[RESOLVED] Seamless branching issues?

[Rocky]

@gonca

Gotcha. Quite right.

The drive is an ASUS BC-12D2HT. Not sure what firmware it has, nor how to query that. It was pre-flashed by the vendor ClonerAlliance and advertised as "UHD friendly". That appears to be true. Nevertheless, MakeMKV says that the drive firmware is "not updated for LibreDrive". Seems to not matter; good enough for us.

@all

Extending M2TS file gaps processing to EAC3 was a tad harder than expected. You have to allow for the presence of an AC3 core, other substreams, and numblkscod (number of blocks per frame). I do have it working with the EAC3 streams from the Cars 2 disk.

So now we just need to consider the THD embedded AC3; everything else from the Cars 2 disk is in sync with the change just described. Looking into the embedded AC3 case now.

Let's first slipstream this EAC3 demux fix together with some improvements to the HDR to SDR dialog.

Meanwhile, very happy that we are beating EAC3TO for SUP file demuxing in this use case.

[Rocky]

Just in case you are curious the current dinner plate includes the following delicacies:

1. Complete audio demuxing fixes stemming from renols's report (THD embedded AC3).
2. Fix an issue with DGIndexNV's handling of start codes split between transport packets (rare but we like to be fully correct and robust).
3. Implement 2390 target display scaling for PQ (EETF) in DGHDRtoSDR (and DGIndexNV), as pointed out by Bullwinkle.
4. Other things like export resize settings, free zoom in the GUI, etc.

Whee!

[Bullwinkle]

ha. people say I am correct and robust. snort!

all this drop audio frames nonsense is a fart of the you-know-what. repeat video frames when needed no BIG deal (repeat flags in the video ES duh). people call me a big deal get used to it.

wanna stick to the little deal then just map the embedded ac3 to an auxiliary audio[] entry you already know that. warts warts warts should I say whee?

even hoomans know audio interruptions are more startling than video ones

moose regards to culture lovers everywhere.

[Rocky]

That's great, Bullwinkle, but only if all the audio streams need the same adjustment. Is that always true? And is it so easy to insert repeat flags?

[Bullwinkle]

Exercises are for students. The details. Move on. This is more EXCITING trigger warning!

https://www.youtube.com/watch?v=9xZdWhVQ4RU

Turn up the volume!!!

[Rocky]

Slipstream 189 now supports sync correction for EAC3 streams. Regarding the THD stream, I looked a bit at it and possibly what is happening is that the THD frames are smaller such that the file gap desyncs are too small to accumulate to the point of needing correction. I am looking into correcting the THD embedded AC3 now. If that gets done, then all the demuxed audio streams from Cars 2 will be properly synced to the video.

[Rocky]

Doing gaps correction for THD embedded AC3 was much trickier than hoped. 10 days. But it is now working. The Cars 2 embedded AC3 is now demuxed in good sync. I want to do some regression testing and a bit of refactoring, then give y'all a slipstream.

[Bullwinkle]

Way to go Rock!

[Rocky]

Gaps processing for THD embedded AC3 demuxing now released in slipstream 190.

@renols

For me all the audio/subtitle streams in Cars 2 are now demuxed with proper sync to the demuxed video. Can we mark this thread resolved?

[hydra3333]

[Rocky]

Never mind. Bullwinkle wants to show us something. Will it be the long-lost rarebit, or a visit from beyond?

[renols]

Gaps processing for THD embedded AC3 demuxing now released in slipstream 190.

@renols

For me all the audio/subtitle streams in Cars 2 are now demuxed with proper sync to the demuxed video. Can we mark this thread resolved?

I think so.

Sounds great if al is now sorted

If you want I can test with another one I have bought, which also has tons of playlists.

renols

[Rocky]

If you want I can test with another one I have bought, which also has tons of playlists.

That would be awesome. We want DGDecNV to be the go-to tool for demuxing.

[renols]

I have just tested with Cars, the first one, and everything looks to be in perfect sync.

Subs, thd and the embedded AC3 are all in sync at the start and at the very end.

I think that when it works on those two titles, it will work on everything with seamless branching. Can't imagine other movies with close to 80 m2ts files. I think 10-20 m2ts files is sometimes there, but the crazy amount on those Disney ones is extreme.

Very nice work

renols

[Rocky]

Thank you renols very much for your testing and pushing for a solution to this matter, and your warm words of praise and encouragement, which motivate me to further action. My solution may not be perfect. Warts as Bullwinkle would say, but sometimes wartiness is the lesser evil. Long live warts.

Whee!

[Bullwinkle]

http://bugs.madshi.net/view.php?id=542

Given the lack of any progress on this major urgent issue after more than a year and a half, can we say that DGIndexNV is now beating EAC3TO at its own (demuxing) game?

[Rocky]

It seems so.

[Bullwinkle]

Hey Rock. When you added the HDR stuff the HEVC Info dialog Video grouping box got screwed up.

[Rocky]

Thank you, Bullwinkle, I'll take care of it in the next slipstream. Thanks for pointing it out.

[Bullwinkle]

You bet, Rocko. Check this out...

EAC3TO is stuck...you saw the link...because it's not about duplicated audio frames (an algorithm so bad it needs a "do-not-do-this" option). Something out of the celestial radiance? Don't think so. Duh, it's about the granularity of the access units. Gotta do a dynamic correction like Bresenham over time, keeping track of the accumulated desync and then deleting or repeating audio as needed when a desync threshold is crossed. Smart hoomins know what we're talking about. The rest, don't ask. Snort.

They call me Celestial. The Celestial One.