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

[pharmacist]

I have made a new A4 target with pseudo-scrambled patches for the CR30 with 300 patches and 21 grey steps. I will upload it later with a rudimentary how-to manual if I have some time, because it is rather busy at work now.

 

[Epatcola]

Update to target and profile generating batch files and scripts.

The ti3 generation script didn't properly handle targets with more than 26 columns.

Also included is a csv file from ColorQC 2 for a 728 patch target. You can use this to go through the process of ti3 file and profile generation and checking.

 

[superjudge]

Long time listener, first time caller. I tweaked the batch files in Epatcola's AWK scripts to generate 10mm patches using -a1.25 and set it to create 25 patch columns on letter-sized paper. Made it WAY easier to scan patches for me after trying repeatedly with 8mm patches. I was able to fit 450 patches on one sheet of letter, and it resulted in a profiles that I'm honestly VERY happy with. I was getting a lot of jaggy edges on the visualization for the 300x8mm patch scans.

 

[pharmacist]

you're welcome. Happy to see another person able to generate a printer profile using the CR30 colorimeter.

 

[Epatcola]

Long time listener, first time caller. I tweaked the batch files in Epatcola's AWK scripts to generate 10mm patches using -a1.25 and set it to create 25 patch columns on letter-sized paper. Made it WAY easier to scan patches for me after trying repeatedly with 8mm patches.

The hole in the CR30 is 4.2mm so 8mm patches allow only about 2mm of misalignment. I think you did well to measure 10mm patches without a guide.

A couple more things I learnt. Elsewhere someone asked about the CR30 light source. It is 3 bright white LEDs focused to a small patch on the measurement surface. I can't measure their spectrum. The three looks the same to my eyes, but, I suppose they could be different to help fill in a full spectrum.

If you look up the hole you can see the sensor is a 4 x 4 array of somethings. I assume they are photodiodes with color filters. The CR30 provides 31 spectral readings so I guess 16 are direct from this array and another 15 are interpolations between pairs.

I only have a vague understanding of argyll profile generation. I wondered why the ti3 file would need both measured Lab and spectral values. Turns out it doesn't. I removed the Lab values from a ColorQC 2 produced .csv file and slightly modified the awk script to produce a new ti3 file. Profiles generated by colprof from original and new ti3 files were very similar but not identical.

I guess colprof uses Lab values (and the specified illuminant) if present and the small profile difference was down to ColorQC 2 interpreting the spectral values slightly differently to calculate Lab values.

 

[pharmacist]

The hole in the CR30 is 4.2mm so 8mm patches allow only about 2mm of misalignment. I think you did well to measure 10mm patches without a guide.

A couple more things I learnt. Elsewhere someone asked about the CR30 light source. It is 3 bright white LEDs focused to a small patch on the measurement surface. I can't measure their spectrum. The three looks the same to my eyes, but, I suppose they could be different to help fill in a full spectrum.

If you look up the hole you can see the sensor is a 4 x 4 array of somethings. I assume they are photodiodes with color filters. The CR30 provides 31 spectral readings so I guess 16 are direct from this array and another 15 are interpolations between pairs.

I only have a vague understanding of argyll profile generation. I wondered why the ti3 file would need both measured Lab and spectral values. Turns out it doesn't. I removed the Lab values from a ColorQC 2 produced .csv file and slightly modified the awk script to produce a new ti3 file. Profiles generated by colprof from original and new ti3 files were very similar but not identical.

I guess colprof uses Lab values (and the specified illuminant) if present and the small profile difference was down to ColorQC 2 interpreting the spectral values slightly differently to calculate Lab values.

Is the difference visible in the final printer output using this printer or not and if so can you guide me to your new awk workflow (link ?). Many thanks for your fantastic work you already have done here.

 

[Epatcola]

Is the difference visible in the final printer output using this printer or not and if so can you guide me to your new awk workflow (link ?). Many thanks for your fantastic work you already have done here.

I just inspected the profiles. The only measurement set I have is from a 728 patch pain paper target so not worth printing anything with it. I attach ti3 files with and without Lab values so you can generate your own profiles to inspect/compare.

I actually just deleted the LAB columns in the .csv file and edited the ti3 file generated by the original script changing NUMBER_OF_FIELDS from 39 to 36 and removing the 3 LAB fields in the DATA_FORMAT line. Awk script changes would be similarly trivial. The script writes SAMPLE_ID SAMPLE_LOC RGB_R RGB_G RGB_B values then appends a line from the .csv file for each patch.

I have an i1studio and have already profiled the printers and papers I have. I got and have been messing with the CR30 for interest/fun. Maybe I will get some new paper or see if I can make a better B&W profile for some papers and have a real use for the CR30.

 

[superjudge]

The hole in the CR30 is 4.2mm so 8mm patches allow only about 2mm of misalignment. I think you did well to measure 10mm patches without a guide.

A couple more things I learnt. Elsewhere someone asked about the CR30 light source. It is 3 bright white LEDs focused to a small patch on the measurement surface. I can't measure their spectrum. The three looks the same to my eyes, but, I suppose they could be different to help fill in a full spectrum.

If you look up the hole you can see the sensor is a 4 x 4 array of somethings. I assume they are photodiodes with color filters. The CR30 provides 31 spectral readings so I guess 16 are direct from this array and another 15 are interpolations between pairs.

I only have a vague understanding of argyll profile generation. I wondered why the ti3 file would need both measured Lab and spectral values. Turns out it doesn't. I removed the Lab values from a ColorQC 2 produced .csv file and slightly modified the awk script to produce a new ti3 file. Profiles generated by colprof from original and new ti3 files were very similar but not identical.

I guess colprof uses Lab values (and the specified illuminant) if present and the small profile difference was down to ColorQC 2 interpreting the spectral values slightly differently to calculate Lab values.

I did make myself a guide, I just felt like I was getting slight blends too often, so I went to 10mm `-a1.25` for the sake of increasing accuracy at the expense of limiting myself to 450 patches per letter-sized page.

I also, with the guide design, wanted a lot of hand rest area to help keep things from sliding around.

 

[superjudge]

I made a bat file to process a dropped CSV file into an ICC profile. Make sure you edit the first few lines of the bat to indicate where gawk.exe, genti3.awk (or genti3semic.awk), your target's ti2 file, and you can add your printer brand and model.

It'll copy the the ti2 and csv file so the basenames match, create the ti3, ask you to verify your printer brand/model, give a display name for the profile, and generate the ICC with a name based on that data (or some other name you specify).

After it creates the ICC, it'll ask to clean up the working files for the generation.

 

[Epatcola]

I made a bat file to process a dropped CSV file into an ICC profile.

Thanks for sharing. I don't know if you followed older threads, but, I have posted that as far as I can see for the ET-8550 all glossy and lustre or semi-glossy media settings are the same except ultra glossy which has a best quality setting that I think needs its own profiles.

I don't think there is any point using more than these media types and settings for profiling any 'real' photo paper.

Ultra glossy High
Ultra glossy Best
Matte High
VFA

All my ET-8550 profiles are annotated with UGH, UGB, MH, or VFA.

The photo quality ink jet media setting might be different, but, I don't know what kind of (other) photo paper you might want to use it for.