Advanced search

On other websites (11427)

• How to pass multiple RTP options to ffmpeg?

  6 November 2017, by MSalters

  Building my own RTSP server for FFMPEG, so I’m executing ffmpeg as a child process.

  The problem I currently have is that I’m adding multicast support, and an RTSP client may add ttl to the RTSP Transport line. No problem so far, as ffmpeg supports that. But exactly how do I pass it? The documented URL format is

   rtp://hostname[:port][?option=val...]

  That’s not the sort of definition you should write if you want to pass a Comp.Sci class. The ellipsis suggests you can pass more than one parameter, but not how. And I need not just ttl= but also localrtpport=.

  I suppose I could follow HTTP conventions and assume that they intended [?option=val[&option=val]*] but I can’t find an authoritative source for that.

  Asked elsewhere but unanswered there too.

• FFmpeg Capture remote screen

  29 August 2018, by ricardo

  I am trying capture a remote screen via ffmpeg and x11grab, I run in my machine this command

  xpra start-desktop :20 --start-child=fluxbox

  this machine have the ip 192.168.1.15

  and in the remote capture I try run this

  /usr/local/bin/ffmpeg -f pulse -server 192.168.1.15 -i tarjeta01.monitor -f x11grab -framerate 25 -r 25 -i 192.168.1.15:20.0 -fflags nobuffer -f rtp -c:v h264_nvenc -preset llhp -profile:v baseline -level 4 -delay 0 -b:v 1500k -threads 4 -cbr 1 -r 25 -an udp://:5008 -f rtp -vn -c:a libopus -ar 48000 -ac 2 -ab 96k -application lowdelay -compression_level 0 -frame_duration 2.5 -cutoff 20000 -vbr constrained udp://localhost:5006

  I make xhost + and xhost + 192.168.1.16 in the machine where run xpra but always receive this error

  [x11grab @ 0x23e4ea0] Cannot open display 192.168.1.15:20, error 1.
  
  192.168.1.15:20: Input/output error

  I try xpra, Xephyr and xfvb but always get the same error

  Thanks for all best regards

• Python - check subprocess activity every n seconds

  22 November 2017, by hlabor

  I’m trying to make python script (currently on windows) which will open some sub-processes (which will run infinitely) and script should periodically check do all of opened sub-processes still work correctly. So it should be done with while loop, I guess.

  The sub-processes are about FFMPEG livestreaming.

  The problem is when I do time.sleep(n) in my loop, because then every FFMPEG livestream stops, so I suppose time.sleep affect on all of child subprocesses.

  I have no idea how to make it work.

  Here is my python code:

  import os, time, sys, datetime, smtplib, configparser, logging, subprocess, psutil
  
  import subprocess
  
  
  
  def forwardudpstream(channel_number, ip_input, ip_output):
  
      try:
  
          ffmpeg_command = 'ffmpeg -i udp://' + ip_input + ' -vcodec copy -acodec copy -f mpegts "udp://' + ip_output + '?pkt_size=1316"'
  
          ffmpeg_output = subprocess.Popen(ffmpeg_command, stdout=subprocess.PIPE, stderr=subprocess.STDOUT, shell=False)
  
          return str(ffmpeg_output.pid)
  
      except:
  
          print ("Exception!")
  
          return '0'
  
  
  
  while True:
  
      configuration = 'config.ini'
  
      channel_list_file = 'CHANNEL_LIST.conf'
  
      pid_folder = "D:\\Forward_UDP_Stream\\pids\\"
  
      channel_list = [line.rstrip('\n') for line in open(channel_list_file)]
  
      for line in channel_list:
  
          if not line.startswith('#') and ('|' in line):
  
              channel_number, ip_input, ip_output = line.split('|')
  
              print('----------')
  
              print("Channel number = ", channel_number)
  
              print("IP Input = ", ip_input)
  
              print("IP Output = ", ip_output)
  
              pid_file_found = False
  
              print("Checking if pid file exists...")
  
              for pidfile in os.listdir(pid_folder):
  
                  if pidfile.startswith(channel_number + '-'):
  
                      print("Pid file is found for this channel.")
  
                      pid_file_found = True
  
                      pid = int(pidfile.split('-')[1].split('.')[0])
  
                      print("PID = ", str(pid))
  
                      print("Checking if corresponding process is active...")
  
                      if not psutil.pid_exists(pid):
  
                          print("Process is not active.")
  
                          print("Removing old pid file.")
  
                          os.remove(pid_folder + pidfile)
  
                          print("Starting a new process...")
  
                          pid_filename = channel_number + '-' + forwardudpstream(channel_number, ip_input, ip_output) + '.pid'
  
                          pid_file = open(pid_folder + pid_filename, "a")
  
                          pid_file.write("Process is running.")
  
                          pid_file.close()
  
                      else:
  
                          print("Process is active!")
  
                      break
  
              if pid_file_found == False:
  
                  print("Pid file is not found. Starting a new process and creating pid file...")
  
                  pid_filename = channel_number + '-' + forwardudpstream(channel_number, ip_input, ip_output) + '.pid'
  
                  pid_file = open(pid_folder + pid_filename, "a")
  
                  pid_file.write("Process is running.")
  
                  pid_file.close()
  
              time.sleep(10)

  Here is my CHANNEL_LIST.conf file example:

  1|239.1.1.1:10000|239.1.1.2:10000
  
  2|239.1.1.3:10000|239.1.1.4:10000

  Perhaps there is some other solution to put waiting and sub-processes to work together. Does anyone have an idea?

  UPDATE:
  I finally make it work when I removed stdout=subprocess.PIPE part from the subprocess command.
  Now it looks like this:

  ffmpeg_output = subprocess.Popen(ffmpeg_command, stderr=subprocess.STDOUT, shell=False)

  So now I’m confused why previous command was making a problem...?
  Any explanation?