Recherche avancée

Sur d’autres sites (8960)

• avformat/matroskaenc : Write data directly into dynamic buffers

  26 novembre 2019, par Andreas Rheinhardt

  avformat/matroskaenc : Write data directly into dynamic buffers
  This avoids copying the data in small chunks (1024B) into
  
  the dynamic buffer's small buffer before finally writing them
  
  into the "big" buffer.
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavformat/matroskaenc.c

• avcodec/flashsv : Avoid deflating data

  11 mars 2022, par Andreas Rheinhardt

  avcodec/flashsv : Avoid deflating data
  Currently priming the zlib decompressor involves compressing
  
  data directly after having decompressed it and decompressing
  
  it again in order to set the "dictionary" and to initialize
  
  the adler32-checksum. Yet this is wasteful and can be simplified
  
  by synthetizing the compressed data via non-compressed blocks.
  This reduced the amount of allocations for the decoding part
  
  of fate-vsynth1-flashsv2, namely from
  
  total heap usage : 9,135 allocs, 9,135 frees, 376,503,427 bytes allocated
  
  to
  
  total heap usage : 2,373 allocs, 2,373 frees, 14,144,083 bytes allocated
  Signed-off-by : Andreas Rheinhardt <andreas.rheinhardt@outlook.com>
  

  • [DH] libavcodec/flashsv.c

• What does "Error in the pull function" mean while executing ffmpeg command ?

  8 avril 2022, par Alex Rypun

  I'm trying to generate a video from multiple video sources (https urls).

  


  If I use say 8 video sources it works as expected. But if I need say 30 sources, it starts to work as expected, but at some moment a lot of errors appear.

  


  

  


  I believe the main one is Error in the pull function..

  


  Such a video-source appears in the output file but not a full requested part (e.g., I set trim=0:5, but ffmpeg generates only 3 seconds).

  


  I tried different videos, different cropped parts, loaded sources to aws s3, but can't identify a "bad case", it's reproduced with different options.

  


  My command looks like this :

  


  ffmpeg -y
-i "https://player.vimeo.com/external/399541658.hd.mp4?s=5432956383527af00bb74b41c120bbb84ff3ac5e&profile_id=175&oauth2_token_id=57447761"
...
-i "https://player.vimeo.com/external/464966383.hd.mp4?s=648aa8277259c499f1d05d6330f9922932c080a6&profile_id=175&oauth2_token_id=57447761"
-i "https://test.com/development/videos/dc3c32cf9a194db1bb52/sources/5q5e3j2h57.mp3"
-filter_complex "[0:v]trim=0:5.64,setpts=PTS-STARTPTS,scale=1920x1080,setdar=16/9[v0];[1:v]trim=0:8,setpts=PTS-STARTPTS,scale=1920x1080,setdar=16/9,loop=0:32767[v1]; ... [10:v]trim=0:13,setpts=PTS-STARTPTS,scale=1920x1080,setdar=16/9[v10];[v0] ... [v10] concat=n=11:v=1 [video];[11:a] atrim=0:99.29 [audio]"
-vsync 2 -map [video] -map [audio] output.mp4


  


  When I downloaded all sources to my local machine and tried to use them in the ffmpeg command it worked fine without errors.

  


  Also, I figured out that it depends on the CPU. When I restricted my CPU the errors appeared earlier. And some commands might produce errors with a restricted CPU but finished successfully with no CPU restriction.

  


  As I understand, for some reason, ffmpeg fails to receive a particular part of a file data, but I believe it should retry to receive lost packages (it's https).

  


  And I don't understand, how it could be related to CPU or sources quantity.

  


  By not understanding the core of evil, I can think about 2 ways (both ugly) :

  


  

  1. to download all sources and generate the final video from local files ;

  


  2. to chunk sources, generate intermediate outputs for each chunk (and save locally), and concatenate them into the final one.

  


  


  Any better suggestions or at least where to seek ?