Printer Profile Validation: M1 vs. M2 and ΔE2000 Evaluation

[jctay]

Just an update with the latest version of ChromIQ with the bug fix for the truncated values of i1Profiler.

I have scanned the same target again twice: the first with i1Profiler and the target generated with the i1Profiler engine (neutralize greys at 50 procent) and again with ChromIQ (with the ArgyllCMS i1Pro driver loaded) and the target generated by ChromIQ with standard settings:

View attachment 19718

This time the white point of both profiles are exactly the same. A slight difference in gamut volume (most probably by slight scanning deviations within the margin of error)

Softproof with the ChromIQ version 4.3.20 scanned and ChromIQ generated target:

View attachment 19721

Target with the new exported txf file from ChromIQ 4.2.20 for i1Profiler scanning and profile generating by i1Profiler:

View attachment 19722

The difference is now negligible and no slight yellowish hue in the ChromIQ profile softproofing like previously.

Digitization Resolution 100x100 and with wider pixel sampling

 

[jctay]

OK, I just have succesfully scanned a ChromIQ target with i1Profiler using the dual scan method and it works like a charm.

View attachment 19700

View attachment 19701

I have moved the letters of the row in such a way it is not scanned by the spectro (using the standard ChromIQ setting a big black letter is on the headway of the scanning eye). I can now safely start from the scanning jig white over the white paper edge and end on the other side of the scanning jig white without any errors:

View attachment 19702

The 3 page target (13x18 cm) with 702 patches is considered a one long page in i1Profiler.

View attachment 19703

Now I am saving the scanned data in mxf format for ChromIQ to generate the target:

View attachment 19704
I have generated a target too with i1profiler and with ChromIQ using the same data from the mxf-file and it is interested there is a slight difference (but negligible) in the white point, which is interesting, but to be expected because of the slight difference how those 2 engines handle perceptual intent.

View attachment 19705


Here a softproof (relative colorimetric with BP=on), first picture with the i1Profiler generated target:

View attachment 19706

The softproof using the ChromIQ generated target (using the standard settings):

View attachment 19707

Hi @pharmacist

Could you upload the following files for checking?

Please ensure that all three files come from the same project in which the MXF file was created.

1. ChromIQ .it2 file
2. i1Profiler .mxf file
3. ChromIQ .it3 file (generated after importing the MXF file)

TIA!


( I'm troubleshooting the chart layout because the MXF file fails to import into ChromIQ for measurement. )

 

[pharmacist]

Hi @pharmacist

Could you upload the following files for checking?

Please ensure that all three files come from the same project in which the MXF file was created.

1. ChromIQ .it2 file
2. i1Profiler .mxf file
3. ChromIQ .it3 file (generated after importing the MXF file)

TIA!


( I'm troubleshooting the chart layout because the MXF file fails to import into ChromIQ for measurement. )

Here you are. The version v2 is done with the latest 4.3.20 version of ChromIQ (ti3/mxf) with the bug fix for the problem with truncated versions.

Note: the older version is scanned only with i1Profiler dual scan method and from this mxf I made 2 profiles with i1Profiler proprietary profile generator + imported in ChromIQ version 4.3.19 and converted to ti3 (which is a converted mxf from the earlier i1Profiler measurement), so actually the mxf and ti3 should be the same data.

The newer version (v2) has two measurements (so independently): the mxf with dual scan with i1Profiler and the ti3 is generated within the ChromIQ program version 4.3.20 and gives the v2.ti3 file. This session had 2 independent scan values.

This is you want to take into account for a fair comparison. The ti2 file is the same for both sessions (v1 and v2)

 

[jctay]

Here you are. The version v2 is done with the latest 4.3.20 version of ChromIQ (ti3/mxf) with the bug fix for the problem with truncated versions.

Note: the older version is scanned only with i1Profiler dual scan method and from this mxf I made 2 profiles with i1Profiler proprietary profile generator + imported in ChromIQ version 4.3.19 and converted to ti3 (which is a converted mxf from the earlier i1Profiler measurement), so actually the mxf and ti3 should be the same data.

The newer version (v2) has two measurements (so independently): the mxf with dual scan with i1Profiler and the ti3 is generated within the ChromIQ program version 4.3.20 and gives the v2.ti3 file. This session had 2 independent scan values.

This is you want to take into account for a fair comparison. The ti2 file is the same for both sessions (v1 and v2)

Thank you.


Group 1
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target.mxf
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target.ti3 (Sequential numbering)
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target_v2.mxf

Where
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target.ti3 is converted from
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target.mxf
using Measurement Condition M2.


Group 2
TPI_Satin_240_gsm_Epson_ET-18100_EPG_standard_ChromIQ_target.ti2 (with SAMPLE_LOC)
TPI_Glossy_240_gsm_Epson_ET-18100_EPG_Standard_ChromIQ_target_v2.ti3 (with same SAMPLE_LOC)

No MXF file for this group


Left : Group 1
Right: Group 2

This is therefore a special project designed to make the .ti3 file compatible with ChromIQ Measurement for import.

 

[jctay]

Create a ChromIQ Project from an MFX File

Would it be feasible to incorporate the following features into ChromIQ?

1. Direct MXF File Import
. Allow users to import MXF files directly to create a ChromIQ project.

2. ICC Profile Creation and Verification
. Support ICC profile creation and profile verification, consistent with the functionality available in other ChromIQ projects.

This would allow existing i1Profiler users to immediately benefit from ChromIQ's functionality.

Thank you.

 

[pharmacist]

Create a ChromIQ Project from an MFX File

Would it be feasible to incorporate the following features into ChromIQ?

1. Direct MFX File Import
. Allow users to import MFX files directly to create a ChromIQ project.

2. ICC Profile Creation and Verification
. Support ICC profile creation and profile verification, consistent with the functionality available in other ChromIQ projects.

This would allow existing i1Profiler users to immediately benefit from ChromIQ's functionality.

Thank you.

Hello you mean mxf import ? It is already imlemented in version 4.3.20 and I used it to create my ChromIQ generated profiles using the readings scanned with iProfiler without the conversion to cgat txt spectral data first.

 

[jctay]

Hello you mean mxf import ? It is already imlemented in version 4.3.20 and I used it to create my ChromIQ generated profiles using the readings scanned with iProfiler without the conversion to cgat txt spectral data first.

View attachment 19732

My understanding is that ChromIQ currently requires an .mxf file to be converted into a .ti3 file and then imported into the Measurement module before an ICC printer profile can be created. This workflow requires the entire sequence of steps to be completed, from patch chart creation through to measurement setup.

What I have in mind is a different approach: allow users to create a ChromIQ project directly from an .mxf or .ti3 file, bypassing the Measurement phase entirely, and proceed directly to ICC profile creation.

 

[pharmacist]

Click on the left upper icon load file

And create your profile directly from the mxf file.

 

[jctay]

Click on the left upper icon load file

View attachment 19733

And create your profile directly from the mxf file.

Thanks for pointing that out.

I'll try again using the .pxf file from the exports folder

 

[jctay]

ChromIQ 4.3.18 is out, and with it an updated step-by-step guide for measuring a printed ChromIQ sheet in i1Profiler as well. It replaces my earlier guide for 4.3.17: the most important change is that you now turn the sheet the other way.

@jctay, your idea started all of this, and @pharmacist, thank you for testing it so quickly: you found the spacers problem and reported that 7 mm patches read fine with an i1Pro 2. Both are fixed or included in 4.3.18.

What changed since the first guide

• Turn the sheet a quarter turn CLOCKWISE, not anticlockwise. The wide clip border (the margin on the left of the upright sheet, where the ChromIQ logo is printed) is then at the top, where the clip of the ruler's backing board holds the paper. Strip A is the top row, and the strip labels are on the right. If you set the clip border to the right in Create Chart, you turn the sheet anticlockwise instead, so that the border is at the top too. Charts made with 4.3.17 keep working: turn those sheets the way their own note says.
• Any paper size can qualify now, portrait or landscape, with any number of patches per strip. The patches must be 7 to 25 mm long along the strip, for the i1Pro 2 and the i1Pro 3 alike. The rest stays as before: every strip full, no spacers, and the whole chart must fit on one i1Profiler page, which is at most 1270 mm tall (about 150 strips of 8 mm). A4, Letter, A4 landscape, 4 x 6 inch and 5 x 7 inch were compared cell by cell with i1Profiler 3.8.8's own chart, using the i1Pro 3 file. On every chart, and especially on small paper, check the grid i1Profiler shows (step 9 below).
• No more false refusal for "Edge spacers". In 4.3.17 a chart with Spacers set to None got no i1Profiler file when the box "Edge spacers (bracket each strip)" was still ticked, although that box prints nothing on such a chart. That was pharmacist's case (504 patches, 8 x 10 mm). 4.3.18 looks at the sheet itself, so such a chart now gets its files.
• The i1Pro 2 file is now written from 7 mm patches, like the i1Pro 3 file (4.3.17 wrote it only from 10 mm). To be honest about it: the i1Pro 2 file has still not been opened in i1Profiler, because i1Profiler on our test Mac offers no i1Pro 2 option. It follows the same rules as the i1Pro 3 file, but it is unchecked there.
• "Prioritise chart area" charts: on A4 and Letter portrait they qualify too, as long as they have no spacers and full strips (for example 24 strips of 27 patches on A4). On taller paper their strips are longer than one i1Profiler row can hold, and the note says so and tells you to use "Prioritise patch size, then fit to page" instead.

Before you start, please keep in mind:

• It is a beta option, off unless you switch it on.
• It works only for the i1Pro 2 and the i1Pro 3, read by hand along the ruler, and only for RGB charts without spacers. Not for the original i1Pro, the i1Pro 3 Plus, the ColorMunki, the i1iO, the i1iSis or any other instrument.
• The layout was checked patch for patch in i1Profiler 3.8.8 with the i1Pro 3 file: for the chart in this guide, i1Profiler's chart and the turned ChromIQ sheet have all 504 of 504 patches in the same place. Nobody has measured a real sheet yet, because I have no i1Pro 2 or i1Pro 3.
• The import pictures at the end use a test file, not a real measurement.

1. Update to ChromIQ 4.3.18

Download it here: https://github.com/itsab1989/ChromIQ/releases/tag/v4.3.18

2. Switch the option on

Open Preferences (the Preferences button at the top right, or ⌘, on a Mac and Ctrl+, on Windows), go to the Beta tab and tick "Also write an i1Profiler chart that matches the printed ChromIQ sheet (i1Pro 2 and i1Pro 3, beta)". Press OK. (If you switched it on in 4.3.17, it is still on.)

View attachment 19666

The ⓘ button next to it explains everything in detail. First what the option is for, and which way to turn the sheet:

View attachment 19667

Which instruments and which charts, and what happens to charts made with 4.3.17:

View attachment 19668

Where the files go, and how the measurement comes back into ChromIQ:

View attachment 19669

Measurements without device values, the beta warning, and the default:

View attachment 19670

3. Which charts qualify

ChromIQ writes the i1Profiler files only for a chart that i1Profiler can lay out exactly like the printed sheet:

• an RGB chart made with "Use the ChromIQ layout engine instead of printtarg", for the instrument i1Pro / i1Pro 2 / i1Pro 3;
• Spacers: None, and every strip full;
• patches 7 to 25 mm long along the strip (for the i1Pro 2 and the i1Pro 3 alike), on any paper size, portrait or landscape;
• strips at most 241.5 mm long (ChromIQ's own strips end at 240 mm unless you set a longer strip length limit);
• small enough for one i1Profiler page, which is at most 1270 mm tall: about 150 strips when they are 8 mm wide, about 100 when they are 12 mm wide.

If a chart does not qualify, ChromIQ writes only the note, which says why, with the real numbers, plus one line in the Create Chart log.

4. Make the chart

On 1. Create Chart, choose MANUAL. This is the chart I used, and it qualifies for both the i1Pro 2 and the i1Pro 3:

• Use the ChromIQ layout engine instead of printtarg: ticked
• Instrument: i1Pro / i1Pro 2 / i1Pro 3
• Paper: A4 (210 × 297 mm) Portrait, Pages: 1
• Create layout: Prioritise patch size, then fit to page
• Patch size (mm): 8.0 × 10.0 (8 mm across the strip, 10 mm along it)
• the patch count left on Auto, so the page is filled with full strips

View attachment 19671

Then open Expert Options and set Spacers: None:

View attachment 19672

Press GENERATE CHART. You get one A4 sheet with 504 patches: 21 strips of 24 patches, 8.04 × 9.99 mm on screen, no spacers, and the margin check says Margins: OK.

View attachment 19673

5. Where the files are

The Create Chart log shows the new files being written:

View attachment 19674

They are in the run's exports folder (for example ~/ChromIQ/<project name>/runs/run1/exports/):

• <name>-i1profiler-turned-i1Pro2.txf: the i1Profiler test chart for the i1Pro 2
• <name>-i1profiler-turned-i1Pro3.txf: the same for the i1Pro 3
• <name>-i1profiler-turned.json: the order file ChromIQ needs to bring the measurement back. Please do not edit or move it.
• <name>-i1profiler-turned-how-to.txt: a note with every step for exactly this chart, including which way to turn the sheet and the grid i1Profiler must show

Here is the note for this chart (opened in a plain text window). The beta warning, the instruments and steps 1 to 5; step 2 says CLOCKWISE, and step 5 names the grid, 24 columns, 21 rows, 1 page, patch width 10.00 mm:

View attachment 19675

Steps 6 to 8, and which i1Profiler row is which ChromIQ strip (row 1 is strip A):

View attachment 19676

The File Guide lists the files too (the ? button at the top right, then Where are my files? (folder guide)):

View attachment 19677
View attachment 19678

6. Print from ChromIQ as usual

Print the chart on 2. Print Chart, exactly as you always do. Do not print i1Profiler's chart: the whole point is to measure ChromIQ's sheet.

7. Turn the sheet a quarter turn CLOCKWISE

Turn the printed sheet a quarter turn clockwise, so that the clip border (the wide margin on the left of the upright sheet, where the ChromIQ logo is printed) is at the top, where the clip of the ruler's backing board holds the paper. The strip labels are now on the right, and strip A is the top row. Every ChromIQ strip is now one i1Profiler row, read from left to right (that is from the strip's last patch to its first, which is fine: the files are written for exactly this).

View attachment 19679

Only if you set the clip border to the right in Create Chart: turn that sheet anticlockwise instead (strip A is then the bottom row). And a sheet printed from a 4.3.17 chart: turn it the way its own note says. When in doubt, step 2 of the note for your chart is always right.

8. Load the .txf in i1Profiler

On i1Profiler's start page, set User Mode to Advanced, and under Application Settings set Default display device to the family of the instrument you measure with (the i1Pro 2 family for an i1Pro 2, the i1Pro 3 family for an i1Pro 3). This matters: with the other family, i1Profiler reads the patch size on the wrong scale and the rows come out wrong. Then choose Printer, Profiling.

View attachment 19680

Press Next on the Patch Set page. On the Test Chart page, under Page Data, press Load and choose the .txf for your instrument. When i1Profiler asks "Would you like to copy this file to the assets folder?", either answer is fine. If you still have a 4.3.17 file for this chart loaded, load the new one: the files of 4.3.18 are written for the clockwise turn.

View attachment 19681

9. Check the grid

i1Profiler must now show exactly the grid the note names. For this chart the note says: 24 columns, 21 rows, 1 page (Page 1 of 1), patch width 10.00 mm. If it shows anything else, stop: the rows would not match the sheet. Please do this check on every chart, and especially on small paper such as 4 x 6 or 5 x 7 inch.

View attachment 19682

(This is i1Profiler's own chart for the i1Pro 3 file of this very chart, saved with its "Save as..." button. Row 1 is strip A, and all 504 cells match the turned sheet. i1Profiler names its patches by its own columns and rows, A to X and 1 to 21, so its "A1" is not ChromIQ's A1. That is expected.)

10. Measure

Press Next to the Measurement page and choose a measurement mode:

• i1Pro 2: Single Scan (M0), or Dual scan (M0,M1,M2 and OBC) if you want M1 and M2 as well (each row is then read twice, forwards and back);
• i1Pro 3: Single scan (M0,M1,M2 and OBC), which gives M0, M1 and M2 in one pass.

Calibrate, and measure the rows from top to bottom, each one from left to right, starting and ending on the white paper beside the row. i1Profiler checks every row against the colours it expects and asks you to read it again if it does not fit. Please always do that: it is the check that catches a row read in the wrong place, and ChromIQ cannot see a misplaced row afterwards. Leave i1Profiler's Scramble patches setting off, and measure the whole chart.

11. Save the measurement

When every row is measured, press Save under Page Data on the Measurement page and save it as an .mxf file (the format this was checked with). A CGATS text export that includes the RGB values works the same way. One without device values is refused for this chart, because only the RGB values tell ChromIQ which patch each reading belongs to.

12. Import in ChromIQ

Open the run, go to 3. Measure, press IMPORT, choose the file with the folder button beside Measurement File (from i1Profiler), and press IMPORT MEASUREMENT. ChromIQ checks that the device values in the file are this chart's, entry for entry, and then puts every reading on its own patch. A file that does not match, or holds only part of the rows, is refused and nothing is changed. If the file holds more than one measurement condition, ChromIQ uses M2, then M1, then M0.

The next three pictures use a test file, not a real measurement: it was generated from this chart's i1Pro 3 .txf the way ChromIQ's own tests build one.

View attachment 19683
View attachment 19684

The log line to look for is "This is a measurement of the turned i1Profiler chart, so ChromIQ put every reading on its own patch of ... in the turned order":

View attachment 19685

After that you can build the profile on 4. Build Profile as usual.

Please tell me how it goes

This is still the part we could not test: a real sheet, a real i1Pro 2 or i1Pro 3, and a real i1Profiler export. If you try it, I would love to hear whether every row reads in i1Profiler, whether the grid matched the note (please tell me the paper size and patch size if it did not), whether the import goes through, and anything that was unclear. If someone with an i1Pro 2 can load the i1Pro 2 file in i1Profiler and check the grid, that would be especially helpful. Feedback is very welcome, good or bad.

After several day of trying , I finally managed to get it working.
Actually, it's quite straightforward.

Incredible work on this! Thank you.

@itsab1989 If you need the .mxf file, please let me know.