cast paper). You might upload your results already so I can have a look at tomorrow.

 

[x64]

That's the thing, "testprofile v2.csv" is hardcoded in "cr30_to_ti3.py" and is only used when running cr30_to_ti3.py directly without adding --csv, --ti1, and --out arguments (what you did), but when running build_profile.py the config gets used instead. I patched that in version 0.1.1 to avoid confusion.

I was able to generate profiles using those files. While doing so I noticed some other improvements (e.g. using the ti2 instead of the ti1, reading the header more dynamically) that will make the script more robust when inputs change. I'll work on it a bit more, expect another version later today. Your exact workflow already works with version 0.1.0 and should be able to produce accurate ti3 files and forward them to colorprof to directly generate a profile, with a correct config and running either:
python3 src\build_profile.py
python src\build_profile.py
or when you use python launcher:
py -3 src\build_profile.py

 

[pharmacist]

OK many thanks for the update. I am now at work. Could you already upload the profile that is generated from the CR30 scanned csv file, so I can have a look at in iccview.de. It would be better to use the ti2 file, because sometimes reordering the patches (to scrambe patch layout) is better for scanning.

 

[x64]

Here is the profile generated with the ti2 and CSV of the 228-target set you shared before. This is not the action glossy set because that one didn't include a ti2.

 

[pharmacist]

ok, I thought I did upload the ti2 file with the set. I will have a look at when I am at home this evening and also will have look at your latest CR30 ti3 dispensary later on to see if I can generate the profile using the ti2 file and my scanned CR30 csv file.

 

[x64]

I just finalized 0.2.1 with a lot of improvements and simplification. The workflow and output remains the same but will not decide anymore what data to pass to the ti3: it simply takes all data you provided in the CSV.

We now also take some header items from the ti2 and include them in the ti3 like a normal ti3 would do. You can set this in the config so you can define what header items you want to pass to the ti3.

 

[pharmacist]

Many thanks @x64 : the only thing is the cr30_to_ti3.py does not function well but the build_profile.py immediately gave me the ti3 file.... My first impression is the profiles generated with the cr30 is very similar in volume and size and shape compared to the ones created with the i1pro1 spectro.....to confirm this I will do some extras tests but my first impression: the cr30 is very capable in producing good printer profiles; which is a fantastic discovery to be a poor man solution for creating printer profiles. Most probably it can be used too for screen calibration as well.....the only con is the limitation of spot reading only. I will publish my findings later in with some extra tests and comparisons with different papers and printers to come to a conclusion.....but the first impression is very promising indeed.

 

[pharmacist]

Today I was able to do 3 different combinations with 2 printers and 2 types of papers:

Epson SC-P900 with Calumet Matte 230 gsm photo paper
Canon TS705 with Action Glossy 210 gsm cast photo paper
Epson ET-8550 with Action Glossy 210 gsm cast photo paper

Using a 228-patch target that is printed on 2 sizes: one large for the CR30 colorimeter and a smaller version with separation bars for the i1Pro2.

This is the CR30 colorimeter that costs about €55-60 (depending on the offers at aliexpress,the price can fluctuate but 55-120 euro, so look carefully at the offers because the price differs greatly). Also be sure to choose the CR30 als the CR10 and the CR20 are less accurate.

The CR30 patches are read with the ColorQC2 program that registers the LAB-values and also the spectral data from 360 to 780 nm in steps of 10 nm and is a LED lamp based colorimeter:

The data is exported in excel .xlsx and later converted to a .csv (comma separated) file in excel, because there is bug in the ColorQC2 program does not export well.

With help of the python batch files of @x64 (again many thanks for the help) I could generate a argyllcms compatible .ti3 file out from the corresponding .ti2 matrix file (228-target.ti2)

For control/comparison I also printed exactly the same target (but much smaller patches and with separation bars) to be scanned with a vintage i1Pro1 (lamp: tungsten):

Below a sample with on top the small patch i1Pro target and below the exactly same target with large (14.7x14.7 mm) patches for the CR30 colorimeter:

Here same print screens in iccview.de (below you can download the 3 profile sets: each set is for the same paper/printer combination but with 2 different profiles of which one is read with the i1Pro1 and the other with the CR30 colorimeter for side by side comparison in iccview.de)

Have a look the white point values and the gamut volume between the two twin profiles: they are very similar in value and shape/size. You can download the profile sets below for a 3D view yourself for comparison.

But the proof is in the eating of the pudding, so I printed on my Epson ET-8550 the same test picture with above with the CR30 colorimeter and below with the vintage i1Pro1 spectrophotometer:

The most striking thing is the CR30 (LED) scanned profile produces more neutral black and white prints compared to the vintage i1Pro1 (tungsten: maybe yellowing of the lamp ?). The overal print quality is very similar between both prints using the profiles generated using 2 types of devices.

The best way is to print using both targets. @Ink stained Fingers : I know you have the ET-8550 printer too and also uses the Action Glossy 210 gsm cast paper of the action for tests. You might have look at those profile too (see below).

Conclusion: with some programming (many thanks for @x64 ) and careful reading each patch it is feasible to create satisfactory printer profiles. For this test I am using a small patch set of 228 patches, which should be sufficient to create a good printer profile (likewise the Spyderprint also uses a 225 patch "High Quality Target").

 

[Ink stained Fingers]

Yes, I'm running the ET-8550 as well, and I have some sheets of the Action Glossy available which I profiled some time ago. I'm using the 106 inks of the ET-7750 just for pricing reasons, I'll check and compare the profile with yours by tomorrow.

 

[pharmacist]

I will already upload my new target that has 247 patches of which 21 grey steps included for those interested. It fills up the whole A4 area with 14x14 mm patches completely.