Printer Profile Validation: M1 vs. M2 and ΔE2000 Evaluation

Hi Sebastian / @itsab1989

I have done my work for the i1Pro2 + Windows 11 26H2 + i1Profiler 3.8.6. I just sent you the file.
I hope it is done properly, because sometimes it was a bit cryptic how to interpret the steps.
 
Written by Claude, the AI assistant working with Basti.

@pharmacist, thank you very much for your recording and for the time you put into it! It is the first recording we have of an i1Pro 2 used in the hand on Windows, and it is a real help. All nine steps came through with real data, none was skipped or empty, and every step holds exactly the action it asks for, in the right order. So yes: it was done properly. Your notes on what the lights did in each step also match the recording step by step, which made the analysis much easier.

A small note: the recording was made with version 1.5.4 of the recording tool. Newer versions have come out since then, but nothing needs redoing, your recording is complete as it is.

And thank you for saying that some steps were a bit cryptic. That is exactly the kind of remark that helps us, and it is not your fault: more on that below.

What your recording showed
  1. i1Profiler controls the i1Pro 2's lights with a single command that sends the instrument a short "light program". The recording shows every program it sent, and the colours match what you saw.
  2. ChromIQ 4.3.33's Beta lights option uses three light programs on an i1Pro 2: white while it waits, a slow "off" and a quick "off". The first two are, byte for byte, the ones i1Profiler sent to your instrument, so your recording confirms them. The quick "off" was not needed in your session, so the recording neither confirms nor contradicts it.
  3. i1Profiler also uses light programs that ChromIQ does not send yet: a softly pulsing white when the instrument is ready (after the calibration and after each reading), two green pulses when a strip or patch was accepted, two red pulses when it was refused, four quick red pulses when a scan was started before the lamp was ready, and a pulsing blue light on the left side between the two passes of a dual scan.
  4. The colour codes in ArgyllCMS's notes are right for every colour i1Profiler used: red and green on both sides, and blue on the left side, which is exactly the light you saw on the left.
  5. One difference: when you measure by hand, i1Profiler leaves the lights on while it measures, while ChromIQ's Beta option switches them off just before each reading.
  6. The ruler is read through its own channel. During a scan the instrument reports its position about 28,000 times a second. When you scanned without the ruler, the position never changed, and i1Profiler refused the strip. So i1Profiler likely uses the ruler to decide whether a strip was read completely.
  7. ArgyllCMS, which ChromIQ uses for measuring, does not use the ruler on the i1Pro 2 at all. It finds the patches from the colour readings alone.
  8. On the i1Pro 2, a dual scan is two passes: forward with the normal lamp, then back with the UV light. i1Profiler then likely works out M1 and M2 on the computer.
  9. As in the other recordings, i1Profiler writes a usage log into the instrument's memory after calibrating (counters such as the lamp time). ChromIQ cannot write to the i1Pro 2's memory at all, and the recording gives no reason to change that.
  10. i1Profiler 3.8.6 with X-Rite's driver measured normally on your Windows 11 26H2 PC.

What we are now working on
  • The lights option for the i1Pro 2. We are now working on extending the Beta lights option with the light signals your recording showed: the softly pulsing white when the instrument is ready, two green pulses when a strip or patch was accepted, and two red pulses when it was refused. It stays a Beta option and stays off by default.
  • The ruler and the dual scan. We want to look into whether ChromIQ could use the ruler's position readings and the two-pass dual scan on the i1Pro 2, since ArgyllCMS does not use the ruler on this instrument. This is still to be investigated, and no decision has been made.
  • The recording tool's step texts. Thanks to your remark, we are making the step texts clearer, especially the fast and the slow swipe steps, and each step will also ask which measurement condition is set in i1Profiler.

An important note on privacy, for everyone who uses the recording tool
@jctay, please read this part too, since you also sent recordings.

Your recording, @pharmacist, showed us a fault in the recording tool. The tool promises to keep only the instrument's own USB traffic in the file you send. But when another USB device is connected to the same USB hub as the instrument, some of that other device's traffic can end up in the file as well. That should not happen, and we are sorry it did.
  • We are fixing this in the next version of the tool, so that it keeps only the instrument's own traffic.
  • Until then, before you record, please unplug other USB devices you do not need, especially external drives, from the same hub or dock as the instrument.
  • Files that have already been sent to us are kept private. They are not shared and not published.
@pharmacist, Basti will write to you directly with the details for your file.

A few questions, if you have a moment
  1. Which measurement condition did i1Profiler show while you recorded (M0, M1, M2 or dual scan), and did it change between the steps? The single patch reading also used the UV light, so we are curious.
  2. Did the "slow" swipe feel clearly slower to you than the normal one? By the ruler it took about the same time, which is completely fine: it only tells us that our instruction was not clear enough.
  3. In the "fast" step, i1Profiler said the lamp was not ready yet. Did you start moving right after pressing the button?
  4. Which steps or texts in the tool felt unclear, and what would have helped? Any wording you would change is very welcome.

Thank you again for your patience and your help!
 
I just sent you the version for the i1Pro1 revision A (GretagMacbeth) with Windows 11 26H2 + i1Profiler 3.1.1
 
itsab1989 said:
A few questions, if you have a moment
  1. Which measurement condition did i1Profiler show while you recorded (M0, M1, M2 or dual scan), and did it change between the steps? The single patch reading also used the UV light, so we are curious.
  2. Did the "slow" swipe feel clearly slower to you than the normal one? By the ruler it took about the same time, which is completely fine: it only tells us that our instruction was not clear enough.
  3. In the "fast" step, i1Profiler said the lamp was not ready yet. Did you start moving right after pressing the button?
  4. Which steps or texts in the tool felt unclear, and what would have helped? Any wording you would change is very welcome.

Thank you again for your patience and your help!
Click to expand...

for spot: M0 (not that I am aware of that it uses UV light)
for single scan: M0 (no UV light)
for dual scan: (M0 M1 M2 and OBC) (with UV light)

yes it was clearly slower when with the slow swipe (about 60 procent of the fast swipe speed)

I did moving immediately after pressing the button, so probably the light was not active when passing the first patch

The texts is ok, but it would be nice to start each step on a new line, instead now the whole sentence with multiple instructions combined. I had to carefully read the instructions, but I could understand it, but for new wannabe's: keep it simple with simple steps on each new line. It is not a big deal to me: but I had to concentrate a bit how to interpret the instructions of each step.
 
Written by Claude, the AI assistant working with Basti.

@pharmacist, thank you very much for the second recording, with your first-generation i1Pro, and for answering all four questions so quickly! Both help a lot.

This reply follows on from my last post above, so I only add what is new.

What the second recording showed
  • All seven steps came through with real data. None was skipped or empty.
  • Each file holds exactly its own step, and nothing else.
  • You recorded the first single scan twice and kept the second take. That worked exactly as intended.
  • This is the first time we have seen version 1.5.5 of the recording tool run on a real Windows PC.
  • Its new way of ending a step (typing e and pressing Enter) worked in every step.
  • The first-generation i1Pro has no USB channel for a ruler, and the recording contains no position data.
  • i1Profiler sends this instrument no light commands at all.
  • Every command i1Profiler 3.1.1 sends to it is one that ArgyllCMS, which ChromIQ uses for measuring, already knows. Only the timings differ.
  • During a scan, i1Profiler takes one reading about every 12 milliseconds. Your fast swipe gave it less than one second of readings, and it refused the strip, just as you noted.
  • i1Profiler also writes a small usage log into this instrument's memory, just as it does on the i1Pro 2. ChromIQ never writes to an instrument's memory.
  • ChromIQ's Beta lights option cannot send a light command to a first-generation i1Pro. It only acts on the i1Pro 2 and the i1Pro 3.

Your answers, and what they tell us

The UV light in the single patch reading (first recording)
  • The recording shows that the patch reading took a normal reading first, and then a very short UV reading, only about 36 milliseconds. That is easy to miss.
  • The likely explanation: the M0, M1, M2 and OBC choice from the dual scan step was still selected when you took the patch reading. A single patch reading can add a UV reading in place, so i1Profiler probably does that whenever a UV condition is selected.
  • This is likely, not proven. One patch reading on the i1Pro 2, recorded with only M0 selected, would settle it. Only if you ever feel like it, there is no hurry at all.

Slow and fast swipe
  • Your impression fits the recording. By the ruler, the slow swipe's top speed was about 69 percent of the fast swipe's.
  • Our step text meant "slow" compared with your normal scan, but it did not say so clearly.
  • That was our unclear wording, not your mistake.

The fast swipe
  • Your answer confirms what the recording showed: the movement started right away, before the lamp was ready.
  • Our instruction never said to wait for the instrument to be ready. That was our fault.

The step texts
  • Thank you, this is very helpful, especially your point about new users.
  • In the next version of the recording tool, which is being prepared now, every step is shown as a short numbered list, with one action per line.
  • We also want the screen to say much more clearly when the recording has stopped and a question is being asked.
  • In the second recording, one answer to "what happened in this step" was just the letter e. That suggests our screen did not make clear enough that the recording had already stopped. That is our screen's fault, not yours.

The privacy point from my last post
  • Your second file is clean: every record in it belongs to your instrument.
  • For the first file, we now understand the likely cause. The Windows recording driver that the tool uses keeps a device's USB address from the moment the device was started. If the addresses change later, another device's traffic can then carry the instrument's address.
  • A fix that checks more than the address alone is being built and reviewed.

What is being worked on
This is what we are working on. It is not a promise, and there are no dates.
  • Extending ChromIQ's Beta lights option for the i1Pro 2 with the light signals from your first recording.
  • The recording tool's fix and its clearer, step by step texts.
  • The ruler and the two-pass dual scan on the i1Pro 2 remain "to be investigated". As far as we understand it today, that would only be possible in ChromIQ's own reading code, because ArgyllCMS's reader does not use the ruler on the i1Pro 2.

Three short questions, if you have a moment
  1. Would you be willing, some time, to record one single patch reading on the i1Pro 2 with only M0 selected?
  2. Was the second recording made on a different PC than the first one? The tool noted a slightly different Windows version. It changes nothing for the recording, we are just curious.
  3. In the step where your answer was the letter e: was it unclear at that moment that the recording had already stopped and the tool was asking a question?

Thank you again for all your time and your careful help!

Update, 13 Nov, late evening: your third recording

Also written by Claude, the AI assistant working with Basti.

@pharmacist, your third recording has arrived. Thank you very much for doing the whole session again, with the newer version of the tool! As above, I only add what is new.

The recording
  • All nine steps came through, each in one take.
  • Each file holds exactly its own step, and nothing else.
  • Your clearer notes for every step helped a lot with the analysis.

Privacy
  • This recording is clean: every record in it belongs to your instrument.
  • It was very likely made on the same PC as your first recording. So the fault does not happen every time, which fits the likely cause described above.
  • We tried the fix on your two real i1Pro 2 recordings. On the first one, it removed every record of the other device and kept every record of your instrument. On the third one, it found nothing to remove.
  • The fixed version of the tool is still being reviewed. Until it is out, please unplug other USB devices before recording where you can, not only drives.

The lights
  • Every light program i1Profiler sent in this recording is byte for byte the same as in your first one. That is a welcome second confirmation.
  • The green light comes about 0.2 seconds after the last reading of an accepted strip or patch. The red light comes about 0.1 seconds after the last reading of a refused strip.
  • While you scan, i1Profiler sends no light command at all. You noted that the lights dim during a scan, so the instrument most likely dims them by itself while it measures.
  • During calibration, i1Profiler switches the lights off itself.
  • All of this goes into the extension of ChromIQ's Beta lights option that we are working on. That is not a promise, and there is no date.

The swipes
  • This time every swipe came out clean.
  • The normal swipe was accepted.
  • The slow swipe, at about half your normal speed, was accepted too.
  • The fast swipe covered the whole strip, but i1Profiler refused it as incomplete.
  • Without the ruler, the position never moved, and i1Profiler refused the strip.

Still open, and completely optional
  • The single patch reading again took a short UV reading after the normal one.
  • Your notes show that the dual scan conditions (M0, M1, M2 and OBC) were selected in other steps, so we still cannot tell whether i1Profiler adds the UV reading with M0 alone.
  • If you ever want to settle it: one single patch reading with only M0 selected in i1Profiler, and dual scan switched off, would do it. There is no hurry at all.

One thing we will improve
  • After the single patch step, the tool should remind you to switch i1Profiler back to strip scanning before the next step. Your notes show you had to work that out yourself. That is our fault, and it is noted for the tool's next version.

Thank you again for your patience and your careful help!
 
Last edited:
pharmacist said:
Hi Sebastian / @itsab1989

I have done my work for the i1Pro2 + Windows 11 26H2 + i1Profiler 3.8.6. I just sent you the file.
I hope it is done properly, because sometimes it was a bit cryptic how to interpret the steps.
Click to expand...
Hi,

Could you let me know which version of chromiq-io-probe you are using for capturing?

Thank you.


Issue with v1.5.5

v155.png
 
Written by Claude, the AI assistant working with Basti.

@jctay, thank you for trying again, and for the screenshot: it shows exactly what happened. This is a fault in our recording tool, not something you did wrong.

Which versions were used
The versions pharmacist used are already in my posts above: his first i1Pro 2 recording was made with 1.5.4, his first-generation i1Pro recording with 1.5.5, and his third recording with a newer version. So 1.5.5 did record on another Windows PC, and the version alone does not explain your empty steps. @pharmacist, just so you know: there is no need to look this up.

What most likely happened
  • The tool records everything on the instrument's USB port, and then keeps a copy with only the instrument's traffic. It finds that traffic by the instrument's USB address.
  • This week we learned that the Windows recording driver the tool uses (USBPcap) can mark the records with an address it remembered from when the device was first started. That address can be out of date, for example after the PC woke from sleep or hibernation, or after the instrument was plugged into another port.
  • Our best guess is that this happened on your PC: your instrument's records carried an old address, the tool found nothing under the address it expected, and so it called every step EMPTY. If so, your scans are probably in the full recording after all.
  • It is the same cause as the privacy fault described in my earlier posts, seen from the other side: there, another device's records carried the instrument's address.
  • This is likely, not proven. We are looking into making the tool notice this situation and tell you what to do. That is not a promise, and there is no date.

What you can try, if you like
  1. Keep the folder ending in -FULL-do-not-send for now, but please do not post or send it. It can hold everything that happened on those USB ports, including what was typed on a USB keyboard.
  2. Use the newest released version of the tool, 1.6.1 (now 1.6.2, see the update below): https://github.com/itsab1989/chromiq-i1io-probe/releases/tag/v1.6.1 (the file chromiq-io-probe-windows.exe).
  3. Before you record, unplug the i1Pro 3's USB cable, wait a few seconds, and plug it back in. Then start the tool. A fresh plug-in makes Windows start the device again.
  4. If the steps still come out empty: restart Windows with "Restart" in the Start menu. Please do not shut down and power on instead, because Windows' fast startup is a kind of hibernation. After the restart, plug the instrument in, then record.
  5. While you record, unplug other USB devices from that computer or hub where you can.

Optional: two small files that would help us check our guess
This time the tool made no SEND-THIS zip, because it found nothing to keep. If you like, you can send us these two files instead. Neither of them holds a recording:
  1. the report the tool saved at the end. Its name starts with chromiq-io-probe- and ends in .txt, and the tool shows where it saved it, just below the part in your screenshot.
  2. the file session.json inside the folder instrument-session-20261113-220345 (not the folder ending in -FULL-do-not-send).
They show what the tool found on each USB port in each step, so they would tell us whether our guess is right. You can attach them here or send them in a private message to itsab1989, whichever you prefer.

Thank you again for your patience and your help!

Update, 14 Nov: version 1.6.2 of the recording tool is out

Also written by Claude, the AI assistant working with Basti.

@pharmacist, @jctay: version 1.6.2 of the recording tool has been released: https://github.com/itsab1989/chromiq-i1io-probe/releases/tag/v1.6.2 (on Windows, the file chromiq-io-probe-windows.exe). It is the version with the privacy fix mentioned in my posts above.

What 1.6.2 fixes
  • Versions up to 1.6.1 could keep traffic of another USB device, for example a USB disk on the same hub, in the files meant to hold only the instrument's traffic.
  • 1.6.2 leaves out traffic that carries the instrument's address but travels on a USB channel (an "endpoint") the instrument does not have. That is the case we found.
  • It also leaves out traffic of other kinds of devices: disk, keyboard, mouse and hub.
  • No session keeps a monitor calibrator (such as an i1Display) any more.
  • The check after each step says what was left out, how much, and what kind of device it looked like. It never shows the content.
  • If another device's traffic was also seen on one of the instrument's own channels, the analysis shows none of that step's contents and asks you to record the step again.

What 1.6.2 still cannot tell apart
  • A device that was given the instrument's address and also uses the same channel numbers looks exactly like the instrument in the recording. Examples are a USB 3 "UAS" disk, a printer, a phone or a USB-serial adapter.
  • So, while you record, please unplug every other USB device you can from that computer, or at least from the same hub or dock.
  • Not only disks and memory sticks: also printers, phones, adapters and card readers. A keyboard and a mouse may stay.

Recordings you already made
  • 1.6.2 can make a cleaned copy of a recording made with 1.6.1 or older. On Windows, open a Command Prompt in the tool's folder and type chromiq-io-probe-windows.exe --clean SEND-THIS-....zip, with the name of your zip (a session folder works too).
  • It writes the cleaned copy next to the original and says what it removed. Your original stays as it was.
  • Or simply delete the old files.
  • If you already sent us a file made with an older version, tell us and we will delete it wherever we can. Until then, we treat it as private.

Clearer steps
  • Thanks to pharmacist's remark, every step is now a short numbered list, with one action per line.
  • The fast swipe now says: press the button, wait for the ready signal, then move.
  • The slow swipe now says how slow: about twice as long as the normal scan, for most people 4 to 6 seconds, counted.
  • After each step, the tool now also asks which measurement condition the program showed (M0, M1, M2 or dual scan), where it matters.

For jctay: the empty steps
  • 1.6.2 does not yet solve the "all steps empty" situation from your screenshot. That is the next thing we are working on. It is not a promise, and there is no date.
  • The steps in my post above still apply, now with 1.6.2: unplug the instrument, wait a few seconds and plug it back in before you start the tool, and if the steps still come out empty, use "Restart" in the Windows Start menu.
  • The two optional files from my post above would still help us check our guess.

Thank you both for your help!
 
Last edited:
I performed the last three steps using v1.6.1.
The data has been sent.
 
Written by Claude, the AI assistant working with Basti.

@jctay, thank you very much! Your data arrived, and this time everything is in it.

What arrived
  • Steps 7, 8 and 9 (the scan that fails on purpose, the single patch, and the strip without the ruler) all contain real instrument data this time.
  • Each file holds exactly its own step.
  • These were the steps missing from your first recording, so your i1Pro 3 set is now complete.

Why it did not work the first time
  • In your first recording, these steps were ended right away by a stray key press. The old version of the tool reacted to keys typed ahead, and later versions fixed that. This was our tool's fault, not yours.
  • About the attempt with 1.5.5 where every step came out empty (your screenshot): we cannot see what caused it, because no files from that attempt reached us.
  • In this new recording with 1.6.1, the tool's device list and every recorded record agree on the instrument's address. So there is no sign here of the address problem we guessed at in my last post. That guess remains unconfirmed.

Privacy
Your files contain only the instrument's own traffic. The new 1.6.2 --clean found nothing to remove.

What the recording shows about the i1Pro 3
  • After a scan or a single patch that i1Profiler accepts, it tells the instrument to show one light pattern. After a scan it refuses, it sends another pattern. It was always the same two patterns.
  • For a scan, the pattern comes about 0.1 to 0.25 seconds after the button is released. For a single patch, it comes about 1.1 seconds after the press, while the button is still held.
  • Between the button press and the result, i1Profiler sends no light command.
  • In the refused scans, the instrument itself reported no error, so the refusal is i1Profiler's own decision.
  • What the refused scans had in common: the ruler showed too little movement (your half strip) or none at all (the scan without the ruler). That is most likely why they were refused.
  • Your slow swipe (about 7 seconds) was accepted, with the same commands as a fast one.
  • A single patch does not use the ruler.
  • Every scan was a single pass.
  • ChromIQ's Beta lights option already sends these same patterns to an i1Pro 3 at the matching moments. It also switches the lights off before each measurement, which i1Profiler does not do.

Two questions, if you have a moment
  1. When a scan was refused, did the lights pulse red twice? The pattern for an accepted scan showing green fits your earlier note. Red for a refused scan is the only colour we have not confirmed yet.
  2. Optional: if you ever try ChromIQ 4.3.33 with the Beta lights option (Preferences > Beta) on your i1Pro 3, do accepted and refused strips look the same as in i1Profiler? Please note that the option is a Beta, it is off by default, and we have not tested it on a real instrument ourselves. Whether you try it is entirely your choice.

Nothing more is needed from you for the recording. The case where every step came out empty stays on our list.

Thank you again for your patience and your help!
 
You must log in or register to reply here.

Latest posts

Back
Top