Recherche avancée

Sur d’autres sites (5255)

• Merge "Experimental change to help with ARNR problem."

  14 janvier 2011, par Paul Wilkins

  Merge "Experimental change to help with ARNR problem."

• Problem with ffmpeg rtp stream to janus webrtc

  13 juin 2020, par XPModder

  I am trying to use ffmpeg and janus-gateway to stream video in the local network. I am piping the h264 video directly in to ffmpeg and from there it gets transferred to janus as an rtp stream. Janus then does the rest.

  



  The problem is, that when I try to open the stream using the streamingtest html page included in janus, I can select the stream, but I never get to see anything. On the console where I started janus, it throws multiple errors starting with : "SDP missing mandatory information"

  



  After starting ffmpeg it shows the SDP in console :

  



  v=0
o=- 0 0 IN IP4 127.0.0.1
s=No Name
c=IN IP4 127.0.0.1
t=0 0
a=tool:libavformat 58.20.100
m=video 8004 RTP/AVP 96
a=rtpmap:96 H264/90000
a=fmtp:96 packetization-mode=1; sprop-parameter-sets=J2QAKKwrQDwBE/LAPEiagA==,KO4BNywA; profile-level-id=640028


  



  The command I use to start ffmpeg is the following :

  



  ffmpeg -loglevel quiet -y -i - -c:v copy -preset veryfast -f rtp rtp://127.0.0.1:8004


  



  I assume that I am missing something in the ffmpeg command and therefore there is some kind of problem with the SDP.

  



  Edit : I looked a little further into this and actually looked at the code for janus at the lines that throw the error. Then I traced that to the origin of the error.

  



  Apparently the SDP is supposed to contain some authorization in for of this :

  



  a=... ice-ufrag=?; ice-pwd=?


  



  So how do I find out what this username and password are and how do I get it in to the SDP from ffmpeg ?

  



  Or how do I disable this entirely ?

  


• problem with changing the frames in an video with opencv

  5 septembre 2023, par Avizipi

  I have a code that runs in the last few months without any problem. The code shows an image from a video with extra data from the relevant db. using the keyboard you can jump around some frames and get better insights. However, in the last week changing the frame of the video stopped working.

  


  I built a small script that show the problem.

  


  import cv2

vid_cap_reader_path = "vid_230830_122746_C.avi"
cap_rgb = cv2.VideoCapture(vid_cap_reader_path)

k = 0
while 1:
    if k == ord("q"):
        break
    if k == ord("b"):
        before = cap_rgb.get(cv2.CAP_PROP_POS_FRAMES)
        cap_rgb.set(cv2.CAP_PROP_POS_FRAMES, before + 30.0)
        after = cap_rgb.get(cv2.CAP_PROP_POS_FRAMES)
        print(f"before = {before} after = {after}")
    ret, img = cap_rgb.read()
    if ret:
        cv2.imshow("bla", img)
        k = cv2.waitKey(0)
    else:
        break


  


  This is what is printed to the terminal :

  


  [mpeg4 @ 0x1d87400] warning: first frame is no keyframe
before = 36.0 after = 0.0
before = 67.0 after = 0.0
[mpeg4 @ 0x1d87400] warning: first frame is no keyframe
before = 40.0 after = 0.0
[mpeg4 @ 0x1d87400] warning: first frame is no keyframe


  


  The video itself doesn't "jump" for the next frame as expected and also most of the properties that opencv have return 0. I tested also CAP_PROP_FRAME_COUNT, and CAP_PROP_POS_MSEC without any lack, although the msec flag did give non-zero results which make no sense.

  


  It is also important to mention that I am building this application in a docker file and I am not using a specific version for my requirements. The tests here are with opencv-python==4.7.0.68. Also, the videos are working fine while playing them on VLC.

  


  I was looking around the net for a similar problem from the recent time and didn't find much except this github issue that may be similar. I started to downgrade opencv one by one without any luck.

  


  Maybe someone here can help with this issue, or give me a lead to doing something different.