(2013-08-22, 14:18)fritsch Wrote: [ -> ]The commit fixes 5.1 LPCM over HDMI.
I got confused for your lengthy text. Please retry with the latest version, when the rebase has happened. 5.1 Flac will be send as AC3 then - so no resample will work either - as it is a transcode codec. But yeah - report back after the new version with all the fixes is in the ppa.
Ok, thanks for the clarification; I will try the latest Xvba PPA build when that is available. Just to be clear on my side as well, both 5.1 LPCM and 5.1 FLAC is being sent as AC3. This works (perfectly) only if the sync option is set to 'Audio clock'. When 'Video clock (resample audio)' or 'Video clock (drop/dupe audio)' is set you get artifacts (once a second with the first, around 1-2 each minute for the last).
btw.–do you have an idea what the cause for the segfaults might be? I could test this with the official branch, but I assume that the Alpha Gotham builds are still based on SDL code (last I tested, which was pretty recent they still were). If so they are unusable for my dual-monitor setup which is pretty flawless at this point–judder/tearing all gone: as well as sync issues taken care of. The video is silky smooth and even better than what I previously had when running XBMC on OS X.
Concerning the segfaults, try to run it in gdb and see where it crashes. Backtraces are welcome.
Edit: Video clock (drop / dupe audio) is my default setting. So that makes me wonder. Repost that logfile please.
Edit2:
Quote:22:21:50 T:139794643572608 DEBUG: Window Manager Name: Mutter (Muffin)
Reproduce from fluxbox or another non compositing window manager.
This is _not_ related to the window manager. Re-read my post.
You cannot sync to display, if some compositing wm implements uses triple buffering. And in that case it seems, that vaapi gets serious problems (see your log).
(2013-07-18, 14:09)Robotica Wrote: [ -> ]Hi.. I've seen the efforts AMD (and many others) are putting in HSA. It looks really great for htpc usage. Anyone with a professional opinion about this in relation to xbmc?
Some
tech news which shows that XBMC to benefit AMD hardware, xbmc needs to be (heavily?) optimized to support their new hUMA tech.
My issue looks a lot like the one reported in:
http://forum.xbmc.org/showthread.php?tid=165253 and for which the following git pull request is associated:
https://github.com/xbmc/xbmc/pull/2815
The referenced bug relates to an uninitialized member in the VDPAU code, and I am wondering if the same could hold true for VAAPI or if this is just a coincidence.
I performed some extensive testing on this today & yesterday– re-ripped a feature, watched it through to the end (with VAAPI acceleration), no issues. Also looked at nearly every file in my small but growing MKV library, again with VAAPI enabled (starting/stopping ~30 video files without issues until trying the 'problematic one', which consistently produces a segfault). Not always the first time, but always within 4-5 tries. Also re-ripped this disc to see if it would behave any differently, which (of course) it didn't. When disabling VAAPI (which works ok for every other video), this video works fine.
I did run XBMC through gdb, but this will have to be redone as it did not produce much useable output (not following the threads), nor produce a stacktrace which would give any indication as to what is failing. GDB syntax welcome.
Including it here anyway.
Code:
(gdb) run
Starting program: /usr/lib/xbmc/xbmc.bin "-fs"
warning: no loadable sections found in added symbol-file system-supplied DSO at 0x7ffff7ffa000
[Thread debugging using libthread_db enabled]
Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
[New Thread 0x7fffe7128700 (LWP 6969)]
[New Thread 0x7fffe6927700 (LWP 6970)]
[New Thread 0x7fffe58fd700 (LWP 6973)]
[Thread 0x7fffe58fd700 (LWP 6973) exited]
[New Thread 0x7fffe58fd700 (LWP 6975)]
[New Thread 0x7fffd7fff700 (LWP 6976)]
[New Thread 0x7ffff7fac700 (LWP 6979)]
[New Thread 0x7fffd4b47700 (LWP 6980)]
[New Thread 0x7fffc7fff700 (LWP 6981)]
[New Thread 0x7fffc77fe700 (LWP 6982)]
[New Thread 0x7fffc6ffd700 (LWP 6983)]
[New Thread 0x7fffc67fc700 (LWP 6984)]
[New Thread 0x7fffc5ffb700 (LWP 6985)]
[New Thread 0x7fffc57fa700 (LWP 6986)]
[New Thread 0x7fffaffff700 (LWP 6987)]
[New Thread 0x7fffaf7fe700 (LWP 6988)]
[New Thread 0x7fffaeffd700 (LWP 6989)]
[New Thread 0x7fffae7fc700 (LWP 6990)]
[New Thread 0x7fffadffb700 (LWP 6991)]
[New Thread 0x7fffd4346700 (LWP 6992)]
[Thread 0x7fffc7fff700 (LWP 6981) exited]
[Thread 0x7fffc67fc700 (LWP 6984) exited]
[Thread 0x7fffc5ffb700 (LWP 6985) exited]
[Thread 0x7fffc6ffd700 (LWP 6983) exited]
[New Thread 0x7fffc6ffd700 (LWP 6993)]
[Thread 0x7fffd4346700 (LWP 6992) exited]
[Thread 0x7fffc6ffd700 (LWP 6993) exited]
[New Thread 0x7fffc6ffd700 (LWP 6998)]
[New Thread 0x7fffc5ffb700 (LWP 6999)]
[New Thread 0x7fffc67fc700 (LWP 7000)]
[Thread 0x7fffc67fc700 (LWP 7000) exited]
[New Thread 0x7fffc67fc700 (LWP 7001)]
[Thread 0x7fffc67fc700 (LWP 7001) exited]
[New Thread 0x7fffc67fc700 (LWP 7002)]
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
[New Thread 0x7fffc7fff700 (LWP 7003)]
[New Thread 0x7fff924f8700 (LWP 7009)]
[New Thread 0x7fff91cf7700 (LWP 7010)]
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
[Thread 0x7fffaffff700 (LWP 6987) exited]
[Thread 0x7fffc6ffd700 (LWP 6998) exited]
[Thread 0x7fffaeffd700 (LWP 6989) exited]
[Thread 0x7fffc5ffb700 (LWP 6999) exited]
[Thread 0x7fff91cf7700 (LWP 7010) exited]
[Thread 0x7fff924f8700 (LWP 7009) exited]
[Thread 0x7fffc67fc700 (LWP 7002) exited]
[New Thread 0x7fffc67fc700 (LWP 7032)]
[Thread 0x7fffc67fc700 (LWP 7032) exited]
[Thread 0x7fffc7fff700 (LWP 7003) exited]
[New Thread 0x7fffc7fff700 (LWP 7034)]
[New Thread 0x7fffc67fc700 (LWP 7037)]
[New Thread 0x7fff924f8700 (LWP 7038)]
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
[New Thread 0x7fff91cf7700 (LWP 7039)]
[New Thread 0x7fffaeffd700 (LWP 7040)]
[New Thread 0x7fff931fd700 (LWP 7041)]
libva info: VA-API version 0.34.0
libva info: va_getDriverName() returns 0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so
libva info: Found init function __vaDriverInit_0_34
libva info: va_openDriver() returns 0
Program received signal SIGSEGV, Segmentation fault.
[Switching to Thread 0x7fffaeffd700 (LWP 7040)]
0x0000000000d9d875 in _aligned_free(void*) ()
(gdb) backtrace
#0 0x0000000000d9d875 in _aligned_free(void*) ()
#1 0x00000000008f4a76 in CDVDDemuxUtils::FreeDemuxPacket(DemuxPacket*) ()
#2 0x0000000000ef3520 in CDVDMsgDemuxerPacket::~CDVDMsgDemuxerPacket() ()
#3 0x0000000000ef3559 in CDVDMsgDemuxerPacket::~CDVDMsgDemuxerPacket() ()
#4 0x0000000000eda679 in IDVDResourceCounted<CDVDMsg>::Release() ()
#5 0x0000000000ee8a0f in CDVDPlayerVideo::Process() ()
#6 0x00000000011087e0 in CThread::Action() ()
#7 0x0000000001108a92 in CThread::staticThread(void*) ()
#8 0x00007ffff55cff8e in start_thread () from /lib/x86_64-linux-gnu/libpthread.so.0
#9 0x00007ffff1378e1d in clone () from /lib/x86_64-linux-gnu/libc.so.6
I am running a recent version of testing (13.0-ALPHA6 Git:fe7d93a) and have noticed video playback is not quite as smooth as it might be, this is visable on some pans/zooms etc. Most of the content I watch are SD/HD mpg files recorded by a MythTV backend although the same phenomena occurs on ripped DVDs. All the files are played back via an NFS share on a wired connection.
Not sure if it is related but on the info screen I have noticed that the w(fps value skips about between 48.xx and 51.xx rather than being steady at 50.00 which I thought should perhaps be the case. I think the various video playback settings are at the recommended values.
A debug log is at
http://xbmclogs.com/show.php?id=49261
I have looked through this thread and not found any clues, any suggestions?
thank you, Paul
you are running a rather old version of July 28th. This one had a bug choosing the correct resolution:
Resolution 1360x768 @ 60.02 - Full Screen (16) not a very good match for fps 50.000 with 2:3 cadence
Can somebody give me a short update which drivers work for me? I'm running latest (two weeks ago) xbmcbuntu and planning to update to fernetmentas latest builds (Gotham) + vdr for the (hopefully) best PVR / Timeshift experience. If I can handle it I would like to add spotyxbmc2.
But I'm still confused which driver to use.
So please correct me if you have any other suggestions for my goals. Thank you!
What GPU are you gonna use? Nvidia,AMD or Intel? And what model?
I would suggest a fanless Nvidia-card build if I were you as that is the one that is the most hassle free AFAIK.
Shame on me, forgot to mention my specs: xbmcbuntu is running on a AMD A6-5400 (I think there's an AMD 7540D integrated). Well, I guess the APU is the important part you need to know, right?
Driver depends on your Ubuntu release. If you are on 12.10 Quantal, use what the apt repo has. If you are on 12.04, use the Catalyst 12.10. If you are on 13.04 - you won't be that lucky and need a patched 13.6 driver, see some pages back.
xbmcbuntu 12.2 is based on quantal, so your fglrx should be fine.