Dolby Vision now possible through MP4 Mux.

[RESET_9999]

So does the Shield only have the bug with the BL using mastering display DCI-P3 for HDR10?
If so there's really nothing we can do. Most titles are mastered in DCI-P3.

Thankfully it's the other way around, MDP P3 BL is fine.
I will do more tests tonight but regardless of the rpu, it looks like as soon as the HDR10 BL was mastered on a bt2020 monitor, the bug happens.
I'll check some bt2020 DV movies too. Such as the Matrix.

The original pattern you shared as "multi pattern DV P8.ts" is BT.2020, though. It didn't have L9 metadata.

yep, I did the first version with CMV2.9. So since that one doesn't have an MDP mismatch, it could confirms my suspicion that the shield has problems with movies mastered on a BT2020 display?
I redid the pattern because I based the dv analysis on maxcll 1000, when the clip actually is 10 000nits. So L1/L2 were incorrect. I guess I went too fast for the color settings in CMV4.0 for the second version

Code: Select all

{
    "Level9": {
        "length": 1,
        "source_primary_index": 2
    }
}

thank you, but with the madvr generated DV from a BT2020 MDP video, L9 defaults to P3 ?
could that be automated?

Code: Select all

        {
          "Level9": {
            "length": 1,
            "source_primary_index": 0
          }

[quietvoid]

thank you, but with the madvr generated DV from a BT2020 MDP video, L9 defaults to P3 ?
could that be automated?

You can use a JSON config along with a madVR measurement file.

[deadchip12]

Could someone briefly explain what were being discussed in the last few pages? The x700/x800m2 bluray player seem like the ultimate devices for playing back dolby vision files, so what else need to be discussed? I'm just curious, not complaining or anything. Are people trying to make it work on other devices?

[RESET_9999]

.

Ok, I did more testing. It turns out the problem is the x265.exe settings used for the BL regardless of the RPU

• The original pattern from the spears and munsil disc = bug on the shield

• Re-encoded with x265.exe in multiples ways: in BT2020 or P3, chroma type 2 or not, default or custom slow preset settings, --uhd-bd or not = bug on the shield

• Re-encoded in resolve (just regular HDR10, dont even need to be 8.1) with any RPU = no bug

Very weird, I give up...

Could someone briefly explain what were being discussed in the last few pages? The x700/x800m2 bluray player seem like the ultimate devices for playing back dolby vision files, so what else need to be discussed? I'm just curious, not complaining or anything. Are people trying to make it work on other devices?

Just trying to figure out the shield colors issues.

[quietvoid]

Ok, I did more testing. It turns out the problem is the x265.exe settings used for the BL regardless of the RPU

• The original pattern from the spears and munsil disc = bug on the shield

• Re-encoded with x265.exe in multiples ways: in BT2020 or P3, chroma type 2 or not, default or custom slow preset settings, --uhd-bd or not = bug on the shield

• Re-encoded in resolve (just regular HDR10, dont even need to be 8.1) with any RPU = no bug

That doesn't make much sense.

[RESET_9999]

That doesn't make much sense.

but it's exactly what's happening. Feel free to provide your version of the pattern and I'll try.

[quietvoid]

That doesn't make much sense.

but it's exactly what's happening. Feel free to provide your version of the pattern and I'll try.

You can try this: https://0x0.st/oBi5.mkv

However the file I used (multi pattern DV P8.ts) is not the same as the Resolve encode. It has more artifacts.
So either Resolve processed the image or something else.

[RESET_9999]

However the file I used (multi pattern DV P8.ts) is not the same as the Resolve encode. It has more artifacts.
So either Resolve processed the image or something else.

thank you
Just tried and the same bug as all the files I did

[rkjnsn]

Okay, another couple of questions. First, I've seen reference to converting P5 to P8 to create "hybrid" streams using the HDR10 BL from the UHD blu-ray combined with the Dolby Vision data available via streaming. Is this ~always the best approach where possible, or are there cases where the blu-ray has scenes encoded at different brightnesses than the streaming base layer, making a madVR measurement the better choice?

Second, if I have a UHD blu-ray of Mulan (2020) along with a Disney+ subscription, is there any way I can capture the RPU metadata myself? E.g., can it be extracted from the streamed data, or is it DRM'd along with the video stream? Is there any tool that can capture the RPU by snooping on a TV-led DV HDMI signal? If not, is there a repository of DV RPU metadata extracted from such sources, or would I have to find a full copy of the streaming version somewhere and extract the RPU from that?

[quietvoid]

thank you
Just tried and the same bug as all the files I did

Okay, last attempt: https://0x0.st/oB-i.mkv

This is just the Resolve file reencoded.

[RESET_9999]

Okay, last attempt: https://0x0.st/oB-i.mkv
This is just the Resolve file reencoded.

no bug for that one.
So it means there's an issue in the source that Resolve somehow manages to fix, and it's only affecting the Shield.
So weird...

[quietvoid]

no bug for that one.
So it means there's an issue in the source that Resolve somehow manages to fix, and it's only affecting the Shield.
So weird...

Not necessarily, I had changed an encoding setting.

Same settings with the non-Resolve video: https://0x0.st/oB-c.mkv
If that still has issues, then not sure.

The only difference in the Resolve image is that I can't seem to match chroma alignment, and it has less banding in the blue part.

[RESET_9999]

Same settings with the non-Resolve video: https://0x0.st/oB-c.mkv
If that still has issues, then not sure.
The only difference in the Resolve image is that I can't seem to match chroma alignment, and it has less banding in the blue part.

bug is back lol
I wonder what magic resolve does.

[quietvoid]

bug is back lol

I guess this is my actual last attempt then: https://0x0.st/oB-B.mkv
It doesn't make sense that the original test pattern causes issues.

[RESET_9999]

same issue...

I tried NVenc, same issue.
I tried FFmpeg to prores to x265, same issue.