Recherche avancée

Sur d’autres sites (4846)

• avcodec/pthread_frame : Don't update the first thread ctx before freeing

  18 mai 2022, par Andreas Rheinhardt

  avcodec/pthread_frame : Don't update the first thread ctx before freeing
  Currently, ff_frame_thread_free() uses the last worker thread
  
  to updates the first worker thread via update_context_from_thread()
  
  immediately before freeing all these worker threads. This is
  
  a remnant of the time in which the first worker was special.
  
  (E.g. the first worker shared its AVCodecInternal with the public
  
  AVCodecContext.)
  But these times are over (none of the uses of is_copy matter
  
  for ff_frame_thread_free()) ; nowadays the only thing that
  
  update_context_from_thread() does is referencing a few
  
  buffers/frames and replacing them with other references instead.
  
  These new references will then be freed immediately thereafter
  
  when the first worker thread is freed. Ensuring that the code is
  
  free of double-frees is achieved by using reference-counted structures
  
  (or in case of AVChannelLayouts : by giving each worker its own copy).
  Some archaeology :
  
  a) Updating the first worker thread from the last one used
  
  has been done since frame-threading was added in
  
  37b00b47cbeecd66bb34c5c7c534d016d6e8da24.
  
  b) The precursor to ff_mpv_common_end() checked for is_copy
  
  before freeing pictures (i.e. it only freed them for the first
  
  worker thread).
  
  c) Commits c2dfb1e37cc72bf144545c4410a4621cbff5c4b1 and
  
  e33811bd2686411233cb0eb4a4ee45eb99d7e736 modified the
  
  update_thread_context function of the H.264 decoder
  
  so that it could fail before calling ff_mpeg_update_thread_context().
  
  d) This led to a double free/an assert violation with a H.264
  
  sample for which ff_mpeg_update_thread_context() is not reached
  
  for the final update_context_from_thread(). Commit
  
  a6e4796fbf0aa9b13451a8ef917ecc4e80d1d272 added code to fix this
  
  sample.
  
  e) This issue was fixed (even with the last mentioned commit reverted)
  
  when the H.264 decoder was deMpegEncContextized in commit
  
  b7fe35c9e50e1701274364adf7280bf4a02b092b (merging commit
  
  2c541554076cc8a72e7145d4da30389ca763f32f).
  
  f) mpegvideo.c stopped using is_copy when it was switched to refcounted
  
  frames in 759001c534287a96dc96d1e274665feb7059145d.
  
  g) 1f4cf92cfbd3accbae582ac63126ed5570ddfd37 removed the init_thread_copy
  
  callbacks ; now no FFCodec.close callback checks for is_copy at all
  
  any more.
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavcodec/pthread_frame.c

• How to fix EXT-X-MAP:URI in m3u8 when specifying absolute hls_fmp4_init_filename using ffmpeg ?

  29 février 2020, par Ryan

  Today I found this blog post teaching me about fragmented mp4s (fmp4s).

  Using the ffmpeg docs, I’ve made great progress.

  I run this :

  '/usr/bin/ffmpeg' '-y' '-i' '/myproject/storage/app/sample_media/2020-02-27/Sample Videos 5.mp4' \
  
  '-c:v' 'libx264' '-s:v' '1920x1080' \
  
   '-crf' '20' '-sc_threshold' '0' '-g' '48' '-keyint_min' '48' \
  
  '-hls_list_size' '0' '-hls_time' '10' '-hls_allow_cache' '0' '-b:v' '4889k' \
  
  '-maxrate' '5866k' '-hls_segment_type' 'fmp4' \
  
  '-hls_fmp4_init_filename' '/myproject/public/storage/000000002/init.mp4' \
  
  '-hls_segment_filename' '/myproject/public/storage/000000002/list_1080p_%04d.m4s' \
  
  '-hls_key_info_file' '/myproject/hls_hls.keyInfo' '-strict' '-2' '-threads' '12' \
  
  '/myproject/public/storage/000000002/list_1080p.m3u8'

  And then I use hls.js on my webpage to try to show the video.

  Initially, it won’t load.

  But then I can get the video to load properly if I refresh the page after making this edit to list_1080p.m3u8 :

  Change from :

  #EXT-X-MAP:URI="/myproject/public/storage/000000002/init.mp4"

  To :

  #EXT-X-MAP:URI="init.mp4"

  How can I change my ffmpeg command so that it knows the absolute path of where to save the init.mp4 but also knows to write in the m3u8 only the filename init.mp4 without any path ?

  (I’ve tried using -hls_fmp4_init_filename init.mp4 without the abosolute path, but then it creates the init.mp4 at the root, which also doesn’t work. I need each video’s init.mp4 to be in its own folder.)

• How do I accurately extend the duration of a tiny video clip ?

  18 avril 2018, par joeycato

  I noticed that ffmpeg doesn’t always scale the duration of a clip to the desired value ( even accounting for milliseconds of precision loss.) For example, whenever I attempt to scale the duration of this tiny 67ms clip to 7 seconds :

  ffmpeg -i input.avi -filter:v "setpts=(7/0.067)*PTS" -an output.avi    

  the resulting video is only 3.5 seconds in duration.

  Not really sure why this is off by multiple seconds, but I suspect it has something to do with the input clip itself ( which is only 2 frames @ 29.97 fps ) So far, I’ve discovered that if I scale the duration to a much larger value first, then scale back from that to 7 seconds, it works fine :

  ffmpeg -i input.avi -filter:v "setpts=(1000)*PTS" -an input_scaled_high.avi
  
  ffmpeg -i input_scaled_high.avi -filter:v "setpts=(7/33.400033)*PTS" -an output.avi

  But I’m hoping I don’t need to resort to an intermediate file. Is it possible to solve this with a single ffmpeg call ?

  If it helps, here is the video file I’m working with : https://drive.google.com/drive/folders/1hI5Xo6kfAfMd8ZylM6XrPX5fuO88gncE?usp=sharing

  Note : I should mention that I don’t need sound at all in this experiment ( hence the -an )