【ffmpeg-libav-tutorial】ffmpeg的hello world

FFmpeg的hello world

ffmpeg的hello world并不是打印hello world，而是打印视频的信息。比如他的容器格式、长度、分辨率、音频通道，最后我们将会解码一些帧并将他们保存为图片

ffmpeg的架构

首先先了解一下ffmpeg的架构以及他的组件如何与其他组件进行通信，如图是视频解码的过程：

首先，我们需要将视频文件加载到AVFormatContext结构体，实际上，它并没有全部加载而是只读取了他的头部信息。

一旦我们加载了容器的最小标头，我们就能够访问他的流（也就是通常所说的视频和音频数据），每路流可以通过AVStream这个结构体进行访问。

假设我们的视频有两路流，分别是aac编码的音频流和h264编码的avc视频流。从每路流中我们可以提取中一片片的数据包，并加载到avpacket中。

这些数据包包含的就是已经被编码过的视频数据，为了进行解码，我们需要将这些数据传递给对应的解码器AVCodec。

AVCodec解码后，将封装到AVFrame中，这些数据就是经过解码后的帧数据。

代码演练

首先，我们需要申请AVFormatContent的内存用于存储视频文件的信息。

AVFormatContext *pFormatContext = avformat_alloc_context();

接下来，我们将通过打开视频文件，并读取他的头部信息，并将视频的最小信息赋值给AVFormatContext结构体（通常这个时候codec还是未打开的状态）。实现这个操作的函数avformat_open_input，这个参数需要传递四个参数：AVFormatContext、文件名称和两个可选参数：输入格式AVInputFormat(如果为NULL，ffmpeg会猜测它的格式) 以及AVDictionary（用来给解封装器的参数）。

avformat_open_input(&pFormatContext, filename, NULL, NULL);

打开后，就能够打印出这个视频文件所包含的信息。

printf("Format %s, duration %lld us", pFormatContext->iformat->long_name,
pFormatContext->duration);

接下来，为了能访问流信息，我们需要从媒体中读取数据。这里需要用函数avformat_find_stream_info来实现。操作后，pFormatContext->nb_streams会保存着视频文件中包含流的数量，当要访问某路流的时候，pFormatContext->streams[i]就能实现对流的访问，并保存在AVStream结构体中。

avformat_find_stream_info(pFormatContext,  NULL);
for (int i = 0; i < pFormatContext->nb_streams; i++)
{
  //
}

对于每路流，我们需要保存他的编码器参数，这里用到了AVCodecParameters,这一结构体。

AVCodecParameters *pLocalCodecParameters = pFormatContext->streams[i]->codecpar;

保存后，就可以打印一些codec的信息。

// specific for video and audio
if (pLocalCodecParameters->codec_type == AVMEDIA_TYPE_VIDEO) {
  printf("Video Codec: resolution %d x %d", pLocalCodecParameters->width, pLocalCodecParameters->height);
} else if (pLocalCodecParameters->codec_type == AVMEDIA_TYPE_AUDIO) {
  printf("Audio Codec: %d channels, sample rate %d", pLocalCodecParameters->channels, pLocalCodecParameters->sample_rate);
}
// general
printf("\tCodec %s ID %d bit_rate %lld", pLocalCodec->long_name, pLocalCodec->id, pCodecParameters->bit_rate);

有了codec信息后，我们需要将这些信息放入申请的AVCodecContext结构体中。这里用到了avcodec_parameters_to_context.这个函数。
一旦存储完codec内容后，就需要打开codec，方便以后解码。将用到avcodec_open2这个函数

AVCodecContext *pCodecContext = avcodec_alloc_context3(pCodec);
avcodec_parameters_to_context(pCodecContext, pCodecParameters);
avcodec_open2(pCodecContext, pCodec, NULL);

接下来，就能从流中读取数据包，并对他们进行解码成帧的操作了。
这里首先，先申请两块内存块AVPacket(存储packet)和AVFrame(存储解码后的帧)。

AVPacket *pPacket = av_packet_alloc();
AVFrame *pFrame = av_frame_alloc();

我们使用av_read_frame中流中读取数据包，直到流中的数据包被读完。

while (av_read_frame(pFormatContext, pPacket) >= 0) {
  //...
}

接下来，读取的数据包就可以去解码了。旧版本的ffmpeg可以用avcodec_decode_video2函数去解码，而在新版本中，将通过avcodec_send_packet这个函数将原始packet送去解码器进行解码。

avcodec_send_packet(pCodecContext, pPacket);

解码后的数据，将通过avcodec_receive_frame进行接收存储到前面申请的AVFrame，这里就能得到一些该帧的信息。

avcodec_receive_frame(pCodecContext, pFrame);
printf("Frame %c (%d) pts %d dts %d key_frame %d [coded_picture_number %d, display_picture_number %d]",
    av_get_picture_type_char(pFrame->pict_type),
    pCodecContext->frame_number,
    pFrame->pts,
    pFrame->pkt_dts,
    pFrame->key_frame,
    pFrame->coded_picture_number,
    pFrame->display_picture_number
);

最后，我们能够将解码后的帧保存成简单的图片。在pFrame->data数组中分别保存着该帧图片的Y、Cr、Cb三个通道，这里可以仅保存灰度的Y通道部分。

save_gray_frame(pFrame->data[0], pFrame->linesize[0], pFrame->width, pFrame->height, frame_filename);

static void save_gray_frame(unsigned char *buf, int wrap, int xsize, int ysize, char *filename)
{
    FILE *f;
    int i;
    f = fopen(filename,"w");
    // writing the minimal required header for a pgm file format
    // portable graymap format -> https://en.wikipedia.org/wiki/Netpbm_format#PGM_example
    fprintf(f, "P5\n%d %d\n%d\n", xsize, ysize, 255);

    // writing line by line
    for (i = 0; i < ysize; i++)
        fwrite(buf + i * wrap, 1, xsize, f);
    fclose(f);
}

这样就能得到一个2MB左右的灰度图片

源代码

#include 
#include 
#include 
#include 
#include 
#include 
#include 

// print out the steps and errors
static void logging(const char *fmt, ...);
// decode packets into frames
static int decode_packet(AVPacket *pPacket, AVCodecContext *pCodecContext, AVFrame *pFrame);
// save a frame into a .pgm file
static void save_gray_frame(unsigned char *buf, int wrap, int xsize, int ysize, char *filename);

int main(int argc, const char *argv[])
{
  logging("initializing all the containers, codecs and protocols.");
	av_register_all();
    avformat_network_init();
  // AVFormatContext holds the header information from the format (Container)
  // Allocating memory for this component
  // http://ffmpeg.org/doxygen/trunk/structAVFormatContext.html
  AVFormatContext *pFormatContext = avformat_alloc_context();
  if (!pFormatContext) {
    logging("ERROR could not allocate memory for Format Context");
    return -1;
  }

  logging("opening the input file (%s) and loading format (container) header", argv[1]);
  // Open the file and read its header. The codecs are not opened.
  // The function arguments are:
  // AVFormatContext (the component we allocated memory for),
  // url (filename),
  // AVInputFormat (if you pass NULL it'll do the auto detect)
  // and AVDictionary (which are options to the demuxer)
  // http://ffmpeg.org/doxygen/trunk/group__lavf__decoding.html#ga31d601155e9035d5b0e7efedc894ee49
  if (avformat_open_input(&pFormatContext, argv[1], NULL, NULL) != 0) {
    logging("ERROR could not open the file");
    return -1;
  }

  // now we have access to some information about our file
  // since we read its header we can say what format (container) it's
  // and some other information related to the format itself.
  logging("format %s, duration %lld us, bit_rate %lld", pFormatContext->iformat->name, pFormatContext->duration, pFormatContext->bit_rate);

  logging("finding stream info from format");
  // read Packets from the Format to get stream information
  // this function populates pFormatContext->streams
  // (of size equals to pFormatContext->nb_streams)
  // the arguments are:
  // the AVFormatContext
  // and options contains options for codec corresponding to i-th stream.
  // On return each dictionary will be filled with options that were not found.
  // https://ffmpeg.org/doxygen/trunk/group__lavf__decoding.html#gad42172e27cddafb81096939783b157bb
  if (avformat_find_stream_info(pFormatContext,  NULL) < 0) {
    logging("ERROR could not get the stream info");
    return -1;
  }

  // the component that knows how to enCOde and DECode the stream
  // it's the codec (audio or video)
  // http://ffmpeg.org/doxygen/trunk/structAVCodec.html
  AVCodec *pCodec = NULL;
  // this component describes the properties of a codec used by the stream i
  // https://ffmpeg.org/doxygen/trunk/structAVCodecParameters.html
  AVCodecParameters *pCodecParameters =  NULL;
  int video_stream_index = -1;

  // loop though all the streams and print its main information
  for (int i = 0; i < pFormatContext->nb_streams; i++)
  {
    AVCodecParameters *pLocalCodecParameters =  NULL;
    pLocalCodecParameters = pFormatContext->streams[i]->codecpar;
    logging("AVStream->time_base before open coded %d/%d", pFormatContext->streams[i]->time_base.num, pFormatContext->streams[i]->time_base.den);
    logging("AVStream->r_frame_rate before open coded %d/%d", pFormatContext->streams[i]->r_frame_rate.num, pFormatContext->streams[i]->r_frame_rate.den);
    logging("AVStream->start_time %" PRId64, pFormatContext->streams[i]->start_time);
    logging("AVStream->duration %" PRId64, pFormatContext->streams[i]->duration);

    logging("finding the proper decoder (CODEC)");

    AVCodec *pLocalCodec = NULL;

    // finds the registered decoder for a codec ID
    // https://ffmpeg.org/doxygen/trunk/group__lavc__decoding.html#ga19a0ca553277f019dd5b0fec6e1f9dca
    pLocalCodec = avcodec_find_decoder(pLocalCodecParameters->codec_id);

    if (pLocalCodec==NULL) {
      logging("ERROR unsupported codec!");
      return -1;
    }

    // when the stream is a video we store its index, codec parameters and codec
    if (pLocalCodecParameters->codec_type == AVMEDIA_TYPE_VIDEO) {
      video_stream_index = i;
      pCodec = pLocalCodec;
      pCodecParameters = pLocalCodecParameters;

      logging("Video Codec: resolution %d x %d", pLocalCodecParameters->width, pLocalCodecParameters->height);
    } else if (pLocalCodecParameters->codec_type == AVMEDIA_TYPE_AUDIO) {
      logging("Audio Codec: %d channels, sample rate %d", pLocalCodecParameters->channels, pLocalCodecParameters->sample_rate);
    }

    // print its name, id and bitrate
    logging("\tCodec %s ID %d bit_rate %lld", pLocalCodec->name, pLocalCodec->id, pCodecParameters->bit_rate);
  }
  // https://ffmpeg.org/doxygen/trunk/structAVCodecContext.html
  AVCodecContext *pCodecContext = avcodec_alloc_context3(pCodec);
  if (!pCodecContext)
  {
    logging("failed to allocated memory for AVCodecContext");
    return -1;
  }

  // Fill the codec context based on the values from the supplied codec parameters
  // https://ffmpeg.org/doxygen/trunk/group__lavc__core.html#gac7b282f51540ca7a99416a3ba6ee0d16
  if (avcodec_parameters_to_context(pCodecContext, pCodecParameters) < 0)
  {
    logging("failed to copy codec params to codec context");
    return -1;
  }

  // Initialize the AVCodecContext to use the given AVCodec.
  // https://ffmpeg.org/doxygen/trunk/group__lavc__core.html#ga11f785a188d7d9df71621001465b0f1d
  if (avcodec_open2(pCodecContext, pCodec, NULL) < 0)
  {
    logging("failed to open codec through avcodec_open2");
    return -1;
  }

  // https://ffmpeg.org/doxygen/trunk/structAVFrame.html
  AVFrame *pFrame = av_frame_alloc();
  if (!pFrame)
  {
    logging("failed to allocated memory for AVFrame");
    return -1;
  }
  // https://ffmpeg.org/doxygen/trunk/structAVPacket.html
  AVPacket *pPacket = av_packet_alloc();
  if (!pPacket)
  {
    logging("failed to allocated memory for AVPacket");
    return -1;
  }

  int response = 0;
  int how_many_packets_to_process = 8;

  // fill the Packet with data from the Stream
  // https://ffmpeg.org/doxygen/trunk/group__lavf__decoding.html#ga4fdb3084415a82e3810de6ee60e46a61
  while (av_read_frame(pFormatContext, pPacket) >= 0)
  {
    // if it's the video stream
    if (pPacket->stream_index == video_stream_index) {
    logging("AVPacket->pts %" PRId64, pPacket->pts);
      response = decode_packet(pPacket, pCodecContext, pFrame);
      if (response < 0)
        break;
      // stop it, otherwise we'll be saving hundreds of frames
      if (--how_many_packets_to_process <= 0) break;
    }
    // https://ffmpeg.org/doxygen/trunk/group__lavc__packet.html#ga63d5a489b419bd5d45cfd09091cbcbc2
    av_packet_unref(pPacket);
  }

  logging("releasing all the resources");

  avformat_close_input(&pFormatContext);
  avformat_free_context(pFormatContext);
  av_packet_free(&pPacket);
  av_frame_free(&pFrame);
  avcodec_free_context(&pCodecContext);
  return 0;
}

static void logging(const char *fmt, ...)
{
    va_list args;
    fprintf( stderr, "LOG: " );
    va_start( args, fmt );
    vfprintf( stderr, fmt, args );
    va_end( args );
    fprintf( stderr, "\n" );
}

static int decode_packet(AVPacket *pPacket, AVCodecContext *pCodecContext, AVFrame *pFrame)
{
  // Supply raw packet data as input to a decoder
  // https://ffmpeg.org/doxygen/trunk/group__lavc__decoding.html#ga58bc4bf1e0ac59e27362597e467efff3
  int response = avcodec_send_packet(pCodecContext, pPacket);

  if (response < 0) {
    logging("Error while sending a packet to the decoder: %s", av_err2str(response));
    return response;
  }

  while (response >= 0)
  {
    // Return decoded output data (into a frame) from a decoder
    // https://ffmpeg.org/doxygen/trunk/group__lavc__decoding.html#ga11e6542c4e66d3028668788a1a74217c
    response = avcodec_receive_frame(pCodecContext, pFrame);
    if (response == AVERROR(EAGAIN) || response == AVERROR_EOF) {
      break;
    } else if (response < 0) {
      logging("Error while receiving a frame from the decoder: %s", av_err2str(response));
      return response;
    }

    if (response >= 0) {
      logging(
          "Frame %d (type=%c, size=%d bytes) pts %d key_frame %d [DTS %d]",
          pCodecContext->frame_number,
          av_get_picture_type_char(pFrame->pict_type),
          pFrame->pkt_size,
          pFrame->pts,
          pFrame->key_frame,
          pFrame->coded_picture_number
      );

      char frame_filename[1024];
      snprintf(frame_filename, sizeof(frame_filename), "%s-%d.pgm", "frame", pCodecContext->frame_number);
      // save a grayscale frame into a .pgm file
      save_gray_frame(pFrame->data[0], pFrame->linesize[0], pFrame->width, pFrame->height, frame_filename);

      av_frame_unref(pFrame);
    }
  }
  return 0;
}

static void save_gray_frame(unsigned char *buf, int wrap, int xsize, int ysize, char *filename)
{
    FILE *f;
    int i;
    f = fopen(filename,"w");
    // writing the minimal required header for a pgm file format
    // portable graymap format -> https://en.wikipedia.org/wiki/Netpbm_format#PGM_example
    fprintf(f, "P5\n%d %d\n%d\n", xsize, ysize, 255);

    // writing line by line
    for (i = 0; i < ysize; i++)
        fwrite(buf + i * wrap, 1, xsize, f);
    fclose(f);
}