Kodi Community Forum
How to Install XBMC PVR Xvba for AMD/Nvidia/Intel GPUs - Printable Version

+- Kodi Community Forum (https://forum.kodi.tv)
+-- Forum: Support (https://forum.kodi.tv/forumdisplay.php?fid=33)
+--- Forum: General Support (https://forum.kodi.tv/forumdisplay.php?fid=111)
+---- Forum: Linux (https://forum.kodi.tv/forumdisplay.php?fid=52)
+---- Thread: How to Install XBMC PVR Xvba for AMD/Nvidia/Intel GPUs (/showthread.php?tid=116996)



- grad - 2011-12-21

fritsch Wrote:There are really a lot of settings in xorg.conf. Could you backup this file an retest with:
sudo aticonfig --initial

For your log I see that you see builds with the Reference Clock Stuff. There are still some bugs, we, Fernetmenta especially, try to track down.

If you use wsnipex ppa packages (not the testing ppa). These are normally more stable, cause they are tested more extensive.

Thanks

Well I came from the initial aticonfig and added switches when things werent working, cause but I'll try this. As for the build: I am using the std wsnipex ppa! *confused*


- fritsch - 2011-12-21

grad Wrote:Well I came from the initial aticonfig and added switches when things werent working, cause but I'll try this. As for the build: I am using the std wsnipex ppa! *confused*

Sorry for confusing you. Everything should be okay. Lets see what will happen after --initial.


- grad - 2011-12-21

fritsch Wrote:Sorry for confusing you. Everything should be okay. Lets see what will happen after --initial.

Thanks for unconfusing me Smile Unfortunately though, things remain the same with --initial. The error is now pretty consistently showing when switching back(!) from the correctly rendered and mode-switched fullscreen to the menu. So the first mode switch works alright, but it doesn't properly switch back. Since I definitely saw the reverse effect at some time also, I'm rather thinking this is some nasty glitch somewhere. (FWIW, I also once tried the xbmc option to insert a pause upon mode switching, but that didn't help either.)

To make my situation worse, I tried the Android remote control app last night and now I'm completely hooked to XBMC. Pure awesomeness :-)


- fritsch - 2011-12-21

grad Wrote:Thanks for unconfusing me Smile Unfortunately though, things remain the same with --initial. The error is now pretty consistently showing when switching back(!) from the correctly rendered and mode-switched fullscreen to the menu. So the first mode switch works alright, but it doesn't properly switch back. Since I definitely saw the reverse effect at some time also, I'm rather thinking this is some nasty glitch somewhere. (FWIW, I also once tried the xbmc option to insert a pause upon mode switching, but that didn't help either.)

To make my situation worse, I tried the Android remote control app last night and now I'm completely hooked to XBMC. Pure awesomeness :-)

Hard to debug. Let`s see if some setting is wrong somewhere in the deep of the X stuff. Could you try booting with the following Live Image?:
http://xbmcfreak.binkey.nl/XBMCbuntu-v14.zip

It is only to test modesetting issues. If this succeeds, we have to look what software version, etc. you are using.

Bye

Edit: Problem is being worked on - just wait some hours, i think testing the live iso is not needed then.


- wsnipex - 2011-12-21

grad Wrote:Thanks for unconfusing me Smile Unfortunately though, things remain the same with --initial. The error is now pretty consistently showing when switching back(!) from the correctly rendered and mode-switched fullscreen to the menu. So the first mode switch works alright, but it doesn't properly switch back. Since I definitely saw the reverse effect at some time also, I'm rather thinking this is some nasty glitch somewhere. (FWIW, I also once tried the xbmc option to insert a pause upon mode switching, but that didn't help either.)

To make my situation worse, I tried the Android remote control app last night and now I'm completely hooked to XBMC. Pure awesomeness :-)

This is the same behavior I'm seeing on my system.
for me there are 2 workarounds:
1.) stop the video manually before it ends
2.) toggle to window mode and back


- grad - 2011-12-21

fritsch Wrote:Edit: Problem is being worked on - just wait some hours, i think testing the live iso is not needed then.

Thanks guys, I'll be standing by! As expected the issue shows up with the live image as well.


- fritsch - 2011-12-21

grad Wrote:Thanks guys, I'll be standing by! As expected the issue shows up with the live image as well.

Sent you a pm with a link to testing packages and description of known bugs.


- grad - 2011-12-21

fritsch Wrote:Sent you a pm with a link to testing packages and description of known bugs.

Thanks a million, this seems to fix the bug indeed!

However, I had to deactivate "Sync playback to display". I only activated "Adjust display refresh rate". When both were activated, the was a horrible constant orgy of flickering going on, I couldn't tell if it actually switched modes constantly or if was bad rendering. Is this to be expected? Will keep on testing this...


- Robotica - 2011-12-21

Great to see all this development for AMD hardware! Does this also include hardware acc. for Flash?


- fritsch - 2011-12-21

grad Wrote:Thanks a million, this seems to fix the bug indeed!

However, I had to deactivate "Sync playback to display". I only activated "Adjust display refresh rate". When both were activated, the was a horrible constant orgy of flickering going on, I couldn't tell if it actually switched modes constantly or if was bad rendering. Is this to be expected? Will keep on testing this...

Wargh i forgot the most important in my mail: Please enable TEARFREE, as fglrx will ignore vblank stuff, if this is not enabled.


- fritsch - 2011-12-21

Robotica Wrote:Great to see all this development for AMD hardware! Does this also include hardware acc. for Flash?

Nope, sorry flash is totally closed source, nothing can be done here.


- Robotica - 2011-12-21

fritsch Wrote:Nope, sorry flash is totally closed source, nothing can be done here.

ah no problem, CPU is strong enough untill Adobe and AMD work out how to access UVD3.


- Robotica - 2011-12-21

I am not sure if this is the right place to ask but......

I noticed a patch related to AMD/ATI hardware and shaders, which was never committed (need rebase). Maybe this is a quick win: http://trac.xbmc.org/ticket/8771

if not, just ignore this post.


- grad - 2011-12-21

fritsch Wrote:Wargh i forgot the most important in my mail: Please enable TEARFREE, as fglrx will ignore vblank stuff, if this is not enabled.

Yeah, that's better! :-) Thanks again for all your efforts! Actually 24p output seems to work just fine over here, I'm not getting any kind of visible delay.


- fritsch - 2011-12-21

This option works more or less with tearfree enabled at the moment. But we want to avoid that tearfree must be used, cause of laggy desktop performance and other side effects, it is not bugfree.

Currently we are searching and destroying bugs in this new method of sync.

Prepare for christmas *hehe*