AVS Forum banner

861 - 880 of 881 Posts

·
Registered
Joined
·
3,226 Posts
I having a strange anomaly with HDR content and only seems to be games. Any white text or icons over a Black or Grey I can actually watch the warm tone of white letters/icon shift to a blue tint. Letters almost appear to have an oil slick appearance. This only seems to happen on HDR games and I cannot replicate it if I switch the console to SDR.

Adjusting gamma seems to help the text/icons not have this blue shine to them.

It's driving me crazy. I don't know if this is color accuracy related, greyscale related, fald related or what. It's maddening.
I posted them on reddit, here's a link https://www.reddit.com/r/VIZIO_Official/comments/iwtb48 Edit


Turns out the link doesn't work, in just a moment I'll have them here. But what you could possibly be seeing is contrast D65 issues.

Sent from my LGMP450 using Tapatalk
 

·
Registered
Joined
·
3,226 Posts
Would you share you settings? Im assuming it's for your Vizio? I have the same tv as you. I would love to give yours a try even though I know it can be panel dependant. Always looking for new tweaks.
The first settings involve going to the factory menu, so be sure to write down or take a picture of your current factory calibration settings. You have to be on a HDMI port in order for the menu to show up.

On your remote with the TV on, press Menu then 1,9,9,9 and return(this button is directly below the channel rocker. Once in the menu scroll down to color temperature.

Change Gain settings to R 504 G 484 B 512(all of them could be 512, so Blue doesnt need to be touched)

Offset settings R 4, G 2, B 2

When done press menu and the factory menu will close. Start a Dolby Vision stream(Vudu etc preferably SmartCast). Once the DV badge appears, go to calibration and color tuner. Skip to 11 point white balance and change the 5% to 60% and input R 13 G 0 B 12.

For some reason Netflix, Vudu, and Disney+ auto turns on CC which affects the picture.

If your display is close to mine in intolerances, then you are in for a visual treat.

Sent from my LGMP450 using Tapatalk
 

·
Registered
Joined
·
460 Posts
You could try the Mehanik patterns. But if you want to do 20pt white balance, it will depend on whether the pattern steps match up with your TV's steps.
I did a full HCFR sweep for Rec2020 to see what it would look like using the Mehanik HCFR 10% window patterns.

I'm a little confused on what to measure, the sustained "stable" luminance or should I only go for "peaks".
The 55Q8DN starts bright and then gets dimmer on each pattern. Usually was most bright at the start of each pattern when the text was visible and then when the text went away it dimmed noticeably.
I did measurements on the sustained levels but that meant all measurements are way to low on luminance across the board really. ~50% from "reference" in general.
The "peak brightness" levels seem to be closer to reference but not really hitting it either.
For example 50% Greyscale has a sustained level ~50nits but might peak shortly for 70nits at the start when the text is there for the target RGB value/data.

The patterns was supposed to be 1000nits with MaxFALL400 but I could only get 650-700nits peak right now with current settings at best.

Samples for "reference". (Contrast 50 & Backlight 50, Rec2020 colourspace, R-G-B 100-100-100 Y-C-M 50'50-50'50-50'50)
3039974
3039975
 

·
Registered
Joined
·
1,426 Posts
I did a full HCFR sweep for Rec2020 to see what it would look like using the Mehanik HCFR 10% window patterns.

I'm a little confused on what to measure, the sustained "stable" luminance or should I only go for "peaks".
The 55Q8DN starts bright and then gets dimmer on each pattern. Usually was most bright at the start of each pattern when the text was visible and then when the text went away it dimmed noticeably.
I did measurements on the sustained levels but that meant all measurements are way to low on luminance across the board really. ~50% from "reference" in general.
The "peak brightness" levels seem to be closer to reference but not really hitting it either.
For example 50% Greyscale has a sustained level ~50nits but might peak shortly for 70nits at the start when the text is there for the target RGB value/data.

The patterns was supposed to be 1000nits with MaxFALL400 but I could only get 650-700nits peak right now with current settings at best.
Hmm, rtings.com measured ~1200 nits peak on a 55Q8FN, I would assume the Q8DN would be comparable. HCFR charts are foreign to me, and I don't see the actual measured luminance values in your screenshot, but the delta Luminance chart doesn't look good. For HDR I'd go for measuring just after the pattern starts displaying. I don't know if local dimming of the text will influence your results unduly. You could try measuring the darkest grayscale patches both with the text visible and removed, to see if there's a difference. If the text is causing problems, a small $ alternative might be the Diversified Video Solutions patterns, which have less text in a different position; it also has two "Samsung 2018 code values" grayscale sets for HCFR, but I don't see a description of exactly what that means. Another small $ alternative might be the Spears & Munsil UHD HDR Benchmark disk, which has 10% windows with no text.
 

·
Registered
Joined
·
460 Posts
Hmm, rtings.com measured ~1200 nits peak on a 55Q8FN, I would assume the Q8DN would be comparable. HCFR charts are foreign to me, and I don't see the actual measured luminance values in your screenshot, but the delta Luminance chart doesn't look good. For HDR I'd go for measuring just after the pattern starts displaying. I don't know if local dimming of the text will influence your results unduly. You could try measuring the darkest grayscale patches both with the text visible and removed, to see if there's a difference. If the text is causing problems, a small $ alternative might be the Diversified Video Solutions patterns, which have less text in a different position; it also has two "Samsung 2018 code values" grayscale sets for HCFR, but I don't see a description of exactly what that means. Another small $ alternative might be the Spears & Munsil UHD HDR Benchmark disk, which has 10% windows with no text.
I'm aware and am baffled myself by the lack of peak luminance.
I tried to correct the error and got quite the way using the 10% window. 0-65% was tracking within ~DeltaE 2~
But it wasn't right as far as I could tell. Needed wonky settings like maximum 50/50/50 RGB for 20-point for 5-70% values with a +2Gamma +1 Brightness to come even close.
Peak was still ~650nits overall but not sustainable for the 10% Window pattern.

The pattern isn't just working out for the FALD behaviour I guess.

I tried out some other verification patterns and they didn't look right after all those increases. Did some measurement samples which mentioned target luminance for greys and it was too much.

The Q8DN just gets too dim using the 10% window pattern. Only peak values might work but those are hard to get a correct measurement on, they fluctuate too much to be reliable to get a proper measurement from at the current setup I was using. (they are all over the place for greyscale)
 

·
Registered
Joined
·
47 Posts
I'm aware and am baffled myself by the lack of peak luminance.
I tried to correct the error and got quite the way using the 10% window. 0-65% was tracking within ~DeltaE 2~
But it wasn't right as far as I could tell. Needed wonky settings like maximum 50/50/50 RGB for 20-point for 5-70% values with a +2Gamma +1 Brightness to come even close.
Peak was still ~650nits overall but not sustainable for the 10% Window pattern.

The pattern isn't just working out for the FALD behaviour I guess.

I tried out some other verification patterns and they didn't look right after all those increases. Did some measurement samples which mentioned target luminance for greys and it was too much.

The Q8DN just gets too dim using the 10% window pattern. Only peak values might work but those are hard to get a correct measurement on, they fluctuate too much to be reliable to get a proper measurement from at the current setup I was using. (they are all over the place for greyscale)
Which display type you picking in HCFR?? Try "non-refresh" display type
 

·
Registered
Joined
·
1,426 Posts
The Q8DN just gets too dim using the 10% window pattern. Only peak values might work but those are hard to get a correct measurement on, they fluctuate too much to be reliable to get a proper measurement from at the current setup I was using. (they are all over the place for greyscale)
Does HCFR work with madTPG in HDR mode? Have you tried that?
 

·
Registered
Joined
·
460 Posts
Which display type you picking in HCFR?? Try "non-refresh" display type
I had used various ones as I had not gotten a hint on which would be correct anywhere, I was using BG-R LED with best result but after your suggestion I redid my SDR Calibration using "non-refresh" again it might be better or not. Quite similar to what I got in the end with BG-R LED, the difference isn't something you recall seeing as different.
The White LED & RG-phosphor hadn't produced pleasing results at all.
I'll be using "non-refresh" for now until something better comes about.

Does HCFR work with madTPG in HDR mode? Have you tried that?
I had missed that was possible. Will try when I want to go at it again. Been messing around a little too much to go at it right away with no results to show for it.
 

·
Registered
Joined
·
460 Posts
Does HCFR work with madTPG in HDR mode? Have you tried that?
Why not suggest this in the first place! LOL! Works just fine. So much easier! [HCFR can do it all with madTPG added]

Anyway the result isn't much different from the HDR10 Video patterns. But this is so much faster and convenient!
Sadly the TV seems to be the issue with it's HDR mode not meeting expectations or specifications like this.

I might have noted a issue with the TV producing a 80-920 range for HDR rather than the standard 64-940range. (A 16-(235) Video level compression might be occurring with HDR output)
Still speculation but it's not doing as it should as it is. (internal USB playback has same issue as PC input in general but might be slighly better for HDR)

3040902
 

·
Registered
Joined
·
1,426 Posts
Anyway the result isn't much different from the HDR10 Video patterns.
Thanks for showing the Luminance graph. The shape is basically the same as what HDTVTest measured for a 55Q8DN, just off by 2x in actual values. To rule out a problem with your meter, have you tried doing an ambient light measurement with the diffuser on, near a very bright light bulb, to see if you can read over 650 nits? Does HCFR have a pattern insertion/black frame insertion feature, that would allow you to configure a black pattern to be display for several seconds before the intended pattern is displayed and measured? If so, have you tried using that? Do the luminance measurements change if you vary the Local Dimming setting? It might also be useful to try measuring 100% luminance white in other window sizes (1%, 2%, 5%, 25%) for each of the Local Dimming settings, to see what variation there is. Another stab in the dark, has the Service Menu ever been tinkered with?
I might have noted a issue with the TV producing a 80-920 range for HDR rather than the standard 64-940range. (A 16-(235) Video level compression might be occurring with HDR output)
Still speculation but it's not doing as it should as it is. (internal USB playback has same issue as PC input in general but might be slighly better for HDR)
I'm sure Vincent would have caught a problem like that in his review. Maybe your set is faulty? In the Mehanik patterns, do the Test Footage images and the various ramps in Colors->Color Ramps->MaxCLL 1000nits look reasonable from USB?
 

·
Registered
Joined
·
460 Posts
Thanks for showing the Luminance graph. The shape is basically the same as what HDTVTest measured for a 55Q8DN, just off by 2x in actual values. To rule out a problem with your meter, have you tried doing an ambient light measurement with the diffuser on, near a very bright light bulb, to see if you can read over 650 nits? Does HCFR have a pattern insertion/black frame insertion feature, that would allow you to configure a black pattern to be display for several seconds before the intended pattern is displayed and measured? If so, have you tried using that? Do the luminance measurements change if you vary the Local Dimming setting? It might also be useful to try measuring 100% luminance white in other window sizes (1%, 2%, 5%, 25%) for each of the Local Dimming settings, to see what variation there is. Another stab in the dark, has the Service Menu ever been tinkered with?

I'm sure Vincent would have caught a problem like that in his review. Maybe your set is faulty? In the Mehanik patterns, do the Test Footage images and the various ramps in Colors->Color Ramps->MaxCLL 1000nits look reasonable from USB?
The Internal Player is basically identical for the Black Level, but you had a little extra visibility. rather than crush @ 80 with defaults PC. 80 is actually just barely visible. Just looks like a little better optimized but it still crushed all black from 72 below as on PC side unless tinkering with the output levels.

I tested the other picture modes and Dynamic produces 975nits but it's unusable with above 10000K color temperature and no options to adjust.

Never done a ambient light measurement. Didn't see the need to do one before.

I found that I can use the 20-point to adjust the peak luminance to a degree. I just went and about to see what would happen if I set 100% with 50/50/50 and it increased peak luminance to ~825nits.
So the 20-point has some use to it.
I allready tried other Window sizes and it didn't improve results in general to use smaller or larger than 10%.
4-5% window only peaked with 550nits for example.

I tried a new 25% Window with the 100% 50/50/50 adjustment and got a best yet 930nits.
:D progress

I do have the new i1Display Pro Plus which is rated 2000nits... But as far as I know HCFR hasn't been updated for some time and was suprised it worked without anything specific needed. Only thing is it detects it as a regular i1Display Pro/Colormunki Display. Could the issue be related? The driver expects a 1000nit meter but it's actually 2000nits capable?
Could the issue be here? All my readings are half of what they should be because the driver thinks it's only 1000nits capable?
But i1Profiler for the i1Display Pro Plus reports the same measurments as HCFR for SDR so I guess it's not and issue.
 

·
Registered
Joined
·
1,426 Posts
I tested the other picture modes and Dynamic produces 975nits
OK, so no need to do an ambient light test. Another test you might try is to see what you measure for 100% white SDR with a 10% window, with Local Dimming on High and Backlight at max; rtings.com measured that at bit higher than HDR.

One other random thought, does HCFR have any kind of "Screen Offset" setting, or Luminance multiplier setting, that might have been accidentally configured to be a ~.5 multiplier?

I do have the new i1Display Pro Plus which is rated 2000nits... But as far as I know HCFR hasn't been updated for some time and was suprised it worked without anything specific needed. Only thing is it detects it as a regular i1Display Pro/Colormunki Display. Could the issue be related?
Best to ask that question in the main HCFR thread.
 

·
Registered
Joined
·
47 Posts
Seems very suspicious to have to make that extreme adjustment. One other thought, when all else fails, have you tried a factory reset?
Thoth with your Samsung is your HDMI black Level set to low for HDR ??
Do you think increase the dynamic iris latency time to 500ms would help with the readings??
 

·
Registered
Joined
·
1,426 Posts
Thoth with your Samsung is your HDMI black Level set to low for HDR ??
Do you think increase the dynamic iris latency time to 500ms would help with the readings??
On a Samsung, the HDMI Black Level setting only matters if the source is RGB, and the only RGB-based calibration I do any more is 8-bit RGB Full SDR PC-mode Game mode for PC gaming; for HDR non-PC-mode Game mode and for both SDR and HDR Movie mode, I use YCbCr. On a Samsung, HDMI Black Level of Low is correct for RGB Limited, and for RGB Full that's really tunneling RGB Limited, and Normal is correct for true RGB Full.

I'm guessing dynamic iris latency is an HCFR setting, not projector dynamic iris? Sorry, not familiar with HCFR. If it's a delay between pattern generation and measurement, I'd say best to experiment to see what gives you repeatable readings while still measuring peak rather than sustained brightness, if they are significantly different for your display.
 

·
Registered
Joined
·
3,226 Posts
I did a calibration using the Spears and Munsil UHD HDR Benchmark disc, and the results are very revealing and suggested that most are watching a not just inaccurate picture but one that is really off the mark in gamut mapping, tone mapping, linear tracking of EOTF PQ curve on the disc side.
 

·
Registered
Joined
·
61 Posts
Im having some big issues with calibrating HDR my tv is an oled b8 im using colormunki display device with hcf software, my issue is i can calibrate the 2pt white balance high 100% and 30% low no problem but when it comes to doing the full greyscale calibration the HDR pattern steps dont match up with my TV's steps ive tried numorous amount of different HDR patterns and no luck also when i do a full sweep in hcfr my luminance levels are much higher than the target luminance levels not sure what is happening ive tried mehanik HDR patterns for my oled b8 but when i adjust the rgb in my tv 21pt it does not corrispond with the patterns codes any help thanks
 

·
Registered
Joined
·
5,325 Posts
You really need to do more reading. This has been discussed in a thread here within the past few days.

Stick with the 2-point in HDR. You have to know the exact points used AND have a pattern generator that can generate them on the fly to do the multipoint. Since the points used change from year to year and model to model, and apparently also change if you move away from the stock contrast and brightness settings, it's nearly impossible for disks and pattern sets to have all the patterns needed for each model..

By the way, please do not cross-post the same question in multiple threads. It's considered poor forum etiquette.
 
861 - 880 of 881 Posts
Top