Hi all,

We're having a hard time calibrating an XM311K for 1000-nit P3-D65 ST.2084.

We use Resolve as a TPG and use an Admesy PCM2-100 probe-matched to a CRI CR-250 (at display's native gamut). Everything is connected to an HP ProBook 450 G10 running Windows 11 Pro 25H2. It is wired via Ethernet to our gigabit LAN and so is the Resolve TPG system and the XM311K for LUT upload.

My workflow thus far has been following the guides/videos from FSI, selecting on the display:
ColorSystem: LightSpace
LUT Mode: Advanced
Range: Full 0-1023 (we use Full throughout our facility)
Gamut: P3-ST2084

Using ColourSpace, I then upload unity LUTs to the front 1D, 3D, and back 1D positions for the P3-ST2084 slot. If I take some Grey Ramp RGB (33) measurements now, the display appears to have a Gamma 2.2 curve. I save the profile and generate a LUT with it using "Source: Rec709 Gamma 2.2".

I would then upload this LUT to the back 1D position, and then run another profile using Cube Based (17^3). From that profile I generate a LUT using "Source: ST2084 P3 D65 1000nits". I upload this to the 3D position.

If I then run validations, I get wildly varying dE ITP depending on my validation patch sets, whether it's Memory Colours RGBCMY, another Cube Based, or Edge Sweep. I never get below an average dE ITP of around 6 and sometimes it.
On top of this, if I compare my calibration to the LUTs installed on "ColorSystem: Default" (factory-installed LUTs loaded), the results are better on the factory calibration.

After being exhausted by so much testing, I'm reaching out to confirm a few things and see if anyone has some experience with this:

• Is the Admesy PCM2-100 supported by ColourSpace 1.0.1.2308?
• Should I be using "Force simple matrix match" when probe matching and Is there a way to quickly compare measurements between an active probe and the reference probe?
• Do I need to be using "Limit Range" when running my back 1D or 3D profiles? The display typically reads a max luma of around 970, which calculates to Code Value: 766.
• Is it odd that the display, with unity LUTs, has a Gamma 2.2 response curve even though it's set to P3-ST2084?
• I've also done this workflow but using Gamma 2.4 for the back 1D LUT instead, and get similar results.

Thanks all

I would request assistance from FSI, as most of your questions are specific to the monitor.

Steve

Hi, use 33p Grayscale and generate the 1D BACK LUT with modified DCI P3D65 with 2.2 gamma.

Check the calibrated nits at your peak output, and modify the PQ P3D65 colorspace to match the nits you will measure.

Find the code level for those nits, use the limit range to profile, generate the LUT, and upload it to the 3D LUT only.

https://www.flandersscientific.com/calibration/XM/xm311k-cal-cs.pdf

Hello,

If you have a need to build a LUT and follow Ted's advice, you should do 33 nodes with 10-bit through Resolve. This can be done using CS Patch Translator the Windows/macOS app I built to push 10-bit colour patches through Resolve from ColourSpace.

If you are on macOS then I have some beta software you can test that will drive the UltraStudio device directly, going to put a post up next week looking for beta testers. Send me a message here: https://aulwa.com/Contact