FFmpeg
 All Data Structures Namespaces Files Functions Variables Typedefs Enumerations Enumerator Macros Groups Pages
demuxing_decoding.c
Go to the documentation of this file.
1 /*
2  * Copyright (c) 2012 Stefano Sabatini
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  * Demuxing and decoding example.
26  *
27  * Show how to use the libavformat and libavcodec API to demux and
28  * decode audio and video data.
29  * @example demuxing_decoding.c
30  */
31 
32 #include <libavutil/imgutils.h>
33 #include <libavutil/samplefmt.h>
34 #include <libavutil/timestamp.h>
35 #include <libavformat/avformat.h>
36 
39 static int width, height;
40 static enum AVPixelFormat pix_fmt;
42 static const char *src_filename = NULL;
43 static const char *video_dst_filename = NULL;
44 static const char *audio_dst_filename = NULL;
45 static FILE *video_dst_file = NULL;
46 static FILE *audio_dst_file = NULL;
47 
48 static uint8_t *video_dst_data[4] = {NULL};
49 static int video_dst_linesize[4];
50 static int video_dst_bufsize;
51 
52 static int video_stream_idx = -1, audio_stream_idx = -1;
53 static AVFrame *frame = NULL;
54 static AVPacket pkt;
55 static int video_frame_count = 0;
56 static int audio_frame_count = 0;
57 
58 /* The different ways of decoding and managing data memory. You are not
59  * supposed to support all the modes in your application but pick the one most
60  * appropriate to your needs. Look for the use of api_mode in this example to
61  * see what are the differences of API usage between them */
62 enum {
63  API_MODE_OLD = 0, /* old method, deprecated */
64  API_MODE_NEW_API_REF_COUNT = 1, /* new method, using the frame reference counting */
65  API_MODE_NEW_API_NO_REF_COUNT = 2, /* new method, without reference counting */
66 };
67 
68 static int api_mode = API_MODE_OLD;
69 
70 static int decode_packet(int *got_frame, int cached)
71 {
72  int ret = 0;
73  int decoded = pkt.size;
74 
75  *got_frame = 0;
76 
77  if (pkt.stream_index == video_stream_idx) {
78  /* decode video frame */
79  ret = avcodec_decode_video2(video_dec_ctx, frame, got_frame, &pkt);
80  if (ret < 0) {
81  fprintf(stderr, "Error decoding video frame (%s)\n", av_err2str(ret));
82  return ret;
83  }
84  if (video_dec_ctx->width != width || video_dec_ctx->height != height ||
85  video_dec_ctx->pix_fmt != pix_fmt) {
86  /* To handle this change, one could call av_image_alloc again and
87  * decode the following frames into another rawvideo file. */
88  fprintf(stderr, "Error: Width, height and pixel format have to be "
89  "constant in a rawvideo file, but the width, height or "
90  "pixel format of the input video changed:\n"
91  "old: width = %d, height = %d, format = %s\n"
92  "new: width = %d, height = %d, format = %s\n",
94  video_dec_ctx->width, video_dec_ctx->height,
95  av_get_pix_fmt_name(video_dec_ctx->pix_fmt));
96  return -1;
97  }
98 
99  if (*got_frame) {
100  printf("video_frame%s n:%d coded_n:%d pts:%s\n",
101  cached ? "(cached)" : "",
103  av_ts2timestr(frame->pts, &video_dec_ctx->time_base));
104 
105  /* copy decoded frame to destination buffer:
106  * this is required since rawvideo expects non aligned data */
108  (const uint8_t **)(frame->data), frame->linesize,
109  pix_fmt, width, height);
110 
111  /* write to rawvideo file */
113  }
114  } else if (pkt.stream_index == audio_stream_idx) {
115  /* decode audio frame */
116  ret = avcodec_decode_audio4(audio_dec_ctx, frame, got_frame, &pkt);
117  if (ret < 0) {
118  fprintf(stderr, "Error decoding audio frame (%s)\n", av_err2str(ret));
119  return ret;
120  }
121  /* Some audio decoders decode only part of the packet, and have to be
122  * called again with the remainder of the packet data.
123  * Sample: fate-suite/lossless-audio/luckynight-partial.shn
124  * Also, some decoders might over-read the packet. */
125  decoded = FFMIN(ret, pkt.size);
126 
127  if (*got_frame) {
128  size_t unpadded_linesize = frame->nb_samples * av_get_bytes_per_sample(frame->format);
129  printf("audio_frame%s n:%d nb_samples:%d pts:%s\n",
130  cached ? "(cached)" : "",
131  audio_frame_count++, frame->nb_samples,
133 
134  /* Write the raw audio data samples of the first plane. This works
135  * fine for packed formats (e.g. AV_SAMPLE_FMT_S16). However,
136  * most audio decoders output planar audio, which uses a separate
137  * plane of audio samples for each channel (e.g. AV_SAMPLE_FMT_S16P).
138  * In other words, this code will write only the first audio channel
139  * in these cases.
140  * You should use libswresample or libavfilter to convert the frame
141  * to packed data. */
142  fwrite(frame->extended_data[0], 1, unpadded_linesize, audio_dst_file);
143  }
144  }
145 
146  /* If we use the new API with reference counting, we own the data and need
147  * to de-reference it when we don't use it anymore */
148  if (*got_frame && api_mode == API_MODE_NEW_API_REF_COUNT)
149  av_frame_unref(frame);
150 
151  return decoded;
152 }
153 
154 static int open_codec_context(int *stream_idx,
155  AVFormatContext *fmt_ctx, enum AVMediaType type)
156 {
157  int ret, stream_index;
158  AVStream *st;
160  AVCodec *dec = NULL;
161  AVDictionary *opts = NULL;
162 
163  ret = av_find_best_stream(fmt_ctx, type, -1, -1, NULL, 0);
164  if (ret < 0) {
165  fprintf(stderr, "Could not find %s stream in input file '%s'\n",
167  return ret;
168  } else {
169  stream_index = ret;
170  st = fmt_ctx->streams[stream_index];
171 
172  /* find decoder for the stream */
173  dec_ctx = st->codec;
174  dec = avcodec_find_decoder(dec_ctx->codec_id);
175  if (!dec) {
176  fprintf(stderr, "Failed to find %s codec\n",
178  return AVERROR(EINVAL);
179  }
180 
181  /* Init the decoders, with or without reference counting */
183  av_dict_set(&opts, "refcounted_frames", "1", 0);
184  if ((ret = avcodec_open2(dec_ctx, dec, &opts)) < 0) {
185  fprintf(stderr, "Failed to open %s codec\n",
187  return ret;
188  }
189  *stream_idx = stream_index;
190  }
191 
192  return 0;
193 }
194 
195 static int get_format_from_sample_fmt(const char **fmt,
196  enum AVSampleFormat sample_fmt)
197 {
198  int i;
199  struct sample_fmt_entry {
200  enum AVSampleFormat sample_fmt; const char *fmt_be, *fmt_le;
201  } sample_fmt_entries[] = {
202  { AV_SAMPLE_FMT_U8, "u8", "u8" },
203  { AV_SAMPLE_FMT_S16, "s16be", "s16le" },
204  { AV_SAMPLE_FMT_S32, "s32be", "s32le" },
205  { AV_SAMPLE_FMT_FLT, "f32be", "f32le" },
206  { AV_SAMPLE_FMT_DBL, "f64be", "f64le" },
207  };
208  *fmt = NULL;
209 
210  for (i = 0; i < FF_ARRAY_ELEMS(sample_fmt_entries); i++) {
211  struct sample_fmt_entry *entry = &sample_fmt_entries[i];
212  if (sample_fmt == entry->sample_fmt) {
213  *fmt = AV_NE(entry->fmt_be, entry->fmt_le);
214  return 0;
215  }
216  }
217 
218  fprintf(stderr,
219  "sample format %s is not supported as output format\n",
220  av_get_sample_fmt_name(sample_fmt));
221  return -1;
222 }
223 
224 int main (int argc, char **argv)
225 {
226  int ret = 0, got_frame;
227 
228  if (argc != 4 && argc != 5) {
229  fprintf(stderr, "usage: %s [-refcount=<old|new_norefcount|new_refcount>] "
230  "input_file video_output_file audio_output_file\n"
231  "API example program to show how to read frames from an input file.\n"
232  "This program reads frames from a file, decodes them, and writes decoded\n"
233  "video frames to a rawvideo file named video_output_file, and decoded\n"
234  "audio frames to a rawaudio file named audio_output_file.\n\n"
235  "If the -refcount option is specified, the program use the\n"
236  "reference counting frame system which allows keeping a copy of\n"
237  "the data for longer than one decode call. If unset, it's using\n"
238  "the classic old method.\n"
239  "\n", argv[0]);
240  exit(1);
241  }
242  if (argc == 5) {
243  const char *mode = argv[1] + strlen("-refcount=");
244  if (!strcmp(mode, "old")) api_mode = API_MODE_OLD;
245  else if (!strcmp(mode, "new_norefcount")) api_mode = API_MODE_NEW_API_NO_REF_COUNT;
246  else if (!strcmp(mode, "new_refcount")) api_mode = API_MODE_NEW_API_REF_COUNT;
247  else {
248  fprintf(stderr, "unknow mode '%s'\n", mode);
249  exit(1);
250  }
251  argv++;
252  }
253  src_filename = argv[1];
254  video_dst_filename = argv[2];
255  audio_dst_filename = argv[3];
256 
257  /* register all formats and codecs */
258  av_register_all();
259 
260  /* open input file, and allocate format context */
261  if (avformat_open_input(&fmt_ctx, src_filename, NULL, NULL) < 0) {
262  fprintf(stderr, "Could not open source file %s\n", src_filename);
263  exit(1);
264  }
265 
266  /* retrieve stream information */
267  if (avformat_find_stream_info(fmt_ctx, NULL) < 0) {
268  fprintf(stderr, "Could not find stream information\n");
269  exit(1);
270  }
271 
273  video_stream = fmt_ctx->streams[video_stream_idx];
274  video_dec_ctx = video_stream->codec;
275 
276  video_dst_file = fopen(video_dst_filename, "wb");
277  if (!video_dst_file) {
278  fprintf(stderr, "Could not open destination file %s\n", video_dst_filename);
279  ret = 1;
280  goto end;
281  }
282 
283  /* allocate image where the decoded image will be put */
284  width = video_dec_ctx->width;
285  height = video_dec_ctx->height;
286  pix_fmt = video_dec_ctx->pix_fmt;
288  width, height, pix_fmt, 1);
289  if (ret < 0) {
290  fprintf(stderr, "Could not allocate raw video buffer\n");
291  goto end;
292  }
294  }
295 
299  audio_dst_file = fopen(audio_dst_filename, "wb");
300  if (!audio_dst_file) {
301  fprintf(stderr, "Could not open destination file %s\n", audio_dst_filename);
302  ret = 1;
303  goto end;
304  }
305  }
306 
307  /* dump input information to stderr */
308  av_dump_format(fmt_ctx, 0, src_filename, 0);
309 
310  if (!audio_stream && !video_stream) {
311  fprintf(stderr, "Could not find audio or video stream in the input, aborting\n");
312  ret = 1;
313  goto end;
314  }
315 
316  /* When using the new API, you need to use the libavutil/frame.h API, while
317  * the classic frame management is available in libavcodec */
318  if (api_mode == API_MODE_OLD)
319  frame = avcodec_alloc_frame();
320  else
321  frame = av_frame_alloc();
322  if (!frame) {
323  fprintf(stderr, "Could not allocate frame\n");
324  ret = AVERROR(ENOMEM);
325  goto end;
326  }
327 
328  /* initialize packet, set data to NULL, let the demuxer fill it */
329  av_init_packet(&pkt);
330  pkt.data = NULL;
331  pkt.size = 0;
332 
333  if (video_stream)
334  printf("Demuxing video from file '%s' into '%s'\n", src_filename, video_dst_filename);
335  if (audio_stream)
336  printf("Demuxing audio from file '%s' into '%s'\n", src_filename, audio_dst_filename);
337 
338  /* read frames from the file */
339  while (av_read_frame(fmt_ctx, &pkt) >= 0) {
340  AVPacket orig_pkt = pkt;
341  do {
342  ret = decode_packet(&got_frame, 0);
343  if (ret < 0)
344  break;
345  pkt.data += ret;
346  pkt.size -= ret;
347  } while (pkt.size > 0);
348  av_free_packet(&orig_pkt);
349  }
350 
351  /* flush cached frames */
352  pkt.data = NULL;
353  pkt.size = 0;
354  do {
355  decode_packet(&got_frame, 1);
356  } while (got_frame);
357 
358  printf("Demuxing succeeded.\n");
359 
360  if (video_stream) {
361  printf("Play the output video file with the command:\n"
362  "ffplay -f rawvideo -pix_fmt %s -video_size %dx%d %s\n",
365  }
366 
367  if (audio_stream) {
369  int n_channels = audio_dec_ctx->channels;
370  const char *fmt;
371 
372  if (av_sample_fmt_is_planar(sfmt)) {
373  const char *packed = av_get_sample_fmt_name(sfmt);
374  printf("Warning: the sample format the decoder produced is planar "
375  "(%s). This example will output the first channel only.\n",
376  packed ? packed : "?");
377  sfmt = av_get_packed_sample_fmt(sfmt);
378  n_channels = 1;
379  }
380 
381  if ((ret = get_format_from_sample_fmt(&fmt, sfmt)) < 0)
382  goto end;
383 
384  printf("Play the output audio file with the command:\n"
385  "ffplay -f %s -ac %d -ar %d %s\n",
386  fmt, n_channels, audio_dec_ctx->sample_rate,
388  }
389 
390 end:
391  avcodec_close(video_dec_ctx);
393  avformat_close_input(&fmt_ctx);
394  if (video_dst_file)
395  fclose(video_dst_file);
396  if (audio_dst_file)
397  fclose(audio_dst_file);
398  if (api_mode == API_MODE_OLD)
399  avcodec_free_frame(&frame);
400  else
401  av_frame_free(&frame);
403 
404  return ret < 0;
405 }