As an example of the extremely limited HDR calibration offered by TV manufacturers for UHD/HDR/WCG, here is LG's documentation for UHD Alliance 'Greyscale Tracking'.
As can be seen this guide has no 'colour' calibration component at all - it is just setting the greyscale (as the document title states), and setting the white point to D65.
The TV makes the assumption that all UHD material is graded to P3 primaries (which is presently true, although there is no guarantee to that fact, and in reality is an incorrect step, as all UHD consumer sources have the material mapped into Rec2020 gamut as the delivery container!), and then assumes that the internal colour management of the TV is good enough to correctly map all volumetric colour space accurately.
And as all Home Cinema enthusiasts know, no TV manufacturers ever gets that right with SDR TVs. So why should UHD/HDR/WCG be any different?
Another example of the limited HDR calibration offered by TV manufacturers can be seen here with the original Dolby Vision Calibration Guide for Vizio displays.
Click the image to the right to download the PDF guide.
The use of 'Golden Reference' files has since been abandoned, with the introduction of Dolby Vision Calibration 2.0. However the reason for the abandonment is a very good example of the issues facing manual HDR calibration, The unit-to-unit variation of TVs of the same exact model from the same manufacturer vary too much for a single 'golden reference' to be a valid target across the board.
The guide describes nothing more than a 'setup' of the display through greyscale measurements, and a validation that the display's peak colour values matched against the 'Golden Reference' file for the Vizio display.
The 'Golden Reference' file simply contains the expected maximum primary and secondary colour gamut plots, and the 'calibration' process is simply to match the displays (if possible) to these targets, and then assumes that the internal colour management of the TV is good enough to correctly map all volumetric colour space accurately.
Again, as all Home Cinema enthusiasts know, no TV manufacturers ever gets that right with SDR TVs. So why should UHD/HDR/WCG be any different?
As described above, the concept of the 'Golden Reference' file ultimately failed, due to the issues mentioned here. The unit to unit variation in TV manufacturing meant peak colour values of the Golden Reference could not be attained by the TVs they were designed for, with Dolby Vision Calibration 2.0 replacing it.
Dolby Vision Calibration 2.0 is a better approach, as it disables the display's PQ EOTF, but maintains the display in peak luma mode.
The display can then be manually calibrated as normal for grey scale, targeting the screen's default 2.2 gamma curve, and measurements made for the peak RGBW primaries.
The peak RGBW values are then saved to a Dolby Vision reference file, and uploaded into the TV via USB. The internal Dolby Vision processing with the display then maps the display's HDR capabilities to the target PQ EOTF and Rec2020 colour space.
While still somewhat short of true 3D LUT based calibration, it is a viable manual approach to HDR calibration.
Note: as mentioned previously, the underlying calibration of HDR10, HDR10+, and Dolby Vision is identical, as all use the same target colour space PQ EOTF and Rec2020 Gamut - the same calibration 'should' work for all standards.
An alternative HDR metadata insertion is the HDfury Integral4K60, which can be used within the HDMI signal path from the LightSpace CMS laptop to the HDR UHDTV, injecting the required metadata into the signal path.
Using the HDfury Integral4K60 makes use of the free, in-built Light Space CMS Patch Generator for direct display profiling, using the HDfury's own control program to pre-set the required HDR metadata settings.
With this configuration any HDR display can be profiled with LightSpace CMS.
Using the HDFury GUI add the required custom HDR10 Metadata.
(Always tick the REC.2020 flag)
And set the required custom AVI InfoFrame.
For For PC/Laptop with HDMI 1.x and PGenerator, set output RGB-Data and patch scale 16-235 while for AccuPel and DVDO AVLab set RGB-Video output with no patch scale.
For DVDO AVLab, as an additional option RGB-Video 2160p24 8bit output can be used
For PC/laptop with HDMI 2.0 output, RGB-Data and patch scale 16-235 can be used
Note: When doing manual HDR measurements do not leave any patch static of the display for any extended period of time, as most HDR TVs have aggressive ABL (Auto Brightness Limiting, or Average Brightness Limiting, depending on who you talk to) and ASBL (Auto Static Brightness Limiting, or Average Static Brightness Limiting).
ABL will immediately reduce the display's peak luminance output when the average brightness of the displayed image is high, which is why small (10% by area) patches are used for profiling.
ASBL reduces the displays luminance output when a static image is seen for more than a few seconds.
Both these effects are required for HDR displays to reduce power consumption (to required legal levels in many countries), and to protect the display from overheating, as well as screen burn-out.
Further, and more importantly, OLED displays can suffer serious, and irreversible image 'burn-in' at high-brightness. While there is no issue at SDR levels, at HDR brightness levels there are real potential issues, as burn-in happens very, very fast - you have been warned!
LightSpace CMS provides two options for Soft Roll-Off/Tone Mapping, with the Soft Roll-Off option being a Light Illusion derived process, with extended user controls, while the BT2390 Tone Mapping option uses fixed parameters, based on the selected colour processing - R'G'B, ICtCp, Y'C'bC'r, or YRGB.
Obviously, for aesthetic viewing any desired options/values can be used, and again show the issue with the HDR concept, as often what are really 'inaccurate' values generate more pleasing results - assuming the EOTF can be user varied!
For true HDR accuracy there is only one option. The display must match, or better, the mastering display's peak luma, and NO Tone Mapping/Soft Roll Off should be used!
The configuration of 'Soft Roll Off/Tone Mapping' is something that can drastically alter the perception of any HDR display, as different roll off values are required for the different peak luma values used when grading source material. One size will not fit all, due to the issues outlined above. This is one of the problems with HDR TVs without the ability to alter the EOTF - they are attempting to use one Roll Off setting for all source material.
If the viewing display matches, or exceeds, the peak luma of the mastering display there is no need for any soft roll-off/tone mapping, and it should not be used.
As we say often when describing the way LightSpace works, there are no 'set' ways to do anything, and that goes for LUT Concatenation.
The obvious way, if you have a LightSpace CMS license with the LUT Manipulation tools, is to use the 'Add' function to add together two LUTs, having first 'Saved' the second LUT (the LUT to be added to the first LUT) via the 'Export' function. With the first LUT held within LightSpace simply select the Add function and navigate to the previously saved LUT.
A less obvious, but actually extremely useful method, is to use the 'LUT Image' function, and 'Save' the second LUT as a tif or dpx image. You can then use the 'LUT Preview function' to open the saved LUT Image, and apply the first LUT to it, saving the resulting LUT Image as a new LUT Image. To save the new LUT Image right click on the image and select 'Save As'.
When the new LUT Image is 'Opened' within LightSpace you will have a the concatenated result!
Again, what is key to concatenating LUTs is the order of addition. Swapping the order will generated different results!
For additional information see the LUT Image page of the website.
Yet another alternative option is to combine the LUTs via the 'Convert Colour Space' menu. If there is a LUT already within LightSpace you can use the 'Use existing' to combine the LUT you are about to make with the LUT already within LightSpace.
The 'Apply to the Image' and 'Apply to the Data' both do the same function as far as the resulting LUT is concerned. The 'Image/Data' concept refers to any 'Reference Image' that may or may not be held within the LUT Image. See: LUT Image for more info.
Note: depending on the profile data set being used within Convert Colour Space there may be some variations in the final result compared to a pure 'Add' process. That is to be expected. But, there is always another way...
If using a LUT box with multiple LUT capabilities the different LUTs can be applied separately within the LUT box, rather than being concatenated.