Dolby Vision now possible through MP4 Mux.

[RESET_9999]

https://slow.pics/c/eUPDkTHU

Barbie 2023
Why scene cuts different in HDR10+ and DV version?

I'm not sure but...

In Dolby vision, sometimes the flashes (single frame) can have their own scene cut which doesn't seem to cause any problem at playback (I tested that with my own rpu).
In HDR10plus, I read somewhere that the metadata can only change at a real scene cut or otherwise, it can cause issues.

[skull88]

https://slow.pics/c/eUPDkTHU

Barbie 2023
Why scene cuts different in HDR10+ and DV version?

I'm not sure but...

In Dolby vision, sometimes the flashes (single frame) can have their own scene cut which doesn't seem to cause any problem at playback (I tested that with my own rpu).
In HDR10plus, I read somewhere that the metadata can only change at a real scene cut or otherwise, it can cause issues.

More logistically, the HDR10+ could have been provided by different source for the streaming platform or requested for streaming only, as it is likely upcoming UHD BD will only have DV from Warner Bros.

[quietvoid]

It's usually HDR10+ that has more random changes in the metadata, like flashes.
Dolby Vision always relies on the shots and flashes don't cause new metadata because the shots are predetermined when editing/grading.

I don't think HDR10+ actually lets you align everything to shots, the process is too automated.

[RESET_9999]

Well, I have definitely seen flashes as a single shot in retail DV source...
The Bad Guys animation movie is one I remember because it confused me when I did my P8 hybrid remux.

[axeyou]

*** I think I found an issue with the x800m2:

1- DTS-HD 2.0 plays only as lossy DTS (core)

Is this bug still present with the X800M2? I tried playing a DTS-HD MA 2.0 track and my receiver reported DTS-HD MA, not DTS. However, I have no clue if the X800M2 is doing something weird like wrapping the lossy DTS core in a DTS-HD MA container on the fly. How did you find this bug? Is there a test file with a silent core to easily verify what is being played?
Thanks!

it might be a thing with my AVR. The osd reports only DTS while with 5.1/7.1 tracks it reports DTS-HD.
It could also be related to the network playback , I don't think I ever tried via USB.

Interesting. I’m playing through network as well. Both the X800M2 and my AVR report DTS-HD MA on a 2.0 track.

Well, I have definitely seen flashes as a single shot in retail DV source...

Yes. I’ve seen that as well on my own rip of Arrow’s Cinema Paradiso DV UHD disc.

[skull88]

Well, I have definitely seen flashes as a single shot in retail DV source...
The Bad Guys animation movie is one I remember because it confused me when I did my P8 hybrid remux.

Yeah, I've also seen that. I'm guessing (some? many?) DoVi shot lists are at least initially coming from an automatic algo pass (especially since v4.0 from Dolby), as it would save anyone doing the grading a huge amount of time/labour, even if they do a second manual pass after to delete/add further, and they either just leave the very short "flash" shots in or don't even notice them, lol.

[speeddemon]

Well, I have definitely seen flashes as a single shot in retail DV source...
The Bad Guys animation movie is one I remember because it confused me when I did my P8 hybrid remux.

There are even flashes in some HDR10 releases... The 2017 US Lionsgate UHD BD of La La Land (2016) even has flashes in its HDR10 BL. Only the 2022 IT Eagle UHD BD of La La Land (2016) fixes this issue in its Base Layer (it also has P7 FEL), but the Base Layer of that release appears to be extremely problematic and requires the FEL for expanded brightness, and the L2 Trims to correct the saturation.

[RESET_9999]

Interesting. I’m playing through network as well. Both the X800M2 and my AVR report DTS-HD MA on a 2.0 track.

Then it must be my AVR (its quite old) and maybe it was the same on my x700 and I just never noticed.
I know that the ffmpeg DDP 1536kbps track doesn't work on the x800m2 and I'm certain they worked on the x700. The EAE 7.1 DDP tracks play fine.

2.0 DTS-HD:


5.1 DTS-HD:

[Trees]

Does anyone know if it's possible to get 10k nit with CM v4.0 like these?





I'm asking because whatever mastering display I choose with 3-1 (Professional CM v4.0 Dolby Algo) I get lesser results, maximum 1k, 2k or 4k nits. 10k nits doesn't seem available. Is there a reason for this or?

Do the guys above have access to tools we don't or does Dovi_scripts just not allow 10k nits (yet)?

[Mrmorrison]

Claiming that a P7 FEL is the same if you convert it to a P8 is wrong, I understand.

Can you explain to me what exactly is lost from this operation?

[RESET_9999]

Does anyone know if it's possible to get 10k nit with CM v4.0 like these?
I'm asking because whatever mastering display I choose with 3-1 (Professional CM v4.0 Dolby Algo) I get lesser results, maximum 1k, 2k or 4k nits. 10k nits doesn't seem available. Is there a reason for this or?
Do the guys above have access to tools we don't or does Dovi_scripts just not allow 10k nits (yet)?

The metadata cannot be 10k if the content is not 10k. The Algo measures each pixel's brightness and produces metadata accordingly.

[RESET_9999]

Claiming that a P7 FEL is the same if you convert it to a P8 is wrong, I understand.

Can you explain to me what exactly is lost from this operation?

look at the comparisons in the red section that says:

Can only be played on a FEL device and can't be converted to P8 without baking FEL into EL

https://docs.google.com/spreadsheets/d/ ... edit#gid=0

[Trees]

Does anyone know if it's possible to get 10k nit with CM v4.0 like these?
I'm asking because whatever mastering display I choose with 3-1 (Professional CM v4.0 Dolby Algo) I get lesser results, maximum 1k, 2k or 4k nits. 10k nits doesn't seem available. Is there a reason for this or?
Do the guys above have access to tools we don't or does Dovi_scripts just not allow 10k nits (yet)?

The metadata cannot be 10k if the content is not 10k. The Algo measures each pixel's brightness and produces metadata accordingly.

But according to the plot, they are 10k nits:





Even though the original plots are like this:





How are they achieving 10k nit peaks with a CM v4.0 conversion (CM v4.0 according to the plots)? I've tied to replicate it but the plots don't go above 4k nits. Here are the MediaInfo outputs:

Code: Select all

Complete name                            : /media/user/WD 1 TB/Videos/The Grand Tour/Series 5/S05E01 - A Scandi Flick missing chapters 2160p HDR10p Amazon Prime Web-DL.mkv
Format                                   : Matroska
Format version                           : Version 4
File size                                : 10.8 GiB
Duration                                 : 1 h 39 min
Overall bit rate                         : 15.4 Mb/s
Movie name                               : A Scandi Flick
Encoded date                             : UTC 2023-09-07 15:54:25
Writing application                      : mkvmerge v75.0.0 ('Goliath') 64-bit
Writing library                          : libebml v1.4.4 + libmatroska v1.7.1

Video
ID                                       : 1
Format                                   : HEVC
Format/Info                              : High Efficiency Video Coding
Format profile                           : Main 10@L5@Main
HDR format                               : SMPTE ST 2094 App 4, Version 1, HDR10+ Profile B compatible
Codec ID                                 : V_MPEGH/ISO/HEVC
Duration                                 : 1 h 39 min
Bit rate                                 : 14.8 Mb/s
Width                                    : 3 840 pixels
Height                                   : 2 160 pixels
Display aspect ratio                     : 16:9
Frame rate mode                          : Constant
Frame rate                               : 25.000 FPS
Color space                              : YUV
Chroma subsampling                       : 4:2:0
Bit depth                                : 10 bits
Bits/(Pixel*Frame)                       : 0.071
Stream size                              : 10.3 GiB (96%)
Title                                    : 2160p HDR10+ Amazon Prime Web-DL
Language                                 : English
Default                                  : No
Forced                                   : No
Color range                              : Limited
Color primaries                          : BT.2020
Transfer characteristics                 : PQ
Matrix coefficients                      : BT.2020 non-constant
Mastering display color primaries        : Display P3
Mastering display luminance              : min: 0.0001 cd/m2, max: 1000 cd/m2
Maximum Content Light Level              : 999 cd/m2
Maximum Frame-Average Light Level        : 745 cd/m2

Code: Select all

Complete name                            : /media/user/WD 1 TB/Videos/The Grand Tour/Series 5/S05E02 - Eurocrash missing chapters 2160p HDR10p Amazon Prime Web-DL.mkv
Format                                   : Matroska
Format version                           : Version 4
File size                                : 11.5 GiB
Duration                                 : 1 h 46 min
Overall bit rate                         : 15.4 Mb/s
Movie name                               : Eurocrash
Encoded date                             : UTC 2023-09-07 16:01:55
Writing application                      : mkvmerge v75.0.0 ('Goliath') 64-bit
Writing library                          : libebml v1.4.4 + libmatroska v1.7.1

Video
ID                                       : 1
Format                                   : HEVC
Format/Info                              : High Efficiency Video Coding
Format profile                           : Main 10@L5@Main
HDR format                               : SMPTE ST 2094 App 4, Version 1, HDR10+ Profile B compatible
Codec ID                                 : V_MPEGH/ISO/HEVC
Duration                                 : 1 h 46 min
Bit rate                                 : 14.8 Mb/s
Width                                    : 3 840 pixels
Height                                   : 2 160 pixels
Display aspect ratio                     : 16:9
Frame rate mode                          : Constant
Frame rate                               : 25.000 FPS
Color space                              : YUV
Chroma subsampling                       : 4:2:0
Bit depth                                : 10 bits
Bits/(Pixel*Frame)                       : 0.071
Stream size                              : 11.0 GiB (96%)
Title                                    : 2160p HDR10+ Amazon Prime Web-DL
Language                                 : English
Default                                  : No
Forced                                   : No
Color range                              : Limited
Color primaries                          : BT.2020
Transfer characteristics                 : PQ
Matrix coefficients                      : BT.2020 non-constant
Mastering display color primaries        : Display P3
Mastering display luminance              : min: 0.0001 cd/m2, max: 1000 cd/m2
Maximum Content Light Level              : 1000 cd/m2
Maximum Frame-Average Light Level        : 680 cd/m2

What's going on here?

[RESET_9999]

But according to the plot, they are 10k nits:

Well you're comparing madvr vs dolby tunning 1.

You have control of the L1 tunning at line 95 in the bat file, change it to 0. default is 1 and the best tunning.
if you use the old legacy mode (0), this will probably get you that frame at 10knits but you shouldn't do that.

:: choose the L1 analysis tuning... Choices are: (old=0 - Most Highlight Detail/Most Mapping=1(default) - More Highlight Detail/More Mapping=2 - Balanced=3 - Less Highlight Detail/Less Mapping=4 - Least Highlight Detail/Least Mapping=5
set L1_Tuning=1

[Trees]

Well you're comparing madvr vs dolby tunning 1.

You have control of the L1 tunning at line 95 in the bat file, change it to 0. default is 1 and the best tunning.
if you use the old legacy mode (0), this will probably get you that frame at 10knits but you shouldn't do that.

:: choose the L1 analysis tuning... Choices are: (old=0 - Most Highlight Detail/Most Mapping=1(default) - More Highlight Detail/More Mapping=2 - Balanced=3 - Less Highlight Detail/Less Mapping=4 - Least Highlight Detail/Least Mapping=5
set L1_Tuning=1

I don't follow. Are you saying The.Grand.Tour.S05E01/02.2023.2160p.AMZN.WEB-DL.DDP5.1.DoiVi.HBRID.P8 are MadVR generated? How do you know that? The plots say CM v4.0 Dolby Vision. Do you think they chose different L1 analysis tuning to achieve "10k nits"? Is it really 10k nits or is that decided by the pixel brightness relative to other pixel brightnesses? I really don't understand what you mean, I'm not sure how it correlates with my question, to be honest with you. I realise I have alot to learn, but I just don't understand what you mean.