FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
muxing.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2003 Fabrice Bellard
3  *
4  * Permission is hereby granted, free of charge, to any person obtaining a copy
5  * of this software and associated documentation files (the "Software"), to deal
6  * in the Software without restriction, including without limitation the rights
7  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
8  * copies of the Software, and to permit persons to whom the Software is
9  * furnished to do so, subject to the following conditions:
10  *
11  * The above copyright notice and this permission notice shall be included in
12  * all copies or substantial portions of the Software.
13  *
14  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
15  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
16  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
17  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
18  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
19  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
20  * THE SOFTWARE.
21  */
22 
23 /**
24  * @file
25  * libavformat API example.
26  *
27  * Output a media file in any supported libavformat format. The default
28  * codecs are used.
29  * @example muxing.c
30  */
31 
32 #include <stdlib.h>
33 #include <stdio.h>
34 #include <string.h>
35 #include <math.h>
36 
37 #include <libavutil/avassert.h>
39 #include <libavutil/opt.h>
40 #include <libavutil/mathematics.h>
41 #include <libavutil/timestamp.h>
42 #include <libavformat/avformat.h>
43 #include <libswscale/swscale.h>
45 
46 #define STREAM_DURATION 10.0
47 #define STREAM_FRAME_RATE 25 /* 25 images/s */
48 #define STREAM_PIX_FMT AV_PIX_FMT_YUV420P /* default pix_fmt */
49 
50 #define SCALE_FLAGS SWS_BICUBIC
51 
52 // a wrapper around a single output AVStream
53 typedef struct OutputStream {
55 
56  /* pts of the next frame that will be generated */
57  int64_t next_pts;
59 
62 
63  float t, tincr, tincr2;
64 
67 } OutputStream;
68 
69 static void log_packet(const AVFormatContext *fmt_ctx, const AVPacket *pkt)
70 {
71  AVRational *time_base = &fmt_ctx->streams[pkt->stream_index]->time_base;
72 
73  printf("pts:%s pts_time:%s dts:%s dts_time:%s duration:%s duration_time:%s stream_index:%d\n",
74  av_ts2str(pkt->pts), av_ts2timestr(pkt->pts, time_base),
75  av_ts2str(pkt->dts), av_ts2timestr(pkt->dts, time_base),
76  av_ts2str(pkt->duration), av_ts2timestr(pkt->duration, time_base),
77  pkt->stream_index);
78 }
79 
80 static int write_frame(AVFormatContext *fmt_ctx, const AVRational *time_base, AVStream *st, AVPacket *pkt)
81 {
82  /* rescale output packet timestamp values from codec to stream timebase */
83  av_packet_rescale_ts(pkt, *time_base, st->time_base);
84  pkt->stream_index = st->index;
85 
86  /* Write the compressed frame to the media file. */
87  log_packet(fmt_ctx, pkt);
88  return av_interleaved_write_frame(fmt_ctx, pkt);
89 }
90 
91 /* Add an output stream. */
92 static void add_stream(OutputStream *ost, AVFormatContext *oc,
93  AVCodec **codec,
94  enum AVCodecID codec_id)
95 {
97  int i;
98 
99  /* find the encoder */
100  *codec = avcodec_find_encoder(codec_id);
101  if (!(*codec)) {
102  fprintf(stderr, "Could not find encoder for '%s'\n",
103  avcodec_get_name(codec_id));
104  exit(1);
105  }
106 
107  ost->st = avformat_new_stream(oc, *codec);
108  if (!ost->st) {
109  fprintf(stderr, "Could not allocate stream\n");
110  exit(1);
111  }
112  ost->st->id = oc->nb_streams-1;
113  c = ost->st->codec;
114 
115  switch ((*codec)->type) {
116  case AVMEDIA_TYPE_AUDIO:
117  c->sample_fmt = (*codec)->sample_fmts ?
118  (*codec)->sample_fmts[0] : AV_SAMPLE_FMT_FLTP;
119  c->bit_rate = 64000;
120  c->sample_rate = 44100;
121  if ((*codec)->supported_samplerates) {
122  c->sample_rate = (*codec)->supported_samplerates[0];
123  for (i = 0; (*codec)->supported_samplerates[i]; i++) {
124  if ((*codec)->supported_samplerates[i] == 44100)
125  c->sample_rate = 44100;
126  }
127  }
130  if ((*codec)->channel_layouts) {
131  c->channel_layout = (*codec)->channel_layouts[0];
132  for (i = 0; (*codec)->channel_layouts[i]; i++) {
133  if ((*codec)->channel_layouts[i] == AV_CH_LAYOUT_STEREO)
135  }
136  }
138  ost->st->time_base = (AVRational){ 1, c->sample_rate };
139  break;
140 
141  case AVMEDIA_TYPE_VIDEO:
142  c->codec_id = codec_id;
143 
144  c->bit_rate = 400000;
145  /* Resolution must be a multiple of two. */
146  c->width = 352;
147  c->height = 288;
148  /* timebase: This is the fundamental unit of time (in seconds) in terms
149  * of which frame timestamps are represented. For fixed-fps content,
150  * timebase should be 1/framerate and timestamp increments should be
151  * identical to 1. */
152  ost->st->time_base = (AVRational){ 1, STREAM_FRAME_RATE };
153  c->time_base = ost->st->time_base;
154 
155  c->gop_size = 12; /* emit one intra frame every twelve frames at most */
156  c->pix_fmt = STREAM_PIX_FMT;
157  if (c->codec_id == AV_CODEC_ID_MPEG2VIDEO) {
158  /* just for testing, we also add B frames */
159  c->max_b_frames = 2;
160  }
161  if (c->codec_id == AV_CODEC_ID_MPEG1VIDEO) {
162  /* Needed to avoid using macroblocks in which some coeffs overflow.
163  * This does not happen with normal video, it just happens here as
164  * the motion of the chroma plane does not match the luma plane. */
165  c->mb_decision = 2;
166  }
167  break;
168 
169  default:
170  break;
171  }
172 
173  /* Some formats want stream headers to be separate. */
174  if (oc->oformat->flags & AVFMT_GLOBALHEADER)
176 }
177 
178 /**************************************************************/
179 /* audio output */
180 
181 static AVFrame *alloc_audio_frame(enum AVSampleFormat sample_fmt,
182  uint64_t channel_layout,
183  int sample_rate, int nb_samples)
184 {
186  int ret;
187 
188  if (!frame) {
189  fprintf(stderr, "Error allocating an audio frame\n");
190  exit(1);
191  }
192 
193  frame->format = sample_fmt;
194  frame->channel_layout = channel_layout;
195  frame->sample_rate = sample_rate;
196  frame->nb_samples = nb_samples;
197 
198  if (nb_samples) {
199  ret = av_frame_get_buffer(frame, 0);
200  if (ret < 0) {
201  fprintf(stderr, "Error allocating an audio buffer\n");
202  exit(1);
203  }
204  }
205 
206  return frame;
207 }
208 
209 static void open_audio(AVFormatContext *oc, AVCodec *codec, OutputStream *ost, AVDictionary *opt_arg)
210 {
211  AVCodecContext *c;
212  int nb_samples;
213  int ret;
214  AVDictionary *opt = NULL;
215 
216  c = ost->st->codec;
217 
218  /* open it */
219  av_dict_copy(&opt, opt_arg, 0);
220  ret = avcodec_open2(c, codec, &opt);
221  av_dict_free(&opt);
222  if (ret < 0) {
223  fprintf(stderr, "Could not open audio codec: %s\n", av_err2str(ret));
224  exit(1);
225  }
226 
227  /* init signal generator */
228  ost->t = 0;
229  ost->tincr = 2 * M_PI * 110.0 / c->sample_rate;
230  /* increment frequency by 110 Hz per second */
231  ost->tincr2 = 2 * M_PI * 110.0 / c->sample_rate / c->sample_rate;
232 
234  nb_samples = 10000;
235  else
236  nb_samples = c->frame_size;
237 
239  c->sample_rate, nb_samples);
241  c->sample_rate, nb_samples);
242 
243  /* create resampler context */
244  ost->swr_ctx = swr_alloc();
245  if (!ost->swr_ctx) {
246  fprintf(stderr, "Could not allocate resampler context\n");
247  exit(1);
248  }
249 
250  /* set options */
251  av_opt_set_int (ost->swr_ctx, "in_channel_count", c->channels, 0);
252  av_opt_set_int (ost->swr_ctx, "in_sample_rate", c->sample_rate, 0);
253  av_opt_set_sample_fmt(ost->swr_ctx, "in_sample_fmt", AV_SAMPLE_FMT_S16, 0);
254  av_opt_set_int (ost->swr_ctx, "out_channel_count", c->channels, 0);
255  av_opt_set_int (ost->swr_ctx, "out_sample_rate", c->sample_rate, 0);
256  av_opt_set_sample_fmt(ost->swr_ctx, "out_sample_fmt", c->sample_fmt, 0);
257 
258  /* initialize the resampling context */
259  if ((ret = swr_init(ost->swr_ctx)) < 0) {
260  fprintf(stderr, "Failed to initialize the resampling context\n");
261  exit(1);
262  }
263 }
264 
265 /* Prepare a 16 bit dummy audio frame of 'frame_size' samples and
266  * 'nb_channels' channels. */
268 {
269  AVFrame *frame = ost->tmp_frame;
270  int j, i, v;
271  int16_t *q = (int16_t*)frame->data[0];
272 
273  /* check if we want to generate more frames */
274  if (av_compare_ts(ost->next_pts, ost->st->codec->time_base,
275  STREAM_DURATION, (AVRational){ 1, 1 }) >= 0)
276  return NULL;
277 
278  for (j = 0; j <frame->nb_samples; j++) {
279  v = (int)(sin(ost->t) * 10000);
280  for (i = 0; i < ost->st->codec->channels; i++)
281  *q++ = v;
282  ost->t += ost->tincr;
283  ost->tincr += ost->tincr2;
284  }
285 
286  frame->pts = ost->next_pts;
287  ost->next_pts += frame->nb_samples;
288 
289  return frame;
290 }
291 
292 /*
293  * encode one audio frame and send it to the muxer
294  * return 1 when encoding is finished, 0 otherwise
295  */
297 {
298  AVCodecContext *c;
299  AVPacket pkt = { 0 }; // data and size must be 0;
300  AVFrame *frame;
301  int ret;
302  int got_packet;
303  int dst_nb_samples;
304 
305  av_init_packet(&pkt);
306  c = ost->st->codec;
307 
308  frame = get_audio_frame(ost);
309 
310  if (frame) {
311  /* convert samples from native format to destination codec format, using the resampler */
312  /* compute destination number of samples */
313  dst_nb_samples = av_rescale_rnd(swr_get_delay(ost->swr_ctx, c->sample_rate) + frame->nb_samples,
315  av_assert0(dst_nb_samples == frame->nb_samples);
316 
317  /* when we pass a frame to the encoder, it may keep a reference to it
318  * internally;
319  * make sure we do not overwrite it here
320  */
321  ret = av_frame_make_writable(ost->frame);
322  if (ret < 0)
323  exit(1);
324 
325  /* convert to destination format */
326  ret = swr_convert(ost->swr_ctx,
327  ost->frame->data, dst_nb_samples,
328  (const uint8_t **)frame->data, frame->nb_samples);
329  if (ret < 0) {
330  fprintf(stderr, "Error while converting\n");
331  exit(1);
332  }
333  frame = ost->frame;
334 
335  frame->pts = av_rescale_q(ost->samples_count, (AVRational){1, c->sample_rate}, c->time_base);
336  ost->samples_count += dst_nb_samples;
337  }
338 
339  ret = avcodec_encode_audio2(c, &pkt, frame, &got_packet);
340  if (ret < 0) {
341  fprintf(stderr, "Error encoding audio frame: %s\n", av_err2str(ret));
342  exit(1);
343  }
344 
345  if (got_packet) {
346  ret = write_frame(oc, &c->time_base, ost->st, &pkt);
347  if (ret < 0) {
348  fprintf(stderr, "Error while writing audio frame: %s\n",
349  av_err2str(ret));
350  exit(1);
351  }
352  }
353 
354  return (frame || got_packet) ? 0 : 1;
355 }
356 
357 /**************************************************************/
358 /* video output */
359 
361 {
362  AVFrame *picture;
363  int ret;
364 
365  picture = av_frame_alloc();
366  if (!picture)
367  return NULL;
368 
369  picture->format = pix_fmt;
370  picture->width = width;
371  picture->height = height;
372 
373  /* allocate the buffers for the frame data */
374  ret = av_frame_get_buffer(picture, 32);
375  if (ret < 0) {
376  fprintf(stderr, "Could not allocate frame data.\n");
377  exit(1);
378  }
379 
380  return picture;
381 }
382 
383 static void open_video(AVFormatContext *oc, AVCodec *codec, OutputStream *ost, AVDictionary *opt_arg)
384 {
385  int ret;
386  AVCodecContext *c = ost->st->codec;
387  AVDictionary *opt = NULL;
388 
389  av_dict_copy(&opt, opt_arg, 0);
390 
391  /* open the codec */
392  ret = avcodec_open2(c, codec, &opt);
393  av_dict_free(&opt);
394  if (ret < 0) {
395  fprintf(stderr, "Could not open video codec: %s\n", av_err2str(ret));
396  exit(1);
397  }
398 
399  /* allocate and init a re-usable frame */
400  ost->frame = alloc_picture(c->pix_fmt, c->width, c->height);
401  if (!ost->frame) {
402  fprintf(stderr, "Could not allocate video frame\n");
403  exit(1);
404  }
405 
406  /* If the output format is not YUV420P, then a temporary YUV420P
407  * picture is needed too. It is then converted to the required
408  * output format. */
409  ost->tmp_frame = NULL;
410  if (c->pix_fmt != AV_PIX_FMT_YUV420P) {
412  if (!ost->tmp_frame) {
413  fprintf(stderr, "Could not allocate temporary picture\n");
414  exit(1);
415  }
416  }
417 }
418 
419 /* Prepare a dummy image. */
420 static void fill_yuv_image(AVFrame *pict, int frame_index,
421  int width, int height)
422 {
423  int x, y, i, ret;
424 
425  /* when we pass a frame to the encoder, it may keep a reference to it
426  * internally;
427  * make sure we do not overwrite it here
428  */
429  ret = av_frame_make_writable(pict);
430  if (ret < 0)
431  exit(1);
432 
433  i = frame_index;
434 
435  /* Y */
436  for (y = 0; y < height; y++)
437  for (x = 0; x < width; x++)
438  pict->data[0][y * pict->linesize[0] + x] = x + y + i * 3;
439 
440  /* Cb and Cr */
441  for (y = 0; y < height / 2; y++) {
442  for (x = 0; x < width / 2; x++) {
443  pict->data[1][y * pict->linesize[1] + x] = 128 + y + i * 2;
444  pict->data[2][y * pict->linesize[2] + x] = 64 + x + i * 5;
445  }
446  }
447 }
448 
450 {
451  AVCodecContext *c = ost->st->codec;
452 
453  /* check if we want to generate more frames */
454  if (av_compare_ts(ost->next_pts, ost->st->codec->time_base,
455  STREAM_DURATION, (AVRational){ 1, 1 }) >= 0)
456  return NULL;
457 
458  if (c->pix_fmt != AV_PIX_FMT_YUV420P) {
459  /* as we only generate a YUV420P picture, we must convert it
460  * to the codec pixel format if needed */
461  if (!ost->sws_ctx) {
462  ost->sws_ctx = sws_getContext(c->width, c->height,
464  c->width, c->height,
465  c->pix_fmt,
467  if (!ost->sws_ctx) {
468  fprintf(stderr,
469  "Could not initialize the conversion context\n");
470  exit(1);
471  }
472  }
473  fill_yuv_image(ost->tmp_frame, ost->next_pts, c->width, c->height);
474  sws_scale(ost->sws_ctx,
475  (const uint8_t * const *)ost->tmp_frame->data, ost->tmp_frame->linesize,
476  0, c->height, ost->frame->data, ost->frame->linesize);
477  } else {
478  fill_yuv_image(ost->frame, ost->next_pts, c->width, c->height);
479  }
480 
481  ost->frame->pts = ost->next_pts++;
482 
483  return ost->frame;
484 }
485 
486 /*
487  * encode one video frame and send it to the muxer
488  * return 1 when encoding is finished, 0 otherwise
489  */
491 {
492  int ret;
493  AVCodecContext *c;
494  AVFrame *frame;
495  int got_packet = 0;
496 
497  c = ost->st->codec;
498 
499  frame = get_video_frame(ost);
500 
501  if (oc->oformat->flags & AVFMT_RAWPICTURE) {
502  /* a hack to avoid data copy with some raw video muxers */
503  AVPacket pkt;
504  av_init_packet(&pkt);
505 
506  if (!frame)
507  return 1;
508 
509  pkt.flags |= AV_PKT_FLAG_KEY;
510  pkt.stream_index = ost->st->index;
511  pkt.data = (uint8_t *)frame;
512  pkt.size = sizeof(AVPicture);
513 
514  pkt.pts = pkt.dts = frame->pts;
515  av_packet_rescale_ts(&pkt, c->time_base, ost->st->time_base);
516 
517  ret = av_interleaved_write_frame(oc, &pkt);
518  } else {
519  AVPacket pkt = { 0 };
520  av_init_packet(&pkt);
521 
522  /* encode the image */
523  ret = avcodec_encode_video2(c, &pkt, frame, &got_packet);
524  if (ret < 0) {
525  fprintf(stderr, "Error encoding video frame: %s\n", av_err2str(ret));
526  exit(1);
527  }
528 
529  if (got_packet) {
530  ret = write_frame(oc, &c->time_base, ost->st, &pkt);
531  } else {
532  ret = 0;
533  }
534  }
535 
536  if (ret < 0) {
537  fprintf(stderr, "Error while writing video frame: %s\n", av_err2str(ret));
538  exit(1);
539  }
540 
541  return (frame || got_packet) ? 0 : 1;
542 }
543 
545 {
546  avcodec_close(ost->st->codec);
547  av_frame_free(&ost->frame);
548  av_frame_free(&ost->tmp_frame);
549  sws_freeContext(ost->sws_ctx);
550  swr_free(&ost->swr_ctx);
551 }
552 
553 /**************************************************************/
554 /* media file output */
555 
556 int main(int argc, char **argv)
557 {
558  OutputStream video_st = { 0 }, audio_st = { 0 };
559  const char *filename;
561  AVFormatContext *oc;
562  AVCodec *audio_codec, *video_codec;
563  int ret;
564  int have_video = 0, have_audio = 0;
565  int encode_video = 0, encode_audio = 0;
566  AVDictionary *opt = NULL;
567 
568  /* Initialize libavcodec, and register all codecs and formats. */
569  av_register_all();
570 
571  if (argc < 2) {
572  printf("usage: %s output_file\n"
573  "API example program to output a media file with libavformat.\n"
574  "This program generates a synthetic audio and video stream, encodes and\n"
575  "muxes them into a file named output_file.\n"
576  "The output format is automatically guessed according to the file extension.\n"
577  "Raw images can also be output by using '%%d' in the filename.\n"
578  "\n", argv[0]);
579  return 1;
580  }
581 
582  filename = argv[1];
583  if (argc > 3 && !strcmp(argv[2], "-flags")) {
584  av_dict_set(&opt, argv[2]+1, argv[3], 0);
585  }
586 
587  /* allocate the output media context */
588  avformat_alloc_output_context2(&oc, NULL, NULL, filename);
589  if (!oc) {
590  printf("Could not deduce output format from file extension: using MPEG.\n");
591  avformat_alloc_output_context2(&oc, NULL, "mpeg", filename);
592  }
593  if (!oc)
594  return 1;
595 
596  fmt = oc->oformat;
597 
598  /* Add the audio and video streams using the default format codecs
599  * and initialize the codecs. */
600  if (fmt->video_codec != AV_CODEC_ID_NONE) {
601  add_stream(&video_st, oc, &video_codec, fmt->video_codec);
602  have_video = 1;
603  encode_video = 1;
604  }
605  if (fmt->audio_codec != AV_CODEC_ID_NONE) {
606  add_stream(&audio_st, oc, &audio_codec, fmt->audio_codec);
607  have_audio = 1;
608  encode_audio = 1;
609  }
610 
611  /* Now that all the parameters are set, we can open the audio and
612  * video codecs and allocate the necessary encode buffers. */
613  if (have_video)
614  open_video(oc, video_codec, &video_st, opt);
615 
616  if (have_audio)
617  open_audio(oc, audio_codec, &audio_st, opt);
618 
619  av_dump_format(oc, 0, filename, 1);
620 
621  /* open the output file, if needed */
622  if (!(fmt->flags & AVFMT_NOFILE)) {
623  ret = avio_open(&oc->pb, filename, AVIO_FLAG_WRITE);
624  if (ret < 0) {
625  fprintf(stderr, "Could not open '%s': %s\n", filename,
626  av_err2str(ret));
627  return 1;
628  }
629  }
630 
631  /* Write the stream header, if any. */
632  ret = avformat_write_header(oc, &opt);
633  if (ret < 0) {
634  fprintf(stderr, "Error occurred when opening output file: %s\n",
635  av_err2str(ret));
636  return 1;
637  }
638 
639  while (encode_video || encode_audio) {
640  /* select the stream to encode */
641  if (encode_video &&
642  (!encode_audio || av_compare_ts(video_st.next_pts, video_st.st->codec->time_base,
643  audio_st.next_pts, audio_st.st->codec->time_base) <= 0)) {
644  encode_video = !write_video_frame(oc, &video_st);
645  } else {
646  encode_audio = !write_audio_frame(oc, &audio_st);
647  }
648  }
649 
650  /* Write the trailer, if any. The trailer must be written before you
651  * close the CodecContexts open when you wrote the header; otherwise
652  * av_write_trailer() may try to use memory that was freed on
653  * av_codec_close(). */
654  av_write_trailer(oc);
655 
656  /* Close each codec. */
657  if (have_video)
658  close_stream(oc, &video_st);
659  if (have_audio)
660  close_stream(oc, &audio_st);
661 
662  if (!(fmt->flags & AVFMT_NOFILE))
663  /* Close the output file. */
664  avio_closep(&oc->pb);
665 
666  /* free the stream */
668 
669  return 0;
670 }