Announcement

Collapse
No announcement yet.

AMD's UVD2-based XvBA Finally Does Something On Linux

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

  • pretty sure it was mentioned somewhere that xvba would likely break with 9.11

    Comment


    • Does anyone else experience freezing in mplayer-vaapi? About halfway through a long video, mplayer will freeze the system. Someone mentioned earlier it is due to mplayer going into sleep mode? Has anyone any ideas how to resolve this issue? Other than the freezing and OSD issues, mplayer-vaapi works great!

      Comment


      • Originally posted by non7top View Post
        I got mplayer crashing. Both current svn mplayer and the one provided by script checkout-patch-build.sh.

        Gentoo users can try mplayer ebuild from arcon overlay with vaapi USE flag.

        Code:
        Opening video decoder: [ffmpeg] FFmpeg's libavcodec codec family
        [VD_FFMPEG] VA API accelerated codec.
        [VD_FFMPEG] Trying pixfmt=0.
        Unsupported PixelFormat -1
        Could not find matching colorspace - retrying with -vf scale...
        Opening video filter: [scale]
        The selected video_out device is incompatible with this codec.
        Try appending the scale filter to your filter list,
        e.g. -vf spp,scale instead of -vf spp.
        [VD_FFMPEG] Trying pixfmt=1.
        Unsupported PixelFormat -1
        Could not find matching colorspace - retrying with -vf scale...
        Opening video filter: [scale]
        The selected video_out device is incompatible with this codec.
        Try appending the scale filter to your filter list,
        e.g. -vf spp,scale instead of -vf spp.
        Unsupported PixelFormat -1
        Unsupported PixelFormat -1
        Selected video codec: [ffh264] vfm: ffmpeg (FFmpeg H.264)
        
        
        Starting playback...
        [h264 @ 0x8987120]decoding to PIX_FMT_NONE is not supported.
        [h264 @ 0x8987120]decode_slice_header error
        [h264 @ 0x8987120]no frame!
        Error while decoding frame!
        [h264 @ 0x8987120]decoding to PIX_FMT_NONE is not supported.
        [h264 @ 0x8987120]decode_slice_header error
        [h264 @ 0x8987120]no frame!
        Error while decoding frame!
        
        Too many audio packets in the buffer: (4099 in 4123594 bytes).
        Maybe you are playing a non-interleaved stream/file or the codec failed?
        For AVI files, try to force non-interleaved mode with the -ni option.
        
        FATAL: Could not initialize video filters (-vf) or video output (-vo).
        I was getting the same error too, at first (Unsupported PixelFormat -1). I had compiled mplayer-vaapi, but not yet installed it on my system. The resolve this error I had to point mplayer to the right codecs.conf file like this:

        Code:
        ./mplayer -codecs-file etc/codecs.conf -vo vaapi:gl -va vaapi <uri>
        with etc/codecs.conf the location of the codecs.conf file modified by the vaapi-patch.

        Comment


        • Originally posted by gururise View Post
          Does anyone else experience freezing in mplayer-vaapi? About halfway through a long video, mplayer will freeze the system. Someone mentioned earlier it is due to mplayer going into sleep mode? Has anyone any ideas how to resolve this issue? Other than the freezing and OSD issues, mplayer-vaapi works great!
          Which revision of catalyst you have ?
          best regards,

          Comment


          • @herriejr

            Maybe you had another codecs file in your system. Usually it is built-in - try removing the system wide one. But i see no reason to manually compile it. As debian package installed via my script it works fine. Just use the scriptname without -1 in the name or disable the "svn up" which currently does not work.

            Comment


            • Originally posted by hepi View Post
              I have integrated radeon 3100 (780v) and it doesn't have UVD (officially). I made some registry tweaks on windows(http://home.comcast.net/~exdeus/ati-hd2x00/) and now I'm able to play fullhd movies with low cpu usage, even with the newest flash 10.1 beta. So the question is 780v = 780g and changes are done only by software or it uses shaders to decode?
              Now THAT is an interesting observation.
              How about it? Anyone know the answer here?

              I know that it is very common for hardware manufacturers to use the same hardware under multiple different names. Its much cheaper to just disable features on high-end hardware in order to achieve a low-end device than it is to have two actually manufacture multiple parts. For example, the AMD triple-core CPUs are identical to the quad-core parts, just with one of the cores disabled.

              Since one of my machines has that same 780v/3100 as yours, the answer to this question would be very interesting to me -- more academically though, the CPU on it is WAY more than sufficient to decode anything I can come up with...

              Comment


              • aac encoding support?

                @Kano
                What would be the easiest way to add aac encoding support with a XvBA enabled mplayer?

                Comment


                • Originally posted by droidhacker View Post
                  Now THAT is an interesting observation.
                  How about it? Anyone know the answer here?

                  I know that it is very common for hardware manufacturers to use the same hardware under multiple different names. Its much cheaper to just disable features on high-end hardware in order to achieve a low-end device than it is to have two actually manufacture multiple parts. For example, the AMD triple-core CPUs are identical to the quad-core parts, just with one of the cores disabled.

                  Since one of my machines has that same 780v/3100 as yours, the answer to this question would be very interesting to me -- more academically though, the CPU on it is WAY more than sufficient to decode anything I can come up with...
                  True, except that this disabled parts could be unworking parts. If you are lucky, it just works by reactivating the de-activated part, but it can also just crash for unknown reason, just because there is something wrong in the chip silicon.
                  It's like overclocking : it can work. It often works should I say. But I can also fail.

                  But, for this particular point, perhaps are they just the same chips, and the difference is only made for marketing purposes. That doesn't change anything on the interest of the topic though.

                  Comment


                  • Originally posted by RealNC View Post
                    I can't see any flickering here, neither with Gtk+ nor with anything else.
                    I take that back. There's corruption in some places. I opened a thread about it since it's off-topic in this one:

                    http://www.phoronix.com/forums/showthread.php?t=20333

                    Comment


                    • Originally posted by Fixxer_Linux View Post
                      True, except that this disabled parts could be unworking parts. If you are lucky, it just works by reactivating the de-activated part, but it can also just crash for unknown reason, just because there is something wrong in the chip silicon.
                      It's like overclocking : it can work. It often works should I say. But I can also fail.

                      But, for this particular point, perhaps are they just the same chips, and the difference is only made for marketing purposes. That doesn't change anything on the interest of the topic though.
                      Except in this case we're talking about having a defective UVD component. JBQ mentioned in another thread that the target market for the 780V is "business" users, and that for this market, it is *preferable* to have the UVD disabled (presumably to keep people from watching movies when they should be working and/or perception of the part as being for common users). We don't see any other GPUs missing the UVD part besides the 780V/3100, so I suspect that rather than being the "defective" parts, it is more likely to simply be disabled -- and at that, probably by software based on some hardware identifier that marks it as non-UVD.

                      Comment


                      • Originally posted by RealNC View Post
                        I take that back. There's corruption in some places. I opened a thread about it since it's off-topic in this one:

                        http://www.phoronix.com/forums/showthread.php?t=20333
                        that is the punishment for using crap (adobe reader) :P

                        Comment


                        • Don't be silly :P It also happens with Gtk apps if I "tray" them; there's black and flickering with I click the tray icon to pop them up again.

                          Comment


                          • @xzero2

                            You need libfaac-dev installed. Then just recompile.

                            Comment


                            • Originally posted by RealNC View Post
                              Don't be silly :P It also happens with Gtk apps if I "tray" them; there's black and flickering with I click the tray icon to pop them up again.
                              as I said, don't use crap

                              Comment


                              • Originally posted by Kano View Post
                                @herriejr

                                Maybe you had another codecs file in your system. Usually it is built-in - try removing the system wide one. But i see no reason to manually compile it. As debian package installed via my script it works fine. Just use the scriptname without -1 in the name or disable the "svn up" which currently does not work.
                                Hello Kano,
                                - reason for manually compiling it is that I also have applied the dshowserver-patch for mplayer
                                - problem with the codecs.conf was indeed another codecs file in my system that was being used by default (for the dshowserver patch to work correctly). Therefore I thought - since it was the same error message - that this could be a possible solution. It indeed works fine with the built-in one (for vaapi).

                                Thanks for your answers and remarks.

                                Comment

                                Working...
                                X