FFmpeg  3.1.3
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 
37 static AVFormatContext *fmt_ctx = NULL;
39 static int width, height;
40 static enum AVPixelFormat pix_fmt;
41 static AVStream *video_stream = NULL, *audio_stream = NULL;
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 /* Enable or disable frame reference counting. You are not supposed to support
59  * both paths in your application but pick the one most appropriate to your
60  * needs. Look for the use of refcount in this example to see what are the
61  * differences of API usage between them. */
62 static int refcount = 0;
63 
64 static int decode_packet(int *got_frame, int cached)
65 {
66  int ret = 0;
67  int decoded = pkt.size;
68 
69  *got_frame = 0;
70 
71  if (pkt.stream_index == video_stream_idx) {
72  /* decode video frame */
73  ret = avcodec_decode_video2(video_dec_ctx, frame, got_frame, &pkt);
74  if (ret < 0) {
75  fprintf(stderr, "Error decoding video frame (%s)\n", av_err2str(ret));
76  return ret;
77  }
78 
79  if (*got_frame) {
80 
81  if (frame->width != width || frame->height != height ||
82  frame->format != pix_fmt) {
83  /* To handle this change, one could call av_image_alloc again and
84  * decode the following frames into another rawvideo file. */
85  fprintf(stderr, "Error: Width, height and pixel format have to be "
86  "constant in a rawvideo file, but the width, height or "
87  "pixel format of the input video changed:\n"
88  "old: width = %d, height = %d, format = %s\n"
89  "new: width = %d, height = %d, format = %s\n",
91  frame->width, frame->height,
92  av_get_pix_fmt_name(frame->format));
93  return -1;
94  }
95 
96  printf("video_frame%s n:%d coded_n:%d pts:%s\n",
97  cached ? "(cached)" : "",
99  av_ts2timestr(frame->pts, &video_dec_ctx->time_base));
100 
101  /* copy decoded frame to destination buffer:
102  * this is required since rawvideo expects non aligned data */
104  (const uint8_t **)(frame->data), frame->linesize,
105  pix_fmt, width, height);
106 
107  /* write to rawvideo file */
109  }
110  } else if (pkt.stream_index == audio_stream_idx) {
111  /* decode audio frame */
112  ret = avcodec_decode_audio4(audio_dec_ctx, frame, got_frame, &pkt);
113  if (ret < 0) {
114  fprintf(stderr, "Error decoding audio frame (%s)\n", av_err2str(ret));
115  return ret;
116  }
117  /* Some audio decoders decode only part of the packet, and have to be
118  * called again with the remainder of the packet data.
119  * Sample: fate-suite/lossless-audio/luckynight-partial.shn
120  * Also, some decoders might over-read the packet. */
121  decoded = FFMIN(ret, pkt.size);
122 
123  if (*got_frame) {
124  size_t unpadded_linesize = frame->nb_samples * av_get_bytes_per_sample(frame->format);
125  printf("audio_frame%s n:%d nb_samples:%d pts:%s\n",
126  cached ? "(cached)" : "",
127  audio_frame_count++, frame->nb_samples,
129 
130  /* Write the raw audio data samples of the first plane. This works
131  * fine for packed formats (e.g. AV_SAMPLE_FMT_S16). However,
132  * most audio decoders output planar audio, which uses a separate
133  * plane of audio samples for each channel (e.g. AV_SAMPLE_FMT_S16P).
134  * In other words, this code will write only the first audio channel
135  * in these cases.
136  * You should use libswresample or libavfilter to convert the frame
137  * to packed data. */
138  fwrite(frame->extended_data[0], 1, unpadded_linesize, audio_dst_file);
139  }
140  }
141 
142  /* If we use frame reference counting, we own the data and need
143  * to de-reference it when we don't use it anymore */
144  if (*got_frame && refcount)
145  av_frame_unref(frame);
146 
147  return decoded;
148 }
149 
150 static int open_codec_context(int *stream_idx,
151  AVCodecContext **dec_ctx, AVFormatContext *fmt_ctx, enum AVMediaType type)
152 {
153  int ret, stream_index;
154  AVStream *st;
155  AVCodec *dec = NULL;
156  AVDictionary *opts = NULL;
157 
158  ret = av_find_best_stream(fmt_ctx, type, -1, -1, NULL, 0);
159  if (ret < 0) {
160  fprintf(stderr, "Could not find %s stream in input file '%s'\n",
162  return ret;
163  } else {
164  stream_index = ret;
165  st = fmt_ctx->streams[stream_index];
166 
167  /* find decoder for the stream */
169  if (!dec) {
170  fprintf(stderr, "Failed to find %s codec\n",
172  return AVERROR(EINVAL);
173  }
174 
175  /* Allocate a codec context for the decoder */
176  *dec_ctx = avcodec_alloc_context3(dec);
177  if (!*dec_ctx) {
178  fprintf(stderr, "Failed to allocate the %s codec context\n",
180  return AVERROR(ENOMEM);
181  }
182 
183  /* Copy codec parameters from input stream to output codec context */
184  if ((ret = avcodec_parameters_to_context(*dec_ctx, st->codecpar)) < 0) {
185  fprintf(stderr, "Failed to copy %s codec parameters to decoder context\n",
187  return ret;
188  }
189 
190  /* Init the decoders, with or without reference counting */
191  av_dict_set(&opts, "refcounted_frames", refcount ? "1" : "0", 0);
192  if ((ret = avcodec_open2(*dec_ctx, dec, &opts)) < 0) {
193  fprintf(stderr, "Failed to open %s codec\n",
195  return ret;
196  }
197  *stream_idx = stream_index;
198  }
199 
200  return 0;
201 }
202 
203 static int get_format_from_sample_fmt(const char **fmt,
204  enum AVSampleFormat sample_fmt)
205 {
206  int i;
207  struct sample_fmt_entry {
208  enum AVSampleFormat sample_fmt; const char *fmt_be, *fmt_le;
209  } sample_fmt_entries[] = {
210  { AV_SAMPLE_FMT_U8, "u8", "u8" },
211  { AV_SAMPLE_FMT_S16, "s16be", "s16le" },
212  { AV_SAMPLE_FMT_S32, "s32be", "s32le" },
213  { AV_SAMPLE_FMT_FLT, "f32be", "f32le" },
214  { AV_SAMPLE_FMT_DBL, "f64be", "f64le" },
215  };
216  *fmt = NULL;
217 
218  for (i = 0; i < FF_ARRAY_ELEMS(sample_fmt_entries); i++) {
219  struct sample_fmt_entry *entry = &sample_fmt_entries[i];
220  if (sample_fmt == entry->sample_fmt) {
221  *fmt = AV_NE(entry->fmt_be, entry->fmt_le);
222  return 0;
223  }
224  }
225 
226  fprintf(stderr,
227  "sample format %s is not supported as output format\n",
228  av_get_sample_fmt_name(sample_fmt));
229  return -1;
230 }
231 
232 int main (int argc, char **argv)
233 {
234  int ret = 0, got_frame;
235 
236  if (argc != 4 && argc != 5) {
237  fprintf(stderr, "usage: %s [-refcount] input_file video_output_file audio_output_file\n"
238  "API example program to show how to read frames from an input file.\n"
239  "This program reads frames from a file, decodes them, and writes decoded\n"
240  "video frames to a rawvideo file named video_output_file, and decoded\n"
241  "audio frames to a rawaudio file named audio_output_file.\n\n"
242  "If the -refcount option is specified, the program use the\n"
243  "reference counting frame system which allows keeping a copy of\n"
244  "the data for longer than one decode call.\n"
245  "\n", argv[0]);
246  exit(1);
247  }
248  if (argc == 5 && !strcmp(argv[1], "-refcount")) {
249  refcount = 1;
250  argv++;
251  }
252  src_filename = argv[1];
253  video_dst_filename = argv[2];
254  audio_dst_filename = argv[3];
255 
256  /* register all formats and codecs */
257  av_register_all();
258 
259  /* open input file, and allocate format context */
260  if (avformat_open_input(&fmt_ctx, src_filename, NULL, NULL) < 0) {
261  fprintf(stderr, "Could not open source file %s\n", src_filename);
262  exit(1);
263  }
264 
265  /* retrieve stream information */
266  if (avformat_find_stream_info(fmt_ctx, NULL) < 0) {
267  fprintf(stderr, "Could not find stream information\n");
268  exit(1);
269  }
270 
271  if (open_codec_context(&video_stream_idx, &video_dec_ctx, fmt_ctx, AVMEDIA_TYPE_VIDEO) >= 0) {
272  video_stream = fmt_ctx->streams[video_stream_idx];
273 
274  video_dst_file = fopen(video_dst_filename, "wb");
275  if (!video_dst_file) {
276  fprintf(stderr, "Could not open destination file %s\n", video_dst_filename);
277  ret = 1;
278  goto end;
279  }
280 
281  /* allocate image where the decoded image will be put */
282  width = video_dec_ctx->width;
283  height = video_dec_ctx->height;
284  pix_fmt = video_dec_ctx->pix_fmt;
286  width, height, pix_fmt, 1);
287  if (ret < 0) {
288  fprintf(stderr, "Could not allocate raw video buffer\n");
289  goto end;
290  }
291  video_dst_bufsize = ret;
292  }
293 
296  audio_dst_file = fopen(audio_dst_filename, "wb");
297  if (!audio_dst_file) {
298  fprintf(stderr, "Could not open destination file %s\n", audio_dst_filename);
299  ret = 1;
300  goto end;
301  }
302  }
303 
304  /* dump input information to stderr */
305  av_dump_format(fmt_ctx, 0, src_filename, 0);
306 
307  if (!audio_stream && !video_stream) {
308  fprintf(stderr, "Could not find audio or video stream in the input, aborting\n");
309  ret = 1;
310  goto end;
311  }
312 
313  frame = av_frame_alloc();
314  if (!frame) {
315  fprintf(stderr, "Could not allocate frame\n");
316  ret = AVERROR(ENOMEM);
317  goto end;
318  }
319 
320  /* initialize packet, set data to NULL, let the demuxer fill it */
321  av_init_packet(&pkt);
322  pkt.data = NULL;
323  pkt.size = 0;
324 
325  if (video_stream)
326  printf("Demuxing video from file '%s' into '%s'\n", src_filename, video_dst_filename);
327  if (audio_stream)
328  printf("Demuxing audio from file '%s' into '%s'\n", src_filename, audio_dst_filename);
329 
330  /* read frames from the file */
331  while (av_read_frame(fmt_ctx, &pkt) >= 0) {
332  AVPacket orig_pkt = pkt;
333  do {
334  ret = decode_packet(&got_frame, 0);
335  if (ret < 0)
336  break;
337  pkt.data += ret;
338  pkt.size -= ret;
339  } while (pkt.size > 0);
340  av_packet_unref(&orig_pkt);
341  }
342 
343  /* flush cached frames */
344  pkt.data = NULL;
345  pkt.size = 0;
346  do {
347  decode_packet(&got_frame, 1);
348  } while (got_frame);
349 
350  printf("Demuxing succeeded.\n");
351 
352  if (video_stream) {
353  printf("Play the output video file with the command:\n"
354  "ffplay -f rawvideo -pix_fmt %s -video_size %dx%d %s\n",
357  }
358 
359  if (audio_stream) {
361  int n_channels = audio_dec_ctx->channels;
362  const char *fmt;
363 
364  if (av_sample_fmt_is_planar(sfmt)) {
365  const char *packed = av_get_sample_fmt_name(sfmt);
366  printf("Warning: the sample format the decoder produced is planar "
367  "(%s). This example will output the first channel only.\n",
368  packed ? packed : "?");
369  sfmt = av_get_packed_sample_fmt(sfmt);
370  n_channels = 1;
371  }
372 
373  if ((ret = get_format_from_sample_fmt(&fmt, sfmt)) < 0)
374  goto end;
375 
376  printf("Play the output audio file with the command:\n"
377  "ffplay -f %s -ac %d -ar %d %s\n",
378  fmt, n_channels, audio_dec_ctx->sample_rate,
380  }
381 
382 end:
383  avcodec_free_context(&video_dec_ctx);
385  avformat_close_input(&fmt_ctx);
386  if (video_dst_file)
387  fclose(video_dst_file);
388  if (audio_dst_file)
389  fclose(audio_dst_file);
390  av_frame_free(&frame);
392 
393  return ret < 0;
394 }
static int decode_packet(int *got_frame, int cached)
static enum AVPixelFormat pix_fmt
This structure describes decoded (raw) audio or video data.
Definition: frame.h:184
misc image utilities
int av_image_alloc(uint8_t *pointers[4], int linesizes[4], int w, int h, enum AVPixelFormat pix_fmt, int align)
Allocate an image with size w and h and pixel format pix_fmt, and fill pointers and linesizes accordi...
static AVFormatContext * fmt_ctx
enum AVCodecID codec_id
Specific type of the encoded data (the codec used).
Definition: avcodec.h:3922
int size
Definition: avcodec.h:1581
static int audio_frame_count
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
Definition: avcodec.h:1877
static int open_codec_context(int *stream_idx, AVCodecContext **dec_ctx, AVFormatContext *fmt_ctx, enum AVMediaType type)
static AVStream * video_stream
attribute_deprecated int avcodec_decode_audio4(AVCodecContext *avctx, AVFrame *frame, int *got_frame_ptr, const AVPacket *avpkt)
Decode the audio frame of size avpkt->size from avpkt->data into frame.
static AVPacket pkt
static int audio_stream_idx
AVCodec.
Definition: avcodec.h:3542
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented...
Definition: avcodec.h:1786
Format I/O context.
Definition: avformat.h:1325
static FILE * audio_dst_file
enum AVSampleFormat sample_fmt
audio sample format
Definition: avcodec.h:2418
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
static int video_stream_idx
timestamp utils, mostly useful for debugging/logging purposes
#define AV_NE(be, le)
Definition: common.h:50
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:268
AVStream ** streams
A list of all streams in the file.
Definition: avformat.h:1393
int avcodec_parameters_to_context(AVCodecContext *codec, const AVCodecParameters *par)
Fill the codec context based on the values from the supplied codec parameters.
static AVFrame * frame
uint8_t * data
Definition: avcodec.h:1580
static uint8_t * video_dst_data[4]
signed 32 bits
Definition: samplefmt.h:62
int av_sample_fmt_is_planar(enum AVSampleFormat sample_fmt)
Check if the sample format is planar.
void av_dump_format(AVFormatContext *ic, int index, const char *url, int is_output)
Print detailed information about the input or output format, such as duration, bitrate, streams, container, programs, metadata, side data, codec and time base.
int av_find_best_stream(AVFormatContext *ic, enum AVMediaType type, int wanted_stream_nb, int related_stream, AVCodec **decoder_ret, int flags)
Find the "best" stream in the file.
int width
width and height of the video frame
Definition: frame.h:236
void av_free(void *ptr)
Free a memory block which has been allocated with av_malloc(z)() or av_realloc(). ...
static int video_dst_linesize[4]
static const char * audio_dst_filename
#define AVERROR(e)
Definition: error.h:43
static AVStream * audio_stream
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
#define av_ts2timestr(ts, tb)
Convenience macro, the return value should be used only directly in function arguments but never stan...
Definition: timestamp.h:76
static AVCodecContext * video_dec_ctx
struct AVDictionary AVDictionary
Definition: dict.h:90
const char * av_get_sample_fmt_name(enum AVSampleFormat sample_fmt)
Return the name of sample_fmt, or NULL if sample_fmt is not recognized.
void av_image_copy(uint8_t *dst_data[4], int dst_linesizes[4], const uint8_t *src_data[4], const int src_linesizes[4], enum AVPixelFormat pix_fmt, int width, int height)
Copy image in src_data to dst_data.
unsigned 8 bits
Definition: samplefmt.h:60
#define FFMIN(a, b)
Definition: common.h:96
AVCodecContext * avcodec_alloc_context3(const AVCodec *codec)
Allocate an AVCodecContext and set its fields to default values.
int width
picture width / height.
Definition: avcodec.h:1836
#define av_err2str(errnum)
Convenience macro, the return value should be used only directly in function arguments but never stan...
Definition: error.h:119
static int refcount
static const char * video_dst_filename
Stream structure.
Definition: avformat.h:876
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames...
Definition: frame.h:248
#define FF_ARRAY_ELEMS(a)
Definition: common.h:100
int coded_picture_number
picture number in bitstream order
Definition: frame.h:285
attribute_deprecated int avcodec_decode_video2(AVCodecContext *avctx, AVFrame *picture, int *got_picture_ptr, const AVPacket *avpkt)
Decode the video frame of size avpkt->size from avpkt->data into picture.
AVSampleFormat
Audio sample formats.
Definition: samplefmt.h:58
void avcodec_free_context(AVCodecContext **avctx)
Free the codec context and everything associated with it and write NULL to the provided pointer...
int sample_rate
samples per second
Definition: avcodec.h:2410
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:215
int main(int argc, char **argv)
main external API structure.
Definition: avcodec.h:1649
AVCodec * avcodec_find_decoder(enum AVCodecID id)
Find a registered decoder with a matching codec ID.
void av_packet_unref(AVPacket *pkt)
Wipe the packet.
int av_dict_set(AVDictionary **pm, const char *key, const char *value, int flags)
Set the given entry in *pm, overwriting an existing entry.
static int get_format_from_sample_fmt(const char **fmt, enum AVSampleFormat sample_fmt)
AVMediaType
Definition: avutil.h:191
static const char * src_filename
int avcodec_open2(AVCodecContext *avctx, const AVCodec *codec, AVDictionary **options)
Initialize the AVCodecContext to use the given AVCodec.
int av_read_frame(AVFormatContext *s, AVPacket *pkt)
Return the next frame of a stream.
static AVCodecContext * audio_dec_ctx
static int video_frame_count
void av_frame_unref(AVFrame *frame)
Unreference all the buffers referenced by frame and reset the frame fields.
const char * av_get_media_type_string(enum AVMediaType media_type)
Return a string describing the media_type enum, NULL if media_type is unknown.
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:198
int av_get_bytes_per_sample(enum AVSampleFormat sample_fmt)
Return number of bytes per sample.
Main libavformat public API header.
enum AVSampleFormat av_get_packed_sample_fmt(enum AVSampleFormat sample_fmt)
Get the packed alternative form of the given sample format.
int avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options)
Read packets of a media file to get stream information.
signed 16 bits
Definition: samplefmt.h:61
static AVCodecContext * dec_ctx
static FILE * video_dst_file
void av_init_packet(AVPacket *pkt)
Initialize optional fields of a packet with default values.
void avformat_close_input(AVFormatContext **s)
Close an opened input AVFormatContext.
int channels
number of audio channels
Definition: avcodec.h:2411
const char * av_get_pix_fmt_name(enum AVPixelFormat pix_fmt)
Return the short name for a pixel format, NULL in case pix_fmt is unknown.
int avformat_open_input(AVFormatContext **ps, const char *url, AVInputFormat *fmt, AVDictionary **options)
Open an input stream and read the header.
int height
Definition: frame.h:236
static int video_dst_bufsize
AVCodecParameters * codecpar
Definition: avformat.h:1228
static int height
int stream_index
Definition: avcodec.h:1582
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:231
AVPixelFormat
Pixel format.
Definition: pixfmt.h:60
This structure stores compressed data.
Definition: avcodec.h:1557
void av_register_all(void)
Initialize libavformat and register all the muxers, demuxers and protocols.
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:241
static int width