Recherche avancée

Sur d’autres sites (11164)

• Save and re-stream RSTP video as straight UDP

  30 septembre 2013, par user1701362

  I am trying to write a program that will connect to a RTSP video source and redirect the video data to another location using UDP. I am also saving the RTSP packets to be able to replay the video stream at a latter moment in time as well. Right now my program can connect to the RTSP video stream and redirect and save, but when I try to look at the redirected video I get nothing using VLC.

  Currently the program just strips out the datagram from the RTSP video packets it receives in its open UDP socket and re-sends them using this code using the boost asio library.

  newVideoSocket->send_to(&dg.data[0], dg.data.size() ,Endpoint);

  When I look at the traffic using Wireshark I see that it is actually sending the data to the new address and it is recognized as a UDP packet, but when I try and view the video using VLC nothing happens. The video stream is Mpeg4 with the video encoded as H.264 and VLC can play it.

  I have tried to connect to the redirected stream as UDP and as RTP at both multicast and unicast addresses but have had no success. Do I need to add or take something out of the datagram before I resend it ? Or is it something wrong with how I am tring to view it in VLC ? Thanks for the help.

• How to match precompiled static library's debugging symbols to source code with Xcode 5 ?

  21 septembre 2013, par Chris Ballinger

  I have to compile FFmpeg for iOS using an external build script, but when I am debugging I see assembler if I delve too deep into libavformat library functions :

  0x109a73:  cmpl   $0, 1192(%ebp)
  
  0x109a7a:  jns    0x109a86                  ; mov_write_header + 198 at movenc.c:3539
  
  0x109a7c:  movl   $1, 1192(%ebp)
  
  0x109a86:  movl   16(%ebp), %eax
  
  0x109a89:  cmpl   $0, 84(%eax)
  
  0x109a8d:  movl   %edx, %ecx
  
  0x109a8f:  jne    0x109ad9                  ; mov_write_header + 281 at movenc.c:3548
  
  0x109a91:  testb  $2, 48(%ecx)
  
  0x109a95:  jne    0x109ac1                  ; mov_write_header + 257 at movenc.c:3541

  There are clear debugging symbols left behind which would lead me to believe that there should be some way to tell Xcode the location of this source code to allow easier debugging.

  Is this even possible ?

  edit : I found a related question here No symbols/source for external library in Xcode 4

  So when I run xcrun dwarfdump libavformat.a | grep "\.c" I get a bunch of results like this :

  AT_decl_file( "libavformat/movenc.c" )

  So I tried putting the relevant source files in folders relative to both the .a file and my .xcodeproj file, but neither of those methods worked. Full source code is available here : https://github.com/openwatch/livestreamer-ios

  edit2 : I found another question about how to set the source mapping for lldb but I'm not quite sure what to do... LLDB equivalent of gdb "directory" command for specifying source search path ?

  It seems like I might need to do

  (lldb) settings set target.source-map libavformat/movenc.c /path/to/libavformat/movenc.c 

  for every file I need ?

• How is video decoding corruption be debugged ?

  17 septembre 2013, par TopGunCoder

  I just started working for a new company and my new role demands that I help debug the video corruption that they are receiving through decoding frames. As much as I intend on digging down deep into the code and looking into the specifics of my problem, it made me think about video debugging in general.

  Since handling videos is very new to me, the whole process seems pretty complex and it seems there are a lot of places for corruption to present itself. The way I see it there is at least three places where corruption could pop up (barring memory corruption from the machine) :

  • Transporting the data before it is decoded
  • decoding implementation that perpetuates corruption once it is encountered, or is all together incorrect (Which seems to be my problem)
  • Transportation to the monitor(which seems unlikely but possible)

  So what i'm really curious about is if/how people debug their video streams to determine the location of any potential corruption they are encountering. I'm sure there is no sure fire method but I am curious to see what problems are even possible and how they can be identified and triaged.

  P.S. - I'm not sure of the differences between different decoding methods but, if this question seems too vague maybe it helps to mention I am using ffmpeg and avcodec_decode_video2 for the decoding.