AVS Forum banner

1561 - 1580 of 1591 Posts

·
Registered
Joined
·
1,055 Posts
Fix LGD Panel Calibration Issues

Use default settings for Contrast/Color and perform the White Balance 109% White Balance calibration using Service Menu controls (note to not press RESET to the White Balance screen)

Since the problem is different per panel, the required adjustments for Contrast/Color are not the same for all users.

You can fix/improve the factory problem by taking multiple Profile of Primary Color and compare RGB separation charts using different Contrast settings (84, 83, 82, 81, 80) and then Color (50, 51, 52, 53, 54, 55).

Most of the time, for the correction of the panel factory calibration issue, it will require to set Contrast @ 80 and Color at 54 or 52 or 50, but each unit will need different values.

What is RGB Separation?

ColourSpace CMS - Next Generation Calibration... Thread

ColourSpace CMS - Next Generation Calibration... Thread
Yes, these are the "fixes" I was referring to in an earlier post. Excellent explanation and summary, as always!!!
 

·
Registered
Joined
·
11,761 Posts
Yes, these are the "fixes" I was referring to in an earlier post. Excellent explanation and summary, as always!!!
I had this issue on my 77”. Ted helped me thru it as always . I ended up at contrast 80 and color 55 . Lut came out great .


77CX YTTV AppleTv 4K Sony X800M2 Yamaha V685
 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,563
LG OLEDs with 03.11.05 FW

After two months from the latest FW release, LG released today a new FW for 2020 OLED TVs (03.11.05).

That firmware is not probably available yet using TV automatic updates, but using USB can update your TV.

I have just tested to see if something is fixed from all these known bugs LG knows/tracking.

I have to say, nothing is fixed, as LG seems not to be interested in the calibration features of the TV.

iTPG is still not matching external iTPG (it will never match, just checked with Murideo Six-G, the most significant deviation appears in DoVi mode).

SDR/HDR10 iTPG is still displaying the full-field magenta flashes.

DoVi is still using the 'old' iTPG, so LightSpace/ColorSpace users have to use the 0-255 patch scale.

DoVi tone mapping (not related to calibration, as its affecting units before calibration) is still producing problems in near black.

The same recommendations I have posted 3.10.20 FW are valid for 03.11.05

See you next year.

Congratulations to LG for adding new bugs to CX, problems C9 were not suffering.

When users are using external iTPG, the only fix required is related to DoVi Tone Mapping.

iTPG of C9 is not producing magenta flashes and no C7/C8/C9 is sufferting from DoVi issue in near black (not related with calibration)
LG 2020 OLEDs with 03.11.10 FW

Two weeks after the 03.11.05 FW release, it's now available the 03.11.10 FW.

Now DoVi iTPG is producing magenta flashes.

All iTPG modes now produce magenta flashes, while in 03.11.05, we had magenta flashes to SDR/HDR10 only.

DoVi iTPG still needs a different patchscale than SDR/HDR10 iTPG.

No improvements regarding the DoVi Tone Mapping in near black.

So 3.10.20 FW added magenta flashes to SDR/HDR10, and 3.11.10 added magenta flashes between patch generation to DoVi iTPG also.

Now it's a good time from LG to see how to remove all these magenta full-field flashes as probably they are come out of ideas of which bug to introduce.
 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,564
Wait......the DoVi iTPG has a magenta background because the whole USB DoVi Playback is now broken.

LG DoVi MP4 playback from USB...instread of a full field gray background, it has a purple background now....

Any DoVi MP4 file from USB has broken colors.

NetFlix DoVi is working.

Does someone in LG perform any basic test before they released FW?

Be aware and ignore the installation of 03.11.10 FW by disabling the automatic check for TV updates.







 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,566 (Edited)

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,567
LG 2020 OLEDs with 03.11.26 FW

Be aware and ignore the installation of 03.11.26 FW by disabling the automatic check for TV updates.

It will make your DoVi MP4 playback from USB totally useless.

I'm wondering what kind of testing LG is performing before an FW release, including their beta testers.
 

·
Registered
Joined
·
165 Posts
About Service Menu WB pre-calibration adjustment :
If I use SDR mode , what gamma value should I choose (2.4 or 2.2 ?) and what color gamut should I choose (wide or auto?)?
65CX...
 

·
Registered
Joined
·
1,007 Posts
2.2 wide
 
  • Like
Reactions: Dreamer2018

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,570
About Service Menu WB pre-calibration adjustment :
If I use SDR mode , what gamma value should I choose (2.4 or 2.2 ?) and what color gamut should I choose (wide or auto?)?
65CX...
Hi,

Use 2.2 Gamma.

The colorspace should not be selectable before starting White Balance Service Menu Calibration.

See:

 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,571
I had this issue on my 77”. Ted helped me thru it as always . I ended up at contrast 80 and color 55 . Lut came out great .


77CX YTTV AppleTv 4K Sony X800M2 Yamaha V685
Your panel is a great example to demonstrate the LGD factory panel calibration issue, which is not related to WB SM calibration of LGE factory calibration.

After running the regular 3D LUT profiling procedure, Pre-Cal WB SM, and then 3D LUT profiling using ColourSpace, you had a lot of unusual errors.

For a ''true'' evaluation of a display's performance, we always recommend the users run a 10-Point Cube (1000 color points) verification.

Let see your 1000p verification using Contrast 85 and Color 50 (default panel settings).

Here is the 3D CIE Chart, orange dots represent dE above 1, while red dots dE above 2.3:



You had many errors to edge colors (clipping), and regular ColorChecker SG measurement run will never locate these errors as it will not measure any edge color!

Looking at the 2D CIE chart also:



Here was your 1000p 'cake' dE chart from the 'Genious Team':




While you had all that colors clipped, testing with many of my patterns for color clipping, you saw all flashing bars to all patterns you tested, so we used RGB Separation chart to fix your panel issue.

After some testing we did, examining different values of Contrast/Color, we found that for your particular panel, Contrast 80 and Color 55 were the best settings that resolve your panel factory calibration.

After setting Contrast to 80 and Color to 55, you peformed a new 3D LUT profiling with the same patchset/settings, and you had these results:



You had only 1 orange error, to the darkest shade of magenta, and no other error above 1dE.

Here was your 1000p 'cake' dE chart from the 'Genious Team':



For the solution to that problem, I have posted all the required details here:

 

·
Registered
Joined
·
453 Posts
Gents. I have a new i1 Display Pro Plus which I have used for my initial calibration runs. I have access to an i1 Pro 1 rev D. I know this is an older spectro. Is it worthwhile to use this to create a profile or am I better off just sticking with the new colorimeter?
 

·
Registered
Joined
·
11,761 Posts
Your panel is a great example to demonstrate the LGD factory panel calibration issue, which is not related to WB SM calibration of LGE factory calibration.

After running the regular 3D LUT profiling procedure, Pre-Cal WB SM, and then 3D LUT profiling using ColourSpace, you had a lot of unusual errors.

For a ''true'' evaluation of a display's performance, we always recommend the users run a 10-Point Cube (1000 color points) verification.

Let see your 1000p verification using Contrast 85 and Color 50 (default panel settings).

Here is the 3D CIE Chart, orange dots represent dE above 1, while red dots dE above 2.3:



You had many errors to edge colors (clipping), and regular ColorChecker SG measurement run will never locate these errors as it will not measure any edge color!

Looking at the 2D CIE chart also:



Here was your 1000p 'cake' dE chart from the 'Genious Team':




While you had all that colors clipped, testing with many of my patterns for color clipping, you saw all flashing bars to all patterns you tested, so we used RGB Separation chart to fix your panel issue.

After some testing we did, examining different values of Contrast/Color, we found that for your particular panel, Contrast 80 and Color 55 were the best settings that resolve your panel factory calibration.

After setting Contrast to 80 and Color to 55, you peformed a new 3D LUT profiling with the same patchset/settings, and you had these results:



You had only 1 orange error, to the darkest shade of magenta, and no other error above 1dE.

Here was your 1000p 'cake' dE chart from the 'Genious Team':



For the solution to that problem, I have posted all the required details here:

Still enjoying this calibration :)


77CX YTTV AppleTv 4K Sony X800M2 Yamaha V685
 
  • Like
Reactions: ConnecTEDDD

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,574
Gents. I have a new i1 Display Pro Plus which I have used for my initial calibration runs. I have access to an i1 Pro 1 rev D. I know this is an older spectro. Is it worthwhile to use this to create a profile or am I better off just sticking with the new colorimeter?
Hi,

i1PRO1 units were EOL from 2011. It's unknown the condition of the meter.

You can perform a test to find you.

Calibrate the 100% White using the ISF-Dark with your i1Display PRO Plus and the LG WRGB OLED table, for 100 nits.

Then calibrate 100% White using the ISF-Bright with your i1PRO1 for 100 nits.

Display the Grayscale Ramp of my disk and swap between these two picture modes to see which calibration looks more 'white' for your eyes.

 

·
Registered
Joined
·
34 Posts
ConnecTEDDD
liberator72

Thanks for your advices!
They are, as usual, very important and useful.

The last two days I spent in various experiments with ColorSpace, which is new to me,
(I purchased a license for only five days ago),
which I was just running on OLED 55CX.
I will write about the results very soon.

Important question.
I often see this "1000p 'cake' dE chart from the 'Genious Team" in this discussion.

Is it possible to download this spreadsheet somewhere to be able to
lay out the results in such a beautiful performance?
 

·
Registered
Joined
·
1,055 Posts
Important question.
I often see this "1000p 'cake' dE chart from the 'Genious Team" in this discussion.
That particular chart is the brain child of @Anger.miki.

The spreadsheet itself is not available as a download. But you can contact him directly to discuss what is required to generate a verification summary (complete with “cake chart”) by sending him a private message.
 

·
Registered
Joined
·
34 Posts
OK. Thanks for the explanation.
By the way, at least there was a button in LightSpace,
which showed some useful statistics about the quality of the profile.
It was called "Profile Data".

I didn't find it in ColorSpace.
Has this feature really disappeared in CS, or is it well hidden somewhere?
 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,578
By the way, at least there was a button in LightSpace,
which showed some useful statistics about the quality of the profile.
It was called "Profile Data".

I didn't find it in ColorSpace.
Has this feature really disappeared in CS, or is it well hidden somewhere?
It hasn't developed yet; now, the team is working on reporting.
 

·
Registered
Joined
·
34 Posts
So OLED 55CX and, new to me, ColorSpace.

1) First of all, I checked the screen for the "screwed factory calibration" that TED mentioned.
RGB separation turned out to be quite normal, even with the default Contrast 85 / Color 50

2) Next I made a 17x3 cube. Of course, observing all the numerous rituals, drawing attention to endless subtleties, details, etc.
this entertaining process.
The Expert DARK mode was used.
The target was 140 cd / m2 peak brightness, Gamma 2.35
Also, as a starting point, I used the standard LG 2.4 gamma preset, not 2.2 as usual.

3) You can see the results in the images below.
The overall results are not bad.
I only see three points out of 1000 where dE2000 is greater than 1.0
All three are below dE 1.5
Unfortunately ColorSpace does not allow me to see more detailed error statistics.

5) However, it was not possible to get the perfect grayscale.
  • One thousand-point control shows one GS-point exceeding dE 1.0
  • GOL (Gray Only Large) - shows at least three such points.
At the same time, several more gs-points dangerously approach 1.0

Interestingly, the most problematic points this time are in the lower, darker part of the grayscale.

6) Here is the link to 1000p control bcs file.

P.S.
For some reason, I can't attach screenshots in this thread.
So here's a link to the photoset.
 

·
Registered
Joined
·
9,365 Posts
Discussion Starter #1,580
So OLED 55CX and, new to me, ColorSpace.

1) First of all, I checked the screen for the "screwed factory calibration" that TED mentioned.
RGB separation turned out to be quite normal, even with the default Contrast 85 / Color 50

2) Next I made a 17x3 cube. Of course, observing all the numerous rituals, drawing attention to endless subtleties, details, etc.
this entertaining process.
The Expert DARK mode was used.
The target was 140 cd / m2 peak brightness, Gamma 2.35
Also, as a starting point, I used the standard LG 2.4 gamma preset, not 2.2 as usual.

3) You can see the results in the images below.
The overall results are not bad.
I only see three points out of 1000 where dE2000 is greater than 1.0
All three are below dE 1.5
Unfortunately ColorSpace does not allow me to see more detailed error statistics.

5) However, it was not possible to get the perfect grayscale.
  • One thousand-point control shows one GS-point exceeding dE 1.0
  • GOL (Gray Only Large) - shows at least three such points.
At the same time, several more gs-points dangerously approach 1.0

Interestingly, the most problematic points this time are in the lower, darker part of the grayscale.

6) Here is the link to 1000p control bcs file.

P.S.
For some reason, I can't attach screenshots in this thread.
So here's a link to the photoset.
1) Can you upload a Profile of Primary only before the profiling?

2) The reason we suggest to use the 2.2 is for not letting any regular menu LG processing to affect the panel response.

It will degrade your picture, and it's not related to dE.

The panel is calibrated (as a panel) for 2.2 gamma.

100 nits should be the target luminance for SDR for 100% White, so 123 nits @ 109% SuperWhite calibrated.

Pre-Calibrate/Profile with 16-255 and verify the post-calibration with 16-235.

3) The 1000 points you are measuring doesn't contain any primary or secondary color!

Where you found that 1000p verification run?

I haven't sent such an incomplete patchset to any user.

You have 1.37dE to 3.7% Gray, the maximum Grayscale error, at such luminance levels, no display is perfect.

It seems that you are following the instructions we have available.

Which patches have you used for these results?

I don't see any problem with your Grayscale.

Grayscale is not important for a WRGB display, as the volumetric performance is more important.

It's the reason we don't recommend the users to just use Grayscale and ColorChecker; typical measurements run not capable of gathering any data of the real performance; they don't contain any edge color.

WRGB panel will not use RGB subpixels to create a Gray color.

It will use W and another two from RGB, so one primary color subpixel will be OFF.

I don't understand your commend about 100-Point grayscale evaluation.

What do you mean that you have 1 point above 1dE, and that is dangerous?

The TV has a 33-Point Cube 3D LUT, and you measured 17-Point Grayscale to generate the 33-Point.

But when you measured for verification 100-Point Grayscale, you found only 1 point above 1dE, and that was dangerous?

How is it possible for the 100-Point Grayscale to have 1 point above 1dE and the Gray Large (which is 33-Point GS) to have 3 points above 1dE?

6) Post a proper 1000p file, to see your RGB separation also.

Your patchset is 1001 patches but its 11-Point (fake 11-Point) as 1-point of the edge colors is missing.

1000 color patchset should be 10-Point Cube.

See how many colors are missing compared to a 10-Point Cube; see that a thousand points are missing from edges:



Which patchset have you used for profiling?
 
1561 - 1580 of 1591 Posts
Top