Recherche avancée

Sur d’autres sites (13968)

• FFMPEG/FFPLAY : Where does AVCodecContext contents get filled

  16 décembre 2013, par Zax

  I have integrated a custom decoder into FFMPEG's Multimedia framework. However i still feel i have missed out a lot of things during the integration. As a result, when i'm trying to render the output data using FFPLAY, the resolution that it is assuming is wrong for my stream.

  The command that i use to render the stream is :

  ./ffplay -vcodec myCustDec -i BasketBallPass_416x240.bin

  So when the above command is executed, on my terminal what it is displaying is :

  Stream #0:0: Video: bintext, pal8, 1280x23168, 25 tbr, 25 tbn, 25 tbc

  From whatever i know, bintext is the parser my decoder is using, then the resolution it is assuming is 1280x23168 (which is wrong).

  Then the FFPLAY's window pops up with a window of size 416x240, with a distorted image in it. What i think is that the decoder is decoding and putting the YUV420 data into the necessary buffer with proper width and height initialization of the AVFrame.

  However due to the resolution that it is displaying in the command line i.e. 1280x23168, the output is distorted. This thing i can confirm because, when i dump the output frame, i'm able to get proper YUV dumped data which can be played in any YUV player.

  The same command when i tryout with some existing codec inside the frame work, below are its dumps :

  Command :

  ./ffplay -vcodec h264 input.m4v

  treminal displays :

  Stream #0:0: Video: h264, yuv420p, 416x240, 25 fps, 25 tbr, 1200k tbn, 25 tbc

  As we see here, the framework is getting the resolution right.

  Now my query is :

  In which place is the resolution of this frames being dumped in the code and should i define my own probe function? can this problem be because of the probe/demuxer? 

• Using FFMPEG in an OS X application

  31 mars 2016, par Sokco

  so I’ve been wanting to make a real-time live streaming application. Essentially, the application would send the microphone feed from an Xcode application to a website where it can be viewed in real-time. Is FFMPEG the best solution for this ? How would I go about doing this ? If that’s too broad, then how do I use the FFMPEG framework in an OS X objective-c application ?

• FFMPEG : Video decoder integration giving segmentation fault

  5 décembre 2013, par Zax

  I have integrated a decoder to FFMPEG multimedia framework. I have referred this how to post to integrate the decoder.

  How to integrate a codec to FFMPEG

  I have followed the above post and integrated my custom decoder which decodes an elementary stream into a YUV420 data.

  However, When i give an input command as shown below, i get the following messages in sequence :

  ./ffmpeg -vcodec myCustomDec -i input_416x240.bin opt.yuv

  For integrating a codec, we need to define 3 functions and assign them to AVCodec's structures function pointers .init, .close and .decode.

  I have given a printf statement in myCustomDec's demuxer's probe function, init function, decode function and close function.

  The output i get after executing the above command is :

  myCustomDecoder probe function Entered
  
  myCustomDecoder demux successful
  
  Codec 0x73736d77 is not in the full list.
  
  Init is Entered
  
   Last message repeated 1 times
  
  Decode entered
  
  Segmentation fault (core dumped)

  So, this means that i'm getting segmentation fault after after entering decode function. My decode function prototype is as shown below :

  static int myCustom_decode_frame(AVCodecContext *avctx, void *data,int *got_frame_ptr, AVPacket *avpkt)

  In this function what i actually do is :

  AVFrame *frame     = data;
  
  printf("\nDecode entered\n");
  
  //Here i call decode frame function
  
  
  
  //Here i call a function to fetch YUV data
  
  
  
  //now i make the pointer that is pointing to YUV buffer data to point to variable of type AVFrame *frame
  
  
  
  //set *got_frame to 1

  So my expectation here is that once i map my YUV buffer pointer to the AVFrame *frame variable, the FFMPEG multimedia framework will take care of dumping this yuv frame to opt.yuv which is specified in the command. My question is : Is my assumption correct that FFMPEG multimedia frame work takes care of dumping data to output file ?

  Second, Why am i getting a message in the output saying Codec 0x73736d77 is not in the full list.

  And finally, How does the FFMPEG framework get to know the width and height of the input stream that i'm providing, because my parser is embeded in the decode code itself, and there is no connection with FFMPEG.

  Any suggestion regarding the same will be really helpful to me. Please do provide your suggestions. Thanks in advance.