Odd issue with ripped UHD and seamless branching

I’ve ripped my True Romance UHD to my server for playback on my numerous clients and have a weird issue that only seems to affect my coreelec boxes (I have an N2 and an N2+).

On the disc it has 2 versions of the films, and uses seamless branching to allow a player to choose which version to watch. I’ve ripped the director’s cut using makemkv and it plays back fine on all of my Windows and Mac clients with standard kodi 19.1.

For some reason, and this is where it gets very weird, when I play on both coreelec devices it pauses the video at the point where it would’ve then used the branching to play the alternative version of the film. I’ve also confirmed this with another rip of Wrath of Khan that also uses branching on the disc. The film carries on playing as the counter moves and the audio continues but the picture remains paused until I manually skip forward of back.

No anomalies are reported in the coreelec log itself so I’m scratching my head at this weird problem.

There is a Kodi patch for OSMC that is supposed to fix a problem with seamless branching:

From 1f5e34ed5809754b0ec5ed135a487e44c5b68c4e Mon Sep 17 00:00:00 2001
From: tanio99 <tanio99@wolke7.net>
Date: Sun, 1 Sep 2019 09:56:32 +0200
Subject: [PATCH 5/7] avoid flooding the AMLCodec with empty packets (solves
 the BD seamless branching problem)

---
 xbmc/cores/VideoPlayer/DVDCodecs/Video/DVDVideoCodecAmlogic.cpp | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/xbmc/cores/VideoPlayer/DVDCodecs/Video/DVDVideoCodecAmlogic.cpp b/xbmc/cores/VideoPlayer/DVDCodecs/Video/DVDVideoCodecAmlogic.cpp
index 14e11b3..3eac6f9 100644
--- a/xbmc/cores/VideoPlayer/DVDCodecs/Video/DVDVideoCodecAmlogic.cpp
+++ b/xbmc/cores/VideoPlayer/DVDCodecs/Video/DVDVideoCodecAmlogic.cpp
@@ -317,6 +317,9 @@ bool CDVDVideoCodecAmlogic::AddData(const DemuxPacket &packet)
   uint8_t *pData(packet.pData);
   int iSize(packet.iSize);
 
+  if (iSize == 0) // we ignore empty packets
+    return true;
+
   if (pData)
   {
     if (m_bitstream)

It’s weird though as surely the MKV is just 1 file with no seamless branching in it?

@smp should we pick this commit?
Do you have a source of the commit? I do not find the patch anymore at (O/C)SMC

It was there a while ago but I can’t find it now.
I used it in my own CE builds, it doesn’t seem to break anything. So yeah, I think you should consider picking it.

1 Like

This topic was automatically closed 91 days after the last reply. New replies are no longer allowed.