AVS Forum banner

281 - 300 of 1461 Posts

·
Registered
Joined
·
8,913 Posts
Discussion Starter #281
Near Black Issues from poor factory Quality Control

Some LG OLED's have passed the factory QC (Quality Control) with near black clipping issues and shipped to stores, for these panels it will be required to adjust the Brightness to 55-56 or up to 60-61 (in worst case scenario) from the default 50 value for the near black clipping issue to be resolved.



Also it will not work any service menu sub-adjustment related to Brightness when you will use internal 1D LUT capability.

By adjusting the sub-brightness in the service menu it will cause the 1D LUT points to be misaligned at near black, set Sub-brightness to default 128 value.

When you have such issues with near black clipping and you see that 1D LUT profiling can't resolve completely the problem (with Brightness @ 50 setting as required for internal 1D LUT to work) then you can skip using the internal 1D LUT completely and resolve the problem by adjusting Brightness control to whatever values you see that is not providing any clipping (while it will not raise your native panel black level) and then start the 3D LUT profiling measurements.

If you will skip performing 1D LUT profiling, select Warm 2 color temperature, Gamma 2.2 preset, Wide Color Space and the adjust Brightness/Contrast to the value you see that is correct for your TV. Perform also only 100% White (or 109% White) pre-calibration using RGB-High Controls only (keep one color channel at zero and reduce values from other 2 channels) while adjusting OLED Light you will control the luminance output (nits). Don't use RGB balance controls to adjust the Luminance output levels. After these pre-calibration adjustments you can proceed to 3D LUT profiling.

For expert users, its recommended instead of adjusting normal TV menu White Balance settings, to adjust only the White Balance values of color temperature from inside the TV's Service Menu only. The reason for that is because the normal menu White Balance settings combined with White Balance of Service Menu settings to work as offset from the native panel, and when you have active two controls which are performing the same job, this can make the internal processing algorithm more complex, so its better idea one White Balance menu to be used only to adjust from native panel response and not from both menus (normal TV + service) the same time.

White Balance settings of 'Warm' in Service Menu will adjust the 'Warm 2' preset of Normal TV Menu.

White Balance settings of 'Medium' in Service Menu will adjust the 'Warm 1' preset of Normal TV Menu.

White Balance settings of 'Cool' in Service Menu will adjust the 'Cool' preset of Normal TV Menu.

LG's factory pre-calibrate these service menu White Balance color temperature modes using a 80% Gray pattern with 6500K for 'Warm', 9300K for 'Medium' and 11000K for 'Cool'.



Use the EZ-ADJ button of Service Remote to access the LG OLED TV Service Menu, it will ask for a password, type '0413', navigate to '12. White Balance', and select the color temperature you want to adjust.

If you want to adjust for example the 'Medium' of Service Menu, start by setting the default 192 value to R-Gain, G-Gain and B-Gain and 64 value to R-Cut, G-Cut and B-Cut controls.



When R/G/B gain in the OSD is at 192 value, it means that the panel works at its Full Dynamic Range. In order to prevent saturation of Full Dynamic range and data, one of R/G/B channels need to be fixed at 192 value, for pre-calibrating the White Balance, lower only the other two channels.

Be careful with IN-STOP button, don't press it, because it will perform a complete factory reset, it will reset the TV's UTT (number of panel usage hours) counter which will affect the compensation cycle operation.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #282
Differences between YCbCr and RGB-Video Colorspaces of HDMI Input Video Signal

All available LG OLED TV models (2015/2016/2017/2018/2019) are processing differently the colorspace of the HDMI input signal; when you will compare by taking meter measurement using patches between RGB-Video and YCbCr input signal, there will be a measurable difference.

As we know, any available consumer SDR movie content has been encoded using YCbCr colorspace with 4:2:0 chroma subsampling.

Devices usually used for movie playback; like stand-alone Blu-Ray Players, Media Players or Network Streamers, with proper configuration of their video output settings (when it's been disabled any video processing enhancement) they can provide a bit-perfect REC.709 YCbCr 4:2:2 or 4:4:4 colorspace video signal output.

The following animated PNG file will swap with one second interval between the digital pixel error analysis of 5 different Blu-Ray or UltraHD Players with YCbCr 4:4:4 colorspace video output using Ted's LightSpace CMS Calibration Disk Blu-Ray with CalMAN's ColorChecker Classic and 100% Color Gamut patches as measured using DVDO's AVLab TPG Color Checker function:



Notebook or Desktop PC's (not all chipsets/brands, under proper configuration), PGenerator (Raspberry Pi based) or Amazon FireTV Stick using LightSpace Connect; they can provide bit perfect RGB-Video patch generation, they are not able to generate bit-perfect YCbCr patches.

But if you will use RGB-Video patterns for profiling and later playback from a YCbCr video output playback device, there will be some differences coming from some internal TV processing colorspace conversions.

LG TV's during their image processing path they will do many mathematical operations on the video data signal.

Usually these colorspace conversions are performed with a significant higher bit-depth from the actual video data signal depth, as higher bit-precision will be used, as lower will be the chance of seeing visible banding caused by the color quantization.

LG will convert any HDMI input RGB-Video colorspace signal to YCbCr 4:4:4 for applying initially most of the UI (user interface) adjustments.

The UI user setting will be merged in an algorithm which will interpolate values between multiple user settings to manipulate the video signal.



Looks like that there additional features of Expert Controls Menu (Dynamic Contrast, White Balance, CMS etc.) which will require more complex processing so for the TV to significantly reduce the bandwidth and the required processing power it will compress the horizontal chroma resolution in half; chroma subsampling from YCbCr 4:4:4 -> YCbCr 4:2:2; as this will reduce by 33.3% the video data bandwidth.

I believe this is the reason why these OLED TV's will not display full chroma (YCbCr 4:4:4 or RGB-Video) with PC Icon enabled (of HDMI Input) from 24p/30p/50p input signal but only with 60p signal where all the Expert Control menu controls are disabled, so the video signal will bypass that 'Expert Controls Menu' processing step.

When you have the PC Icon enabled on the HDMI input where the pattern generator is connected and configured to output 60p patches, then the HDMI Video Input lag will be reduced. (lower latency)

Before the YCbCr video signal will enter the first LUT table, it will be converted using a standard linear 3x3 colorspace conversion matrix to RGB and after all the LUT processing it will be converted down to 10-bit RGB for the panel.

The following animated PNG file will swap with one second interval between two post-verification result pictures:



The animated picture showcases the difference between 10-bit RGB-Video and 10-bit YCbCr REC.709 patterns using for verification CalMAN's ColorChecker SG patchset (96 colors).

There is a shifting to colors with larger differences to darker colors.

That test has been performed using a Klein K-10A colorimeter with the Accupel DVG-6000 Ultra Plus (with Pro Engineering Option) Reference external hardware pattern generator.

So the ideal solution (for the best image fidelity) is to use an external and bit-perfect YCbCr 4:2:2 or 4:4:4 REC.709 for SDR patch generation solution, to emulate exactly the same HDMI input colorspace you will send to the display when you will playback a movie also, this will avoid the colorspace conversion differences and you will have you full video chain calibrated, including all the colorspace conversions, when you will use a bit-perfect stand-alone player.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #283
Differences between patch generation framerates of HDMI Input Video Signal

Image characteristics are sensitive to different input video signal refreshing rates.

The following animated PNG file will swap with one second interval between two verification results pictures:



The animated picture showcase the difference between 1080i50 and 1080p60 patch generation, 56-Point Grayscale measurements used for that test.

There is a gamma shifting at the bottom and top end of the grayscale.

That test has been performed using a Klein K-10A colorimeter with the Accupel DVG-6000 Ultra Plus (with Pro Engineering Option) Reference external hardware pattern generator.

So the ideal solution (for the best image fidelity) is to use an external and bit-perfect patch generator configure to output 24p, to emulate exactly the same framerate you will send to the display when you will playback a movie also.

For external pattern generator users, it will be easy to change framerate output setting.

For PGenerator (Raspberry Pi based) users, they can change the framerate setting using DeviceControl Interface:



For Amazon FireTV Stick with LightSpace Connect users, the standard display settings menu has only resolution options.

Using Fire OS 5 (or later) you can change the video output framerate (24p, 25p, 30p, 50p, or 60p) by holding REVERSE and UP buttons simultaneously on the FireTV Stick remote for 10 seconds, the device will begin to cycle through additional resolution/framerate options not available in the standard display settings menu.

 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #284
Other Pre-Calibration Notes

1) The LUT upload per each Picture Mode (Cinema, Technicolor Expert, ISF Expert Bright, ISF Expert Dark or Game) it will affect all HDMI inputs globally (or internal apps). But it will be required to copy/transfer the Picture Mode, Expert Controls and Picture Options of the specific HDMI Input used for patch generation to the other HDMI Inputs you will use.

2) To start calibrating, you need to upload UNITY (1:1 Input:Output) LUT’s, to perform that ‘reset’ the user has to press “Reset 1D Lut” “Send Icon” and also “Reset 3D Lut” “Send Icon”. Both LUT tables on LG will be reset to UNITY.

If you will not use the internal 1D LUT, then you shouldn't reset the 1D LUT (by sending UNITY 1D LUT) from LG Template of Device Control Interface, because this action will disable the White Balance calibration controls, something you will need for pre-calibrating 100% or 109% White. Reset only the 3D LUT table.

3) You can adjust only OLED Light while you will pre-calibrate but don’t adjust OLED Light after the profiling.

4) For Video Levels (16-235) or Video Levels Extended (16-255) display profiling, set to your LG TV Black Level @ Low.

5) Before starting any measurement, the display should be working for at least one hour, to stabilize during that warm-up period.

6) Panel luminance will decrease if the input signal remains static for more than two minutes; change the test pattern to refresh the screen luminance.

At high luminance levels, heat generated in OLED panels can result in difficulties achieving repeatable results.

7) When RGB-Video patch generation will be used (16-235) or RGB-Video Extended (16-255) then there no difference if the bit-depth output of the patch generator will be 8 or 10 or 12-bits.

But when YCbCr 4:2:2 or 4:4:4 patch generation will be used, its recommended to set the bit-depth output of the patch generator to 10 or 12-bit (not 8-bit).

Calibration software will send request of 8-bit RGB triplet values with colorspace output flag also to the generator processor and from there the RGB triplet values will be converted to YCbCr colorspace, the 3x3 colorspace conversion matrix (RGB-Video -> YCbCr) calculation will introduce some rounding errors when the output will be 8-bit, so having pre-configured the output as 10 or 12-bit, this will compensate for rounding errors.

There no 1:1 mapping between RGB and YCbCr colorspaces. There YCbCr values with no corresponding RGB values and RBG values with no corresponding YCbCr values.

Only about 60% of all RGB values can be represented in YCbCr space when you will use the same amount of bits for both triplets.

This means the most damage happens in RGB -> YCbCr when you take a 3 * 8-bit RGB triplet, convert and round it back to 3* 8-bits of precision.

For that reason is recommended to set YCbCr colorspace bit-depth output of the pattern generator to 10 or 12-bit.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #285
All above posts cover the 1) and 2) of the following steps which are important for proper configuration/recommendation of hardware/software settings for LG 2018/2019 OLED TV profiling using LightSpace CMS and LG Template for DeviceControl Interface:

1) Connect your LG 2018 OLED with LG Template of DeviceControl Interface.

2) Calibration Notes.


3) Patch Generation Settings.

4) Meter Settings.

5) Pre-Roll Settings.

6) Display Profiling Options.

7) 1D LUT Display Profiling Patchsets.

8) 3D LUT Display Profiling Patchsets.

9) 1D LUT Display Profiling Procedure.

10) 3D LUT Display Profiling Procedure.

I will update once I will complete the other steps, with same way of very detailed instructions which no-one else ever provided for these TV's, in global calibration community.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #286
For quick reference, until I will post further setting recommendations, below you can find the actual AVSForum users post links (with posted settings of software/hardware settings they used) from several succesful profiling using LG 2018 OLED's with LG Template of DeviceControl Interface and LightSpace for LUT generation. Some of them used CalMAN as additional verification tool to cross-check the post-calibration final results which LightSpace has produced/generated (not used CalMAN for calibration/profiling to the charts they posted).

1 - 2 - 3 - 4 - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 - 14 - 15 - 16 - 17 - 18.
 

·
Registered
Joined
·
588 Posts
I've did a bunch of test the last couple days and tried to think outside the box.

I was able to get rid of the high dE error for that specific dark brown color which I had before and @jrref couldn't get rid of.

The results speak for them self for now I'm pretty satisfied and can say that the new hybrid mode is definitely a huge step forward to increase the overall accuracy of the 3D LUT.

I used a set which contains 101 grayscale points. Would love to see other results as well :)

I used CalMAN for cross verification.













LightSpace Calibration Report
https://drive.google.com/file/d/1XvThHeKjuG-2axTSd6qmh0d-WCTiF15f/view?usp=drivesdk
 

·
Registered
Joined
·
406 Posts
@ConnecTEDDD Since one should use YCbCr patterns for signals encoded in YCbCr, would the inverse apply when calibrating an input for a video game console and/or PC (gaming or otherwise). As I believe consoles output in RGB in SDR, as do PC video signals, should one use RGB patterns for picture modes for these? An example of my scenario: UB820 player, PS4 Pro/Xbox One X/Switch/PC, I would then do the following to ensure best accuracy:

ISF Bright/Dark For SDR Movies: YCbCR Patterns
SDR Game Mode: RGB Patterns
Cinema/Technicolor as day/night for PC: RGB Patterns

For HDR everything is in YCbCr so this wouldn't be an issue there. And thanks for the info on how you should make sure to match the source signal on the LG.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #290
@ConnecTEDDD Since one should use YCbCr patterns for signals encoded in YCbCr, would the inverse apply when calibrating an input for a video game console and/or PC (gaming or otherwise). As I believe consoles output in RGB in SDR, as do PC video signals, should one use RGB patterns for picture modes for these? An example of my scenario: UB820 player, PS4 Pro/Xbox One X/Switch/PC, I would then do the following to ensure best accuracy:

ISF Bright/Dark For SDR Movies: YCbCR Patterns
SDR Game Mode: RGB Patterns
Cinema/Technicolor as day/night for PC: RGB Patterns

For HDR everything is in YCbCr so this wouldn't be an issue there. And thanks for the info on how you should make sure to match the source signal on the LG.
Hi,

My recommendation apply to any kind of calibration performed, for 3D LUT or manual calibration, and its not affecting only LG OLED's but other brands/models also.

Since we are aiming to have the best possible results, all details like that one is important to improve the final results.

When its possible, you always have to use the some colorspace during patch generation as you will use later for the content playback.

If we talk for specifically for LG OLED 2018, when you want to watch a movie from the HDMI input, then the ideal is to use YCbCr 4:2:2 or 4:4:4 10/12 bit patch generation and set all your players (Blu-Ray/Media Player/Network Streamer) to have output YCbCr 4:2:2 or 4:4:4 with 10/12-bit.

About patch generation you need at least 10-bit YCbCr 4:2:2 or 4:4:4 (as explained to the '7' here). If it will be 4:2:2 or 4:4:4, it doesn't matter so much since the patterns are solid colors window so there will be not any difference to the up-sampling.

About YCbCr player output setting, between 4:2:2 or 4:4:4 ; 10-12bit... your selection will be related with what device will do better the color up-sampling, the TV or player?...... or if its better to send the resolution as native 1080p for SDR movie or to upscale to 2160p, there you have to perform evaluation to find what device do better up-scalling....for both kind of tests you can use the S&M calibration disk which has colorspace evaluation patterns.

Since content is YCbCr 4:2:0 compressed video, the player has to do at least the vertical upsampling 4:2:0 to 4:2:2, and then see if the TV or the Source is the better choice to do the horizontal upsampling (4:2:2 to 4:4:4).

About your Panasonic UB820 Blu-Ray Player, set it to YCbCr 4:4:4, it will do better job that the LG OLED for sure.

All the for movies which are 24p YCbCr, now for games....

For XBOX One/PS4/Switch, since games need 2160p60, you don't have many choices.

With 2160p60 you are limited to 8-bit with RGB because the bandwidth required for the connection is 17.82 Gbit.

TV's/Players chips for input/output signal with HDMI 2.0 can transfer up to 18 GBit so you can't do 10 bit RGB with 2160p (8-bit is not enough for HDR also)

So select 2160p60 YCbCr for the consoles and probably they will output 4:2:0 with 10 or 12 bit.

HDR10 need 10-bit so you can't transfer 2160p60 RGB 10-bit with 2.0 HDMI chips.

I don't have info about how consoles internally are handling the graphics, the logic says at RGB (so uncompressed) so later the player to the output will compress to YCbCr to be able to output 2160p.

But worth to test if 1080p60 RGB-Video 12-bit output will look better with SDR games (or 2160p60 8bit SDR), and enable also the PC Icon of the LG, for the TV to be able to receive full chroma and bypass the internal conversion to 4:2:2 required for some procesing stages (works for 59.97p or 60p only). The only problem there it will be the gradation because panel with work at 8-bit. But worth to test and see.

Some tests I performed early last year with XBOX One S (I don't remember the FW version), when you were playing an UltraHD HDR 2160p24 movie, with settings: 2160p, 8/10/12 bit...then it will playback outputing with 2160p24 HDR 10bit RGB-Video (doesn't matter if you have selected previously 8/10/12), but generally it had wrong colorspace conversion matrix when I tested (YCbCr 2020 -> RGB chroma bug) With REC.709 YCbCr -> RGB-Video it had no problem, but I found some 'roundnng' digital errors, which was slight different if you were playing a disk or USB, for details see there.

But when you playback 2160p HDR10 movie (Billy Lynn's Long Halftime Walk), which is UltraHD HDR 2160p60....

With settings: 2160p, 12 bit (4:2:2 Enabled) it will playback with 2160p60 HDR 10bit YCbCr 4:2:2.

With settings: 2160p, 12 bit (4:2:2 Disabled) it will playback with 2160p60 HDR 10bit YCbCr 4:2:0.

------------------------------------------

Content: Blu-Ray 1080p60

Settings: 1080p, 8/10/12 bit

It will playback with 1080p60 8/10/12 bit RGB-Video

------------------------------------------

Content: Blu-Ray 1080p24

Settings: 1080p, 8/10/12 bit

It will playback with 1080p24 8/10/12 bit RGB

------------------------------------------

Content: Blu-Ray 1080p24

Settings: 2160p, 8/10/12 bit

It will playback with 2160p24 10bit YCbCr 4:4:4

------------------------------------------

For PC, for desktop.....the LG has a very bad banding when you send PC Full Range signal, so it will look better if you compress your PC VGA output to RGB-Video and use LG with RGB-Video, than using RGB Full from PC VGA and LG with TV Black 'High'... When you will use VGA with RGB-Limited output, use patterns with RGB-Full patterns in that case, since the VGA later will compress them to RGB-Video to its output.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #291
Hi Ted,
I have LG C8 and was wondering if I can completely switch off abl but I guess it is not possible.
Thank you for reply.
Hi,

For SDR, you can have no ABL when you will have peak output of 100% White calibrated with max 150-160 nits.

When you go higher, the panel will be more unstable over the time and drift will be increased.

There image processing circuit with algorithm to detect (checking APL in RGB) and reduce the values it will send to the controller where the RGB 10-bit (3x10-bit) values will be converted to WRGB 10-bit (4x10-bit) for the panel, there others algorithms also to protect internal components from overheating or voltage driving issues, static images, for not consuming too much power etc.

LG 2018 OLEDs can reach and maintain approximately 400 nits with up to 70% APL (Average Picture Level), meaning that they have enough brightness headroom to be comfortably used in bright environments for watching low-to-mid APL content (the most of movies and TV series) for day mode calibration (secondary priority calibration).

The impact of the ABL will be more noticed with high-APL content such as hockey and winter sports, as brighter reduction will start when the picture APL will be increased.

For the day view, 320 nits calibrated with gamma 2.2 looks like a good choice for day viewing, if its comfortable for the eyes, it has to do with the room/lights/windows etc.

But for Reference night night, any Automatic Brightness Limiter (ABL) functions shall not be enabled, so 100 nits should be the calibrated luminance target for SDR.

Before a lot of years, when for HD SDR mastering Sony reference CRT's being used, 100 nits was the reference white also, while some studios in Europe were using lower luminance, 80 nits.

When you go higher from 100 nits with LG WRGB OLED, the volumetric, non-linear and non-additive issues will be increased also, these issues can be corrected only by using a 3D LUT.

To give you one example that may help further, the following pictures shows one LG 65C9 WRGB OLED profiled with 100 nits peak output, measured using 21-Point Cube (9261 measurements).

ColourSpace (not yet released) software used for the graphs.

The target colorspace is the native panel colorspace of the display, so should be 100% accurate if the display had no inherent issues.

This means that you measure the panel to its native mode and then you validate creating a new colorspace based to native primaries/white point/ average gamma you measured.

You can easily see the very large non-linear volumetric errors!



The green points showing measurements that have a sub-1 dE2000 error.

Orange points are between 1 and 2.3 dE2000.

Red points are above 2.3 dE2000.

All the Red points show errors over 2.3 dE.

Looking the 3D CIE xyY graph, the areas of errors are easier to plot:



And with Tangenet lines (plots of the errors, shows where the point should actually be):



And as a Normalized 3D Cube, where the non-linear errors inherent within the display are very easy to spot - the scalloping at the edge for example:



And with Tangent Lines:



Note the way the tangent lines change direction, denoting severe volumetric non-linear errors.

If you have a non-linear system, you need a full 3D model to manage better what you want to do.

The greater the non-linearity, the bigger the size of the 3D LUT required to manage it.

This example above works as an 'early but next generator and truly professional way to evaluate a display at really deep/high level, of its real performance, an example for 'professional reviewers', as for better understanding the real issues, it will require more advanced methodology than using a typical/limited patch-set and classic verification methods.

The above procedure, for LightSpace users, its relatively easy to perform, by profiling the display with a large cube based profile (with the display set to its native, un-calibrated setting), generating a new Color Space with the peak RGB & W values, as well as the average gamma (click to generate report to see the average gamma), and then generating a LUT with the Source as the new color space, and Destination as the actual profile.

The graphs will not be such advanced as ColourSpace but a closer cook of the generated LUT to the 'UNITY' (full cube with equaled spaced dots, when you will look the 3D Cube Graphics LUT Viewer) the better will be the underlying capabilities of the display.
 

·
Registered
Joined
·
406 Posts
Hi,

My recommendation apply to any kind of calibration performed, for 3D LUT or manual calibration, and its not affecting only LG OLED's but other brands/models also.

Since we are aiming to have the best possible results, all details like that one is important to improve the final results.

When its possible, you always have to use the some colorspace during patch generation as you will use later for the content playback.

If we talk for specifically for LG OLED 2018, when you want to watch a movie from the HDMI input, then the ideal is to use YCbCr 4:2:2 or 4:4:4 10/12 bit patch generation and set all your players (Blu-Ray/Media Player/Network Streamer) to have output YCbCr 4:2:2 or 4:4:4 with 10/12-bit.

About patch generation you need at least 10-bit YCbCr 4:2:2 or 4:4:4 (as explained to the '7' here). If it will be 4:2:2 or 4:4:4, it doesn't matter so much since the patterns are solid colors window so there will be not any difference to the up-sampling.

About YCbCr player output setting, between 4:2:2 or 4:4:4 ; 10-12bit... your selection will be related with what device will do better the color up-sampling, the TV or player?...... or if its better to send the resolution as native 1080p for SDR movie or to upscale to 2160p, there you have to perform evaluation to find what device do better up-scalling....for both kind of tests you can use the S&M calibration disk which has colorspace evaluation patterns.

Since content is YCbCr 4:2:0 compressed video, the player has to do at least the vertical upsampling 4:2:0 to 4:2:2, and then see if the TV or the Source is the better choice to do the horizontal upsampling (4:2:2 to 4:4:4).

About your Panasonic UB820 Blu-Ray Player, set it to YCbCr 4:4:4, it will do better job that the LG OLED for sure.

All the for movies which are 24p YCbCr, now for games....

For XBOX One/PS4/Switch, since games need 2160p60, you don't have many choices.

With 2160p60 you are limited to 8-bit with RGB because the bandwidth required for the connection is 17.82 Gbit.

TV's/Players chips for input/output signal with HDMI 2.0 can transfer up to 18 GBit so you can't do 10 bit RGB with 2160p (8-bit is not enough for HDR also)

So select 2160p60 YCbCr for the consoles and probably they will output 4:2:0 with 10 or 12 bit.

HDR10 need 10-bit so you can't transfer 2160p60 RGB 10-bit with 2.0 HDMI chips.

I don't have info about how consoles internally are handling the graphics, the logic says at RGB (so uncompressed) so later the player to the output will compress to YCbCr to be able to output 2160p.

But worth to test if 1080p60 RGB-Video 12-bit output will look better with SDR games (or 2160p60 8bit SDR), and enable also the PC Icon of the LG, for the TV to be able to receive full chroma and bypass the internal conversion to 4:2:2 required for some procesing stages (works for 59.97p or 60p only). The only problem there it will be the gradation because panel with work at 8-bit. But worth to test and see.

Some tests I performed early last year with XBOX One S (I don't remember the FW version), when you were playing an UltraHD HDR 2160p24 movie, with settings: 2160p, 8/10/12 bit...then it will playback outputing with 2160p24 HDR 10bit RGB-Video (doesn't matter if you have selected previously 8/10/12), but generally it had wrong colorspace conversion matrix when I tested (YCbCr 2020 -> RGB chroma bug) With REC.709 YCbCr -> RGB-Video it had no problem, but I found some 'roundnng' digital errors, which was slight different if you were playing a disk or USB, for details see there.

But when you playback 2160p HDR10 movie (Billy Lynn's Long Halftime Walk), which is UltraHD HDR 2160p60....

With settings: 2160p, 12 bit (4:2:2 Enabled) it will playback with 2160p60 HDR 10bit YCbCr 4:2:2.

With settings: 2160p, 12 bit (4:2:2 Disabled) it will playback with 2160p60 HDR 10bit YCbCr 4:2:0.

------------------------------------------

Content: Blu-Ray 1080p60

Settings: 1080p, 8/10/12 bit

It will playback with 1080p60 8/10/12 bit RGB-Video

------------------------------------------

Content: Blu-Ray 1080p24

Settings: 1080p, 8/10/12 bit

It will playback with 1080p24 8/10/12 bit RGB

------------------------------------------

Content: Blu-Ray 1080p24

Settings: 2160p, 8/10/12 bit

It will playback with 2160p24 10bit YCbCr 4:4:4

------------------------------------------

For PC, for desktop.....the LG has a very bad banding when you send PC Full Range signal, so it will look better if you compress your PC VGA output to RGB-Video and use LG with RGB-Video, than using RGB Full from PC VGA and LG with TV Black 'High'... When you will use VGA with RGB-Limited output, use patterns with RGB-Full patterns in that case, since the VGA later will compress them to RGB-Video to its output.
Thanks for the info. I definitely let the UB820 do the upscaling, never even bothered to do a comparison as Panasonic's top of the line processing would easily beat the LG's. The bug on the Xbox UHD Blu-Ray playback was fixed at some point last year btw, not that I ever use it for that with the Panasonic I have. However the Xbox still handles gamma or black levels wrong, as Limited RGB/Full RGB don't seem to make any difference there, whereas you can see the difference clearly on the PS4. Both the PS4 Pro and Xbox One S/X can do 4:2:2 10 bit HDR as long as your cable and tv can handle it (PS4 Pro will do it automatically, where the Xbox has a setting for it) since it's right up against the 18gbps limits of the cable.

Shouldn't 2160p RGB 8 bit (SDR gaming) fit into an HDMI 2.0 signal though instead of converting it to YCbCR? It's only when in HDR 10 bit that the bandwidth is too much, or am I wrong?
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #293
Shouldn't 2160p RGB 8 bit (SDR gaming) fit into an HDMI 2.0 signal though instead of converting it to YCbCR? It's only when in HDR 10 bit that the bandwidth is too much, or am I wrong?
Yes, it can fix but you have to compare if it will look better from 10bit YCbCr for example.

What TV model you have?

2160p60 8-bit is 17.82 Gbit, so it can pass using HDMI 2.0b.
 

·
Registered
Joined
·
472 Posts
Use of parametric gamma for near back and overall feedbacks

I have redone my 1DLUT generation using the new SW (05.10.03).
As this the standard process, as explained by Ted, I could not get what I want, I have thought to same variances, due to the fact that following his standard method, black is crunched.

I am calibrating a 100 Nits. Manually I first find the luminance level that makes me closest to pannel on switch (52) and then flatten 2.4 Gamma and RGB balance.At 05 IR I increase luminance to the best Near Black I can get and then I scale a little 10 IRE accordingly. Here the luminance settings I have used

and how that reflects in measurements:


When I do Autocal (source rec 709), based on the following measures;

I get the following readings
.
The outcome is that black is definitely cruched (i.e. no way to see 2% grey).

My idea is: why not using a parametric gamma to reproduce in autocal, the same result? I took some time, but at the end I have created an excel tool to generate such a parametric gamma. Here how the previous manual calibration appears after applying such parametric gamma:
.
I only touched three gamma levels, increasing gamma, not touching the darkest one, as already indicating the need to boost some luminance.

Done the usual 1DLUT process generation, using that parametric gamma as source, here the outcome:
.
More luminance at low IRE, now 2% can be seen, anyhow the one I have with manual calibration is more visible.

My general comments:


  1. It looks this approach (parametric gamma) is to some extent working. I cannot get the same level of near black (in manual I can see 18, in autocal I can see 20), but anyhow there is some improvment.
  2. But what is concerning me the most, is what I see at high IREs in all Autocal: this is confirmed playing ramps, where color shifting is clearly visible. Manual calibration is OK on that.
  3. I have played various near black patterns and other ones to check if my somehow extreme luminance settings are creating artifacts: up to now nothing worth to be reported.
My conclusion: honestly I was expecting more from this 1DLUT Autocal. What concerns me the most is the color shift at high IRE. I have done some tests with Calman, based on 04.10.55 and I have got results in line with my manual calibration. It looks like the way my pannel reacts is not expected by LS software, and augmented data don't fix it. I am not a fan of Autocal, for its absenze of back-up, but combining its 1DLUTgeneration with LS 3DLUT generation looks, at least on my pannel, the most effective approach.

My next step will be 3DLUT generation, based on my manual calibration


PS: has anybode else noted that when clicking "enable calibration" in Device Control, Gamut is set to Wide and gamma to 2.2 in all Cinema modes?
 
  • Like
Reactions: sillysally

·
Registered
Joined
·
472 Posts
Near Black Issues from poor factory Quality Control

Some LG OLED's have passed the factory QC (Quality Control) with near black clipping issues and shipped to stores, for these panels it will be required to adjust the Brightness to 55-56 or up to 60-61 (in worst case scenario) from the default 50 value for the near black clipping issue to be resolved.

..........

For expert users, its recommended instead of adjusting normal TV menu White Balance settings, to adjust only the White Balance values of color temperature from inside the TV's Service Menu only. The reason for that is because the normal menu White Balance settings combined with White Balance of Service Menu settings to work as offset from the native panel, and when you have active two controls which are performing the same job, this can make the internal processing algorithm more complex, so its better idea one White Balance menu to be used only to adjust from native panel response and not from both menus (normal TV + service) the same time.

White Balance settings of 'Warm' in Service Menu will adjust the 'Warm 2' preset of Normal TV Menu.

White Balance settings of 'Medium' in Service Menu will adjust the 'Warm 1' preset of Normal TV Menu.

White Balance settings of 'Cool' in Service Menu will adjust the 'Cool' preset of Normal TV Menu.

LG's factory pre-calibrate these service menu White Balance color temperature modes using a 80% Gray pattern with 6500K for 'Warm', 9300K for 'Medium' and 11000K for 'Cool'.
Only a note: Obvious that doing that you are also modifying manual calibration you have already done. Further, doing that for HDR calibration with my previous B6, I have found out that modifications in SM medium are affecting also Warm2. I think it's related to the 5 expert modes temperatures (Cold, Medium, Warm1, 2, 4) out of 3 setting in SM. Just a warning as I have not checked on E8.
I have calibrated them with the help of the B6 Service Manual and setting an expert mode at 100 Nits I could get very accurate WP without touching 2 points High.
 
  • Like
Reactions: L30Z3N

·
Registered
Joined
·
2,165 Posts
Each colorspace (REC.709 for home release or DCI-P3 for commercial cinema release) while they use exact the same xy coordinates to create the D65 White Point, each colorspace is using different mixture of primary colors luminance to archive that, while both have 6504K color temperature.

When you use the RGB Balance Chart of a calibration software and see the three (R/G/B) Channels Bars at exact 100% = 0 dE; doesn't mean that it has been used equal percentage of luminance per each color channel.

The calibration software it's doing the normalizing the luminance ratio per primary color internally according to the selected colorspace target options to provide to the user interface a more calibration friendly chart.



REC.709 Colorspace D65 White Point (x: 0.3127 y: 0.329, 6504K) is using Red 21.27%, Green 71.52% and Blue 7.22% luminance per each color channel.

DCI-P3 Colorspace D65 White Point (x: 0.3127 y: 0.329, 6504K) is using Red 22.9%, Green 69.17% and Blue 7.93% luminance per each color channel.

REC.2020 Colorspace D65 White Point (x: 0.3127 y: 0.329, 6504K) is using 26.27%, Green 67.80% and Blue 5.93% luminance per each color channel.
Isn't this because each colorspace is using different red, green, and blue primaries? For example, Rec.709 is using 21.27% of x: .640 y: .330 red, but Rec.2020 is using 26.27% of x: .708 y: .292 red. These aren't different mixtures of the same red, blue, and green, but different mixtures of different reds, blues, and greens.

I'm also assuming that our OLED sets must use a different mixture than any of these to achieve D65 at x: 0.3127 y: 0.239, 6405K, for two reasons. First, the RGB sub pixel elements do not match the primaries for any of these color spaces, falling slightly short of DCI-P3, and well short of Rec.2020. Second, is the integration of the white sub-pixel.
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #298
Isn't this because each colorspace is using different red, green, and blue primaries? For example, Rec.709 is using 21.27% of x: .640 y: .330 red, but Rec.2020 is using 26.27% of x: .708 y: .292 red. These aren't different mixtures of the same red, blue, and green, but different mixtures of different reds, blues, and greens.
Hi,

Different ratio of primary luminance mixture coming from the fact that primaries are different per colorspace.

I'm also assuming that our OLED sets must use a different mixture than any of these to achieve D65 at x: 0.3127 y: 0.239, 6405K, for two reasons. First, the RGB sub pixel elements do not match the primaries for any of these color spaces, falling slightly short of DCI-P3, and well short of Rec.2020. Second, is the integration of the white sub-pixel.
Since each display has different native gamut, what each display is doing to the sub-pixels mixture to de-saturate from native primaries to reach a REC.709 primary colors for example, its unique per each display.

With WRGB OLED after all the processing of the signal, before the panel driving there is a controller (secret sausage) where its converting the 3x 10bit RGB to 4x 10bit RGBW values, but 3 of the 4 pixels are active, LG Electronics talking with more detail and propose also a new 3-matrix meter correction method inside to:

83‐3: Calibration of Colorimeters for RGBW Displays @ SID Symposium Digest of Technical Papers Issue 50 (June 2019).
 

·
Registered
Joined
·
8,913 Posts
Discussion Starter #299
Only a note: Obvious that doing that you are also modifying manual calibration you have already done. Further, doing that for HDR calibration with my previous B6, I have found out that modifications in SM medium are affecting also Warm2. I think it's related to the 5 expert modes temperatures (Cold, Medium, Warm1, 2, 4) out of 3 setting in SM. Just a warning as I have not checked on E8.
I have calibrated them with the help of the B6 Service Manual and setting an expert mode at 100 Nits I could get very accurate WP without touching 2 points High.
Hi, if you check with C8 what color modes each affecting the SM changes in SDR and report back, then I will update the info/quide. ;)
 

·
Registered
Joined
·
472 Posts
SM WB => Mode Temperature

Hi, if you check with C8 what color modes each affecting the SM changes in SDR and report back, then I will update the info/quide. ;)
Here outcome of my testing, done on SM, decreasing for each SM the default RGB by 20 points and verifying the result on Dark Room mode with Oled Light set to 25.
The outcome is ('=>' means 'affects')


  • SM Cold => Cold
  • SM Medium => Medium, Warm1
  • SM Warm => Warm1, Warm2, Warm3
See the attachment for details.
 

Attachments

281 - 300 of 1461 Posts
Top