Recherche avancée

Sur d’autres sites (3310)

• avcodec/hevc_filter : Pass HEVCLocalContext when slice-threading

  29 juin 2022, par Andreas Rheinhardt

  avcodec/hevc_filter : Pass HEVCLocalContext when slice-threading
  The HEVC decoder has both HEVCContext and HEVCLocalContext
  
  structures. The latter is supposed to be the structure
  
  containing the per-slicethread state.
  Yet that is not how it is handled in practice : Each HEVCLocalContext
  
  has a unique HEVCContext allocated for it and each of these
  
  coincides with the main HEVCContext except in exactly one field :
  
  The corresponding HEVCLocalContext.
  
  This makes it possible to pass the HEVCContext everywhere where
  
  logically a HEVCLocalContext should be used.
  This commit stops doing this for lavc/hevc_filter.c ; it also constifies
  
  everything that is possible in order to ensure that no slice thread
  
  accidentally modifies the main HEVCContext state.
  There are places where this was not possible, namely with the SAOParams
  
  in sao_filter_CTB() or with sao_pixels_buffer_h in copy_CTB_to_hv().
  
  Both of these instances lead to data races, see
  
  https://fate.ffmpeg.org/report.cgi?time=20220629145651&slot=x86_64-archlinux-gcc-tsan-slices
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavcodec/hevc_filter.c
  • [DH] libavcodec/hevcdec.c
  • [DH] libavcodec/hevcdec.h

• Prevent suspend event when streaming video via HTML video tag

  24 septembre 2014, par jasongullickson

  I seem to be having the opposite problem of most people who are streaming video using the HTML video tag ; I’m saturating the client with data.

  When playing a long video served via ffserver (webm container) everything works great but eventually the browser (Chrome in this case) will begin throwing "suspend" events. After a number of these ( 50-100), a "stalled" event will fire and playback will stop.

  I believe the problem is that once Chrome has buffered a certain amount of video it goes into "suspend" and stops downloading more data. I’ve tested this theory by throttling the speed at which video data is delivered, and if I keep the delivered frame rate close to the playback rate, I can prevent this from happening, but of course deliberately holding back server performance isn’t ideal.

  What I’m looking for is either a way to suppress this "suspend" behavior altogether, or alternatively a way to respond to the event that prevents the eventual "stalled" state.

  Presumably the browser at some point exits the "suspend" state and begins requesting data again, but I haven’t actually observed this occurring. I’m using a chain of mpeg2 -> ffmpeg -> ffserver to stream the video so if the browser is attempting to resume loading data I don’t see the request in my application. I could use a proxy or a sniffer to watch for the traffic but I would expect that maybe there is an ffserver log that can tell me the same thing ? In any event if it’s attempting to resume the download it’s failing, and there’s no indication server-side that there’s a reason for the request to fail (in fact I can pull up the same video feed from ffserver and see it playing correctly).

  So I feel like I’ve isolated this to a client-side playback issue, and one where the browser is voluntarily giving up on loading the data, but I’m not sure how to convince it to "not do that", or at least attempt to resume when it runs the buffer dry.

• avcodec/hevc_mvs : Pass HEVCLocalContext when slice-threading

  29 juin 2022, par Andreas Rheinhardt

  avcodec/hevc_mvs : Pass HEVCLocalContext when slice-threading
  The HEVC decoder has both HEVCContext and HEVCLocalContext
  
  structures. The latter is supposed to be the structure
  
  containing the per-slicethread state.
  Yet that is not how it is handled in practice : Each HEVCLocalContext
  
  has a unique HEVCContext allocated for it and each of these
  
  coincides except in exactly one field : The corresponding
  
  HEVCLocalContext. This makes it possible to pass the HEVCContext
  
  everywhere where logically a HEVCLocalContext should be used.
  This commit stops doing this for lavc/hevc_mvs.c ; it also constifies
  
  everything that is possible in order to ensure that no slice thread
  
  accidentally modifies the main HEVCContext state.
  Reviewed-by : Anton Khirnov <anton@khirnov.net>
  
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavcodec/hevc_mvs.c
  • [DH] libavcodec/hevcdec.c
  • [DH] libavcodec/hevcdec.h