[Tool] ChromIQ – a macOS and Windows GUI for ArgyllCMS printer profiling (v4.3.0)

[itsab1989]

It is hard to say what exactly went wrong. Some screenshots, the log files and / or the project folder might help to look into this.

Please take the following with a grain of salt because I am not 100% sure on this.
In the Check&Refine tab ChromIQ uses Argylls profcheck to get the data. This checks how well the values you have measured fit within the profile you have created. It is expected that the paper can’t exactly reproduce what you see on the screen but when the profile is created it takes into account your measurement but also does some smoothing and many calculations to correctly weigh all of this data.
If some patches you measured are then really far off that can shift a big part of the profile. And then check and refine tells you that this did not come out as expected. So chances are that when you re-measure the patches with the highes deviation and get them right the next time the whole profile falls more in line with the other measurements you have made. However you can’t say this for sure.
The judgement about the quality comes more or less only from how many patches fall out of the threshold you set in the tool.
And my thought was that when it flags so many patches that you uave to remeasure more then 50% of the chart it might be better to tell the user to start from scratch instead of reading single strips again.
Maybe this recommendation is not the best though.

Regarding how easy a measurement should be this depends a little bit on your skill as well. Of course making margins and patches bigger will make things easier and more forgiving (at least it should). But it helps a lot to move the device slowly, smoothly, always with the same pressure etc.
This sounds easy but I found it takes some practice. And on the flip side if I move my Colormunki too slow I get misreadings as well. Then it has a tendency to put your measurements to the wrong patches. It happened to me that they are off by one position a few times. The measured vs expected overlay and the red outlines helped me to spot those issues.
Truth is I can’t tell you whether this is an issue of only the ChromIQ chartreading engine or Argylls chartread as well. Because chartread is a black box and you can’t visualize its measurement the same way. But since we base our engine on the exact same code I suspect that it is actually the same for both and our engine only makes those errors easier to spot and correct.

 

[Im-back-Arno]

Thanks for the rapid feedback!

This screams linear drive. Amazon has a nice set:

Linear

Add a NEMA 17 Stepper motor and a SK-ZMC02 programmable Controller and we’re in business for about €70 or so.

My i1Pro2 rail is a bit “grabby” since it’s dry plastic sliding over dry aluminum and I did add a tiny drop of trimmer oil already but friction remains dependent on pressure (the head seems to be grabbing the paper just slightly as opposed to hovering just above it) so consistent sliding is not easy. Yes, it seems to require practice. I should just automate it which would be a nice one day job. Just need to find some time (…).

Cheers,

Arno

 

[itsab1989]

But keep in mind that what I said is not set in stone. People with more experience might have a better guess / advice.

If you manage to attach your log file here and / or the project folder this might help Claude to figure things out. I think my subscription runs out in maybe 12 days or so and after that I‘d rather not re-subscribe for a few months. I know I have said this a few times already and then right when the subscription ran out someone reported a bug which I did not want to leave unfixed so I subscribed again…

 

[know]

I then tried a three pager, also 1944 patches with extra skin tones as opposed to neutral grey optimized. This did give me a profile, but again it complained that there were too many large deviations from the measurements so I re-scanned the rows where the largest deviations were and lo and behold I got an excellent profile (at least that’s what it said). During the first measurement I had quite a few rows with a couple of red boxes around the patches and rescanning these rows did not remove the red boxes but apparently did give better measurements which confused me somewhat.

Hi @Im-back-Arno ,
Thanks for testing

I don't know if you read the help text for the "Show patch values on hover" checkbox in the Measure tab. It has changed since we did a lot of changes to detect better large measurement errors. You should probably read it again.

The thresholds for when errors are detected has default settings, but may not suit all users, depending on their printer and other factors. You can try to change the settings in Preferences --> Measurement and see how the highlighting on measured page changes. Also read the help information for the patch error limit and Neighbour check to understand more on the setting of the limits. Rereading patches may confirm a reading, but it may not always turn yellow or green, depending on the result of the re-reading of a patch. However, in many cases they should turn yellow. If both patch error and the neighbour check triggers as red errors, and the re-read only confirms one of the two errors, the patch may still stay red.

If the limits are not well suited for your printer-paper-instrument combination being measured then too many patches may light up. You should then consider changing the limits to prevent trigging on too many correct readings.

Here is an example measured chart with highlighted patches for the following limits:

The yellow patches were re-readings that confirmed the very high deviation from expected values (patch error limit triggered) because the printer cannot reproduce these colors well on my printer.

Here more patches are flagged as red, and they were not re-read by me.

Hover over the highlighted patches to see details about the measurements and the triggered errors that caused the highlighting.

Cheers,
Knut