Announcement

Collapse
No announcement yet.

libvdpau, libva Both Updated Today

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #11
    Maybe the test files crash the system

    Comment


    • #12
      Originally posted by Kano View Post
      Maybe the test files crash the system
      That could be, I was under the impression it actually does something. Also in the When such a benchmark is done I would recommend trying it out on a video file with no audio stream as I notice pulseaudio slaps on some pretty nice overhead when it is used to isolate the cpu usage to video playback only ( or use a OS where pulse is disabled by default).

      BTW, anybody looking for quality highbit rate samples to test play back to see if their current solution is good enough (tm) just google for:

      Ducks.Take.Off.1080p.QHD.CRF25.x264-CtrlHD which is up to 106 mBit (even vdpau can't keep up to this one, even though it is a extreme clip)
      and
      SAMPLE-Cloverfield.2008.1080p.BluRay.DTS.x264-CtrlHD.mkv which is up to 66 mbit (vdpau however has no issues with this even on a 8200 IGP)
      Last edited by deanjo; 20 November 2009, 07:40 PM. Reason: Editted to place sample suggestions and to mark Good Enough as a tradmark of opensource video driver development

      Comment


      • #13
        You can be sure, that H264 L4.1 would be better suited for a test. But even if it decodes it can have got wrong colors. Have got 2 examples for this problem...

        Comment


        • #14
          WTF

          I'm just hate this situation within Linux.
          If you use VGA card from one vendor, that means that you will be complete newby with other vendor.
          I'm one time owner of ATI HD4580 (next time I'll try ATI will be at least few years from now).
          I think that some one like Kronos should publish platform independent and of course vendor independent API for video acceleration. You know ... Like OpenVA
          Or whatever ...
          I know that experimenting is good. And different approaches are good.
          The choice is good.
          But they all useless when you can't build a system that just works.
          Some times, i just want things to work out of the box.
          So long live implemented standards first, and experiments next.
          Experiments will provide base for next standards.

          Comment


          • #15
            @deanjo

            Isn't it a bit weird when the bitrate is 66 mbps max when the input (which was logical bluray) has a max rate of 48 mbps?

            Comment

            Working...
            X