Hi,

the LG calibration guide says that a 3D LUT calibration doesn't make much sense for the HDR mode. One reason that has been mentioned is that due to the WOLED white pixel, colors are getting desaturated with increasing brightness...

The LG G5 has a new WOLED panel technology which provides a higher color volume. Would it now make more sense to do a 3D LUT calibration for the HDR mode? Or is a manual calibration still the best option?

Thank you!

While I don't own a G5, I'm guessing it still has the same limitations as prior generations of WOLED technology, i.e., color luminance cannot keep up with grayscale as brightness increases.

You can easily test this yourself by placing the display in HDR and running the gray and primary ramp sequence in the characterization window and reviewing the RGB separation chart.

If the RGB channels plateau and veer off to the right (indicating a hard clip) well before 100% white, this means the the G5 suffers from the same issue as all prior WOLEDs and HDR color cannot be calibrated accurately via 3D LUT.

The G5 is still a WOLED.
It just uses a dual stack layer to increase the available gamut (a bit).
It does nothing for the issue of desaturation at higher brightness levels due to the white pixel.
This is not a hard-clip issue - it is a 'reduction' in the peak luma the RGB pixels can achieve vs. the white pixel, causing a loss in gamut/saturation.

Steve

Even if we ignore the issue with reduced gamut due to WRGB design, WOLEDs are very unstable to use any form of 3D LUT; even using an external LUT box will not work.

Just to clarify, you mean that WOLEDs are unstable with a 3D LUT in the HDR mode? In the SDR Mode it is fine?

In HDR, the TVs are very unstable; you can do your tests and use the drift plot chart to see how significant the deviation is in HDR.

SDR works great.

Thank you for explaining!

So let's say I like to experiment a bit and want to do a 3D LUT calibration for the HDR10 mode, just to learn how the result compares to a manual calibration. Which steps are different compared to the SDR 3D LUT calibration guide? Or is Colourspace not able to generate a HDR 3D LUT?

Thank you!

Even Portrait/Calman, with their push for "autocal" as THE solution, doesn't do 3D LUTs for HDR. You'd be seriously wasting your time. And with the hundreds or thousands of high-luminance readings necessary for creating a LUT, you run the risk of shortening the life of your set, if not damaging it outright.
Sorry for mentioning a competitor, Steve!

Actually, you would have more luck trying with ColourSpace than with Calamn.

Calman's issue is Calman, which is why they do not provide 3D LUT calibration for FSI's professional HDR monitors that use the same WOLED glass.

ColourSpace can calibrate FSI monitors for HDR via 3D LUTs, and is what FSI use for factory calibration.

The issue with LG TVs is thermal instability with HDR operation, so if you can 'tame' that, there is no reason a 3D LUT could not work.
Using a small Quick Profile could be a good approach, along with a fan pointing at the TV to cool it, etc...

Steve

Nice to know, Steve. Thanks!

Assuming the thermal instability is overcome, how would the HDR 3DLUT be applied on the LG G5?

I have tried and done with success: https://www.avsforum.com/posts/63483284/
The key was an high stabilisation time (3 seconds). Depending the probe (it was a Klein), the stabilisation my be higher (i,e. 5 seconds for I1D3) or shorter, depending probe speed. I will do some study with patch ordering and stabilisation patterns. That with Prometheus has cancelled the need for stabilisation.

ebr9999
Thanks for sharing your process with 'peak brightness' off, I was able to replicate it with ostensibly good results (no obvious artifacts or problems) -- it made for an interesting experiment / learning experience.

Would you recommend for (or against?) using this method in conjunction with a traditional 22pt white balance HDR calibration, perhaps with a Unity LUT loaded in the 3D slot (if it even matters) while adjusting the greyscale balance / gamma 2.2 tracking prior to profiling the display using the Grey/Primary/Secondary ramp?

I also have a fast Admesy probe and was curious about alternative profiling methods, looking forward to test this more if even just as a curiousity.

Update: Enrico's workflow does have a flaw in it.

The initial calibration is to the TV's own native colour space, with an alternative white point.
But, the conversion to Rec2020 uses the default P3 colour space (with AWP) as the Destination, which is incorrect.
The Destination should be the Extracted native colour space the TV has been calibrated to, with the AWP.
NOT the default P3 space with AWP.
It seems to kinda work in Enrico's workflow as the extracted native space of the LG was close to, but not accurate to, P3.

The actual concept that was tried to be used by Enrico has been in the User Guides for a long time, in the Advanced Operation guide, LUT Concatenation section.

Also, using Map Space would do the two-step calibration workflow in a single step...

Steve