[Tool] ChromIQ – a macOS and Windows GUI for ArgyllCMS printer profiling (v4.0.0)

[Alan G]

One reason for more patches, is the size of the print. If one is only printing letter or A4 prints, it more difficult to see color transitions unless you are using a loupe. However, when you move to larger prints color gradations are more important. I have done some office installations with 13 x 19 and 17 x 22 inch prints and an excellent profile is a necessity.

 

[itsab1989]

@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.

 

[pharmacist]

The TC9.18 target is often used by commercial profile companies as the standard target of choice, because it uses an RGB sampling of 9x9x9 and very importantly a good sampling of the near neutrals around the neutral axis. Normally it is often difficult to establish the neutral axis with perfectly neutral greys only, you will need stained grey ramps in several color area's (RGB + CMY) so the software can cube-fit the nearest possible point using these stained greys.
By strengthening the sampling count around this near neutral area with the extended gray target of Spyderprint you can get optimal results without the need of a 2 step method: first with standard target to determine where the exact neutral axis is for your printer/paper/ink combination and a second one to optimize and smoothen the first readings to obtain a final optimized target.

It is for most people more than enough, however if you want really accurate profiles you will need more measurement points and range of RGB= 10x10x10 or even higher like 12x12x12, but for the human eyes the most important area is the area around the neutral axis for which the human eye is the most sensitive for color casts, so taking more sampling points in this critical area is more beneficial than more points in the rest of the color cube for a given patch count.

I just finished printing my special ChromeIQ Colormunki/i1Studio scanning jig:

It is for Landscape scanning of A4/Letter and can be easily printed on a standard 3D printer and is light and sturdy.

If interested I can upload the stl-file.

 

[itsab1989]

Nice! You even added the name on it! The „e“ is extra though

 

[pharmacist]

Oh that is just a cosmetic fault, Can be easily corrected

 

[itsab1989]

Oh that is just a cosmetic fault, Can be easily corrected

View attachment 18702

Believe me - I know that feeling

But since you are already designing custom accesories for the app is it save to assume that you are at least to a good degree happy with it?

 

[crenedecotret]

Edit:
I just read the description of your process in your other thread and I am lacking the knowledge to fully understand what and how you do it


Edit 2:
3.7.33 now supports loading i1Profiler measurement data into the build profile tab by using txt2ti2 when a txt file is selected. Based on the information of @pharmacist here:
https://www.printerknowledge.com/th...-spectro-with-i1profiler-and-argyllcms.17061/

Honestly, I used plain chatGPT to brainstorm ideas with me. I told it I wanted something really easy like colormunki's process and then I had Claude build the code. I can get a decent profile on a colormunki with two sheets of letter paper without using any rulers or mini-patches.

I asked Claude to go through the two-step wizard code and generate an .MD file to describe everything being done (attached, renammed as .TXT). You could probably feed this back to Claude in plan mode and implement the feature in ChromeIQ.

The idea is this
1. Build a one page profile
2. Select RGB values where the initial profile was weaker + some anchor points (like skin tones) + some neutrals from a bank of 10000 or so candidates. The RGB values are run through the profile to figure out what was the Delta E on those. There is also some weighing done to make sure the patches are not too close to the original profile. Note... These RGB values ALONE would probably not build a good profile (this creates a TI1 file)
3. create a target based on the selection done at step 2 (this created a TI2 file and TIFF)
4. Have the user measure the second sheet (this created the TI3 file)
5. combine the TI3 measurements from the initial profile and the extra set of patches.
6. build the profile from the combined measurements
Since the second set tries to improve in weak areas and areas important to human perception, this might be more noticeable than a single profile with preconditioning

There are a few things I decided in WSProfiler.. most advanced options are hidden and my GUI is really built for beginners that dont care about tweaking things. I dont even get rid of the clipboard area for i1pro. You'll probably want to expose features and re-use your own chart reading code.

Maybe a prompt like this will get you there "Read two_step_profiling_explained and make a plan to implement a similar wizard in ChromIQ. Make sure to re-use our chart display and reading code. Ask questions until I tell you I'm satified with our plan"

 

[itsab1989]

@crenedecotret
Is there a version of your app that is running so we could see the process in action with real results?
When I read this it sounds smart to me. But the truth is I have no idea whether it is actually good or not.

 

[Alan G]

I successfully read a 506 patch set on letter size paper with my i1 Pro for a precondition profile. Reading went smoothly with very low errors in each patch. I use -g32 for the gray ramp. I have just printed two pages (1012 patchs with -n64 for the final profile and will read these tomorrow. I can compare this profile with the one I did several years ago with higher patch counts for the pre-condition profile and the final one.

 

[crenedecotret]

@crenedecotret
Is there a version of your app that is running so we could see the process in action with real results?
When I read this it sounds smart to me. But the truth is I have no idea whether it is actually good or not.

You can grab it here. https://github.com/crenedecotret/wsprofiler. It's python so it should work pretty well in Windows or MAC but I only tested it on Linux

The second page visually is interesting when printed out.. many neutral and "natural" colors like skin tones. The algorithm seems to work well. I've never had any real issues with one page profiles except in for some paper/ink combos so yes, it's all very theoretical. Here is what the "robot" had to say about it. It's mostly the "low novelty" that makes a difference. I you build a 400 patch profile and then another 400 patch profile with preconditioning, there is some overlap...


Why This Approach Works Better Than `targen -c`

Argyll's `targen -c precond.icc` generates patches by:
1. Building a uniform grid in the **intermediate profile's PCS space** (Lab)
2. Inverting through the profile to find corresponding RGB values

This tends to concentrate patches near the gamut **surface** and gives fewer
samples to the interior and the neutral axis. It also produces many patches
that are perceptually very close to Pass-1 patches (low novelty).

The wsprofiler approach explicitly addresses these weaknesses:
- Each patch is scored for **novelty** relative to Pass-1 — redundant patches
score near zero
- The **region-coverage** term forces exploration of under-sampled Lab voxels
- The **neutrality bonus** and **luma balance** terms ensure gray-axis and
tonal-range coverage
- The **min_dE gate** prevents clumping even within a single iteration

In practice, this produces a Pass-2 chart that fills perceptual gaps left by
Pass-1, giving the final `colprof` a more uniformly sampled measurement set
and a better profile overall.