[FernetMenta-Master] cachemembuffersize settings in advancedsettings.xml ignored
#1
Video streams are no longer cached in the XBMC 14 builds. They all all read the advancedsettings.xml correctly, but do not use any of the caching settings. Verified the behaviour on several systems.
Has there been a design change on this? Latest version tested is 14.0~git20140406.0501-1fb475d-0saucy

I have directly attached storage to the system running XBMC, but the only way to make 100% sure the player is not I/O starved at some point during a movie is to enable caching.

DEBUG log for non-working cachemembuffersize: http://pastebin.com/pvDSJRBa (XBMC version 14.0~git20140406.0501-1fb475d from the FernetMenta-Master branch.)

DEBUG log for working cachemembuffersize: http://pastebin.com/2qZvGtQs (XBMC version 13.0~git20131003.0500-a662ab2 from the Xvba branch)

Both use the same advancedsettings.xml. Not sure if the debug logs are helpful wrt. this issue, but providing them nonetheless.

btw., this is on a test system with Pulseaudio, so ignore the PA crap in the logs.
Reply
#2
Please reproduce with mainline version and submit a problem report. I did not change anything to this
Reply
#3
Thanks for you reply-I'll do that. I just tested a mainline nightly build and it behaves the same. Sorry for pointing fingers at your code. Smile I'm usually running Xvba or your branch due to the non-working dual-head with mainline–unless it's on a test system.

btw., is this the case for the audio stutter issue as well (which I assume, exists with ActiveAE, and seem to have occurred after the 'keepalive' settings for the audio sinks were introduced)?
Reply
#4
your setting is wrong for Gotham builds, its buffermode instead of alwaysforcebuffer
http://wiki.xbmc.org/index.php?title=HOW...ideo_cache
Reply
#5
(2014-04-08, 18:50)Drag0nFly Wrote: btw., is this the case for the audio stutter issue as well (which I assume, exists with ActiveAE, and seem to have occurred after the 'keepalive' settings for the audio sinks were introduced)?

Keep Audio alive is there since ActiveAE started ... only the settings name was changed later on. I don't see any correlation here.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
#6
@snipex– Smile I was thinking there had to have been a design change going on here. Just as I submitted a 'bug' report for this, I tried your suggestion which resolved it. Thanks a bunch.

Then the only remaining issue is the audio stutter on pause when using passthrough (at least when using S/PDIF). Was this introduced in mainline? I would like to test this, but I cannot since mainline still relies on SDL and I am unable to run it on my system (test system only has HDMI, not S/PDIF)
Reply
#7
Try to produce a minimal log, that shows the issue, please.
First decide what functions / features you expect from a system. Then decide for the hardware. Don't waste your money on crap.
Reply
#8
@fritsch, that is probably the case. Although I had to revert back to an October 2013 (Xvba-)build. I've tested a range of FernetMenta-Master versions from Jan 2014 up until the most recent ones this April, and the stutter is there.
Reply
#9
I don't have any audio stutter on pause with spdif passthrough. Can you give steps on how to reproduce that.
Reply
#10
I've now produced a new debug log with the latest build (Apr 6th), and trimmed off any unnecessary parts. Still left a bit in there, though, as I don't want to remove too much. The one I had in my original bug report (different thread, some pages back) was from March 21st.

The log is in the separate thread I made for the audio stutter on pause issue--
http://forum.xbmc.org/showthread.php?tid...pid1677230

I'll be happy to test any fixes (or config changes, although I've tried quite a few).

@wsnipex–the stutter is reproducible each time, and only occurs with the newer 14-alpha1 builds. When changing the audio device keepalive settings to off (from 1 min), the audio artifacts are worse, which was why I initially suspected this as the culprit.

I also notice that the audio sync delay has to be changed with ActiveAE (seem to remember seeing something about this as well). Normally I'd use a 175ms delay, but it is not needed for DTS streams with Gotham. FLAC 5.1 streams, however, need the 175ms delay set.
(On Frodo the 175ms setting worked for all media (DD TrueHD, DTS HD MA + PXM / FLAC etc.)
Reply

Logout Mark Read Team Forum Stats Members Help
[FernetMenta-Master] cachemembuffersize settings in advancedsettings.xml ignored0