AVS Forum banner

2481 - 2500 of 3036 Posts

·
Registered
Joined
·
400 Posts
Will do. Have you tested with the Contrast Control on LG OLED?

I have found anomalies at 100 contrast in factory mode, including raised black.

At 99 contrast viewing the Dynamic Range Low pattern from S&M I clip at 65, but remove all noise from the absolute black background.

Also, at 100 contrast I'm seeing a color shift compared to 99 contrast. Does Calman Matrix LUT require the contrast at 100 to load the LUT?

I will experiment tonight. I'm getting great results with SDR, but I'm afraid to touch HDR due to the Matrix LUT not being configured properly.

I did try increasing Color until my saturation looked better, but I'm still getting lower color errors using the factory LUT.

I just wish I could do the 1D LUT and still keep the factory HDR 3D LUT.


Do you have a C9? Which firmware are you running?
I unfortunately have a B8, latest firmware.

I'm seeing anomalies with factory settings. I'm not sure if these are creating issues before Calman does its thing.

EDIT: I have access to C9, CX... I can simply view the pattern on the S&M disc and view specific content on all sets to report is 2019 and 2020 have the same behavior.
 

·
Registered
Joined
·
4,297 Posts
I unfortunately have a B8, latest firmware.

I'm seeing anomalies with factory settings. I'm not sure if these are creating issues before Calman does its thing.

EDIT: I have access to C9, CX... I can simply view the pattern on the S&M disc and view specific content on all sets to report is 2019 and 2020 have the same behavior.


You are referring to HDR10 mode correct?
 

·
Registered
Joined
·
400 Posts
I unfortunately have a B8, latest firmware.

I'm seeing anomalies with factory settings. I'm not sure if these are creating issues before Calman does its thing.

EDIT: I have access to C9, CX... I can simply view the pattern on the S&M disc and view specific content on all sets to report is 2019 and 2020 have the same behavior.


You are referring to HDR10 mode correct?
Correct. Dolby Vision does not exhibit the change in black level tied to the Contrast control, and I'm not having any issues with post-cal verify when using Calman to create the Dolby config file.
 

·
Registered
Joined
·
4,297 Posts
Correct. Dolby Vision does not exhibit the change in black level tied to the Contrast control, and I'm not having any issues with post-cal verify when using Calman to create the Dolby config file.


One thing to be aware of is that the B8 has the Alpha 7 processor which has inferior HDR10 color processing compared to the Alpha 9 in C8.
 

·
Registered
Joined
·
588 Posts
I've been using the default values (45/5/25) for SDR in this case.

You posted the values for the pattern insertion not the pattern delay. You find that setting under the source tab of the tpg that you are using.

Btw it has nothing to do with that I tried values from 0.25-1.00 second and it didn't changed anything. This seems to be a video level problem between TPG and LG.
 
  • Like
Reactions: mrtickleuk

·
Registered
Joined
·
24 Posts
I’m referring to pattern delay for the patter generator not pattern insertion. Which meter are you using?

I would suggest at least .75 pattern delay on the source tab of your pattern generator in Calman.
You posted the values for the pattern insertion not the pattern delay. You find that setting under the source tab of the tpg that you are using.

Btw it has nothing to do with that I tried values from 0.25-1.00 second and it didn't changed anything. This seems to be a video level problem between TPG and LG.
Ah, sorry. I've been using the iTPG in my C9 with a 0.25 delay.
Meter is an i1 Display Pro.
 

·
Registered
Joined
·
400 Posts
Correct. Dolby Vision does not exhibit the change in black level tied to the Contrast control, and I'm not having any issues with post-cal verify when using Calman to create the Dolby config file.


One thing to be aware of is that the B8 has the Alpha 7 processor which has inferior HDR10 color processing compared to the Alpha 9 in C8.
Understood, but doing a pre-cal verify in HDR10 before touching the Matrix LUT gives significantly better color checker results, and no issues with real world content

I could see the 17^3 LUT creating issues with large LUT sets compared to an A9, but the issue I'm trying to solve is why the HDR 10 Matrix LUT is necessary.

Not only the undersaturation bug, but the color shift present after uploading any custom 1D LUT. Biggest offenders are Orange, Yellow.

Until resolved I can only recommend Calman for SDR, because any adjustments to the HDR pipeline create more real world issues than they solve.

I want this to work, but it's clear that something is not translating properly.

If I had the same errors out of the box, I'd blame my equipment, but at this point I am trying to see if there's some way to pre-set the LG to allow for proper calibration.

I will be testing the C9 and CX this week with both an external TPG and the C9 IPG.

If the same issues are gone, then I'll be upgrading to the C9.
 

·
Registered
Joined
·
2,165 Posts
For those that are having the clipping issue, can you load a new DV config file that you have edited in notepad so that the Tmin value is zero instead of the measured value?
I'm happy to give that a try, and have already been considering that anyway. Do you have any information on what numerical format(s) DV uses for this value? Since the CalMAN measured/generated number is always in scientific E notation, such as "1E-11" I'm unclear whether it will recognize "0," "0.0," or something else as zero. It would be a great help to know what the factory Dolby Vision Config File uses.
 

·
Registered
Joined
·
4,297 Posts
I'm happy to give that a try, and have already been considering that anyway. Do you have any information on what numerical format(s) DV uses for this value? Since the CalMAN measured/generated number is always in scientific E notation, such as "1E-11" I'm unclear whether it will recognize "0," "0.0," or something else as zero. It would be a great help to know what the factory Dolby Vision Config File uses.


0.0 should work. I won’t be able to test it myself until I go to the office tomorrow.
 

·
Registered
Joined
·
6,626 Posts
I've only tried using the Lightning LUT in my previous calibrations. But before my trial of Lightspace runs out I wanted to take a peak.

Lightning LUT:


60min IRP about 900-1000 points:


Seems like I get the sharkfin artifact as well. It's nothing that I can spot during normal viewing, perhaps test patterns is more revealing.
Now having created 3D LUTs with LS/CS, as Blackjoker points out last time i posted a CM LUT, although the LUT viewer will indicate how good or bad the LUT "may" be, even on a perfect looking LUT you need to look at the Gray Scale and CMS scans. Just because the LUT may look perfect in the viewer doesn't mean the results are good and vice versa. In other words i've created 3DLUTs that looked perfect in the viewer and the gray scale was not very good, imo. Your CM 3DLUT looks very typical with that yellow spike. I got rid of it by increasing the delay time for my meter but i'm still getting non-perfect low blue areas.
 
  • Like
Reactions: Gnaklellac

·
Registered
Joined
·
2,165 Posts
Ok, here we go...

LG OLED65C9AUA Dolby Vision Cinema Picture Mode

CalMAN custom 1D LUT + CalMAN custom DVConfigFile Tmin 1E-11:
DVS Black Clipping 1 (DV profile 5): 3.1%/32 & below bars missing. Barely perceptible/questionable glow from black areas.
DVS Black Clipping 2 (DV profile 5): 3.1%/32 & below bars missing. Barely perceptible/questionable glow from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 65-66 and above visible. UHDA LCD, Pulsar, & UHDA OLED blocks discernible. Barely perceptible/questionable glow from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Soft glow entire screen (Guesstimating .0001 nit or less).

LG factory 1D LUT + CalMAN custom DVConfigFile Tmin 1E-11:
DVS Black Clipping 1 (DV profile 5): 3.1%/32 & below bars missing. No visually detectable light output from black areas.
DVS Black Clipping 2 (DV profile 5): 3.1%/32 & below bars missing. No visually detectable light output from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 67-68 and above visible. UHDA LCD, & Pulsar blocks discernible. UHDA OLED block barely/questionably discernible. No detectable light output from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Black screen, no visually detectable light.

LG factory 1D LUT + LG factory DVConfigFile:
DVS Black Clipping 1 (DV profile 5): Bars down through 1.2%/12 visible. No visually detectable light output from black areas.
DVS Black Clipping 2 (DV profile 5): Bars down through 1.2%/12 visible. No visually detectable light output from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 66-67 and above visible. UHDA LCD, Pulsar, & UHDA OLED blocks discernible. No visually detectable light output from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Black screen, no visually detectable light.

CalMAN custom 1D LUT + LG factory DVConfigFile:
DVS Black Clipping 1 (DV profile 5): Bars down through 1.2%/12 visible. Barely perceptible/questionable glow from black areas.
DVS Black Clipping 2 (DV profile 5): Bars down through 1.2%/12 visible. Barely perceptible/questionable glow from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 65-66 and above visible. UHDA LCD, Pulsar, & UHDA OLED blocks discernible. Barely perceptible/questionable glow from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Soft glow entire screen (Guesstimating .0001 nit or less).

LG factory 1D LUT + CalMAN custom DVConfigFile Tmin 0.0:
DVS Black Clipping 1 (DV profile 5): 3.1%/32 & below bars missing. No visually detectable light output from black areas.
DVS Black Clipping 2 (DV profile 5): 3.1%/32 & below bars missing. No visually detectable light output from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 67-68 and above visible. UHDA LCD, & Pulsar blocks discernible. UHDA OLED block barely/questionably discernible. No detectable light output from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Black screen, no visually detectable light output.

CalMAN custom 1D LUT + CalMAN custom DVConfigFile Tmin 0.0:
DVS Black Clipping 1 (DV profile 5): 3.1%/32 & below bars missing. Barely perceptible/questionable glow from black areas.
DVS Black Clipping 2 (DV profile 5): 3.1%/32 & below bars missing. Barely perceptible/questionable glow from black areas.
S&M DV 2.9 Dynamic Range Low (DV profile 7): Bars 65-66 and above visible. UHDA LCD, Pulsar, & UHDA OLED blocks discernible. Barely perceptible/questionable glow from black areas.
DVS 0% - Black/10% Patch 0 C0 C0 - CalMAN Dolby Vision Workflow (version 5.9.XX): Soft glow entire screen (Guesstimating .0001 nit or less). Possible/uncertain improvement over Tmin 1E-11, but still not completely black as with LG factory 1D LUT. (Seemed like it took longer to see.)

My summary and conclusions from the above:

Near black clipping of DVS/R. Masciola patterns consistently results from the use of a custom DV Config file, and is unaffected by which 1D LUT is used (factory vs. custom). No detectible difference between Tmin = 0.0 vs. 1E-11 in this regard. There is no visible clipping using the dynamic range low pattern provided by Stacey Spears in any case. (Could this be related to some difference between DV Profiles 5 & 7?) Thanks to Stacey for making this available!!!

Elevated black consistently results from use of a custom 1D LUT, and is unaffected by which DV Config file is used (factory vs. custom). Possible slight improvement in this regard by changing Tmin from 1E-11 to 0.0 in the DV Config file.
 

·
Registered
Joined
·
24 Posts
Now having created 3D LUTs with LS/CS, as Blackjoker points out last time i posted a CM LUT, although the LUT viewer will indicate how good or bad the LUT "may" be, even on a perfect looking LUT you need to look at the Gray Scale and CMS scans. Just because the LUT may look perfect in the viewer doesn't mean the results are good and vice versa. In other words i've created 3DLUTs that looked perfect in the viewer and the gray scale was not very good, imo. Your CM 3DLUT looks very typical with that yellow spike. I got rid of it by increasing the delay time for my meter but i'm still getting non-perfect low blue areas.
The post calibration verification looks good (haven't done any other verification apart from that at this time). That being said I'm just looking for a middle ground with good results in relation to effort.

I have the following tools at hand:
LG C9 with iTPG
i1 Display Pro
Calman Home for LG
Vero 4K+ with AVS709 test patterns and Mehanik HDR10 test patterns
Harmony remote which gives me access to the service menu if needed.

So my options are limited to:
Manual 2-pt grayscale (maybe 20-pt) with test patterns from my Vero
Manual CMS adjustments with test patterns from my Vero
Autocal for grayscale and color

Given my rookie level and limited toolset. I should be able to get better results from Autocal?
(Raised blacks in HDR aside I'm subjectively pleased with the results so far, but the OCD is always there :) )

I’m referring to pattern delay for the patter generator not pattern insertion. Which meter are you using?

I would suggest at least .75 pattern delay on the source tab of your pattern generator in Calman.
OK, thanks. Will hopefully have some more time during the weekend to tinker.
Is the >0.75 pattern delay a general recommendation for most scenarios or just when doing a larger amount of messurements?
 

·
Registered
Joined
·
2,888 Posts
I had very good success with Calman on my 77C9 back in December and it still looks amazing. This past weekend I did our new 55C9 that only had about 85 hours on it. I will do it again when it gets closer to 200 hours, but the out of box was not awful and after calibration is looks excellent. I used lightning LUT since this was just an interim calibration. Post calibration results are exception, with the only complaint is the panel has a peak brightness for HDR of around 646 nits. But I will see what happens after a few compensation cycles. I have a iDisplay Pro that I have had a few years and happy with the results.

Note that my Night mode was using a gamma target of 2.3 since it was at 141 cdm at 100 and around 160 at 109 and my day mode used 2.2, whic landed at 187 cdm for 100 and then around 220 for 109.
 

Attachments

·
Registered
Joined
·
6,626 Posts
OK, thanks. Will hopefully have some more time during the weekend to tinker.
Is the >0.75 pattern delay a general recommendation for most scenarios or just when doing a larger amount of messurements?
We are finding that setting a min delay of 0.75 is helping create a more accurate LUT for some reason. Going from 0.5 to 0.75 seconds got rid of that yellow spike for me so you will need to experiment.
 
  • Like
Reactions: KC-Technerd

·
Registered
Joined
·
9,208 Posts
Ok, here we go...

LG OLED65C9AUA Dolby Vision Cinema Picture Mode
big snip!

Thanks for all that work!

What does "DVS" in the phrase "DVS Black Clipping" mean please?

My summary and conclusions from the above:

Near black clipping of DVS/R. Masciola patterns consistently results from the use of a custom DV Config file, and is unaffected by which 1D LUT is used (factory vs. custom). No detectible difference between Tmin = 0.0 vs. 1E-11 in this regard. There is no visible clipping using the dynamic range low pattern provided by Stacey Spears in any case. (Could this be related to some difference between DV Profiles 5 & 7?) Thanks to Stacey for making this available!!!
Yes, this matches what we understood already. That pattern displays fine from the factory, but you get the black clipping on that pattern as soon when you use a DV config file. This means there's a risk that you'll also get the black clipping in content of course, (which is why IIRC people starting checking it after they'd noticed the clipping).

Elevated black consistently results from use of a custom 1D LUT,
Agreed. Again IIRC, this DV blacks issue is a different, and additional, issue to the 1D LUT problem we get with HDR10/HLG .
 

·
Registered
Joined
·
2,165 Posts
What does "DVS" in the phrase "DVS Black Clipping" mean please?
DVS for Diversified Video Solutions, which is Ryan Masciola, to specify the source of the pattern being evaluated.

Yes, this matches what we understood already. That pattern displays fine from the factory, but you get the black clipping on that pattern as soon when you use a DV config file. This means there's a risk that you'll also get the black clipping in content of course, (which is why IIRC people starting checking it after they'd noticed the clipping).
Yes. I wanted to use this opportunity to verify everything again, including what I had previously tested and what was understood already. Tyler specifically requested evaluation of whether editing Tmin from the measured value to zero (Tmin = 1E-11 to Tmin = 0.0) had any effect on this behavior. I found no evidence that it did. What was interesting is that this clipping does not occur on the Spears & Munsil Dynamic Range Low pattern. I suspect this is something to do with the difference in Dolby profiles, or a reaction to the brighter content/higher contrast in the DVS clipping patterns than the Spears & Munsil pattern (or perhaps both). Determining that is likely beyond any testing I'm going to be capable of doing.

Again IIRC, this DV blacks issue is a different, and additional, issue to the 1D LUT problem we get with HDR10/HLG .
Yes, Tyler has stated that. It very well may be. It is certainly independent to the picture mode in question, but it does completely parallel the HDR10/HLG behavior in following the use of a custom 1D LUT. Because I don't know how to completely isolate the loading and use of a custom 1D LUT from a custom 3D LUT in HDR10/HLG, I can't isolate it to resulting from a custom 1D LUT in HDR10/HLG. I'm instead basing that on what Tyler has previously reported. In Dolby Vision I can isolate it to resulting from the use of a custom 1D LUT, because I can fairly easily isolate the C9's use of a custom 1D LUT from its use of a custom DV Config File.
 
  • Like
Reactions: mrtickleuk
2481 - 2500 of 3036 Posts
Top