FFmpeg: How to put encoded media data from one container to another with out re-encoding?

Question

So for example: I have file.mp3, I know that my desired format can play sound with out video (for example FLV) so how to put encoded mp3 data using ffmpeg from mp3 container into flv (where to get articles/code samples on this)?

I mean not from cmd but from C++ using ffmpeg as library. (see tags)

Answer 1

Here is the command to convert .mp3 file to .flv(which does not have any video data).

ffmpeg -i test.mp3 -ab 32k -acodec libmp3lame -ac 1 -ar 44100 audio.flv.

You can execute this command from your program.

If you need help on how to install and use ffmpeg you can go to their site:

http://ffmpeg.org

Thanks,

Mahmud

Answer 2

Have you considered just running ffmpeg from a popen() / system() call from c++?

It's a lot easier than setting up the ffmpeg library, it makes it trivial to multithread (not really an issue in the example) and frees you from any LGPL linking and dll-hell issues.

Answer 3

Here's what you want to do:

AVFormatContext *ptrFormatContext;
int i, videoStream, audioStream;
AVCodecContext *ptrCodecCtxt;
AVCodec *ptrCodec;
AVFrame *ptrFrame;
AVPacket ptrPacket;
int frameFinished;
float aspect_ratio;

AVCodecContext  *aCodecCtx;
AVCodec         *aCodec;

AVCodecContext  *aTargetCodecCtxt;
AVCodecContext  *vTargetCodecCtxt;
AVCodec         *aTargetCodec;
AVCodec         *vTargetCodec;
AVSampleFormat  ptrSampleFormats[2] = {AV_SAMPLE_FMT_S16, AV_SAMPLE_FMT_S32};



audioStream = videoStream = -1;

av_register_all();
avcodec_register_all();

ptrFormatContext = avformat_alloc_context();

if(avformat_open_input(&ptrFormatContext, filename, NULL, NULL) != 0 )
{
    qDebug("Error opening the input");
    exit(-1);
}
if(av_find_stream_info( ptrFormatContext) < 0)
{
    qDebug("Could not find any stream info");
    exit(-2);
}
dump_format(ptrFormatContext, 0, filename, (int) NULL);

for(i=0; i<ptrFormatContext->nb_streams; i++)
{
    switch(ptrFormatContext->streams[i]->codec->codec_type)
    {
    case AVMEDIA_TYPE_VIDEO:
    {
        if(videoStream < 0) videoStream = i;
        break;
    }
    case AVMEDIA_TYPE_AUDIO:
    {
        if(audioStream < 0) audioStream = i;
    }
    }
}
if(audioStream == -1)
{
    qDebug("Could not find any audio stream");
    exit(-3);
}
if(videoStream == -1)
{
    qDebug("Could not find any video stream");
    exit(-4);
}

aCodecCtx = ptrFormatContext->streams[audioStream]->codec;
if( (aCodec = avcodec_find_decoder(aCodecCtx->codec_id)) == NULL)
{
    qDebug("Could not find the audio decoder");
    exit(-5);
}
if( (avcodec_open(aCodecCtx, aCodec)) != 0 )
{
    qDebug("Could not open the audio decoder");
    exit(-6);
}

ptrCodecCtxt = ptrFormatContext->streams[videoStream]->codec;
if( (ptrCodec = avcodec_find_decoder(ptrCodecCtxt->codec_id)) == NULL )
{
    qDebug("Could not find the video decoder");
    exit(-7);
}
if((avcodec_open(ptrCodecCtxt, ptrCodec)) != 0)
{
    qDebug("Could not find any video stream");
    exit(-8);
}

Then some other stuff, mostly irrelevant if you don't want to reencode...

ptrFrame = avcodec_alloc_frame();

while(av_read_frame(ptrFormatContext,&ptrPacket) >= 0)
{
    if(ptrPacket.stream_index == videoStream)
    {
       //do stuff with the package, for eg transcribe it into another output stream..

    }
    else if (ptrPacket.stream_index == audioStream)
    {
       //do stuff with the package, for eg transcribe it into another output stream..
    }
}

Hope that's helpful. The code is however only an excerpt and will not work on its own, but it'll help you get the idea.

Answer 4

ffmpeg -i file.mp3 -acodec copy output.flv