Chinaunix首页 | 论坛 | 博客
  • 博客访问: 2354439
  • 博文数量: 141
  • 博客积分: 3552
  • 博客等级: 中校
  • 技术积分: 4148
  • 用 户 组: 普通用户
  • 注册时间: 2009-07-15 14:39
个人简介

熟悉Linux下程序设计及各种应用程序 熟悉C Language 熟悉Glusterfs、FFmpeg、CDN 系统设计,计算机图形系统设计、分布式程序设计 目前主要研究方向:流媒体

文章分类

全部博文(141)

分类: C/C++

2013-10-03 17:39:51

使用ffmpeg实现转码样例(代码实现)


使用ffmpeg转码主要工作如下:

Demux -> Decoding -> Encoding -> Muxing


其中接口调用如下:


点击(此处)折叠或打开

  1. av_register_all();
  2. avformat_open_input
  3. avformat_find_stream_info
  4. open_codec_context
  5. av_image_alloc
  6. avcodec_alloc_frame
  7. avformat_alloc_output_context2
  8. avcodec_open2
  9. avcodec_alloc_frame
  10. avpicture_alloc
  11. avpicture_alloc
  12. avformat_write_header
  13. av_init_packet
  14. av_read_frame
  15. avcodec_decode_video2
  16. av_image_copy
  17. avcodec_encode_video2
  18. av_interleaved_write_frame
  19. av_write_trailer


下面的代码为主要将视频转码,封装为h264编码格式的mp4文件,音频为mp3,但是主要操作并不处理音频文件。代码如下



点击(此处)折叠或打开

  1. #include <stdlib.h>
  2. #include <stdio.h>
  3. #include <string.h>
  4. #include <math.h>

  5. #include <libavutil/opt.h>
  6. #include <libavutil/mathematics.h>
  7. #include <libavformat/avformat.h>
  8. #include <libswscale/swscale.h>
  9. #include <libswresample/swresample.h>

  10. /* 5 seconds stream duration */
  11. #define STREAM_DURATION 200.0
  12. #define STREAM_FRAME_RATE 25 /* 25 images/s */
  13. #define STREAM_NB_FRAMES ((int)(STREAM_DURATION * STREAM_FRAME_RATE))
  14. #define STREAM_PIX_FMT AV_PIX_FMT_YUV420P /* default pix_fmt */

  15. static AVFormatContext *input_fmt_ctx = NULL;
  16. static int sws_flags = SWS_BICUBIC;
  17. static AVPacket input_pkt;
  18. static int video_stream_idx = -1;
  19. static uint8_t *video_dst_data[4] = {NULL};
  20. static int video_dst_linesize[4];
  21. static AVStream *input_video_stream;
  22. static AVFrame *input_frame;
  23. static AVCodecContext *video_dec_ctx = NULL;
  24. static int video_dst_bufsize;


  25. static int open_codec_context(int *stream_idx,
  26.  AVFormatContext *fmt_ctx, enum AVMediaType type)
  27. {
  28.  int ret;
  29.  AVStream *st;
  30.  AVCodecContext *dec_ctx = NULL;
  31.  AVCodec *dec = NULL;

  32.  ret = av_find_best_stream(fmt_ctx, type, -1, -1, NULL, 0);
  33.  if (ret < 0) {
  34.  return ret;
  35.  } else {
  36.  *stream_idx = ret;
  37.  st = fmt_ctx->streams[*stream_idx];

  38.  /* find decoder for the stream */
  39.  dec_ctx = st->codec;
  40.  dec = avcodec_find_decoder(dec_ctx->codec_id);
  41.  if (!dec) {
  42.  fprintf(stderr, "Failed to find %s codec\n",
  43.  av_get_media_type_string(type));
  44.  return ret;
  45.  }

  46.  if ((ret = avcodec_open2(dec_ctx, dec, NULL)) < 0) {
  47.  fprintf(stderr, "Failed to open %s codec\n",
  48.  av_get_media_type_string(type));
  49.  return ret;
  50.  }
  51.  }

  52.  return 0;
  53. }

  54. /* Add an output stream. */
  55. static AVStream *add_stream(AVFormatContext *oc, AVCodec **codec,
  56.  enum AVCodecID codec_id)
  57. {
  58.  AVCodecContext *c;
  59.  AVStream *st;

  60.  /* find the encoder */
  61.  *codec = avcodec_find_encoder(codec_id);
  62.  if (!(*codec)) {
  63.  fprintf(stderr, "Could not find encoder for '%s'\n",
  64.  avcodec_get_name(codec_id));
  65.  exit(1);
  66.  }

  67.  st = avformat_new_stream(oc, *codec);
  68.  if (!st) {
  69.  fprintf(stderr, "Could not allocate stream\n");
  70.  exit(1);
  71.  }
  72.  st->id = oc->nb_streams-1;
  73.  c = st->codec;

  74.  switch ((*codec)->type) {
  75.  case AVMEDIA_TYPE_AUDIO:
  76.  c->sample_fmt = AV_SAMPLE_FMT_FLTP;
  77.  c->bit_rate = 64000;
  78.  c->sample_rate = 44100;
  79.  c->channels = 2;
  80.  break;

  81.  case AVMEDIA_TYPE_VIDEO:
  82.  c->codec_id = codec_id;

  83.  c->bit_rate = 400000;
  84.  c->time_base.den = STREAM_FRAME_RATE;
  85.  c->time_base.num = 1;
  86.  c->gop_size = 12;
  87.  c->pix_fmt = STREAM_PIX_FMT;
  88.  break;

  89.  default:
  90.  break;
  91.  }

  92.  /* Some formats want stream headers to be separate. */
  93.  if (oc->oformat->flags & AVFMT_GLOBALHEADER)
  94.  c->flags |= CODEC_FLAG_GLOBAL_HEADER;

  95.  return st;
  96. }

  97. /**************************************************************/
  98. /* audio output */

  99. static float t, tincr, tincr2;

  100. static uint8_t **src_samples_data;
  101. static int src_samples_linesize;
  102. static int src_nb_samples;

  103. static int max_dst_nb_samples;
  104. uint8_t **dst_samples_data;
  105. int dst_samples_linesize;
  106. int dst_samples_size;

  107. struct SwrContext *swr_ctx = NULL;

  108. static void open_audio(AVFormatContext *oc, AVCodec *codec, AVStream *st)
  109. {
  110.  AVCodecContext *c;
  111.  int ret;

  112.  c = st->codec;

  113.  /* open it */
  114.  ret = avcodec_open2(c, codec, NULL);
  115.  if (ret < 0) {
  116.  fprintf(stderr, "Could not open audio codec: %s\n", av_err2str(ret));
  117.  exit(1);
  118.  }

  119.  /* init signal generator */
  120.  t = 0;
  121.  tincr = 2 * M_PI * 110.0 / c->sample_rate;
  122.  /* increment frequency by 110 Hz per second */
  123.  tincr2 = 2 * M_PI * 110.0 / c->sample_rate / c->sample_rate;

  124.  src_nb_samples = c->codec->capabilities & CODEC_CAP_VARIABLE_FRAME_SIZE ?
  125.  10000 : c->frame_size;

  126.  ret = av_samples_alloc_array_and_samples(&src_samples_data, &src_samples_linesize, c->channels,
  127.  src_nb_samples, c->sample_fmt, 0);
  128.  if (ret < 0) {
  129.  fprintf(stderr, "Could not allocate source samples\n");
  130.  exit(1);
  131.  }

  132.  /* create resampler context */
  133.  if (c->sample_fmt != AV_SAMPLE_FMT_S16) {
  134.  swr_ctx = swr_alloc();
  135.  if (!swr_ctx) {
  136.  fprintf(stderr, "Could not allocate resampler context\n");
  137.  exit(1);
  138.  }

  139.  /* set options */
  140.  av_opt_set_int (swr_ctx, "in_channel_count", c->channels, 0);
  141.  av_opt_set_int (swr_ctx, "in_sample_rate", c->sample_rate, 0);
  142.  av_opt_set_sample_fmt(swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
  143.  av_opt_set_int (swr_ctx, "out_channel_count", c->channels, 0);
  144.  av_opt_set_int (swr_ctx, "out_sample_rate", c->sample_rate, 0);
  145.  av_opt_set_sample_fmt(swr_ctx, "out_sample_fmt", c->sample_fmt, 0);

  146.  /* initialize the resampling context */
  147.  if ((ret = swr_init(swr_ctx)) < 0) {
  148.  fprintf(stderr, "Failed to initialize the resampling context\n");
  149.  exit(1);
  150.  }
  151.  }

  152.  /* compute the number of converted samples: buffering is avoided
  153.   * ensuring that the output buffer will contain at least all the
  154.   * converted input samples */
  155.  max_dst_nb_samples = src_nb_samples;
  156.  ret = av_samples_alloc_array_and_samples(&dst_samples_data, &dst_samples_linesize, c->channels,
  157.  max_dst_nb_samples, c->sample_fmt, 0);
  158.  if (ret < 0) {
  159.  fprintf(stderr, "Could not allocate destination samples\n");
  160.  exit(1);
  161.  }
  162.  dst_samples_size = av_samples_get_buffer_size(NULL, c->channels, max_dst_nb_samples,
  163.  c->sample_fmt, 0);
  164. }

  165. /* Prepare a 16 bit dummy audio frame of 'frame_size' samples and
  166.  * 'nb_channels' channels. */
  167. static void get_audio_frame(int16_t *samples, int frame_size, int nb_channels)
  168. {
  169.  int j, i, v;
  170.  int16_t *q;

  171.  q = samples;
  172.  for (j = 0; j < frame_size; j++) {
  173.  v = (int)(sin(t) * 10000);
  174.  for (i = 0; i < nb_channels; i++)
  175.  *q++ = v;
  176.  t += tincr;
  177.  tincr += tincr2;
  178.  }
  179. }

  180. static void write_audio_frame(AVFormatContext *oc, AVStream *st)
  181. {
  182.  AVCodecContext *c;
  183.  AVPacket pkt = { 0 }; // data and size must be 0;
  184.  AVFrame *frame = avcodec_alloc_frame();
  185.  int got_packet, ret, dst_nb_samples;

  186.  av_init_packet(&pkt);
  187.  c = st->codec;

  188.  get_audio_frame((int16_t *)src_samples_data[0], src_nb_samples, c->channels);

  189.  /* convert samples from native format to destination codec format, using the resampler */
  190.  if (swr_ctx) {
  191.  /* compute destination number of samples */
  192.  dst_nb_samples = av_rescale_rnd(swr_get_delay(swr_ctx, c->sample_rate) + src_nb_samples,
  193.  c->sample_rate, c->sample_rate, AV_ROUND_UP);
  194.  if (dst_nb_samples > max_dst_nb_samples) {
  195.  av_free(dst_samples_data[0]);
  196.  ret = av_samples_alloc(dst_samples_data, &dst_samples_linesize, c->channels,
  197.  dst_nb_samples, c->sample_fmt, 0);
  198.  if (ret < 0)
  199.  exit(1);
  200.  max_dst_nb_samples = dst_nb_samples;
  201.  dst_samples_size = av_samples_get_buffer_size(NULL, c->channels, dst_nb_samples,
  202.  c->sample_fmt, 0);
  203.  }

  204.  /* convert to destination format */
  205.  ret = swr_convert(swr_ctx,
  206.  dst_samples_data, dst_nb_samples,
  207.  (const uint8_t **)src_samples_data, src_nb_samples);
  208.  if (ret < 0) {
  209.  fprintf(stderr, "Error while converting\n");
  210.  exit(1);
  211.  }
  212.  } else {
  213.  dst_samples_data[0] = src_samples_data[0];
  214.  dst_nb_samples = src_nb_samples;
  215.  }

  216.  frame->nb_samples = dst_nb_samples;
  217.  avcodec_fill_audio_frame(frame, c->channels, c->sample_fmt,
  218.  dst_samples_data[0], dst_samples_size, 0);

  219.  ret = avcodec_encode_audio2(c, &pkt, frame, &got_packet);
  220.  if (ret < 0) {
  221.  fprintf(stderr, "Error encoding audio frame: %s\n", av_err2str(ret));
  222.  exit(1);
  223.  }

  224.  if (!got_packet)
  225.  return;

  226.  pkt.stream_index = st->index;

  227.  /* Write the compressed frame to the media file. */
  228.  ret = av_interleaved_write_frame(oc, &pkt);
  229.  if (ret != 0) {
  230.  fprintf(stderr, "Error while writing audio frame: %s\n",
  231.  av_err2str(ret));
  232.  exit(1);
  233.  }
  234.  avcodec_free_frame(&frame);
  235. }

  236. static void close_audio(AVFormatContext *oc, AVStream *st)
  237. {
  238.  avcodec_close(st->codec);
  239.  av_free(src_samples_data[0]);
  240.  av_free(dst_samples_data[0]);
  241. }

  242. /***********************bbs.ChinaFFmpeg.com****孙悟空***********************/
  243. /* video output */

  244. static AVFrame *frame;
  245. static AVPicture src_picture, dst_picture;
  246. static int frame_count;

  247. static void open_video(AVFormatContext *oc, AVCodec *codec, AVStream *st)
  248. {
  249.  int ret;
  250.  AVCodecContext *c = st->codec;

  251.  /* open the codec */
  252.  ret = avcodec_open2(c, codec, NULL);
  253.  if (ret < 0) {
  254.  fprintf(stderr, "Could not open video codec: %s\n", av_err2str(ret));
  255.  exit(1);
  256.  }

  257.  /* allocate and init a re-usable frame */
  258.  frame = avcodec_alloc_frame();
  259.  if (!frame) {
  260.  fprintf(stderr, "Could not allocate video frame\n");
  261.  exit(1);
  262.  }

  263.  /* Allocate the encoded raw picture. */
  264.  ret = avpicture_alloc(&dst_picture, c->pix_fmt, c->width, c->height);
  265.  if (ret < 0) {
  266.  fprintf(stderr, "Could not allocate picture: %s\n", av_err2str(ret));
  267.  exit(1);
  268.  }

  269.  /* If the output format is not YUV420P, then a temporary YUV420P
  270.   * picture is needed too. It is then converted to the required
  271.   * output format. */
  272.  if (c->pix_fmt != AV_PIX_FMT_YUV420P) {
  273.  ret = avpicture_alloc(&src_picture, AV_PIX_FMT_YUV420P, c->width, c->height);
  274.  if (ret < 0) {
  275.  fprintf(stderr, "Could not allocate temporary picture: %s\n",
  276.  av_err2str(ret));
  277.  exit(1);
  278.  }
  279.  }

  280.  /* copy data and linesize picture pointers to frame */
  281.  *((AVPicture *)frame) = dst_picture;
  282. }

  283. static void write_video_frame(AVFormatContext *oc, AVStream *st)
  284. {
  285.  int ret;
  286.  static struct SwsContext *sws_ctx;
  287.  AVCodecContext *c = st->codec;

  288.  // fill_yuv_image(&dst_picture, frame_count, c->width, c->height);


  289.  if (oc->oformat->flags & AVFMT_RAWPICTURE) {
  290.  /* Raw video case - directly store the picture in the packet */
  291.  AVPacket pkt;
  292.  av_init_packet(&pkt);

  293.  pkt.flags |= AV_PKT_FLAG_KEY;
  294.  pkt.stream_index = st->index;
  295.  pkt.data = dst_picture.data[0];
  296.  pkt.size = sizeof(AVPicture);

  297.  ret = av_interleaved_write_frame(oc, &pkt);
  298.  } else {
  299.  AVPacket pkt = { 0 };
  300.  int got_packet;
  301.  av_init_packet(&pkt);

  302.  /* encode the image */
  303.  ret = avcodec_encode_video2(c, &pkt, frame, &got_packet);
  304.  if (ret < 0) {
  305.  fprintf(stderr, "Error encoding video frame: %s\n", av_err2str(ret));
  306.  exit(1);
  307.  }
  308.  /* If size is zero, it means the image was buffered. */

  309.  if (!ret && got_packet && pkt.size) {
  310.  pkt.stream_index = st->index;

  311.  /* Write the compressed frame to the media file. */
  312.  ret = av_interleaved_write_frame(oc, &pkt);
  313.  } else {
  314.  ret = 0;
  315.  }
  316.  }
  317.  if (ret != 0) {
  318.  fprintf(stderr, "Error while writing video frame: %s\n", av_err2str(ret));
  319.  exit(1);
  320.  }
  321.  frame_count++;
  322. }

  323. static void close_video(AVFormatContext *oc, AVStream *st)
  324. {
  325.  avcodec_close(st->codec);
  326.  av_free(src_picture.data[0]);
  327.  av_free(dst_picture.data[0]);
  328.  av_free(frame);
  329. }

  330. /**************************************************************/
  331. /* media file output */

  332. int main(int argc, char **argv)
  333. {
  334.  const char *filename;
  335.  const char *input_file;
  336.  AVOutputFormat *fmt;
  337.  AVFormatContext *oc;
  338.  AVStream *audio_st, *video_st;
  339.  AVCodec *audio_codec, *video_codec;
  340.  double audio_time, video_time;
  341.  int ret;
  342.  int decoded;
  343.  int got_frame;

  344.  /* Initialize libavcodec, and register all codecs and formats. */
  345.  av_register_all();

  346.  if (argc != 3) {
  347.  printf("usage: %s output_file inputfile\n"
  348.  "API example program to output a media file with libavformat.\n"
  349.  "This program generates a synthetic audio and video stream, encodes and\n"
  350.  "muxes them into a file named output_file.\n"
  351.  "The output format is automatically guessed according to the file extension.\n"
  352.  "Raw images can also be output by using '%%d' in the filename.\n"
  353.  "\n", argv[0]);
  354.  return 1;
  355.  }

  356.  input_file = argv[2];
  357.  filename = argv[1];

  358.  if (avformat_open_input( &input_fmt_ctx, input_file, NULL, NULL) < 0) {
  359.  fprintf(stderr, "Could not open source file %s\n", input_file);
  360.  exit(-1);
  361.  }

  362.  /* retrieve stream information */
  363.  if (avformat_find_stream_info(input_fmt_ctx, NULL) < 0) {
  364.  fprintf(stderr, "Could not find stream information\n");
  365.  exit(1);
  366.  }

  367.  if (open_codec_context(&video_stream_idx, input_fmt_ctx, AVMEDIA_TYPE_VIDEO) >= 0) {
  368.  input_video_stream = input_fmt_ctx->streams[video_stream_idx];
  369.  video_dec_ctx = input_video_stream->codec;
  370.  ret = av_image_alloc(video_dst_data, video_dst_linesize,
  371.  video_dec_ctx->width, video_dec_ctx->height,
  372.  video_dec_ctx->pix_fmt, 1);
  373.  if (ret < 0) {
  374.  fprintf(stderr, "Could not allocate raw video buffer\n");
  375.  }
  376.  video_dst_bufsize = ret;
  377.  }


  378.  input_frame = avcodec_alloc_frame();
  379.  if (!input_frame) {
  380.  fprintf(stderr, "Could not allocate frame\n");
  381.  ret = AVERROR(ENOMEM);
  382.  exit(-1);
  383.  }

  384.  /* allocate the output media context */
  385.  avformat_alloc_output_context2(&oc, NULL, NULL, filename);
  386.  if (!oc) {
  387.  printf("Could not deduce output format from file extension: using MPEG.\n");
  388.  avformat_alloc_output_context2(&oc, NULL, "mpeg", filename);
  389.  }
  390.  if (!oc) {
  391.  return 1;
  392.  }
  393.  fmt = oc->oformat;

  394.  video_st = NULL;
  395.  audio_st = NULL;

  396.  fmt->video_codec = AV_CODEC_ID_H264;
  397.  if (fmt->video_codec != AV_CODEC_ID_NONE) {
  398.  video_st = add_stream(oc, &video_codec, AV_CODEC_ID_H264);
  399.  }
  400.  if (video_st->codec) {
  401.  video_st->codec->width = video_dec_ctx->width;
  402.  video_st->codec->height = video_dec_ctx->height;
  403.  }
  404.  fmt->audio_codec = AV_CODEC_ID_MP3;
  405.  if (fmt->audio_codec != AV_CODEC_ID_NONE) {
  406.  audio_st = add_stream(oc, &audio_codec, AV_CODEC_ID_MP3);
  407.  }

  408.  /* Now that all the parameters are set, we can open the audio and
  409.   * video codecs and allocate the necessary encode buffers. */
  410.  if (video_st)
  411.  open_video(oc, video_codec, video_st);
  412.  if (audio_st)
  413.  open_audio(oc, audio_codec, audio_st);

  414.  av_dump_format(oc, 0, filename, 1);


  415.  /* open the output file, if needed */
  416.  if (!(fmt->flags & AVFMT_NOFILE)) {
  417.  ret = avio_open(&oc->pb, filename, AVIO_FLAG_WRITE);
  418.  if (ret < 0) {
  419.  fprintf(stderr, "Could not open '%s': %s\n", filename,
  420.  av_err2str(ret));
  421.  return 1;
  422.  }
  423.  }

  424.  /* Write the stream header, if any. */
  425.  ret = avformat_write_header(oc, NULL);
  426.  if (ret < 0) {
  427.  fprintf(stderr, "Error occurred when opening output file: %s\n",
  428.  av_err2str(ret));
  429.  return 1;
  430.  }

  431.  av_init_packet(&input_pkt);
  432.  input_pkt.data = NULL;
  433.  input_pkt.size = 0;


  434.  if (frame)
  435.  frame->pts = 0;
  436.  for (;;) {
  437.  /* Compute current audio and video time. */
  438.  audio_time = audio_st ? audio_st->pts.val * av_q2d(audio_st->time_base) : 0.0;
  439.  video_time = video_st ? video_st->pts.val * av_q2d(video_st->time_base) : 0.0;

  440.  if ((!audio_st || audio_time >= STREAM_DURATION) &&
  441.  (!video_st || video_time >= STREAM_DURATION))
  442.  break;

  443.  /* write interleaved audio and video frames */
  444.  if (!video_st || (video_st && audio_st && audio_time < video_time)) {
  445.  write_audio_frame(oc, audio_st);
  446.  } else {
  447.  av_read_frame(input_fmt_ctx, &input_pkt);

  448.  decoded = input_pkt.size;
  449.  if (input_pkt.stream_index == video_stream_idx) {
  450.  /* decode video frame */
  451.  ret = avcodec_decode_video2(video_dec_ctx, input_frame, &got_frame, &input_pkt);
  452.  if (ret < 0) {
  453.  fprintf(stderr, "Error decoding video frame\n");
  454.  return ret;
  455.  }

  456.  if (got_frame) {
  457.  av_image_copy(&dst_picture, video_dst_linesize,
  458.  (const uint8_t **)(input_frame->data), input_frame->linesize,
  459.  video_dec_ctx->pix_fmt, video_dec_ctx->width, video_dec_ctx->height);
  460.  }
  461.  }

  462.  write_video_frame(oc, video_st);
  463.  frame->pts += av_rescale_q(1, video_st->codec->time_base, video_st->time_base);
  464.  }
  465.  }

  466.  av_write_trailer(oc);

  467.  /* Close each codec. */
  468.  if (video_st)
  469.  close_video(oc, video_st);
  470.  if (audio_st)
  471.  close_audio(oc, audio_st);

  472.  if (!(fmt->flags & AVFMT_NOFILE))
  473.  avio_close(oc->pb);

  474.  avformat_free_context(oc);
  475.  av_free(input_frame);
  476.  av_free(video_dst_data[0]);
  477.  avcodec_close(video_dec_ctx);
  478.  return 0;
  479. }

以上代码为从dox/example/muxing.c中修改得到

编译命令如下:

点击(此处)折叠或打开

  1. gcc -g doc/examples/muxing.c -o muxing -lavcodec -lavdevice -lavfilter -lavformat -lavutil -lswscale -lswresample -lpostproc -lx264 -lmp3lame -lz -liconv -lbz2
执行命令如下:

点击(此处)折叠或打开

  1. ./muxing output.mp4 ~/Movies/input.avi
执行后效果如下:

查看转码完成后的多媒体文件的信息:

查看转码后的文件的视频:






阅读(51033) | 评论(0) | 转发(6) |
给主人留下些什么吧!~~