Recherche avancée

Sur d’autres sites (5292)

• How to convert raw H.264 stream from video call to mp4 file

  13 février 2020, par Daniel

  I have captured a SIP point to point video call using wireshark and I used the program ’videosnarf’ on Ubuntu 12.04 to extract the raw H.264 stream from the PCAP. I am having some problems with ffmpeg when trying to convert it to MP4. Below is the ffprobe output :

  [STREAM]
  
  index=0
  
  codec_name=h264
  
  codec_long_name=H.264 / AVC / MPEG-4 AVC / MPEG-4 part 10
  
  codec_type=video
  
  codec_time_base=1/2400000
  
  codec_tag_string=[0][0][0][0]
  
  codec_tag=0x0000
  
  width=704
  
  height=396
  
  has_b_frames=0
  
  sample_aspect_ratio=1:1
  
  display_aspect_ratio=16:9
  
  pix_fmt=yuv420p
  
  level=51
  
  r_frame_rate=25/1
  
  avg_frame_rate=25/1
  
  time_base=1/1200000
  
  start_time=N/A
  
  duration=N/A
  
  [/STREAM]

  Here is the command I used to generate the MP4 :

  ffmpeg -f h264 -i H264-media-3.264 -vcodec copy output.mp4

  When I transfer the file to my Windows PC, I can’t open the file in VLC. When opening in Windows Media Player (not sure the version but it’s on Windows 8) it plays but very quickly (almost as if it’s playing in fast forward. I also tried with :

  ffmpeg -f h264 -i H264-media-3.264 -vcodec copy -r 25 output.mp4

  I am able to play the raw H.264 stream using "MPC-HC" on Windows but I need it in MP4 format.

• Output FFMPEG segments to GCS

  28 janvier 2019, par Terrabyte

  Im trying to output an ffmpeg hls stream to google cloud storage after each segment is created or at least after the entire encoding is done. Doesn anyone know how to accomplish this ?

  This is what i have in my bash script :

  $FFMPEG -hide_banner -i "$infile" \
  
      $PASSVAR \
  
      -y \
  
      -codec:v "$VIDEO_CODEC" \
  
      -codec:a "$AUDIO_CODEC" \
  
      -threads "$NUMTHREADS" \
  
      -map 0:v \
  
      -map 0:a \
  
      -flags \
  
      -global_header \
  
      -f segment \
  
      -segment_list "$playlist_name" \
  
      -segment_time "$SEGLENGTH" \
  
      -segment_format mpeg_ts \
  
      -g "$key_frames_interval" \
  
      -keyint_min "$key_frames_interval" \
  
      $resolution \
  
      $bitrate \
  
      $FFMPEG_ADDITIONAL \
  
      $FFMPEG_FLAGS \
  
      "$OUTPUT_DIRECTORY/$output_name" | gsutil -m mv "$OUTPUT_DIRECTORY/$output_name" gs://bucket-vod-example
  
  }

  does anyone know how to output the files to gsutls as it’s being encoded ? I can make the output the mounted storage point of google cloud storage but encoding directly to it is extremely slow so if possible is there way to encode locally then transfer as it’s being encoded ?

• How OpenGL directly manipulates data on GPU that a video hardware decoded by FFmpeg CUDA

  24 janvier 2019, par gn cavalry

  Ubuntu 18.04
  FFmpeg 4.1
  CUDA 10.0

  What I want to do is using FFmpeg CUDA decode a video, directly peocess data on GPU by GLSL , then encode data using FFmpeg CUDA. Like this :

  Raw Video —> FFmpeg CUDA decoder —> GLSL —> FFmpeg CUDA encoder —> New Video

  Here is the short code :

  ret = avcodec_send_packet(decoder_ctx, pkt);
  
  while (ret >= 0) {
  
      ret = avcodec_receive_frame(decoder_ctx, frame);
  
  
  
      /**
  
       * Now CUDA decoded data is stored on GPU and referenced by frame
  
       * How can GLSL manipulate decoded data on GPU directly, 
  
       * without transfer it to CPU and load to GPU again by glTexImage2D()
  
       */
  
  
  
  
  
      // I don't if the following code is right, if GLSL processed data is still referenced by frame
  
      while (avcodec_send_frame(encoder_ctx, frame)) {
  
          avcodec_receive_packet(encoder_ctx, &enc_pkt);
  
      }
  
  }

  Can anyone help me ? Thanks very much !