|
|
#25 | |
|
Moderator
Join Date: Aug 2005
Posts: 1,327
|
Any problem that appears in both Xv and VDPAU is most likely unrelated to the driver; the two driver code-bases are entirely separate. Perhaps an incorrectly encoded file, a corrupted file, or a HW issue?
|
|
|
|
|
|
|
#26 | |
|
Registered User
Join Date: Jan 2011
Posts: 2
|
Quote:
).But nevertheless, why would it be unlikely to be a driver issue? XV and VDPAU are HW accelerated, right? If some algorithm or timing is wrong, it could miscalculate the pixelvalues and probably set the alpha values wrong because of overflows?! Since I have no clue about the Nvidia driver and the HW acceleration via VDPAU/XV I can just guess. If this also happens with XV, and the code bases are seperate, what could affect both? Just timings of memory/GPU/bus? Some filters/videomixers in the GPU? I wonder why it only happens on bigger dark areas like shadows. But maybe it also happens on bright areas. Since the underlying window is white, it would not be visible. I need to check that, but at the moment I reverted to 256.53 to see if this issue still appears. But of course it could be an HW issue. If it is, there should be a workaround. As far as I understand, mostly the ION has this issue?! File should be ok, was playing fine with X11 video output. Also my Laptop plays it correctly. I'm also not able to check it, since I do not have the hardware at the moment, could take some time. |
|
|
|
|
|
|
#27 | |
|
Registered User
Join Date: Sep 2008
Posts: 19
|
Quote:
It's happening on two systems, both with g210 video cards using DVI. I rolled back to 256, and my problems are gone. I doubt it's a HW issue. |
|
|
|
|
|
|
#28 | |
|
Registered User
Join Date: Sep 2005
Posts: 14
|
I've been seeing exactly this issue with VDPAU for months: about a day and a half after reboot, white or colored sparkles appear to "shine though" black regions, and/or artifacts, always with tearing. Can happen in myth or mplayer.
Here's a clue I stumbled upon by accident: once, when this issue occurred after the normal day and a half, I noticed that I could see the video over my VNC session, which I believe has always been invisible to VNC servers when using VDPAU. I was using x11vnc to serve and Remmina to view. EDIT: I'm on a Geforce 210. Last edited by jafar; 02-26-11 at 11:20 AM. Reason: Add video card info |
|
|
|
|
|
|
#29 |
|
Registered User
Join Date: Nov 2004
Location: Between the keyboard and the chair.
Posts: 490
|
in beta 270.18/270.26 too?
|
|
|
|
|
|
#30 | |
|
Registered User
Join Date: Sep 2005
Posts: 14
|
|
|
|
|
|
|
|
#31 |
|
Registered User
Join Date: Aug 2006
Posts: 3
|
Had that problem, too. Latest Ubuntu release. nvidia-260.19.06, 2.6.35-27-generic #44-Ubuntu SMP Fri Jan 21 17:40:44 UTC 2011 x86_64 GNU/Linux
current xine-lib 1.2 compiled from mercurial repo (today) with xine-ui 0.99.6 and xvdr plugin 1.0.5 against xineliboutput 1.0.5 on server. Problem was: kill xine vdpau playback without proper xine shutdown and xine displays the problem upon next playback (seems also to be valid for x11). Problem goes away with xorg crash/restart. Solution: enable composite extension in xorg.conf. For some reason I had disabled it. |
|
|
|
|
|
#32 | |
|
Registered User
Join Date: Sep 2004
Posts: 29
|
Quote:
on the off-chance it was an OS/Driver issue, I reformatted/re-installed and the problem still continued. On a case-by-case basis (resolution, encoder type etc) I can tune VDPAUBUFFERSIZE to mostly make the problem go away, but what works for one playback type doesn't work for another. 2.6.35.11-83.fc14.i686 NVRM: loading NVIDIA UNIX x86 Kernel Module 260.19.36 Tue Jan 18 17:10:40 PST 2011 |
|
|
|
|
|
|
#33 | |
|
Registered User
Join Date: Aug 2006
Location: Lille
Posts: 15
|
Quote:
edit : composite is off in xorg.conf to prevent tearing. |
|
|
|
|
|
|
#34 |
|
Registered User
Join Date: Sep 2004
Posts: 29
|
Any traction on this issue?
|
|
|
|
![]() |
| Thread Tools | |
|
|