FFmpeg  4.3.9
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 
59 {
60  if (frame->width != width || frame->height != height ||
61  frame->format != pix_fmt) {
62  /* To handle this change, one could call av_image_alloc again and
63  * decode the following frames into another rawvideo file. */
64  fprintf(stderr, "Error: Width, height and pixel format have to be "
65  "constant in a rawvideo file, but the width, height or "
66  "pixel format of the input video changed:\n"
67  "old: width = %d, height = %d, format = %s\n"
68  "new: width = %d, height = %d, format = %s\n",
72  return -1;
73  }
74 
75  printf("video_frame n:%d coded_n:%d\n",
77 
78  /* copy decoded frame to destination buffer:
79  * this is required since rawvideo expects non aligned data */
81  (const uint8_t **)(frame->data), frame->linesize,
82  pix_fmt, width, height);
83 
84  /* write to rawvideo file */
86  return 0;
87 }
88 
90 {
91  size_t unpadded_linesize = frame->nb_samples * av_get_bytes_per_sample(frame->format);
92  printf("audio_frame n:%d nb_samples:%d pts:%s\n",
95 
96  /* Write the raw audio data samples of the first plane. This works
97  * fine for packed formats (e.g. AV_SAMPLE_FMT_S16). However,
98  * most audio decoders output planar audio, which uses a separate
99  * plane of audio samples for each channel (e.g. AV_SAMPLE_FMT_S16P).
100  * In other words, this code will write only the first audio channel
101  * in these cases.
102  * You should use libswresample or libavfilter to convert the frame
103  * to packed data. */
104  fwrite(frame->extended_data[0], 1, unpadded_linesize, audio_dst_file);
105 
106  return 0;
107 }
108 
109 static int decode_packet(AVCodecContext *dec, const AVPacket *pkt)
110 {
111  int ret = 0;
112 
113  // submit the packet to the decoder
114  ret = avcodec_send_packet(dec, pkt);
115  if (ret < 0) {
116  fprintf(stderr, "Error submitting a packet for decoding (%s)\n", av_err2str(ret));
117  return ret;
118  }
119 
120  // get all the available frames from the decoder
121  while (ret >= 0) {
122  ret = avcodec_receive_frame(dec, frame);
123  if (ret < 0) {
124  // those two return values are special and mean there is no output
125  // frame available, but there were no errors during decoding
126  if (ret == AVERROR_EOF || ret == AVERROR(EAGAIN))
127  return 0;
128 
129  fprintf(stderr, "Error during decoding (%s)\n", av_err2str(ret));
130  return ret;
131  }
132 
133  // write the frame data to output file
134  if (dec->codec->type == AVMEDIA_TYPE_VIDEO)
135  ret = output_video_frame(frame);
136  else
137  ret = output_audio_frame(frame);
138 
140  }
141 
142  return ret;
143 }
144 
145 static int open_codec_context(int *stream_idx,
147 {
148  int ret, stream_index;
149  AVStream *st;
150  AVCodec *dec = NULL;
151  AVDictionary *opts = NULL;
152 
153  ret = av_find_best_stream(fmt_ctx, type, -1, -1, NULL, 0);
154  if (ret < 0) {
155  fprintf(stderr, "Could not find %s stream in input file '%s'\n",
157  return ret;
158  } else {
159  stream_index = ret;
160  st = fmt_ctx->streams[stream_index];
161 
162  /* find decoder for the stream */
164  if (!dec) {
165  fprintf(stderr, "Failed to find %s codec\n",
167  return AVERROR(EINVAL);
168  }
169 
170  /* Allocate a codec context for the decoder */
172  if (!*dec_ctx) {
173  fprintf(stderr, "Failed to allocate the %s codec context\n",
175  return AVERROR(ENOMEM);
176  }
177 
178  /* Copy codec parameters from input stream to output codec context */
179  if ((ret = avcodec_parameters_to_context(*dec_ctx, st->codecpar)) < 0) {
180  fprintf(stderr, "Failed to copy %s codec parameters to decoder context\n",
182  return ret;
183  }
184 
185  /* Init the decoders */
186  if ((ret = avcodec_open2(*dec_ctx, dec, &opts)) < 0) {
187  fprintf(stderr, "Failed to open %s codec\n",
189  return ret;
190  }
191  *stream_idx = stream_index;
192  }
193 
194  return 0;
195 }
196 
197 static int get_format_from_sample_fmt(const char **fmt,
198  enum AVSampleFormat sample_fmt)
199 {
200  int i;
201  struct sample_fmt_entry {
202  enum AVSampleFormat sample_fmt; const char *fmt_be, *fmt_le;
203  } sample_fmt_entries[] = {
204  { AV_SAMPLE_FMT_U8, "u8", "u8" },
205  { AV_SAMPLE_FMT_S16, "s16be", "s16le" },
206  { AV_SAMPLE_FMT_S32, "s32be", "s32le" },
207  { AV_SAMPLE_FMT_FLT, "f32be", "f32le" },
208  { AV_SAMPLE_FMT_DBL, "f64be", "f64le" },
209  };
210  *fmt = NULL;
211 
212  for (i = 0; i < FF_ARRAY_ELEMS(sample_fmt_entries); i++) {
213  struct sample_fmt_entry *entry = &sample_fmt_entries[i];
214  if (sample_fmt == entry->sample_fmt) {
215  *fmt = AV_NE(entry->fmt_be, entry->fmt_le);
216  return 0;
217  }
218  }
219 
220  fprintf(stderr,
221  "sample format %s is not supported as output format\n",
222  av_get_sample_fmt_name(sample_fmt));
223  return -1;
224 }
225 
226 int main (int argc, char **argv)
227 {
228  int ret = 0;
229 
230  if (argc != 4) {
231  fprintf(stderr, "usage: %s input_file video_output_file audio_output_file\n"
232  "API example program to show how to read frames from an input file.\n"
233  "This program reads frames from a file, decodes them, and writes decoded\n"
234  "video frames to a rawvideo file named video_output_file, and decoded\n"
235  "audio frames to a rawaudio file named audio_output_file.\n",
236  argv[0]);
237  exit(1);
238  }
239  src_filename = argv[1];
240  video_dst_filename = argv[2];
241  audio_dst_filename = argv[3];
242 
243  /* open input file, and allocate format context */
244  if (avformat_open_input(&fmt_ctx, src_filename, NULL, NULL) < 0) {
245  fprintf(stderr, "Could not open source file %s\n", src_filename);
246  exit(1);
247  }
248 
249  /* retrieve stream information */
250  if (avformat_find_stream_info(fmt_ctx, NULL) < 0) {
251  fprintf(stderr, "Could not find stream information\n");
252  exit(1);
253  }
254 
257 
258  video_dst_file = fopen(video_dst_filename, "wb");
259  if (!video_dst_file) {
260  fprintf(stderr, "Could not open destination file %s\n", video_dst_filename);
261  ret = 1;
262  goto end;
263  }
264 
265  /* allocate image where the decoded image will be put */
270  width, height, pix_fmt, 1);
271  if (ret < 0) {
272  fprintf(stderr, "Could not allocate raw video buffer\n");
273  goto end;
274  }
275  video_dst_bufsize = ret;
276  }
277 
280  audio_dst_file = fopen(audio_dst_filename, "wb");
281  if (!audio_dst_file) {
282  fprintf(stderr, "Could not open destination file %s\n", audio_dst_filename);
283  ret = 1;
284  goto end;
285  }
286  }
287 
288  /* dump input information to stderr */
290 
291  if (!audio_stream && !video_stream) {
292  fprintf(stderr, "Could not find audio or video stream in the input, aborting\n");
293  ret = 1;
294  goto end;
295  }
296 
297  frame = av_frame_alloc();
298  if (!frame) {
299  fprintf(stderr, "Could not allocate frame\n");
300  ret = AVERROR(ENOMEM);
301  goto end;
302  }
303 
304  /* initialize packet, set data to NULL, let the demuxer fill it */
306  pkt.data = NULL;
307  pkt.size = 0;
308 
309  if (video_stream)
310  printf("Demuxing video from file '%s' into '%s'\n", src_filename, video_dst_filename);
311  if (audio_stream)
312  printf("Demuxing audio from file '%s' into '%s'\n", src_filename, audio_dst_filename);
313 
314  /* read frames from the file */
315  while (av_read_frame(fmt_ctx, &pkt) >= 0) {
316  // check if the packet belongs to a stream we are interested in, otherwise
317  // skip it
320  else if (pkt.stream_index == audio_stream_idx)
323  if (ret < 0)
324  break;
325  }
326 
327  /* flush the decoders */
328  if (video_dec_ctx)
330  if (audio_dec_ctx)
332 
333  printf("Demuxing succeeded.\n");
334 
335  if (video_stream) {
336  printf("Play the output video file with the command:\n"
337  "ffplay -f rawvideo -pix_fmt %s -video_size %dx%d %s\n",
340  }
341 
342  if (audio_stream) {
344  int n_channels = audio_dec_ctx->channels;
345  const char *fmt;
346 
347  if (av_sample_fmt_is_planar(sfmt)) {
348  const char *packed = av_get_sample_fmt_name(sfmt);
349  printf("Warning: the sample format the decoder produced is planar "
350  "(%s). This example will output the first channel only.\n",
351  packed ? packed : "?");
352  sfmt = av_get_packed_sample_fmt(sfmt);
353  n_channels = 1;
354  }
355 
356  if ((ret = get_format_from_sample_fmt(&fmt, sfmt)) < 0)
357  goto end;
358 
359  printf("Play the output audio file with the command:\n"
360  "ffplay -f %s -ac %d -ar %d %s\n",
361  fmt, n_channels, audio_dec_ctx->sample_rate,
363  }
364 
365 end:
369  if (video_dst_file)
370  fclose(video_dst_file);
371  if (audio_dst_file)
372  fclose(audio_dst_file);
375 
376  return ret < 0;
377 }
Main libavformat public API header.
#define AV_NE(be, le)
Definition: common.h:50
#define FF_ARRAY_ELEMS(a)
Definition: common.h:100
static uint8_t * video_dst_data[4]
static int audio_stream_idx
static int width
static int output_audio_frame(AVFrame *frame)
static AVCodecContext * video_dec_ctx
int main(int argc, char **argv)
static AVPacket pkt
static AVFormatContext * fmt_ctx
static int audio_frame_count
static int get_format_from_sample_fmt(const char **fmt, enum AVSampleFormat sample_fmt)
static int output_video_frame(AVFrame *frame)
static enum AVPixelFormat pix_fmt
static const char * video_dst_filename
static int video_dst_bufsize
static int video_dst_linesize[4]
static AVStream * video_stream
static int decode_packet(AVCodecContext *dec, const AVPacket *pkt)
static const char * audio_dst_filename
static AVStream * audio_stream
static int open_codec_context(int *stream_idx, AVCodecContext **dec_ctx, AVFormatContext *fmt_ctx, enum AVMediaType type)
static FILE * video_dst_file
static int video_frame_count
static int height
static const char * src_filename
static AVFrame * frame
static AVCodecContext * audio_dec_ctx
static FILE * audio_dst_file
static int video_stream_idx
static AVCodecContext * dec_ctx
int avcodec_open2(AVCodecContext *avctx, const AVCodec *codec, AVDictionary **options)
Initialize the AVCodecContext to use the given AVCodec.
AVCodec * avcodec_find_decoder(enum AVCodecID id)
Find a registered decoder with a matching codec ID.
int avcodec_parameters_to_context(AVCodecContext *codec, const AVCodecParameters *par)
Fill the codec context based on the values from the supplied codec parameters.
AVCodecContext * avcodec_alloc_context3(const AVCodec *codec)
Allocate an AVCodecContext and set its fields to default values.
void avcodec_free_context(AVCodecContext **avctx)
Free the codec context and everything associated with it and write NULL to the provided pointer.
int avcodec_receive_frame(AVCodecContext *avctx, AVFrame *frame)
Return decoded output data from a decoder.
int avcodec_send_packet(AVCodecContext *avctx, const AVPacket *avpkt)
Supply raw packet data as input to a decoder.
void av_packet_unref(AVPacket *pkt)
Wipe the packet.
void av_init_packet(AVPacket *pkt)
Initialize optional fields of a packet with default values.
int av_read_frame(AVFormatContext *s, AVPacket *pkt)
Return the next frame of a stream.
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 avformat_find_stream_info(AVFormatContext *ic, AVDictionary **options)
Read packets of a media file to get stream information.
void avformat_close_input(AVFormatContext **s)
Close an opened input AVFormatContext.
int avformat_open_input(AVFormatContext **ps, const char *url, ff_const59 AVInputFormat *fmt, AVDictionary **options)
Open an input stream and read the header.
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,...
struct AVDictionary AVDictionary
Definition: dict.h:86
#define AVERROR_EOF
End of file.
Definition: error.h:55
#define av_err2str(errnum)
Convenience macro, the return value should be used only directly in function arguments but never stan...
Definition: error.h:119
#define AVERROR(e)
Definition: error.h:43
void av_frame_unref(AVFrame *frame)
Unreference all the buffers referenced by frame and reset the frame fields.
void av_frame_free(AVFrame **frame)
Free the frame and any dynamically allocated objects in it, e.g.
AVFrame * av_frame_alloc(void)
Allocate an AVFrame and set its fields to default values.
void av_free(void *ptr)
Free a memory block which has been allocated with a function of av_malloc() or av_realloc() family.
AVMediaType
Definition: avutil.h:199
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.
@ AVMEDIA_TYPE_AUDIO
Definition: avutil.h:202
@ AVMEDIA_TYPE_VIDEO
Definition: avutil.h:201
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.
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...
int av_sample_fmt_is_planar(enum AVSampleFormat sample_fmt)
Check if the sample format is planar.
int av_get_bytes_per_sample(enum AVSampleFormat sample_fmt)
Return number of bytes per sample.
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.
enum AVSampleFormat av_get_packed_sample_fmt(enum AVSampleFormat sample_fmt)
Get the packed alternative form of the given sample format.
AVSampleFormat
Audio sample formats.
Definition: samplefmt.h:58
@ AV_SAMPLE_FMT_FLT
float
Definition: samplefmt.h:63
@ AV_SAMPLE_FMT_S32
signed 32 bits
Definition: samplefmt.h:62
@ AV_SAMPLE_FMT_U8
unsigned 8 bits
Definition: samplefmt.h:60
@ AV_SAMPLE_FMT_DBL
double
Definition: samplefmt.h:64
@ AV_SAMPLE_FMT_S16
signed 16 bits
Definition: samplefmt.h:61
misc image utilities
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.
AVPixelFormat
Pixel format.
Definition: pixfmt.h:64
main external API structure.
Definition: avcodec.h:526
enum AVPixelFormat pix_fmt
Pixel format, see AV_PIX_FMT_xxx.
Definition: avcodec.h:736
int width
picture width / height.
Definition: avcodec.h:699
enum AVSampleFormat sample_fmt
audio sample format
Definition: avcodec.h:1194
const struct AVCodec * codec
Definition: avcodec.h:535
int sample_rate
samples per second
Definition: avcodec.h:1186
AVRational time_base
This is the fundamental unit of time (in seconds) in terms of which frame timestamps are represented.
Definition: avcodec.h:649
int channels
number of audio channels
Definition: avcodec.h:1187
enum AVCodecID codec_id
Specific type of the encoded data (the codec used).
Definition: codec_par.h:60
AVCodec.
Definition: codec.h:190
enum AVMediaType type
Definition: codec.h:203
Format I/O context.
Definition: avformat.h:1335
AVStream ** streams
A list of all streams in the file.
Definition: avformat.h:1403
This structure describes decoded (raw) audio or video data.
Definition: frame.h:300
int coded_picture_number
picture number in bitstream order
Definition: frame.h:414
int nb_samples
number of audio samples (per channel) described by this frame
Definition: frame.h:366
int64_t pts
Presentation timestamp in time_base units (time when frame should be shown to user).
Definition: frame.h:393
uint8_t * data[AV_NUM_DATA_POINTERS]
pointer to the picture/channel planes.
Definition: frame.h:314
int width
Definition: frame.h:358
int height
Definition: frame.h:358
int linesize[AV_NUM_DATA_POINTERS]
For video, size in bytes of each picture line.
Definition: frame.h:331
int format
format of the frame, -1 if unknown or unset Values correspond to enum AVPixelFormat for video frames,...
Definition: frame.h:373
uint8_t ** extended_data
pointers to the data planes/channels.
Definition: frame.h:347
This structure stores compressed data.
Definition: packet.h:332
int stream_index
Definition: packet.h:357
int size
Definition: packet.h:356
uint8_t * data
Definition: packet.h:355
Stream structure.
Definition: avformat.h:865
AVCodecParameters * codecpar
Codec parameters associated with this stream.
Definition: avformat.h:1012
timestamp utils, mostly useful for debugging/logging purposes
#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