Recherche avancée

Sur d’autres sites (9659)

• FFmpeg efficient capture from rtsp ipcamera

  29 août 2020, par Jax2171

  I need to capture an audio/video rtsp stream uncompressed in a file from ipcamera. Audio (pcm_alaw) and video (h264) must be synchronized. It is necessary that the file does not get corrupted if the camera loses the connection for a few moments (mp4).

  


  At the moment I use the command below, but the ts codec does not support pcm_alaw and therefore the audio is not heard :

  


  ffmpeg -stimeout 2000000 -rtsp_transport tcp -i rtsp://admin:1234@192.168.5.22/h264 -c:v copy -c:a copy -f mpegts -y main.ts


  


  I use the mpegts codec because I need to check the duration of the capture in real time with the command :

  


  ffprobe -i /home/pi/NAS/main.mov -show_entries format=duration -v quiet -of csv="p=0"


  


  If i use mkv or avi its output would be :

  


  N/A


  


  The verification of the duration is important because I capture files of about 3 hours and at my choice I perform some data while the capture is in progress. I prefer not to compress the audio because I have often noticed some asynchrony with respect to the video when cutting.

  


  Thank you.

  


• Sending an mp4 video over the network with avconv command to a raspberry pi in a piwall videowall

  21 juillet 2017, par noswoscar

  I would like to send an mp4 video from a raspbery pi to another raspberry-pi (in order to build a piwall) but the command I am using for sending a video stream (tested with correct outcome on other devices) doesn’t work.

  Details :
  I am using the avconv command from libav.org to send an mp4 video from a "master raspberry pi" through the network to my "slave raspberry pi". The raspberry pi (which is receiving the video) has the Raspbian Jessie Operating System on it. The goal of my experiment is to use the pwomxplayer command to display the video received on the "slave raspberry pi".

  The full command i am using for sending the mp4 video stream is as follows :
  avconv -re -i input.mp4 -vcodec libx264 -f h264 -an udp ://224.0.0.1:1234
  (works when sending to my raspberry pi from my dell laptop)

  It is interesting to note that when using the pre-mentioned avconv command on my dell laptop which has an intel processor and debian OS installed, the video stream is sent properly. However, the same command doesn’t work when sending the video stream from a raspberry pi which has an arm processor and raspbian OS installed.
  I wonder why this is !

  Am I right in thinking the arm processor is not interpreting my command avconv as well as the intel processor ? Or is it more a question of architecture ? Or is my avconv command syntax incorrect ?

  Thank you for your help !

  More info :
  Ip address of master raspberry pi : 192.168.72.10
  Ip address of slave raspberry pi : 192.168.72.11
  I am using the route add -net 224.0.0.0 netmask 224.0.0.0 eth0 command on master pi to send the video stream to all laptops connected to my ethernet port.
  The following command works for files ending in .h264 :
  avconv -re -i input.h264 -vcodec libx264 -f h264 -an udp ://224.0.0.1:1234
  (but I would like to send .mp4 files encoded with h264 not just .h264 files)

• Audio out of sync after merging video chunks extracted with ffmpeg using -c copy and -reset_timestamps 1 [closed]

  8 décembre 2024, par Tree

  I’m splitting a long video (around 2 hours) into smaller chunks using ffmpeg commands like :

  


  ffmpeg -i input.mp4 -ss START_SECOND -t LENGTH -c copy -reset_timestamps 1 -y chunk_X.mp4


  


  Then, I’m concatenating the resulting chunks using a concat list file :

  


  ffmpeg -f concat -safe 0 -i list.txt -c copy -y merged_output.mp4


  


  So the chunks always come from the same file, and have the same codec/size/bitrate.

  


  My goal is to avoid re-encoding the entire video to keep the process fast. However, I’m noticing that after the final merge, the audio in some parts of the merged video is out of sync with the video. The first chunk seems fine, but from the second chunk onward, the audio often drifts out of sync.

  


  What could be causing the audio desynchronization, and how can I correctly extract and merge the chunks without losing A/V sync, while still avoiding a full re-encode ?

  


  I tried to first pass in -ss before -i to cut on keyframes, but I get the same result.