Dolby Vision now possible through MP4 Mux.

[Gatorman]

This may have been answered somewhere within this thread but since makemkv is able to rip the dolby vision layer is it possible to create a dolby vision playable mkv file through mkv or do you have use scripts?

This really depends on the player you plan on viewing your media on. Checkout the signature under RESET_9999.

[RESET_9999]

Hi,
I did some more testing with Flac audio. For some movies after converting trueHD/dts-hd to Flac (i think it applies to lpcm too) the resulting audio track is significantly lower volume.

I'll check it out but I'm pretty sure LPCM is fine.

EDIT: did some tests and the original THD/DTS, flac or lpcm all sound at the same level on my PC.
Anyway, I added -drc_scale 0 in the cli, maybe that will help?

[DjObama]

Hi,
I did some more testing with Flac audio. For some movies after converting trueHD/dts-hd to Flac (i think it applies to lpcm too) the resulting audio track is significantly lower volume.

I'll check it out but I'm pretty sure LPCM is fine.

EDIT: did some tests and the original THD/DTS, flac or lpcm all sound at the same level on my PC.
Anyway, I added -drc_scale 0 in the cli, maybe that will help?

Yes, I checked and on PC there is no problem. However on TV there is a significant diffrence. I just checked another movie - Hero (2002), not even a HDR movie, but Flac plays very quiet compared to dts. Can somebody check the flac playback on their tvs/recievers compared to dts-hd tracks? I read about it a bit and seems that dts format boosts the volume while decoding, but it doesn't happen on pc.

[skull88]

Hi,
I did some more testing with Flac audio. For some movies after converting trueHD/dts-hd to Flac (i think it applies to lpcm too) the resulting audio track is significantly lower volume.

I'll check it out but I'm pretty sure LPCM is fine.

EDIT: did some tests and the original THD/DTS, flac or lpcm all sound at the same level on my PC.
Anyway, I added -drc_scale 0 in the cli, maybe that will help?

Yes, I checked and on PC there is no problem. However on TV there is a significant diffrence. I just checked another movie - Hero (2002), not even a HDR movie, but Flac plays very quiet compared to dts. Can somebody check the flac playback on their tvs/recievers compared to dts-hd tracks? I read about it a bit and seems that dts format boosts the volume while decoding, but it doesn't happen on pc.

Sounds like this is not an issue that should really be troubleshooted on this thread, but on a forum or subreddit relating to your TV or AVR? I recommend you try E-AC3 instead of FLAC, as it is 2nd best alternative if you aren't happy with FLAC playback or adjusting volume on the fly when you watch movies.

[ans40]

Quick question; when generating CMv4.0 content (3-1), how does one determine what Mastering Display to use? I assuming higher isn't necessarily better?

2 options:

1- just use the static metadata from your source HDR10
2- check the brightness first with madvr. If 95% of the movie is under 1000nits, use 1000nits MDL else, 4000.

Option 2 is better because some movies have a static MDL of 4000nits but the actual brightness is under 1000nits.
A 4000nits RPU is a lot darker than a 1000nits one regardless of the content's actual brightness and in my opinion, 4000nits MDL + low nits content is wrong without L2 trim passes because if you look at the generated(or original) trims with a 4000nits analysis, they brighten/compensate a lot for the darkness a 4000 rpu brings.

https://slow.pics/c/uVX5iFOD

Using Resolve to generate DV.

BL HDR10: https://ibb.co/fYWDhzp

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

[skull88]

Quick question; when generating CMv4.0 content (3-1), how does one determine what Mastering Display to use? I assuming higher isn't necessarily better?

2 options:

1- just use the static metadata from your source HDR10
2- check the brightness first with madvr. If 95% of the movie is under 1000nits, use 1000nits MDL else, 4000.

Option 2 is better because some movies have a static MDL of 4000nits but the actual brightness is under 1000nits.
A 4000nits RPU is a lot darker than a 1000nits one regardless of the content's actual brightness and in my opinion, 4000nits MDL + low nits content is wrong without L2 trim passes because if you look at the generated(or original) trims with a 4000nits analysis, they brighten/compensate a lot for the darkness a 4000 rpu brings.

https://slow.pics/c/uVX5iFOD

Using Resolve to generate DV.

BL HDR10: https://ibb.co/fYWDhzp

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

This is for Interstellar, and the way a movie is shot, not just plot peaks, should be taken into account when making the choice. A number of poorly lit scenes really demonstrate that 4K may not be ideal for them, but other scenes are much brighter. I marginally prefer a 1K MDL for generation on this one having watched with both 4K and 1K injected RPUs, but I guess you could try 2K as a compromise, it's honestly a toss-up and personal preference on this title in the end. Oh, and if doing yourself, make sure to correctly set L5/blanking on the IMAX shots vs. main timeline blanking.

[RESET_9999]

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

You should update your script, I added more information that makes the decision easier. My initial recommendation was if 95% is under 1000nits , choose 1000nits and for this movie only 2.4% is over 1k. I personally opted for the 1000nits MDL.
But as @skull88 said, we probably should consider the avg_pq too and I'm still thinking about the % (I'm gathering more information). Also, madvr definitely includes the black bars in its measurement so maxfall (without cropping) will always be lower than it is in reality. see: https://slow.pics/c/pV4NXJMB

BTW, here's my Resolve project for this movie, this will save you a lot of time
https://drive.google.com/drive/u/0/fold ... rHG4XNSy8x

[Gatorman]

I've noticed the default x265 HDR encoder settings batch file has a flag to set --max-luma to 1023. Should this be modified if the source file has a MaxCLL of 4000?

[RESET_9999]

I've noticed the default x265 HDR encoder settings batch file has a flag to set --max-luma to 1023. Should this be modified if the source file has a MaxCLL of 4000?

No, the maximum brightness value in 10bits is 1023
in 12bits it's 4095
in 8bits it's 255

[DjObama]

I'll check it out but I'm pretty sure LPCM is fine.

EDIT: did some tests and the original THD/DTS, flac or lpcm all sound at the same level on my PC.
Anyway, I added -drc_scale 0 in the cli, maybe that will help?

Yes, I checked and on PC there is no problem. However on TV there is a significant diffrence. I just checked another movie - Hero (2002), not even a HDR movie, but Flac plays very quiet compared to dts. Can somebody check the flac playback on their tvs/recievers compared to dts-hd tracks? I read about it a bit and seems that dts format boosts the volume while decoding, but it doesn't happen on pc.

Sounds like this is not an issue that should really be troubleshooted on this thread, but on a forum or subreddit relating to your TV or AVR? I recommend you try E-AC3 instead of FLAC, as it is 2nd best alternative if you aren't happy with FLAC playback or adjusting volume on the fly when you watch movies.

Yes, sorry for introducing chaos. The issue seems to be with my tv, I chcecked pcm audio and it was louder than flac (altough weirdly the dialogue was very quiet). I'll stick to eac3 for now.

I just wonder if dovi_scripts could be configured to keep the flac track for 4-3-1 MP4?

[ans40]

2 options:

1- just use the static metadata from your source HDR10
2- check the brightness first with madvr. If 95% of the movie is under 1000nits, use 1000nits MDL else, 4000.

Option 2 is better because some movies have a static MDL of 4000nits but the actual brightness is under 1000nits.
A 4000nits RPU is a lot darker than a 1000nits one regardless of the content's actual brightness and in my opinion, 4000nits MDL + low nits content is wrong without L2 trim passes because if you look at the generated(or original) trims with a 4000nits analysis, they brighten/compensate a lot for the darkness a 4000 rpu brings.

https://slow.pics/c/uVX5iFOD

Using Resolve to generate DV.

BL HDR10: https://ibb.co/fYWDhzp

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

This is for Interstellar, and the way a movie is shot, not just plot peaks, should be taken into account when making the choice. A number of poorly lit scenes really demonstrate that 4K may not be ideal for them, but other scenes are much brighter. I marginally prefer a 1K MDL for generation on this one having watched with both 4K and 1K injected RPUs, but I guess you could try 2K as a compromise, it's honestly a toss-up and personal preference on this title in the end. Oh, and if doing yourself, make sure to correctly set L5/blanking on the IMAX shots vs. main timeline blanking.

It's kind of remarkable you recognized what movie that was lol. Thanks @skull88 and @RESET_9999 -- checked out the 4k and 1k MDL versions this morning and prefer the 1k.

[ans40]

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

You should update your script, I added more information that makes the decision easier. My initial recommendation was if 95% is under 1000nits , choose 1000nits and for this movie only 2.4% is over 1k. I personally opted for the 1000nits MDL.
But as @skull88 said, we probably should consider the avg_pq too and I'm still thinking about the % (I'm gathering more information). Also, madvr definitely includes the black bars in its measurement so maxfall (without cropping) will always be lower than it is in reality. see: https://slow.pics/c/pV4NXJMB

BTW, here's my Resolve project for this movie, this will save you a lot of time
https://drive.google.com/drive/u/0/fold ... rHG4XNSy8x

Oh, hey. Just as an FYI if I'm seeing this right, your template had the wrong blanking for scene #705 (frame#72080). Your settings file saved me a bunch of time, just ran through and did a quick double-check and caught that .

[skull88]

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

You should update your script, I added more information that makes the decision easier. My initial recommendation was if 95% is under 1000nits , choose 1000nits and for this movie only 2.4% is over 1k. I personally opted for the 1000nits MDL.
But as @skull88 said, we probably should consider the avg_pq too and I'm still thinking about the % (I'm gathering more information). Also, madvr definitely includes the black bars in its measurement so maxfall (without cropping) will always be lower than it is in reality. see: https://slow.pics/c/pV4NXJMB

BTW, here's my Resolve project for this movie, this will save you a lot of time
https://drive.google.com/drive/u/0/fold ... rHG4XNSy8x

Oh, hey. Just as an FYI if I'm seeing this right, your template had the wrong blanking for scene #705 (frame#72080). Your settings file saved me a bunch of time, just ran through and did a quick double-check and caught that .

Good catch, it does need blanking adjustment back to letterboxed for that short scene.

[GyulaZ]

Hi Guys,

I would like to request some help in this never ending story I jumped into.
Long story short, I have an LG C7, and I've been trying to somehow trigger Dolby Vision playback with the appropriate encoding for the past few weeks.
My problem is that I can't tell what is considered appropriate encoding for my given TV in terms of DV playback, and the internet is full of confusing info.

As it seems, this forum could have the solution for my problem, but I'm unable to find the proper answer for the LG C7 directly, as the search engine ignores short words.

Mainly the movies I have access to, uses the following format:

Video codec: HEVC or x265
Dolby Vision, Version 1.0, dvhe.08.06, BL+RPU
SMPTE ST 2086, HDR10 compatible
Single track single layer

Audio is mostly Dolby Digital, in some cases TrueHD Atmos.

I'm using either the built in player via USB from an external drive, or streaming the content through Plex, the issue is the same in both cases.
I already know that DV playback is not triggered from a mkv container, therefore in this case the movie is played back using the HDR layer only.
Also, when I mux it to a mp4 container DV is recognized, but the playback is not starting at all, only circles going round and round after some error message is shown.

I have tools like ffmpeg, dovi_tool, dvmkv2tomp4, but none of them proved to solve my problem.
Please keep in mind that despite I consider myself to be an IT guy, but this topic is pretty new and confusing to me currently.

I would appreciate is someone could point out that what kind of output or encoding I need exactly to meet the requirements of proper Dolby Vision playback on an LG C7.

Thanks for the help in advance!

[valington]

Thoughts on using a 2000nit MDL? It has a decent amount of scenes over 1000 nits. 4000 seems excessive though, so I'm between 1000 and 2000nit MDL for Resolve DV analysis. TIA

You should update your script, I added more information that makes the decision easier. My initial recommendation was if 95% is under 1000nits , choose 1000nits and for this movie only 2.4% is over 1k. I personally opted for the 1000nits MDL.
But as @skull88 said, we probably should consider the avg_pq too and I'm still thinking about the % (I'm gathering more information). Also, madvr definitely includes the black bars in its measurement so maxfall (without cropping) will always be lower than it is in reality. see: https://slow.pics/c/pV4NXJMB

BTW, here's my Resolve project for this movie, this will save you a lot of time
https://drive.google.com/drive/u/0/fold ... rHG4XNSy8x

Hallo RESET please give me you latest tool s pack