Recherche avancée

Sur d’autres sites (5341)

• Lost video stream when streaming using FFmpeg and RTSP camera

  13 février 2019, par Vape

  on the Linux server, I have FFmpeg installed which streams video from Chinese low-cost IP camera to Twitch or Youtube server. After a few hours, the video is not visible but on the server side, the FFmpeg is still running and also the IP camera respond to the "ping" command.

  Here is the script I’m using :

  #
  
  # Camera IP
  
  #
  
  AQUARIUM_CAM_IP="192.168.123.102"
  
  
  
  
  
  #
  
  # Aquarium data file
  
  #
  
  AQUARIUM_DATA_FILE="/run/aquarium-cam/data.txt"
  
  
  
  
  
  #
  
  # FFmpeg parameters
  
  #
  
  FFMPEG_LOG_LEVEL=fatal
  
  
  
  # Bitrate (1000k = 1Mbit/s)  and  encoding speed (affects CPU)  and  number of CPU cores to use
  
  FFMPEG_VBR="1000k"
  
  FFMPEG_QUAL="ultrafast"
  
  FFMPEG_THREADS="1"
  
  
  
  # Streaming source
  
  FFMPEG_CAM_RTSP_SRC="rtsp://${AQUARIUM_CAM_IP}:554/user=admin&password=&channel=1&stream=0.sdp" # Camera source
  
  
  
  # Streaming destination
  
  FFMPEG_TWITCH_STREAM_URL_DST="rtmp://live-ber.twitch.tv/app"  # RTMP stream URL
  
  FFMPEG_TWITCH_KEY=""
  
  
  
  # Data overlay setup
  
  FFMPEG_TEXT_OVERLAY_FONT_PATH="OpenSans-Regular.ttf"
  
  FFMPEG_TEXT_OVERLAY_FONT_SIZE=25
  
  FFMPEG_TEXT_OVERLAY_OFFSET_X=5
  
  FFMPEG_TEXT_OVERLAY_OFFSET_Y=60
  
  FFMPEG_TEXT_OVERLAY_RELOAD=1
  
  FFMPEG_TEXT_OVERLAY_BOX="1"
  
  FFMPEG_TEXT_OVERLAY_BOX_BORDER_WIDTH="5"
  
  FFMPEG_TEXT_OVERLAY_BOX_COLOR="blue@0.5"

  the FFmpeg script :

   ffmpeg \
  
      -loglevel ${FFMPEG_LOG_LEVEL} -f lavfi -i anullsrc \
  
      -rtsp_transport tcp \
  
      -i "${FFMPEG_CAM_RTSP_SRC}" \
  
      -vcodec libx264 -pix_fmt yuv420p -preset ${FFMPEG_QUAL} -g 75 -b:v ${FFMPEG_VBR} \
  
      -vf "\
  
  drawtext=fontfile=${FFMPEG_TEXT_OVERLAY_FONT_PATH}:textfile=${AQUARIUM_DATA_FILE}:\
  
  x=${FFMPEG_TEXT_OVERLAY_OFFSET_X}:y=${FFMPEG_TEXT_OVERLAY_OFFSET_X}:\
  
  reload=${FFMPEG_TEXT_OVERLAY_RELOAD}: \
  
  fontcolor=white:fontsize=${FFMPEG_TEXT_OVERLAY_FONT_SIZE}:\
  
  box=${FFMPEG_TEXT_OVERLAY_BOX}:boxborderw=${FFMPEG_TEXT_OVERLAY_BOX_BORDER_WIDTH}:\
  
  boxcolor=${FFMPEG_TEXT_OVERLAY_BOX_COLOR}"\
  
      -threads ${FFMPEG_THREADS} -bufsize 512k \
  
      -f flv "${FFMPEG_TWITCH_STREAM_URL_DST}/${FFMPEG_TWITCH_KEY}"

  The other strange thing is that when the FFmpeg start and the stream begin the CPU utilization is about 30% in the case when there is no stream but the FFmpeg is still alive the CPU utilization is below 20% or less.

  Any Idea how to resolve this kind of problem ?
  Has FFmpeg some option to terminate if there is no "stream" or if it has lost a connection with a camera ?

• swscale : add two spatially stable dithering methods

  23 mars 2014, par Øyvind Kolås

  swscale : add two spatially stable dithering methods
  Both of these dithering methods are from http://pippin.gimp.org/a_dither/ for
  
  GIF they can be considered better than bayer (provides more gray-levels), and
  
  spatial stability - often more than twice as good compression and less visual
  
  flicker than error diffusion methods (the methods also avoids error-shadow
  
  artifacts of diffusion dithers).
  These methods are similar to blue/green noise type dither masks ; but are
  
  simple enough to generate their mask on the fly. They are still research work
  
  in progress ; though more expensive to generate masks (which can be used in a
  
  LUT) like ’void and cluster’ and similar methods will yield superior results
  

  • [DH] doc/scaler.texi
  • [DH] libswscale/options.c
  • [DH] libswscale/output.c
  • [DH] libswscale/swscale_internal.h
  • [DH] libswscale/utils.c

• canvas to ffmpeg - invalid data error

  12 juillet 2017, par nameless

  I’m curently trying to pipe some raw data directly into ffmpeg. The data comes from a canvas. What I do is the following :

  var imageData = ctx.getImageData(0,0,600,600);
  
  var dataArray = imageData.data;
  
  var rgbArray = [];
  
  for (var i = 0; i < dataArray.length; i+=4) {
  
      rgbArray.push([dataArray[i], dataArray[i+1], dataArray[i+2]])
  
  }
  
  
  
  var rgb24Array = rgbArray.map(function(rgbList){
  
      return (rgbList[0] << 16) | (rgbList[1] << 8) | (rgbList[2])
  
  });

  The ffmpegArgs are set like this :

  var ffmpegArgs = [
  
      '-c:v', 'rawvideo' // input container
  
      '-f', 'rawvideo',
  
      '-pix_fmt', 'rgb24', // input pixel format
  
      '-s', '600x600' //input size
  
      '-i', 'pipe:0', // input source
  
      '-format', 'mp4', // output container format
  
      '-c:v', 'libx264', // output video codec
  
      '-b:v', '2m', // output bitrate
  
      'udp://239.255.123.46:1234' // output destination
  
  ];

  So basically, the imageData I get from the canvas has 4 elements for each pixel (rgba), I first filter out the rgb values and then pack them to send them to ffmpeg, I directly pipe them in like ffmpeg.stdin.write(rgb24Array).

  But I get a TypeError: invalid data and I’m not sure why... I also tried to leave data as it is and use rgba as pix_fmt, but same error.

  The stacktrace of the error is :

  TypeError: invalid data
  
     at Socket.write (net.js:617:11)
  
     at null._repeat (....line with ffmpeg.stdin.write() in it)
  
     at wrapper [as_onTimeout] (timers.js:275:11)
  
     at Timer.listonTimeout (timers.js:92:15)

  Does anybody have a idea where the problem could be ?

  Edit :
  changed some things, first, I now use a Uint8Array :

  var imageData = ctx.getImageData(0,0,600,600);
  
              var srcArray = imageData.data;
  
              var dstArray = new Uint8Array(imageData.width * imageData.height * 3);
  
              for(var i = 0, p = 0; i < srcArray.length; i++) { // i++ skips alpha
  
                dstArray[p++] = srcArray[i++];                   // red comp. and incr.
  
                dstArray[p++] = srcArray[i++];                   // green comp. and incr.
  
                dstArray[p++] = srcArray[i++];                   // blue comp. and incr.
  
              }
  
  
  
              ffmpeg.stdin.write(dstArray);

  Also added -video_size parameter to the ffmpegArgs. But still it’s not working and I get the same errors.