[Tool] ChromIQ – a macOS and Windows GUI for ArgyllCMS printer profiling (v3.0.0 beta 10)

[Alan G]

@Alan G
May I ask why you are seemingly creating so many profiles on a regular basis (you mentioned that you have already created 5 with my app)?
I mean I want to improve this app to a point at which it is a real help for beginners and experts. But to be honest I will then probably not very often use it myself. My printing habits are pretty boring - same printer and two types of paper - not much experimentation.
Pharmacist told me he is mixing his own custom inks and I assume that why he needs profiling pretty oftern. I was just wondering whats your reason for it.

I have half boxes of papers that I do not print on any longer and am using these to play around with the app. Three of the profiles were just one page patch sets to test various settings and give you feedback. The other day I did a profile with a pre-condition profile. Yesterday and today I tested out the higher density patch set with my i1 Pro (506 patches per page at the 0.95 setting). I wish I had the app a month ago as I was testing out three new Canson papers that I might want to use for further use. My work flow was a two page patch set (924) for the preconditioning profile and four pages for the final profile (1848).

Having settled on just a few papers, I don't need to do much profiling.

 

[pharmacist]

I am trying to test the TC9.18 + extended grays target tomorrow. I have already printed the 2 sheets and will scan the targets tomorrow and report back.

 

[itsab1989]

@crenedecotret
Somehow I did not get an email notification about the new posts. I just saw it now by chance.

I have created a beta version with part of what your process does: keeping the first measurement and merging it with the measurments of the second pass with the preconditioning profile. It seems that the merged ti3 is accepted by colprof but it really feels like hacking at this point and needs testing. I don't even know if a profile created with this process (at least at this stage) is better in any way than a regular pre-conditioning pass profile.

But I am sure @pharmacist can spot the difference

https://github.com/itsab1989/ChromIQ/releases/tag/v3.8.0-beta.2

Copied from Claude:

What the feature is (for your testers)

An optional "ChromIQ-style refinement process" (Settings toggle, off by
default). When on, ChromIQ reuses the measurements from an earlier profile to
build a more accurate new one: you pick a previous profile as the Refinement
profile in Create Chart, its measurement data is kept beside the chart, and at
build time it's merged with your fresh measurements so colprof builds from a
larger patch set. The result is saved as …_merged.icc. With the toggle off,
nothing changes.

What testers should focus on

1. Off = unchanged — confirm the normal workflow is identical with the setting
off.
2. Happy path — enable it, pick a prior profile, print/measure the new chart,
tick the new Measure-tab option, build → confirm a clean …_merged.icc.
3. Guided Check & Refine after a merged build — confirm it only flags strips
physically on the printout (never an earlier-run patch), and that re-measure +
rebuild works.
4. Mismatch safety — point at a profile measured with a different
instrument/colour space → expect a clear message and a fallback build from the
new measurements only.

 

[itsab1989]

I have half boxes of papers that I do not print on any longer and am using these to play around with the app. Three of the profiles were just one page patch sets to test various settings and give you feedback. The other day I did a profile with a pre-condition profile. Yesterday and today I tested out the higher density patch set with my i1 Pro (506 patches per page at the 0.95 setting). I wish I had the app a month ago as I was testing out three new Canson papers that I might want to use for further use. My work flow was a two page patch set (924) for the preconditioning profile and four pages for the final profile (1848).

Having settled on just a few papers, I don't need to do much profiling.

Wow, thanks for doing all of this!

 

[crenedecotret]

@crenedecotret
Somehow I did not get an email notification about the new posts. I just saw it now by chance.

I have created a beta version with part of what your process does: keeping the first measurement and merging it with the measurments of the second pass with the preconditioning profile. It seems that the merged ti3 is accepted by colprof but it really feels like hacking at this point and needs testing. I don't even know if a profile created with this process (at least at this stage) is better in any way than a regular pre-conditioning pass profile.

But I am sure @pharmacist can spot the difference

https://github.com/itsab1989/ChromIQ/releases/tag/v3.8.0-beta.1

Copied from Claude:

What the feature is (for your testers)

An optional "ChromIQ-style refinement process" (Settings toggle, off by
default). When on, ChromIQ reuses the measurements from an earlier profile to
build a more accurate new one: you pick a previous profile as the Refinement
profile in Create Chart, its measurement data is kept beside the chart, and at
build time it's merged with your fresh measurements so colprof builds from a
larger patch set. The result is saved as …_merged.icc. With the toggle off,
nothing changes.

What testers should focus on

1. Off = unchanged — confirm the normal workflow is identical with the setting
off.
2. Happy path — enable it, pick a prior profile, print/measure the new chart,
tick the new Measure-tab option, build → confirm a clean …_merged.icc.
3. Guided Check & Refine after a merged build — confirm it only flags strips
physically on the printout (never an earlier-run patch), and that re-measure +
rebuild works.
4. Mismatch safety — point at a profile measured with a different
instrument/colour space → expect a clear message and a fallback build from the
new measurements only.

View attachment 18704

It's not quite the same as my "hacked" process because your second set of patches does not specifically target things that went wrong in the initial profile.

You might not see any visible improvement but there is no point throwing away the information from the initial set. They are valid and perhaps useful patches that you took time to measure. I used to do the same thing in a plain profile editor just because I could It wont make the profile worse

 

[itsab1989]

It's not quite the same as my "hacked" process because your second set of patches does not specifically target things that went wrong in the initial profile.

You might not see any visible improvement but there is no point throwing away the information from the initial set. They are valid and perhaps useful patches that you took time to measure. I used to do the same thing in a plain profile editor just because I could It wont make the profile worse

I am aware that this is not as sophisticated as what you did. I just had not seen your post and did not know how you approached this. But I thought I could at least start with the implementation by doing it this way.
But I still don't know how you produce the second target because somehow it sounds to me like you completely avoid targen in the second run and create the optimized values yourself.

 

[pharmacist]

OK, trying to read row Y (on the following page): a cosmetic error:

next page is row Y, it says: scanning row z:

All rows read, but the marker indicates the previous row being read:

With B/W spacers it is better:

However the target is not optimal, as the non-randomized layout gives problem with the spacers (too less contrast difference so the adjacent patches are seen as one throwing an error of no enough patches read):

I have to manually add spacers to have the spectro see the difference of adjacent patches in order to correct the errors. It was worse with the colored patches (especially the red space bars between red patches):

I will try randomized patches in order to have maximum color difference between adjacent patches. Also the system throws an error when producing the profile, so there is something wrong with the workflow:

 

[pharmacist]

OK now trying randomized patch order in order to get maximum color difference between color patches:

Randomized with B/W spaces no problem at all, thus we should change from non-randomized with B/W spaces as a default workflow:

OK it was succesful:

But the softproofing is horrible, so I think the workflow is broken:

I strongly suggest to remove this workflow. I will upload the standard TC9.18 without the extende grays, because I think this is causing the problems:

 

[pharmacist]

OK now trying the standard TC9.18 target (already having alot of stained near neutral ramps):

Workflow is now perfect (softproofing no longer strange colours):

 

[pharmacist]

The following setting gives exact evenly distributed rows, so maybe the special ChromIQ legenda (printer/date/paper/profile name) can be added to this TC9.18 workflow. @itsab1989 :

Especially the printtarg arguments are important for 2 exactly the same row sizes for both pages (page 1/2 and 2/2):

I suggest to immediately remove the faulty TC9.18 + extended grays as it has some flaws and does not function properly.