Seamless branching on DTS-HD MA

Post Reply
DAE avatar
BDgeek2
Posts: 18
Joined: Tue Jul 07, 2020 10:48 pm

Seamless branching on DTS-HD MA

Post by BDgeek2 »

Hi there!

First of all, congratulations Rocky! As a long time user and great fan of this incredible program, it's getting better an better!

I know seamless branching is a PIA, and I'm sorry if this has been covered before, but I didn't find anything pertaining a DTS-HD MA audio track. I'm also sorry I'm not knowledgeable on technical aspecs, so please pardon, but I'll try my best to explain.

I've done some testing with Ghost Rider: Spirit of Vengeance USA BD. It's a branched disc that uses DTS-HD MA for the main audio track.

Muxing with the latest Makemkv (1.16.5) and than demuxing with eac3to produced the exact same video stream as direct BD demux with DGdemux, but the DTS-HD MA file of DGDemux was slightly smaller than that resulting of Makemkv mux.

I decoded both files on w64 and compared them on audacity. The conclusion was that the time stamps start out the same, but the end the DGDemux on is -10ms ahead of that from Makemkv.

I know it's a negligible difference, but I just like to understand what's been done differently between the two softwares and which resulting file should be most accurate?

Thanks a lot!
User avatar
Rocky
Posts: 3556
Joined: Fri Sep 06, 2019 12:57 pm

Seamless branching on DTS-HD MA

Post by Rocky »

Thank you BDgeek2.

All I can tell you is the apps are using different algorithms for trying to retain audio/video sync in the presence of seamless branching. I don't know what eac3to and/or makemkv are doing internally (they are closed source like DGDemux). 10ms is indeed inconsequential. If you can demonstrate an issue with DGDemux, I'll be happy to investigate, but "it's different from eac3to" is not an issue for me. A significant desync for files demuxed by DGDemux is an issue.

So I am not trying to evade your question. I simply don't know what these 3rd-party apps are doing. I cannot even say with certainty that eac3to actually performs gaps correction for DTS-HD MA.
DAE avatar
BDgeek2
Posts: 18
Joined: Tue Jul 07, 2020 10:48 pm

Seamless branching on DTS-HD MA

Post by BDgeek2 »

Thanks a lot Rocky! I really appreciate your attention!
User avatar
Rocky
Posts: 3556
Joined: Fri Sep 06, 2019 12:57 pm

Seamless branching on DTS-HD MA

Post by Rocky »

Any time, friend!

BTW, I do know that there are problems with eac3to and seamless branching. For example, this issue has been open and unaddressed for a long time. DGDemux handles such streams without issues.

http://bugs.madshi.net/view.php?id=542
DAE avatar
BDgeek2
Posts: 18
Joined: Tue Jul 07, 2020 10:48 pm

Seamless branching on DTS-HD MA

Post by BDgeek2 »

Thanks a lot Rocky!

Out of curiosity, I just demuxed one of the most branched discs I have: John Wick US Blu-ray (correct playlist 0589 or 0019). I used 3 separate tools.

It's a DTHD Atmos disc not a DTS one. But just for the sake of it:
I converted the tracks to w64 and analysed them in Audacity. They all start with the exact same timecodes, but the last audio wave takes place at:

DGDemux 1h40m58.203ms 0ms
MakeMKV 1h40m58.211ms +8ms
TSmuxer 1h40m58.219ms +16ms

Curious that each subsequent tool ads 8ms in regars to the previous one. I think up until a 20ms difference there probably won't be a perceived lip sync issue, but it's interesting to note this difference in methods nonethelss.
User avatar
Rocky
Posts: 3556
Joined: Fri Sep 06, 2019 12:57 pm

Seamless branching on DTS-HD MA

Post by Rocky »

It is indeed interesting. Thank you.
DAE avatar
BDgeek2
Posts: 18
Joined: Tue Jul 07, 2020 10:48 pm

Seamless branching on DTS-HD MA

Post by BDgeek2 »

You're welcome! ;)
Post Reply