Recherche avancée

Sur d’autres sites (5798)

• Setting up multi webcam streaming

  28 septembre 2013, par Moritz Mädler

  we are currently planning to connect all our offices via webcams selectable from a website. So I guess the setup could be something like this :

  Webcam 1 \
  
  Webcam 2 - > Streamserver -> Webserver displaying html5 Video 
  
  Webcam n /

  I've made first tests from streaming from Windows using VLC to a central ffserver, but the results where rather bad (after 1-2 frames the stream stopped).

  Do you have any suggestions on
   which stream software I can use on the client/webcam site (Linux or Windows)
   which software I can use on the server side (Linux) for collecting and recoding streams ?

  Thanks,

  Moritz

• Gstreamer Multiple Source with Single Sink Implementation in PythonGST

  11 octobre 2018, par biswajitGhosh

  I’m new to Gstreamer as well as Python-GST also. I have to collect multiple source raw video stream data into a single sink, I don’t know, if it is possible or not.

  Lets explain more my scenario :

  I have 2 video source server for now, one is my webcam and another is just an mp4 file, I did open those source using following command

  Source1 :

  gst-launch-1.0 v4l2src device=/dev/video0 !
  ’video/x-raw,width=640,height=480’ ! x264enc pass=qual quantizer=20
  tune=zerolatency ! rtph264pay ! udpsink host=127.0.0.1 port=5000

  Source2 :

  gst-launch-1.0 filesrc location = file_name.mp4 !
  ’video/x-raw,width=640,height=480’ ! x264enc pass=qual quantizer=20
  tune=zerolatency ! rtph264pay ! udpsink host=127.0.0.1 port=5000

  I’m trying to send both stream data to localhost 5000 port after encoding with H264.

  For receiving I have a Python Sink Server like this :

  import gi
  
  gi.require_version('Gst', '1.0')
  
  gi.require_version('GstApp', '1.0')
  
  from gi.repository import GObject, Gst, GstApp
  
  
  
  GObject.threads_init()
  
  Gst.init(None)
  
  GST_DEBUG="6"
  
  
  
  class Example:
  
      def __init__(self):
  
          self.mainloop   = GObject.MainLoop()
  
          self.pipeline   = Gst.Pipeline.new("Pipeline") 
  
          self.bus        = self.pipeline.get_bus()
  
          self.bus.add_signal_watch()
  
          self.bus.connect("message", self.on_message)
  
          # self.bus.connect('message::eos', self.on_eos)
  
          # self.bus.connect('message::error', self.on_error)
  
          # self.bus.connect("sync-message::element", self.on_sync_message)
  
  
  
          #Create QUEUE elements
  
          self.queue1     = Gst.ElementFactory.make("queue",          None)
  
          self.queue2     = Gst.ElementFactory.make("queue",          None)
  
  
  
          # Create the elements
  
          self.source     = Gst.ElementFactory.make("udpsrc",         None)
  
          self.depay      = Gst.ElementFactory.make("rtph264depay",   None)
  
          self.parser     = Gst.ElementFactory.make("h264parse",      None)
  
          self.decoder    = Gst.ElementFactory.make("avdec_h264",     None)
  
          self.sink       = Gst.ElementFactory.make("appsink",        None)
  
  
  
          # Add elements to pipeline
  
          self.pipeline.add(self.source)
  
          self.pipeline.add(self.queue1)
  
          self.pipeline.add(self.depay)
  
          self.pipeline.add(self.parser)
  
          self.pipeline.add(self.decoder)
  
          self.pipeline.add(self.queue2)
  
          self.pipeline.add(self.sink)
  
  
  
          # Set properties
  
          self.source.set_property('port', 5000)
  
          self.source.set_property('caps', Gst.caps_from_string("application/x-rtp, encoding-name=H264,payload=96"))
  
          self.sink.set_property('emit-signals', True)
  
          # turns off sync to make decoding as fast as possible
  
          self.sink.set_property('sync', False)
  
          self.sink.connect('new-sample', self.on_new_buffer, self.sink)
  
  
  
  
  
  
  
  
  
      def on_new_buffer(self, appsink, data):
  
          print "exec two..."
  
          appsink_sample = GstApp.AppSink.pull_sample(self.sink)
  
          # with open('example.h264', 'a+') as streamer:
  
          buff = appsink_sample.get_buffer()
  
          size, offset, maxsize = buff.get_sizes()
  
          frame_data = buff.extract_dup(offset, size)
  
          print(size)
  
          # flag, info = buff.map(Gst.MapFlags.READ)
  
          # streamer.write(info.data)
  
          # print(info.size)
  
          return False
  
  
  
      def run(self):
  
          ret = self.pipeline.set_state(Gst.State.PLAYING)        
  
          if ret == Gst.StateChangeReturn.FAILURE:
  
              print("Unable to set the pipeline to the playing state.")
  
              exit(-1)
  
  
  
          self.mainloop.run()
  
  
  
      def kill(self):
  
          self.pipeline.set_state(Gst.State.NULL)
  
          self.mainloop.quit()
  
  
  
      def on_eos(self, bus, msg):
  
          print('on_eos()')
  
          self.kill()
  
  
  
      def on_error(self, bus, msg):
  
          print('on_error():', msg.parse_error())
  
          self.kill()
  
  
  
      def on_message(self, bus, message):
  
          t = message.type
  
          if t == Gst.MessageType.EOS:
  
              print "End of Stream :("
  
              self.kill()
  
          elif t == Gst.MessageType.ERROR:
  
              err, debug = message.parse_error()
  
              print "Error: %s" % err, debug
  
              self.kill()
  
  
  
  
  
      def on_sync_message(self, bus, message):
  
          print message.src
  
  
  
  
  
  example = Example()
  
  example.run()

  Firstly AppSink callback does not working, I don’t know why ? I think I have to make that configuration correct into Python Code. Can any one please help to figure it out ?

  Secondly when I tried with FFMpeg and FFPlay I got so many H264 encoding issues like this :

  

  Main confusion is Can GStreamer handle Multiple Source data into a single sink(I need to distinguished each video frame).

  Thanks a lot.

• Http server live streaming in python

  8 juillet 2015, par George Subi

  I want to send the live output of ffmpeg command to every GET request, but I’m novice in python and can’t solve it. This is my code :

  import subprocess # for piping
  
  from http.server import HTTPServer, BaseHTTPRequestHandler
  
  
  
  class RequestHandler(BaseHTTPRequestHandler):
  
  
  
      def _writeheaders(self):
  
          self.send_response(200) # 200 OK http response
  
          self.send_header('Connection', 'keep-alive')
  
          self.send_header('Content-Type', 'video/mp4')
  
          self.send_header('Accept-Ranges', 'bytes')
  
          self.end_headers()
  
  
  
      def do_GET(self):
  
          self._writeheaders()
  
          global p
  
          DataChunkSize = 10000
  
  
  
          print("starting polling loop.")
  
          while(p.poll() is None):
  
              print("looping... ")
  
              stdoutdata = p.stdout.read(DataChunkSize)
  
              self.wfile.write(stdoutdata)
  
  
  
          print("Done Looping")
  
  
  
          print("dumping last data, if any")
  
          stdoutdata = p.stdout.read(DataChunkSize)
  
          self.wfile.write(stdoutdata)
  
  
  
  if __name__ == '__main__':
  
      command = 'ffmpeg -re -i video.avi -an -vcodec libx264 -vb 448k -f mp4 -movflags dash -'
  
      print("running command: %s" % (command, ))
  
      p = subprocess.Popen(command, stdout=subprocess.PIPE, bufsize=-1, shell=True)
  
      print("Server in port 8080...")
  
      serveraddr = ('', 8080) # connect to port 8080
  
      srvr = HTTPServer(serveraddr, RequestHandler)
  
      srvr.serve_forever()

  This is a prove of concept with a video.avi but the real application is with a live signal video.

  When connect to html5 video tag, play the video well, but when open another connection then play the video from the begining and not from the live moment. Also, when close the web, the server print an error :

  BrokenPipeError: [Errno 32] Broken pipe

  Yes, the pipe is broken when cut the connection is obvious. I think that maybe do it with pipes is not correct, but I don’t know how to do it.

  The goal is to run only one ffmpeg command and the output send to every client connected or will do it. Anybody can help me please ?