Next stop - gstreamer?
Phoronix: Xine Picks Up Support For NVIDIA's VDPAU
It was just a month ago that NVIDIA had introduced the Video Decode and Presentation API for Unix that brought PureVideo-like features to Linux and as our early benchmarks showed this video API did an effective job at offloading video-related tasks to the graphics card that otherwise would be handled by the CPU. Last week we then took a $20 processor and $30 graphics card and managed to play HD videos on Linux quite well when using VDPAU. When NVIDIA had introduced VDPAU they did supply patches that added support for this API to the MPlayer and ffmpeg projects...
http://www.phoronix.com/vr.php?view=Njk0MA
Next stop - gstreamer?
There are a few things about this news article that are bothering me. There is no link to any other source of information about the support for vdpau in xine. Subversion repository is mentioned, but xine-lib uses Mercurial and xine-ui CVS and none of those contain any reference to vdpau. The reader in prompted to look up additional information from the xine web-site, but there isn't the slightest reference to vdpau to be found on that web-site.
Here are some info that is lacking:
Mailing list for xine-vdpau:
http://lists.kafic.ba/pipermail/xine...er/thread.html
The subversion repository:
svn://jusst.de/xine-vdpau
You forgot the irc channel as well
IRC
Freenode, #xine-vdpau
This happens all the time on Phorinix articles.
Precious links are simply missing.
GStreamer at its core is basically just a wrapper around FFmpeg, so GStreamer will get VDPAU support soon after it has been added to FFmpeg's SVN trunk, not before, (just as most other open source projects that are based on FFmpeg).
I personally think that it will probably take a while before FFmpeg gets VDPAU support in its SVN trunk though, this because NVIDIA will first need to make the implementation optional as well as automatic, (so that when enabled it will only be used if the hardware supports it and always fallback to software decoding if the hardware does not support it or fail, plus being able to force disabling the function at compile time as an option), then clean up the code patch, make sure it does not break anything else, and all in all make it a acceptable for FFmpeg's developers to commit to their trunk, (most FFmpeg's developers are perfectionists and are very picky about the code that gets committed to their SVN, ...probably because so many other third-party open source projects depend on FFmpeg).
FFmpeg-Based Projects => http://www.ffmpeg.org/projects.html