Bug 22007 - Video playback artefacts when using Xv
Summary: Video playback artefacts when using Xv
Status: RESOLVED FIXED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2009-05-30 13:14 UTC by Tim Besard
Modified: 2011-01-20 14:12 UTC (History)
3 users (show)

See Also:
i915 platform:
i915 features:


Attachments
Xorg configuration. (311 bytes, application/octet-stream)
2009-05-30 23:39 UTC, Tim Besard
no flags Details
Xorg log. (51.25 KB, application/octet-stream)
2009-05-30 23:42 UTC, Tim Besard
no flags Details
xorg.conf (1.39 KB, application/octet-stream)
2009-07-08 02:07 UTC, Perry3D
no flags Details
Xorg.0.log (69.50 KB, application/octet-stream)
2009-07-08 02:08 UTC, Perry3D
no flags Details
Fix the texture size of the chroma planes in Xv (1.17 KB, patch)
2011-01-20 13:45 UTC, Marton Balint
no flags Details | Splinter Review

Description Tim Besard 2009-05-30 13:14:51 UTC
When playing video using the recently Xv acceleration, I experience some minor artefacts in the image. The artefacts have been occurring for a month or longer, and might have gotten introduced together with the reworked Xv acceleration (I'll attempt a git bisect in a few weeks if nothing gets found at first sight).

When disabling Xv's DMA transfers, the artefacts definitely get worse. It's also remarkable the artefacts only appear in the lower 30% of the screen, and tend to appear around heavily moving objects. A single artefact is almost always a ~360x4 wide rectangle, and if multiple appear they tend to group in blocks.
Disabling Xv altogether resolves the issue completely.

I'm running -radeon from GIT (Arch Linux AUR build), but the same artefacts appear on Ubuntu Jaunty with Tormod's PPA enabled (which also provides up-to-date binaries). Kernel version is 2.6.29, Xorg version is 1.6.1.901. GPU is an HD4870.

Some images:
Corruption with DMA transfers disabled: http://imgur.com/GYZTc.png (clear band of corruption in the lower part of the screen)
Same scene with DMA transfers enabled: http://imgur.com/yCXER.png (almost no corruption, only in the lower right corner)
Another scene with DMA transfers enabled: http://imgur.com/k66ct.png (in this scene only the hand is actively moving, and the non-moving parts are not corrupted)
Comment 1 Alex Deucher 2009-05-30 17:05:30 UTC
Please attach your xorg log and config.
Comment 2 Tim Besard 2009-05-30 23:39:55 UTC
Created attachment 26314 [details]
Xorg configuration.
Comment 3 Tim Besard 2009-05-30 23:42:49 UTC
Created attachment 26315 [details]
Xorg log.

The problem occurs too when using no config at all and let X autodetect. Running x86_64 btw.
Comment 4 Perry3D 2009-07-08 02:06:26 UTC
i have the same problem.
Also on a Arch x68_64. Kernel is 2.6.30.1, xorg is 1.6.1.901 and libdrm is 2.4.11. My hardware is a HD4850.
This only appears with XV. When i switch to X11 the artefacts disappear.
Comment 5 Perry3D 2009-07-08 02:07:42 UTC
Created attachment 27490 [details]
xorg.conf
Comment 6 Perry3D 2009-07-08 02:08:37 UTC
Created attachment 27491 [details]
Xorg.0.log
Comment 7 Perry3D 2009-07-22 01:52:58 UTC
Update to mesa 7.5, libdrm 2.4.12 and xorg 1.6.2. Bug is still there.
Comment 8 Borkd 2009-08-26 16:41:41 UTC
I have this problem as well, take a look at the following screenshot: http://yfrog.com/12imageartefactsp

I am running x86_64 Archlinux.  2.6.30 kernel, Xorg 1.6.3, xf86-video-ati (GIT). DRI and EXA are enabled.

I have a Radeon HD4850.
Comment 9 Marton Balint 2009-10-20 17:35:37 UTC
Same problem here with a HD4850.

The bug can be easily reproduced with small XV sizes. For example:

mplayer movie.avi -vf crop=50:50

It should be noted though, that after using mplayer with a crop size, you will have to use a different crop size next time, to hit the bug again. But if you just select a random number between 10 and 100, you will almost certainly always experience the bug.

And yes, only the bottom 50% of the screen is affected, and it seems that the chroma planes are broken.
Comment 10 Alex Deucher 2010-10-19 17:24:23 UTC
Is this still an issue with newer versions of the driver?
Comment 11 Marton Balint 2010-10-20 13:00:59 UTC
(In reply to comment #10)
> Is this still an issue with newer versions of the driver?

Yes, just tested the latest version of the radeon driver from git, the problem still exists. If you need any more information about the bug please let me know.
Comment 12 Marton Balint 2011-01-20 13:45:05 UTC
Created attachment 42239 [details] [review]
Fix the texture size of the chroma planes in Xv

The attached patch fixes the issue for me.
Comment 13 Alex Deucher 2011-01-20 14:12:50 UTC
Pushed:
f41cf83750ba9a2e0797fdb21ea9104b85ce53aa
thanks!


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.