I have helped a user to calibrate that monitor using ColourSpace, so I wanted to share the findings/results for other ColourSpace users of that monitor.
I had no physical access to the monitor (I will buy one for me in the future), so the user was sending me the measurements I was asking for...
We wanted to see which monitor settings were useful to use for pre-calibration and which controls were useless.
Since the BMD Bi-Converter can process a LUT only with YCbCr 4:2:2 signal, we had to adapt Resolve output/LG input settings for a VIDEO LEGAL calibration.
For the Patch Generation, ColourSpace was controlling Resolve iTPG (Video Levels), BMD Mini Monitor 4K was sending YCbCr 10bit via HDMI cable to BMD Bi-Converter, and another HDMI cable again to LG Monitor.
He had a Klein K-10A colorimeter and a Colorimetry Research CR-250 with ColourSpace PRO.
(The calibration can be performed with ColourSpace LTE also using X-Rite instruments... we used ColourSpace PRO due to high-end meters)
LG 32EP950 is clipping colors using the default settings; it will require reducing the Saturation slider (which works as a global color luminance slider also).
Reducing the color luminance will unclip the colors.
Using ColourSpace's RGB Separation and 3D xyY chart, and by taking multiple readings with -1, -2, -3 with Gray + Primary Ramb measurement, you can find out which Saturation setting will un-clip colors.
We used the new BMD MicroConverter Micro Converter BiDirectional SDI/HDMI 12G; remember that this box does not support HDCP; however, they are available solutions to remove HDCP protection.
It looks like LG doesn't like FULL RANGE RGB 10-bit signaling from Terranex SDI->HDMI 6G, even using HDFury at the output to LG to send custom RGB AVI, it didn't work, signal reported as 422, while it was RGB.
Direct HDMI connection from BMD Mini Monitor 4K can send RGB 10/12 bit, looking the LG panel, but it will degrade RGB Separation.
After that test, I believe this monitor is working correctly with 422 input, and some internal conversions to re-convert RGB input to YCbCr are destroying the RGB Separation.
At this point, I can't recommend RGB input and FULL RANGE for full-range workflow.
VIDEO workflow with 4:2:2 has no issues.
SPD measurement from CR-250:
For users with high-end meters, who are usually saving the meter correction tables to the colorimeter's memory slots, it's not recommended to use that meter profiling method.
It will require advanced meter profiling method; for example, LightSpace is using (like other software) the standard meter profiling method (based on NIST...FCCM).
WRGB OLED (LG consumer) or this LG JOLED will need to use the new meter profiling method called in ColourSpace as FCVM (Four Colour Volumetric Matching)
Using the traditional method, you will have a lot of negative values for Z (of XYZ coordinates); as you can see in the picture below, red is out of the visible spectrum; it has negative values, the math has been improved with FCVM.
FCVM will help to reduce these errors and provide more accurate meter profiling.
FCVM is the default meter profiling method in ColourSpace; it will measure three times the four patches (WRGB) to calculate the correction matrix.
So the recommendation is to use meter profiling inside ColourSpace, not use Klein ChromaSurf or Colorimetry Research software to create a meter correction table inside the meter EPROM....and call that table with ColourSpace.
For the profiling, we used a 17-Point Cube with an anisometric patch sequence, 0.75-sec extra delay to prevent network traffic, injection of 0.35-sec black frames (stabilization).
We used the BMD utility to load a reset 3D LUT.
LUT Generation using Peak Chroma, apply videoscale pass black filter, export to 33-Point Cube BMD LUT format.
REC.709 Gamma 2.4 100 nits was our targets.
We used the BMD utility to upload the calibration 3D LUT.
LG JOLED is a super-stable monitor; below, you can see the drift plot during the 17-Point profiling.
ColourSpace took a white patch reading every 50 other color frames to populate this chart.
Axis is +-3 nits; the color lines represent the changes per color channel.
This performance is amazing for OLED, and I haven't seen other OLED so stable.
ColourSpace is using that data to calculate the correction for the drift during the monitor profiling.
It's a good chart also to see if everything was OK during the profiling.
We used 1000p point verification for the post-calibration verification, 10-point cube, to see how well the monitor is tracking the whole colorspace in multiple levels of saturation/hue/luminance.
The dE2000 average for 1000 colors is insane.... 0.16 !I haven't seen that performance again with any other monitor!
LG JOLED 32EP950 is stable and provides a true reference performance for SDR REC.709.
It has a black level of 0.0001 nits.
The native gamut of LG is larger from P3 colorspace (gray triangle = LG Gamut / white triangle = P3D65), maybe ~99% P3 due to some small desaturation near 100% Green.
BMD MicroConverter Micro Converter BiDirectional SDI/HDMI 12G looks like the most affordable low-cost LUT box right now, for YCbCr 4:2:2 10-bit (input-output) with one slot for 33-Point 3D LUT with HDMI 2.0 input/output.
It looks like the best hardware upgrade for any other consumer OLED without internal LUTs, like Panasonic/Sony OLEDs for professionals (or for consumers, if they will use another hardware to remove the HDCP)
NOTE: About the LUT processing, I will ask the user to check with a different monitor to see if the LUT processing of the 3D LUT is working with RGB-Full Range signal using an LG OLED WRGB, so see if the Blackmagic Design Micro Converter BiDirectional SDI/HDMI 12G can apply the LUT with RGB-Video / RGB-Data input/output. I will update you tomorrow.
