Recherche avancée

Sur d’autres sites (7877)

• Is FFmpegAudioDecoder supposed to reinitialize upon append of new init segment

  14 novembre 2023, par martin

  I am attempting to switch audio tracks but when switching the FFmpegAudioDecoder never reinitializes like it does with video tracks of differing resolutions. I am not certain if this is the intended behavior of FFmpegAudioDecoder and would love to learn more about the expected behavior.

  


  When switching audio tracks I end up calling the following operations :

  


  if sourceBuffer.getIsUpdate() {sourceBuffer.abort()}
sourceBuffer.remove(0-videoDuration)
initSegmentDataStream = fetch init segment of new audio representation
sourceBuffer.appendBuffer(initSegmentDataStream)


  


  These are the Media tab messages from initial video load

  


  ChunkDemuxer
Selected FFmpegAudioDecoder for audio decoding, config: codec: aac, profile: unknown, bytes_per_channel: 2, channel_layout: STEREO, channels: 2, samples_per_second: 48000, sample_format: Signed 16-bit, bytes_per_frame: 4, seek_preroll: 0us, codec_delay: 0, has extra data: false, encryption scheme: Unencrypted, discard decoder delay: false, target_output_channel_layout: STEREO, target_output_sample_format: Unknown sample format, has aac extra data: true
Cannot select DecryptingVideoDecoder for video decoding
Cannot select VDAVideoDecoder for video decoding
Cannot select VpxVideoDecoder for video decoding
Selected Dav1dVideoDecoder for video decoding, config: codec: av1, profile: av1 profile main, level: not available, alpha_mode: is_opaque, coded size: [1280,720], visible rect: [0,0,1280,720], natural size: [1280,720], has extra data: false, encryption scheme: Unencrypted, rotation: 0°, flipped: 0, color space: {primaries:BT709, transfer:BT709, matrix:BT709, range:LIMITED}
Dropping audio frame (DTS 0us PTS -105375us,-62709us) that is outside append window [0us,9223372036854775807us).
Dropping audio frame (DTS 42666us PTS -62708us,-20042us) that is outside append window [0us,9223372036854775807us).
Truncating audio buffer which overlaps append window start. PTS -20041us frame_end_timestamp 22625us append_window_start 0us
Effective playback rate changed from 0 to 1


  


  For comparison this is what I get when appending the init segment of a different video resolution / track

  


  video decoder config changed midstream, new config: codec: av1, profile: av1 profile main, level: not available, alpha_mode: is_opaque, coded size: [1920,1080], visible rect: [0,0,1920,1080], natural size: [1920,1080], has extra data: false, encryption scheme: Unencrypted, rotation: 0°, flipped: 0, color space: {primaries:BT709, transfer:BT709, matrix:BT709, range:LIMITED}



  


  Chrome version : Version 119.0.6045.123 (Official Build)

  


  When appending the new init segment of an audio track I was expecting the FFmpegAudioDecoder to be reinitialized like the Dav1dVideoDecoder does for video tracks

  


• Understanding avio_alloc_context

  1er février 2018, par Javier Ramírez

  How does the avio_alloc_context function work ? Should this function be used only once when starting the stream, or in the main loop ?

  When I use it only when I open my file of my class, in my drawing function (video to OpenGL texture) in main loop, it shows only the first frame, and then my window closes, indicating that there was an error in the function :

  av_read_frame (parameters...);

  I have also noticed that the avio_alloc_context function is used to seek a frame of a video. But is this just to open a file, or is it to do it in the main loop ?

• Trouble getting video stream from Bebop 2 (python, open cv and ffmpeg)

  20 décembre 2018, par Seb A

  I’m currently working with the Bebop 2 drone from parrot and I would like to get the video stream from the bebop and to use the frame I get. To control the bebop I’m using this library : https://github.com/amymcgovern/pyparrot.

  First I’ve tried to use opencv by opening the sdp file used by the bebop to stream the video with this code :

  capture = cv2.VideoCapture('./bebop.sdp')
  
  
  
      while True :
  
          rt,img = capture.read()
  
          cv2.imshow("cam",img)
  
          cv2.waitKey(10)

  here is the content of the sdp file :

  c=IN IP4 192.168.42.1
  
  m=video 55004 RTP/AVP 96
  
  a=rtpmap:96 H264/90000

  but I got this error

  [rtp @ 0000020b890b3300] Protocol 'rtp' not on whitelist 'file,crypto'!
  
  warning: Error opening file (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:808)
  
  warning: ./bebop.sdp (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:809)
  
  False
  
  OpenCV Error: Assertion failed (size.width>0 && size.height>0) in cv::imshow, file C:\projects\opencv-python\opencv\modules\highgui\src\window.cpp, line 331
  
  Traceback (most recent call last):
  
   File "pilotage.py", line 109, in <module>
  
     _main_(args)
  
   File "pilotage.py", line 56, in main
  
     cv2.imshow('drone view',im)
  
  cv2.error: C:\projects\opencv-python\opencv\modules\highgui\src\window.cpp:331: error: (-215) size.width>0 &amp;&amp; size.height>0 in function cv::imshow
  
  </module>

  I’ve seen that this is due to the ffmpeg command used by opencv that does not include the option -protocol_whitelist
  So I tried to get the image directly with ffmpeg by using this command

  "ffmpeg -protocol_whitelist \"file,rtp,udp\"  -i  ./bebop.sdp -r 30 -b:v 800k image_%03d.png &"

  But the images I got were awful and useless.
  Like this one :
  

  So I tried different options to improve the quality but I can’t get it to work.
  Thanks for your help.

  I’m on windows 10 and using latest python version