ASUS have just added 2 new ColourSpace Integrated monitors to their impressive colour accurate monitor range.

PA32UCE and PA27UCGE

At the same time the PA32UCDM is a huge hit with those looking for ultimate colour accuracy, including for colour grading.
https://lightillusion.com/asus_manual.html

Steve

Hi Steve,

I have some confusion regarding ASUS's 3D LUT.

I'm using the PA32UCR-K model, and over the past two to three weeks, I've been trying to calibrate its SDR and HDR modes.

I believe I can confidently say that the 3D LUTs created by ColourSpace are not an issue.

When I validate them using Active LUT or a 3D LUT BOX (BMD), I get excellent results.

However, something strange happens when I upload the 3D LUT to the monitor—it introduces some deviations, especially in grayscale. (I'm using 1D+3D)

To investigate this, I've repeatedly checked and ensured that all interfering options in the monitor's OSD are turned off.

I've also tried uploading Default Unity for all 1D, 3D, and 1D+3D LUTs or using Unity Bypass to clear them before performing the calibration.

Additionally, I attempted direct calibration by connecting the ASUS monitor to a PGenerator.

I even used different instruments—MK550T, i1 Pro 3, and i1 Display Pro—for measurements.

I also experimented with different 3D LUT precision levels: 17³, 21³, and 33³.

Despite all these efforts, the results always change after uploading the 3D LUT to the PA32UCR-K.

Could this mean that there are some unknown internal interferences within the ASUS monitor, preventing it from achieving a good calibration result?

VERIFY WITH BMD

VERIFY WITH UPLOAD TO ASUS

Yes, that simply means there is an issue within the monitor's internal signal path.
All is explained here: https://lightillusion.com/calibration_issues.html
You can even workout (roughly) where the issue is.

You may want to profile, generate, and upload all three LUTs independently.
But I have no idea if that will actually help.

You could also concatenate two LUTs, with the second made with the first LUT uploaded, so the second LUT correctes the distortion cause by the first when upload...

Steve

Thank you for your suggestion. I think I will try it using the method you provided.

Here's a quick update on the results.
After uploading the first LUT to the screen, I got a slightly distorted outcome. AVG ΔE 0.5153.
Using the verification results from the first run, I generated a corrected LUT and uploaded it to the screen again.
This yielded ideal results. AVG ΔE 0.2556.
On my PA32UCR-K, this method was very effective for SDR calibration.

However, it didn't work for HDR.
For HDR, I ended up resolving the issue using an external 3D LUT box.

BEFORE

AFTER

There is no different when doing HDR using the internal LUT or an external LUT Box, so I expect something was wrong with the calibration workflow.

Steve

Thank you, Steve.
I'll look into whether there's an issue with how I'm calibrating HDR.
With my PA32UCR-K, the screen's low-light level always stays at 1.5 nits when using HDR.
This causes significant problems in the shadow areas.

There's an "HDR Effect" option in the monitor settings that brings the low-light levels closer to true black.
However, it's clearly incorrect to enable this setting—once activated, it makes all calibration results even worse.

I'll try testing with another display.
I'm still learning about HDR calibration—thank you for offering some insights; I'll keep experimenting.

I do have a small question:
For HDR calibration, should it always be done in 10-bit, rather than 8-bit?
The PGenerator I'm currently using doesn't seem capable of outputting 10-bit directly—could this be the root of the issue?

PGen does do 10 bit accurately.
But there is no issue using 8 bit.

Steve

ChihHuangLin

PGenerator can generate bit-perfect 10-bit patches with YCbCb 4:4:4, YCbCr 4:2:2, or RGB in SDR/HDR with REC.709 or REC.2020 colorimetry.

PGenerator can generate bit-perfect 8-bit patches with YCbCb 4:4:4 or RGB in SDR/HDR with REC.709 or REC.2020 colorimetry.

Some devices, like JVC projectors or Samsung TVs, need 10-bit patches to correctly trigger the HDR mode and colorimetry.

Resolve TPG or Murideo Six-G or PGenerator (non-Pi4 or P400 boards), work in 8-bit.

(Non-Pi4 or P400 boards can only output 8-bit REC.709 RGB)

Murideo Seven or PGenerator (Pi4/Pi400 board) works in 10-bit.

Thanks Steve and Connected

I will continue to try and learn.

Hi Steve!

Given the spike nature of the panel technology, should I use my X-Rite i1 Pro3 to create the offset matrix or would it be better to rely on the corelation files TED provides? I know I know that neither are great options, but given the instability of the X-Rite i1 Display Pro Colorimeter. I wonder which approach would be the lesse evil approach.

Unfortunately, without a high-end spectro to compare with, that is impossible to say.
You will basically have to try both, and define the better visually...

You can use the Perceptual Match approach to compare the results to a display that does calibrate well with an i1Pro spectro Probe Match.
https://lightillusion.com/perceptual_match_guide.html

Steve