Where can I find modern tutorials for libav, ffmpe

Question

I want to make a quick program in C that will open a video, save each frame as a ppm, and dump motion vectors. All the tutorials I can find are from almost ten years ago and call deprecated or non-existent functions.

Are there any good online resources, websites, videos, or textbooks that cover a modern approach to doing these types of things?

Answer 1

        #include <iostream>
    #include <cstdio>
    #include <cstdlib>
    #include <fstream>
    #include <math.h>

    #define __STDC_CONSTANT_MACROS
    //FFMPEG LIBRARIES
    extern "C"
    {
    #include "libavcodec/avcodec.h"
    #include "libswscale/swscale.h"
    #include "libavutil/opt.h"
    #include "libavutil/common.h"
    #include "libavutil/channel_layout.h"
    #include "libavutil/imgutils.h"
    #include "libavutil/mathematics.h"
    #include "libavutil/samplefmt.h"
    #include "libavutil/time.h"
    #include "libavformat/avformat.h"
    #include "libavformat/avio.h"
    #include "libavfilter/avfilter.h"
    #include "libavdevice/avdevice.h"
    #include "libavfilter/avfiltergraph.h"
    #include "libavfilter/buffersink.h"
    }

    using namespace std;

    void SaveMyFrame(AVFrame *sAVFrame , int swidth, int sheight, int iFrame)
    {
    FILE *pfile;
    char szFilename[32];
    int y;

    sprintf(szFilename , "frame%d.ppm" , iFrame);
    pfile = fopen(szFilename , "wb");
    if(pfile == NULL)
    {
      cout<<"\n\ncould'nt open file";
      return;
    }

    fprintf(pfile , "P6\n%d %d\n255\n" , swidth , sheight );

    for( y=0; y<sheight; y++)
    {
      fwrite(sAVFrame->data[0]+y*sAVFrame->linesize[0] , 1 , swidth*3 , pfile );
    }

    fclose(pfile);
    }

    int CaptureScene(int VideoStreamIndx ,
                       AVFormatContext *bAVFormatContext ,
                       AVCodecContext *bAVCodecContext,
                       AVCodec *bAVCodec )
    {
      AVPacket bAVPacket;
      AVFrame *bAVFrame = NULL;
      bAVFrame = av_frame_alloc();
      AVFrame *bAVFrameRGB = NULL;
      bAVFrameRGB = av_frame_alloc();

    if(bAVFrame == NULL)
    {
      cout<<"\n\nframe alloc failed";
    }

    if(bAVFrameRGB == NULL)
    {
      cout<<"\n\nframe alloc RGB failed";
    }

      int numBytes;
      uint8_t *buffer = NULL;

    numBytes =  av_image_get_buffer_size(AV_PIX_FMT_RGB24 , bAVCodecContext->width,bAVCodecContext->height, 32);  // avpicture_get_size  deprecated

    buffer=(uint8_t *)av_malloc(numBytes*sizeof(uint8_t));

    avpicture_fill((AVPicture *)bAVFrameRGB , buffer , AV_PIX_FMT_RGB24 , bAVCodecContext->width , bAVCodecContext->height);

    int framefinish;
    struct SwsContext *sws_ctx = NULL;
    sws_ctx = sws_getContext( bAVCodecContext->width,
                              bAVCodecContext->height,
                              bAVCodecContext->pix_fmt,
                              bAVCodecContext->width,
                              bAVCodecContext->height,
                              AV_PIX_FMT_RGB24,
                              SWS_BILINEAR,
                              NULL,NULL,NULL);
    int i =0;

    while(av_read_frame(bAVFormatContext,&bAVPacket) >=0)
    {
          if(bAVPacket.stream_index == VideoStreamIndx)
      {
          avcodec_decode_video2(bAVCodecContext , bAVFrame , &framefinish , &bAVPacket);
          if(framefinish)
        {
          // convert image from native format to RGB
          sws_scale(sws_ctx , (uint8_t const* const *)bAVFrame->data ,
          bAVFrame->linesize , 0, bAVCodecContext->height,
          bAVFrameRGB->data , bAVFrameRGB->linesize);
          // save frame to disk
          if(++i <= 100)SaveMyFrame(bAVFrameRGB , bAVCodecContext->width , bAVCodecContext->height , i );

        }

      }

    }

    av_free(bAVFrame);
    av_free(bAVFrameRGB);
    }

    int main()
    {

      avdevice_register_all();
      avcodec_register_all();
      av_register_all();

      char *dev_name = "/dev/video0";

     int VideoStreamIndx = -1;
     AVCodecContext *pAVCodecContext = NULL;
     AVCodec *pAVCodec = NULL;
     AVInputFormat *inputFormat =av_find_input_format("v4l2");
     AVDictionary *options = NULL;
     av_dict_set(&options, "framerate", "20", 0);

     AVFormatContext *pAVFormatContext = NULL;

     if(avformat_open_input(&pAVFormatContext, dev_name, inputFormat, NULL) != 0)
     {
       cout<<"\nError : could'nt open video source\n\n";
       return -1;
     }

     if( avformat_find_stream_info( pAVFormatContext , NULL) < 0)
     {
     cout<<"Error : streams not found";
      return -1;
     }

      av_dump_format(pAVFormatContext , 0 , "/dev/video1" , 0 );

     for(int i=0; i<pAVFormatContext->nb_streams ;i++ )
     {
      if( pAVFormatContext->streams[i]->codec->codec_type == AVMEDIA_TYPE_VIDEO ) // if video stream found then get the index.
      {
        VideoStreamIndx = i;
        break;
      }
     }

    if((VideoStreamIndx) == -1)
    {
      cout<<"Error : video streams not found";
      return -1;
    }

    pAVCodecContext = pAVFormatContext->streams[VideoStreamIndx]->codec;

    pAVCodec = avcodec_find_decoder( pAVCodecContext->codec_id );

    if(pAVCodec == NULL)
    {
     fprintf(stderr,"Unsupported codec !");
     return -1;
    }

    int value = avcodec_open2(pAVCodecContext , pAVCodec , NULL);
    if( value < 0)
    {
      cout<<"Error : Could not open codec";
      return -1;
    }

    int Vwidth , Vheight , videoFPS , videoBaseTime , duration_2 ;
    int sframe , length , Fheight;

    /*
    To fetch/display some media information programatically
    */

    //int64_t duration_1 = pAVFormatContext->duration;
    videoFPS = av_q2d(pAVFormatContext->streams[VideoStreamIndx]->r_frame_rate);
    videoBaseTime = av_q2d(pAVFormatContext->streams[VideoStreamIndx]->time_base);
    Vwidth = pAVFormatContext->streams[VideoStreamIndx]->codec->width;
    Vheight = pAVFormatContext->streams[VideoStreamIndx]->codec->height;
    //duration_2 = (unsigned long)pAVFormatContext->streams[VideoStreamIndx]->duration*(videoFPS*videoBaseTime);

    cout<<"Video FPS :"<<videoFPS;
    cout<<"\n\n width : "<<Vwidth;
    cout<<"\n\n height : "<<Vheight;
    cout<<"\n\n time base"<<videoBaseTime;
    //cout<<"\n\nduration (1): "<<duration_1;
    //cout<<"\n\nduration (2): "<<duration_2;

    CaptureScene( VideoStreamIndx , pAVFormatContext , pAVCodecContext , pAVCodec );

    avcodec_close(pAVCodecContext);
    avformat_close_input(&pAVFormatContext);

    return 0;
    }

This C++ code will help you to get .ppm image from USB (UVC) based camera. it saves in .ppm format. use v4l2 for linux and dshow for windows. instead of camera frame feed , a video also can be served as a input. this code cooked with latest ffmpeg libs.

libavutil 55. 24.100 / 55. 24.100

libavcodec 57. 43.100 / 57. 43.100

libavformat 57. 37.100 / 57. 37.100

libavdevice 57. 0.101 / 57. 0.101

libavfilter 6. 46.100 / 6. 46.100

libswscale 4. 1.100 / 4. 1.100

libswresample 2. 0.101 / 2. 0.101

libpostproc 54. 0.100 / 54. 0.100

Answer 2

I have been working with ffmpeg and libav for several years and also have found no decent recent API-level tutorials. Sometimes I just have to dive into the source to figure out what is going on and how to use it. Also, reading the source for the CLI programs (which use the underlying libraries) can also be informative.

In fact, ffmpeg recommends that you just read source code, either for real OSS projects or their examples. Libav provides Doxygen documentation, but no step-by-step tutorials. IMO this is typical of open source projects: the source may be good, but documentation is often lacking.

Taking a step back, you might consider OpenCV, which is well documented (books exist!) and has an easier API. Lastly, ppm is such a simple format you could probably write a 50 line function to save the pixels yourself, and I found a motion extraction program here.