Dolby Vision now possible through MP4 Mux.

[SamuriHL]

None of it matters anyway. It won't play. I either get playback failed or it plays 2 seconds and hangs. That's with the FEL track. The MEL track as suspected doesn't trigger DV at all and plays in HDR. So much for that fun theory.

[shawnc22]

Ah that's a bummer. Maybe these tracks are just a bit atypical compared to regular movies.

[SamuriHL]

Are you suggesting that 10000 nit Dolby Vision files are somehow non-standard?? Yea, i know. it was worth a try though.

[SamuriHL]

WOA! Ok then. I just fast forwarded a little bit in the video and it plays. Alright then. Maybe we can test this out.

[RESET_9999]

I think something else is wrong with either the single layer el+rpu file or the shield because on the Doctor sleep scene, yes the vertex does reduce the red push but there is still a red tint on the little girl face while on the x700 with a dual-layer profile 7 file, it's totally clean. I would presume it's because of the file because i dont see this problem on streaming apps.


EDIT:

It's not the file.. Just tested it on the x700 and it's identical to the dual layer file. No red artifacts in the face.. Did this issue happen with other movies as well? i dont think its the same issue as the red push, looks more like artifacts.
https://screenshotcomparison.com/comparison/4913

[shawnc22]

I think something else is wrong with either the single layer el+rpu file or the shield because on the Doctor sleep scene, yes the vertex does reduce the red push but there is still a red tint on the little girl face while on the x700 with a dual-layer profile 7 file, it's totally clean. I would presume it's because of the file because i dont see this problem on streaming apps.


EDIT:

It's not the file.. Just tested it on the x700 and it's identical to the dual layer file. No red artifacts in the face.. Did this issue happen with other movies as well? i dont think its the same issue as the red push, looks more like artifacts.
https://screenshotcomparison.com/comparison/4913

Oh I had no idea the x700 supported single track dual layer files. Maybe this is why the nvidia devs stated that getting rid of the BT2020 signal didn't entirely solve the issue and that they're still actively working on it.

[RESET_9999]

Oh I had no idea the x700 supported single track dual layer files.

yes but only the MEL movies. In .TS/mp4 container only of course.

I dont use my shield to watch DV rip because i have a x700 so, did anyone ever noticed something very obvious like that Doctor Sleep scene where you can tell for sure that something is wrong?

The saving private ryan flickering issue is also another big concern to me. too bad it's FEL, I can't test the single layer el+rpu on the x700. But I don't recall that I noticed flickering when I watched the dual layer TS.

[SamuriHL]

I need to dust off the UB820 (literally) and do some real world comparisons because I'm noticing some red tinting in places where it shouldn't be. Sometimes it's almost pink. Having the ability to directly compare the discs on the UB820 with the Shield will be a godsend. I can pause scenes that I am concerned with and directly compare them by switching inputs on the receiver which happens fast. it might be this weekend before I have time to really sit down and do an in depth comparison but I have at least one scene in 2001 that I need to compare asap. Something is definitely wrong. But it only seems to be with file playback. i'm not seeing the same issues with the apps (vudu, netflix etc).

[SamuriHL]

Since I had the Spears and Munsil disc out and put the FEL MKV rip out for the shield, I decided to run a few scene comparisons. Only a few just for spot checking, nothing in depth.

First some observations about my setup. I'm running a C8 that has the service menu white point calibrated to D65. I did NOT calibrate the DV modes, however, they ARE using the SM white point adjustment. Given how accurate the C8 is out of the box, this is about as close to reference as you're going to get with these displays, especially when driving content that is 10k nits. Which brings me to my next observation. At 10k nits, we're obviously well into the aggressively tone mapped curve of the C8. Given that I push about 800 nits max out of that thing, obviously it's HIGHLY tone mapped. I use this scene as a torture test for madvr dynamic tone mapping for HDR. I've not really messed with the Dolby Vision version of it other than to do a quick comparison with HDR. On the shield, I did NOT apply the HD Fury change. This is important to note as you'll soon see.

I was looking for specific artifacts here. The red tinting I was talking about tends to happen with partially white backgrounds. Therefore, I limited my testing to 3 scenes and compared the UB820 to the shield. The shield was set to 2160 res, 23.976Hz refresh rate, 12 bit YCbCr output. The UB820 is set similarly. With that out of the way....

I don't have specific time codes for everything. I know 2 of them off the top of my head. 32 seconds and 45 seconds. The 3 scenes are the fence with snow, the horses in the snow, and the deer.

The fence with snow, I usually look at the detail in the snow itself and along the fence posts. I detected no artifacts in this scene in a direct comparison with the UB820. Color representation was accurate. No red push or tint of any kind was detected.

The horse scene is notoriously nasty on tone mapping algorithms. It was designed that way. I focus on the snow in front of the horses to see the detail within the snow. The horses themselves should be quite details even with all the snow noise around them. And the background should not be washed out, even though it is blurry due to the snow. I detected no artifacts in this scene. Color representation was accurate. No red push or tint of any kind was detected.

The deer is a fun one. I like looking at his nose cause it torture tests your black levels. And his antlers have quite a bit of detail. The detail in his fur around his face is also nice to check out. I'm sure you've detected a theme here. I detected no artifacts in this scene. Color representation was accurate. No red push or tine of any kind was detected.

To be clear, I was not pixel peeping here. But these are scenes I'm excruciatingly familiar with from my madvr tone mapping testing. The FEL MKV gave the exact same picture as my UB820 on these particular scenes. We're going to need to do a way more in depth comparison, obviously, but, the spot check passed. Without any HD Fury fixes. Clearly whatever issue we have is not impacting ALL video.

[shawnc22]

Nice work! This definitely lends more credence to the Nvidia folks' explanation that the BT2020 tag is not solely responsible for the issues and why it's been so hard to narrow down the cause. Hopefully the HD Fury will at least give them a clue of what's going on. Also exciting that you're not seeing any noticeable difference for playback of the FEL tracks! The mystery deepens

Something else I'm also wondering is if forcing LLDV will solve the red push issue completely, or will it still lead to the red tint that @RESET9999 mentioned seeing in the Doctor Sleep scene even after the workaround. Just another wrinkle to the equation.

[SamuriHL]

I agree that far more testing is needed now that we have some initial toys to play with. I didn't get a chance to look at 2001 yet but it's on my list. But so far I'm liking what I see a lot. If we can get this to be consistent across the board and look as good as a stand alone player we're all in heaven. I still would like to know why the mel track wasn't picked up by makemkv. Seems like a possible bug.

Sent from my SM-G975U using Tapatalk

[lexyz]

Haven't check mkv yet, but with old p4 ts I can clearly see red tint in "horses", "deer" and "stopwatch" scenes.
Bravia (player-led DV) + Shield TV (Plex 8.4.0.19231)
AVR reports input signal as BT.2020 12bits Dolby Vision

[shawnc22]

Ah that's the first time I've heard of the issue persisting into LLDV mode. Can you tell if you see the red push in the Doctor Sleep scene here: https://mega.nz/file/YSxhkSLK#fkfy54tYV ... xIOBZLyGR4. It may be hard to tell if you have nothing to compare it to since there is naturally some blushing in the little girl's face, but the Shield just exaggerates it.

Also if you have time, can you look at this scene from Saving Private Ryan and see if you notice the flickering that's happening throughout: https://mega.nz/file/8XZljQgZ#P2xzsRWST ... 5fCW9zZa_8 ?

[lexyz]

shawnc22
Red push is present
SPR flickers

The MEL track as suspected doesn't trigger DV at all and plays in HDR.

Can confirm. MEL DV (20001.ts) -> 1.15.2 mkv plays in HDR
Still plays in DV with old p4 *ts conversion

[lexyz]

Haven't check mkv yet, but with old p4 ts I can clearly see red tint in "horses", "deer" and "stopwatch" scenes.
Bravia (player-led DV) + Shield TV (Plex 8.4.0.19231)
AVR reports input signal as BT.2020 12bits Dolby Vision

Same with mkv and Exoplayer (23/07/2020)