<div dir="ltr">Many thanks!<div><br></div><div>After installing</div><div><br></div><div>apt-get install i965-va-driver* libva-intel-vaapi-driver*</div><div><br></div><div>my NUCDN2820FYKH correctly found</div><div><br></div><div>libva info: VA-API version 0.35.0 and /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so</div><div><br></div><div>and started to smoothly play with vaapisink instead of autovideosink (with the difference that the time overlay does not show up any more; text overlay works OK)</div><div><br></div><div>But my Ubuntu 14.04 workstation with an nVidia card</div><div><br></div><div>reports failure:</div><div><br></div><div><div>libva info: VA-API version 0.35.0</div><div>libva info: va_getDriverName() returns 0</div><div>libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/nouveau_drv_video.so</div><div>libva info: va_openDriver() returns -1</div><div>vaInitialize failed with error code -1 (unknown libva error),exit</div></div><div><br></div><div>and I could not find the latter .so library in any nouveau-related packages. Does it mean that nVidia/nouveau do not support VAAPI? I earlier experienced hang-ups with the nouveau driver and some people suggested that it is a known issue with nouveau. What would be ideal, dream graphics cards to support gstreamer?</div><div><br></div><div><br></div><div><br></div></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Sep 18, 2014 at 4:06 PM, Gwenole Beauchesne <span dir="ltr"><<a href="mailto:gb.devel@gmail.com" target="_blank">gb.devel@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Hi,<br> <div><div class="h5"><br> 2014-09-18 12:13 GMT+02:00 Sergei Vorobyov <<a href="mailto:sergei.vorobyov@facilitylabs.com">sergei.vorobyov@facilitylabs.com</a>>:<br> > Thanks!<br> ><br> > Indeed the problem was caused by the fact I tried to gst-discover via an ssh<br> > connection, my fault. When I do it directly on the NUC, the .wmv file in<br> > question is discovered and played OK... Almost.<br> ><br> > Although the file started to play *reasonably well*, gst-launch continued to<br> > complain about i965_drv_video.so, which indeed was missing:<br> ><br> > fl@nuc3:~/ads$ gst-launch-1.0 filesrc location=621.wmv ! decodebin !<br> > videoconvert ! ximagesink<br> > Setting pipeline to PAUSED ...<br> > Pipeline is PREROLLING ...<br> > libva info: VA-API version 0.35.0<br> > libva info: va_getDriverName() returns 0<br> > libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so<br> > libva info: va_openDriver() returns -1<br> > libva info: VA-API version 0.35.0<br> > libva info: va_getDriverName() returns 0<br> > libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so<br> > libva info: va_openDriver() returns -1<br> > libva info: VA-API version 0.35.0<br> > libva info: va_getDriverName() returns 0<br> > libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so<br> > libva info: va_openDriver() returns -1<br> > Redistribute latency...<br> > Pipeline is PREROLLED ...<br> > Setting pipeline to PLAYING ...<br> > New clock: GstSystemClock<br> > Got EOS from element "pipeline0".<br> > Execution ended after 0:00:05.000180624<br> > Setting pipeline to PAUSED ...<br> > Setting pipeline to READY ...<br> > Setting pipeline to NULL ...<br> > Freeing pipeline ...<br> ><br> > Worse things started to happen when I run apt-get install libva-*, which<br> > installed dri/i965_drv_video.so<br> ><br> > The playback became *real slow* (because of acceleration?) and gst-launch<br> > reported:<br> <br> </div></div>It became slow because...<br> <span class=""><br> > fl@nuc3:~/ads$ gst-launch-1.0 filesrc location=621.wmv ! decodebin !<br> > videoconvert ! ximagesink<br> <br> </span>... you are downloading the VA surface pixels to system memory, then<br> use SW color conversion to RGB format for ximagesink.<br> <div><div class="h5"><br> > Setting pipeline to PAUSED ...<br> > Pipeline is PREROLLING ...<br> > libva info: VA-API version 0.35.0<br> > libva info: va_getDriverName() returns 0<br> > libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so<br> > libva info: Found init function __vaDriverInit_0_35<br> > libva info: va_openDriver() returns 0<br> > Got context from element 'vaapidecode0': gst.vaapi.Display=context,<br> > display=(GstVaapiDisplay)NULL;<br> > Pipeline is PREROLLED ...<br> > Setting pipeline to PLAYING ...<br> > New clock: GstSystemClock<br> > WARNING: from element /GstPipeline:pipeline0/GstXImageSink:ximagesink0: A<br> > lot of buffers are being dropped.<br> > Additional debug info:<br> > gstbasesink.c(2791): gst_base_sink_is_too_late ():<br> > /GstPipeline:pipeline0/GstXImageSink:ximagesink0:<br> > There may be a timestamping problem, or this computer is too slow.<br> > WARNING: from element /GstPipeline:pipeline0/GstXImageSink:ximagesink0: A<br> > lot of buffers are being dropped.<br> > Additional debug info:<br> > gstbasesink.c(2791): gst_base_sink_is_too_late ():<br> > /GstPipeline:pipeline0/GstXImageSink:ximagesink0:<br> > There may be a timestamping problem, or this computer is too slow.<br> > WARNING: from element /GstPipeline:pipeline0/GstXImageSink:ximagesink0: A<br> > lot of buffers are being dropped.<br> > Additional debug info:<br> > gstbasesink.c(2791): gst_base_sink_is_too_late ():<br> > /GstPipeline:pipeline0/GstXImageSink:ximagesink0:<br> > There may be a timestamping problem, or this computer is too slow.<br> > WARNING: from element /GstPipeline:pipeline0/GstXImageSink:ximagesink0: A<br> > lot of buffers are being dropped.<br> > Additional debug info:<br> > gstbasesink.c(2791): gst_base_sink_is_too_late ():<br> > /GstPipeline:pipeline0/GstXImageSink:ximagesink0:<br> > There may be a timestamping problem, or this computer is too slow.<br> > Got EOS from element "pipeline0".<br> > Execution ended after 0:00:05.221524402<br> > Setting pipeline to PAUSED ...<br> > Setting pipeline to READY ...<br> > Setting pipeline to NULL ...<br> > Freeing pipeline ...<br> ><br> > Luckily, when I replaced ximagesink with autovideosink, the playback became<br> > fast and smooth as before reporting:<br> ><br> > fl@nuc3:~/ads$ gst-launch-1.0 filesrc location=621.wmv ! decodebin !<br> > videoconvert ! autovideosink<br> > Setting pipeline to PAUSED ...<br> > libva info: VA-API version 0.35.0<br> > libva info: va_getDriverName() returns 0<br> > libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/i965_drv_video.so<br> > libva info: Found init function __vaDriverInit_0_35<br> > libva info: va_openDriver() returns 0<br> > Pipeline is PREROLLING ...<br> > Got context from element 'vaapidecode0': gst.vaapi.Display=context,<br> > display=(GstVaapiDisplay)NULL;<br> > Pipeline is PREROLLED ...<br> > Setting pipeline to PLAYING ...<br> > New clock: GstSystemClock<br> > Got EOS from element "pipeline0".<br> > Execution ended after 0:00:05.060843602<br> > Setting pipeline to PAUSED ...<br> > Setting pipeline to READY ...<br> > Setting pipeline to NULL ...<br> > Freeing pipeline ...<br> <br> </div></div>And this is fine since, by using autovideosink, you allow for the<br> "vaapisink" element to get autoplugged.<br> <span class=""><br> > Does it mean that Video Acceleration and ximagesink are incompatible? How to<br> > choose an optimal video sink, rather than autovideosink? I tried others like<br> > clutter, v4l2sink. Nothing works except autovideosink.<br> <br> </span>An optimal sink, if you use gstreamer-vaapi, is vaapisink. Having said<br> that, cluttersink & glimagesink shall work equally well too, nowadays.<br> The latter with GStreamer >= 1.4 of course, and recent enough<br> gstreamer-vaapi (git master branch).<br> <span class=""><br> > Which sink is actually selected by autovideosink? I am hesitant of using<br> > autovideosink in the real C application with gst_element_factory_make<br> > ("autovideosink", "videosnk"), because it is more tricky to map this<br> > autovideosink into an X-window. Note this mapping is static with ximagesink<br> > (no callbacks needed)<br> <br> </span>If you know your system supports VA-API, then you could use vaapisink<br> that also supports the GstVideoOverlay interface.<br> <div class="HOEnZb"><div class="h5"><br> > On Wed, Sep 17, 2014 at 2:02 PM, Gwenole Beauchesne <<a href="mailto:gb.devel@gmail.com">gb.devel@gmail.com</a>><br> > wrote:<br> >><br> >> Hi,<br> >><br> >> 2014-09-17 13:51 GMT+02:00 Sergei Vorobyov<br> >> <<a href="mailto:sergei.vorobyov@facilitylabs.com">sergei.vorobyov@facilitylabs.com</a>>:<br> >> > While GStreamer supports the .wmv on nVidia cards, it apparently does<br> >> > not on<br> >> > Intel's built-in graphics cards found in Intel's NUC2820FYKH (see below:<br> >> > the<br> >> > same .wmv file is gst-discovered on the NUC and on a workstation with<br> >> > the<br> >> > Quadro FX 580 card). What are the prospects of getting fully functional<br> >> > drivers/libraries for Intel NUCs in the nearest future?<br> >> ><br> >> > Thanks!<br> >> ><br> >> > PS: Although Intel does not specify it, I found that the NUC2820 has the<br> >> > Ivy<br> >> > Bridge derived HD Graphics with a 756MHz burst core frequency<br> >> ><br> >> > ---------------------------------------------------------<br> >> > fl@nuc3:~/ads$ gst-discoverer-1.0 621.wmv<br> >> > Analyzing file:///home/fl/ads/621.wmv<br> >> > libva info: VA-API version 0.35.0<br> >> > libva info: va_getDriverName() returns 1<br> >> > libva error: va_getDriverName() failed with operation<br> >> > failed,driver_name=i965<br> >> > Done discovering file:///home/fl/ads/621.wmv<br> >><br> >> This means that, a plain `vainfo' call on your nuc3 system will yield<br> >> the same output. If so, this would mostly report an additional error<br> >> like X11 display not available. This means that, if you have been<br> >> connecting to nuc3 with ssh, and that the remote host has X running,<br> >> the correct command would be DISPLAY=:0.0 vainfo, resp.<br> >> gst-discoverer-1.0. Otherwise, initialization of the VA driver, so<br> >> gst-discoverer-1.0 will fail finding out anything, unless you have a<br> >> SW decoder around (gst-libav).<br> >><br> >> However, there normally is correct support for headless displays. So,<br> >> if you have built gstreamer-vaapi with the VA/DRM backend too, and<br> >> your user belongs to the "video" group, or whatever has access to<br> >> /dev/dri/card0, this should fallback to that headless mode.<br> >><br> >> ><br> >> > Topology:<br> >> > container: Advanced Streaming Format (ASF)<br> >> > video: Windows Media Video 9 Screen<br> >> > audio: Windows Media Audio 9<br> >> ><br> >> > Properties:<br> >> > Duration: 0:00:05.460000000<br> >> > Seekable: yes<br> >> > Tags:<br> >> > container format: ASF<br> >> > audio codec: WMA Version 9<br> >> > video codec: Microsoft Windows Media 9<br> >> ><br> >> > --------------------------------------------------------<br> >> > sergei@wsl:~$ gst-discoverer-1.0 621.wmv<br> >> > Analyzing file:///home/sergei/621.wmv<br> >> > Done discovering file:///home/sergei/621.wmv<br> >> ><br> >> > Topology:<br> >> > container: Advanced Streaming Format (ASF)<br> >> > video: Windows Media Video 9 Screen<br> >> > audio: Windows Media Audio 9<br> >> ><br> >> > Properties:<br> >> > Duration: 0:00:05.460000000<br> >> > Seekable: yes<br> >> > Tags:<br> >> > container format: ASF<br> >> > audio codec: WMA Version 9<br> >> > video codec: Microsoft Windows Media 9<br> >> ><br> >> > _______________________________________________<br> >> > gstreamer-devel mailing list<br> >> > <a href="mailto:gstreamer-devel@lists.freedesktop.org">gstreamer-devel@lists.freedesktop.org</a><br> >> > <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br> >> ><br> >><br> >><br> >> Regards,<br> >> --<br> >> Gwenole Beauchesne<br> >> Intel Corporation SAS / 2 rue de Paris, 92196 Meudon Cedex, France<br> >> Registration Number (RCS): Nanterre B 302 456 199<br> >> _______________________________________________<br> >> gstreamer-devel mailing list<br> >> <a href="mailto:gstreamer-devel@lists.freedesktop.org">gstreamer-devel@lists.freedesktop.org</a><br> >> <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br> ><br> ><br> ><br> > _______________________________________________<br> > gstreamer-devel mailing list<br> > <a href="mailto:gstreamer-devel@lists.freedesktop.org">gstreamer-devel@lists.freedesktop.org</a><br> > <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br> ><br> <br> <br> Regards,<br> --<br> Gwenole Beauchesne<br> Intel Corporation SAS / 2 rue de Paris, 92196 Meudon Cedex, France<br> Registration Number (RCS): Nanterre B 302 456 199<br> _______________________________________________<br> gstreamer-devel mailing list<br> <a href="mailto:gstreamer-devel@lists.freedesktop.org">gstreamer-devel@lists.freedesktop.org</a><br> <a href="http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel" target="_blank">http://lists.freedesktop.org/mailman/listinfo/gstreamer-devel</a><br> </div></div></blockquote></div><br></div>