Recherche avancée

Sur d’autres sites (7939)

• Decoding with FFMPEG on Visual Studio 2010

  6 juin 2013, par user2439801

  I just started using FFMPEG with C++ and try to code an audio decoder then write the decoded audio into a file.

  However i'm not sure about which data to write to the output file. As far as i know from looking at the sample codes it seems to be the AVFrame -> data[0].
  But when i try to print it on the consoles, i get some random numbers that are different each time i run the program. And when i try to write this AVFrame->data[0] into a file i keep getting an error.

  So my question is how can i write the decoded audio after i call the function av_codec_decode_audio4 ?

  Below i attached my code and i pass the argument "C :\02.mp3" which is a path for a valid mp3 file on my PC.

  Thank you for your help.

  // TestFFMPEG.cpp : Audio Decoder
  
  //
  
  
  
  #include "stdafx.h"
  
  
  
  #include <iostream>
  
  #include <fstream>
  
  #include <sstream>
  
  
  
  extern "C" {
  
      #include 
  
      #include 
  
      #include 
  
  
  
  }
  
  
  
  using namespace std;
  
  
  
  
  
  int main(int argc, char* argv[])
  
  {
  
  int audioStream = -1;
  
  
  
  AVCodec         *aCodec;
  
  AVPacket        avPkt;
  
  AVFrame         *decode_frame = avcodec_alloc_frame();
  
  
  
  AVCodecContext  *aCodecCtxt;
  
  AVFormatContext *pFormatCtxt = NULL;
  
  
  
  if(argc != 2) {     // Checking  whether there is enough argument
  
      return -1; 
  
  }
  
  
  
  av_register_all();  //Initialize CODEC
  
  avformat_network_init();
  
  av_init_packet (&amp;avPkt);
  
  
  
  
  
   if (avformat_open_input (&amp;pFormatCtxt, argv[1],NULL,NULL)!= 0 ){ //Opening File
  
       return -2;
  
   }
  
  
  
   if(avformat_find_stream_info (pFormatCtxt,NULL) &lt; 0){ //Get Streams Info 
  
       return -3; 
  
   }
  
  
  
   AVStream *stream = NULL;
  
   //av_read_play (pFormatCtxt); //open streams
  
  
  
  
  
   for (int i = 0;  i &lt; pFormatCtxt->nb_streams ; i++) { //Find Audio Stream
  
       if (pFormatCtxt->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO){
  
           audioStream =i;
  
       }
  
   }
  
  
  
   aCodecCtxt = pFormatCtxt ->streams [audioStream]->codec; // opening decoder   
  
   aCodec = avcodec_find_decoder( pFormatCtxt->streams [audioStream] ->codec->codec_id);
  
  
  
   if (!aCodec) {
  
       return -8;
  
   }
  
  
  
   if (avcodec_open2(aCodecCtxt,aCodec,NULL)!=0) {
  
       return -9; 
  
   } 
  
  
  
  int cnt = 0;
  
  
  
  while(av_read_frame(pFormatCtxt,&amp;avPkt) >= 0 ){
  
  
  
      if (avPkt.stream_index == audioStream){
  
          int check = 0; 
  
          int result = avcodec_decode_audio4 (aCodecCtxt,decode_frame,&amp;check, &amp;avPkt);
  
          cout &lt;&lt; "Decoded : "&lt;&lt; (int) decode_frame->data[0] &lt;&lt;", "&lt;&lt; "Check : " &lt;&lt; check &lt;&lt; ", Format :" &lt;&lt; decode_frame->format &lt;&lt;" " &lt;&lt; decode_frame->linesize[0]&lt;&lt; " "&lt;<cnt return="return" acodec="acodec">id;  
  
  } 
  
  </cnt></sstream></fstream></iostream>

• Batch Programming Looping ?

  13 mai 2013, par Michael Barth

  I'm not an expert in batch programming, my only skill is in C\C++. I'm not even sure how to go about this really.

  I've got a bunch of videos like Video1.mp4, Video2.mp4 (that's not their actual names, but they do have numbers indicating beginning and end)

  Basically I have this FFMPEG command that I can use to convert all of these to an aspect ratio of 16:10 that looks something like this :

  "ffmpeg -i "Section 1 Video 1.mp4" -aspect 16:10 OutSection 1 Video 1.mp4"
  
  "ffmpeg -i "Section 1 Video 2.mp4" -aspect 16:10 OutSection 1 Video 2.mp4"
  
  "ffmpeg -i "Section 2 Video 1.mp4" -aspect 16:10 OutSection 2 Video 1.mp4"

  Now instead of writing this command over and over again, is there anyway for me to substitute the numbers for actual variables ? I know that Section 1 ends at Video 27 and Section 2 ends at Video 26 and so on. Basically I need the loop to run 165 times so I don't need to write the command 165 times.

  Or is this beyond the scope of capability of a mere batch program ?

• ffmpeg Video Scan Progress during Frame Extarction

  5 avril 2013, par Isantipov

  I scan input video to extract certain frames using ffmpeg's select filter. The selection is based on a complex creteria and the number of extracted frames can't be predicted (I'm doing scene detection, but this can be something different - e.g. selecting all the I frames, etc).

  What I need is to display the percentage of scanned (decoded) video (e.g. 10% or 90%).

  ---

  I tried several ways to do this parsing console output as people usually do when dealing with encoding, but it doesn't help with the progress of scanning (e.g. Can ffmpeg show a progress bar ? or ffmpeg Progress Bar - Encoding Percentage in PHP)

  ffmpeg -progress sceneProgr.txt -i input.wmv -vsync passthrough -an -vf select='gt(scene\,0.2)',showinfo scene%%05d.png

  The output this produces is as follows :

  <..>    
  
  frame=    0 fps=0.0 q=0.0 size=N/A time=00:00:00.00 bitrate=N/A    
  
  n:0 pts:16550 pts_time:16.55 pos:4205325 fmt:rgb24 sar:0/1 s:640x480 i:P iskey:1 type:I checksum:95895BC9 plane_checksum:[95895BC9]
  
  frame=    1 fps=0.7 q=0.0 size=N/A time=00:00:00.00 bitrate=N/A    
  
  n:1 pts:24591 pts_time:24.591 pos:6685325 fmt:rgb24 sar:0/1 s:640x480 i:P iskey:0 type:P checksum:FF4CC015 plane_checksum:[FF4CC015]
  
  'frame=...' and 'n:...' lines are repeated for each of the extracted frames.

  Both frame=... and n:... lines refer to the numbers in the output and therefore can't be used to calculate progress the way people usually do this (as I can't predict how many frames will be found beforehead and besides, they are not uniformly spread across the input video).

  ---

  If I specify -progress progress.txt parameter, the progress.txt file is as follows :

  frame=5
  
  fps=1.2
  
  stream_0_0_q=0.0
  
  total_size=N/A
  
  out_time_ms=43209833
  
  out_time=00:00:43.209833
  
  dup_frames=0
  
  drop_frames=0
  
  progress=continue
  
  
  
  
  
  frame=6
  
  fps=1.3
  
  stream_0_0_q=0.0
  
  total_size=N/A
  
  out_time_ms=52252200
  
  out_time=00:00:52.252200
  
  dup_frames=0
  
  drop_frames=0
  
  progress=continue
  
  
  
  frame=6
  
  fps=1.2
  
  stream_0_0_q=0.0
  
  total_size=N/A
  
  out_time_ms=52252200
  
  out_time=00:00:52.252200
  
  dup_frames=0
  
  drop_frames=0
  
  progress=continue

  New portion is written approximately every second and refers to the last extracted frame.
  Which is somewhat helpful as out_time refers to the last extracted frames' position in the input video, so I can calculate the progress of the scan from it as

  progress = out_time_ms/total_input_time 

  But this is not ideal as it will be updated only when the new frame which matches the select criteria is extracted. So, If I have a large portion of video with no matching frames, the progress won't change for a lot of time.

  ---

  Wrapping-up :

  I am looking for a way to calculate the progress of video scanning when using select filter.

  Any ideas are strongly appreciated.