Problem/bug with the 'full field pattern insertion' feature

[j82k]

Hi,
I’ve been using displaycal for about two week now, mainly for creating a madVR 3dlut for my LG Oled C8.

Basically I would like to have the full field pattern inserted after every patch. Could this be implemented?

But since that is not possible I set it to the minimum of 1 second.

Doing this will often lead to displaycal not being able to finish the run.
Either it stops while measuring with error messages like:
“The measured luminance response is highly non-monotonic (device or instrument fault). Cannot generate tone response curves.”

Or it just hangs for hours at the calculation phase that happens after the measurements.

This seems to only happen when the full field pattern insertion interval is set to 1 second. It somehow seems to screw up measurements.
The problem already shows during the interactive display adjustment phase where it cycles between measuring white and the inserted full field.

Using the check measurement file feature on the created files shows a ton of wrong measurements.

All of this doesn’t happen when the full field pattern insertion interval is set to a higher value. So either this is a bug or there is something wrong with my setup.

I’m using a i1 Display Pro and Argyll V2.0.2 beta.

Attached are two examples of such failed measurements with full field pattern insertion interval set to 1 second.

Attachments:

You must be logged in to view attached files.

Calibrite Display Pro HL on Amazon  
Disclosure: As an Amazon Associate I earn from qualifying purchases.

[Florian Höch]

This seems to only happen when the full field pattern insertion interval is set to 1 second. It somehow seems to screw up measurements.

Some of the patches have been measured as black or white, which seems to indicate that it’s either an OLED image retention or madTPG issue (if the previous patch bleeds into the next, then it should help to override the default display update delay). Are you using exclusive fullscreen mode or D3D11? The D3D9 or windowed overlay path in madTPG seem to be much more reliable (madTPG with D3D11 has a tendency to flicker), atleast on my system (nVidia).

[j82k]

To rule out the Oled I just tried this on my monitor. It’s an older IPS monitor with low input lag and without any picture processing. Again tons of wrong measurements when using the full field pattern insertion feature at certain interval/duration settings, just like on the Oled.

I’m using windowed or windowed fullscreen d3d11 in madVR but disabling d3d11 didn’t make a difference. I also can’t see any flicker or weirdness when madTPG changes the pattern.
There must be some timing error at certain settings where it measures part of the full field pattern or something like that. Increasing the minimum display update delay by a few hundred ms also didn’t help.

I did some quick tests on my monitor, creating a madVR 3dlut with default settings with only changing the full field pattern setting and reducing the amount of patches.

interval 1s, duration 1s -> no errors
interval 2s, duration 2s -> no errors
interval 2s, duration 1s -> no errors
interval 1s, duration 2s -> tons of measurement errors, 3dlut creation doesn’t even complete
interval 1s, duration 3s -> lots of measurement errors, 3dlut completes but gamut looks like a mess
interval 1s, duration 5s -> lots of measurement errors, 3dlut completes but gamut looks like a mess

Could you try if you can reproduce the problem when using interval 1s, duration 2s? This seems to cause the most errors and I would like to know if the problem is on my end or if this could be a bug.

[Florian Höch]

Increasing the minimum display update delay by a few hundred ms also didn’t help.

Try 1000ms (1s) or more.

[Florian Höch]

Alright, short intervals interfere with patch update delay measurements at the start. Also, an additional delay after each FFP helps stability. This will be fixed in the next update.

[j82k]

Thanks 🙂

When you fix this, could you also allow lower values than 1 second for the interval so that the FFP could basically be inserted after every measured patch?

[Florian Höch]

Replace %ProgramFiles(x86)%\DisplayCAL\lib\library.zip with the attached file (rename the original) and let me know if this works for you.

Attachments:

You must be logged in to view attached files.

[j82k]

Replace %ProgramFiles(x86)%\DisplayCAL\lib\library.zip with the attached file (rename the original) and let me know if this works for you.

Just did some quick testsing with low patch amounts using 0s interval and also 1s interval.
No errors or weird measurements at all, looks like the problem is completely fixed.

Thanks a ton for fixing this so quickly and also for implementing my request of 0 second interval. 🙂
This might be very useful for Oled owners.

• This reply was modified 5 years, 3 months ago by j82k.

[benjolino]

Hi everyone,

I get the same error trying to calibrate my Dell LP2480zx through DaVinci Resolve.

“The measured luminance response is highly non-monotonic (device or instrument fault). Cannot generate tone response curves.” They I1 pro I use works perfectly if I don’t go through Resolve or if I use it with Truecolor, the software from my LG Main monitor.

I tried it several times with different settings. I did it always like this since years and it worked but now it doesn’t. I think I made an update of Displaycal.

I’m on a MAC btw. Can anybody help me?

Best

Ben

• This reply was modified 4 years, 9 months ago by benjolino.

[Florian Höch]

I1 pro

i1 Pro? Or i1 Display Pro? The i1 Pro is really not suitable for prolonged measurement runs because of its sensor black calibration drift.

I’m on a MAC btw

Disable screen saver and powersaving. Also setup Resolve itself correctly (see the guide in the Wiki).

[I.Qu]

Hi, i am having the same problem, did you ever come up with a solution?

[I.Qu]

Is there a solution to this problem? with the i1 display pro

[Author]

Posts