2 Copyright (C) 2013 Paul Brossier <piem@aubio.org>
4 This file is part of aubio.
6 aubio is free software: you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation, either version 3 of the License, or
9 (at your option) any later version.
11 aubio is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with aubio. If not, see <http://www.gnu.org/licenses/>.
21 #include "aubio_priv.h"
25 #include <libavcodec/avcodec.h>
26 #include <libavformat/avformat.h>
27 #if defined(HAVE_SWRESAMPLE)
28 #include <libswresample/swresample.h>
29 #elif defined(HAVE_AVRESAMPLE)
30 #include <libavresample/avresample.h>
32 #include <libavutil/opt.h>
35 // determine whether we use libavformat from ffmpeg or from libav
36 #define FFMPEG_LIBAVFORMAT (LIBAVFORMAT_VERSION_MICRO > 99 )
37 // max_analyze_duration2 was used from ffmpeg libavformat 55.43.100 through 57.2.100
38 #define FFMPEG_LIBAVFORMAT_MAX_DUR2 FFMPEG_LIBAVFORMAT && ( \
39 (LIBAVFORMAT_VERSION_MAJOR == 55 && LIBAVFORMAT_VERSION_MINOR >= 43) \
40 || (LIBAVFORMAT_VERSION_MAJOR == 56) \
41 || (LIBAVFORMAT_VERSION_MAJOR == 57 && LIBAVFORMAT_VERSION_MINOR < 2) \
44 // backward compatibility with libavcodec55
45 #if LIBAVCODEC_VERSION_INT < AV_VERSION_INT(57,0,0)
46 #define HAVE_AUBIO_LIBAVCODEC_DEPRECATED 1
49 #if LIBAVCODEC_VERSION_INT < AV_VERSION_INT(55,28,1)
50 #warning "libavcodec < 56 is deprecated"
51 #define av_frame_alloc avcodec_alloc_frame
52 #define av_frame_free avcodec_free_frame
53 #define av_packet_unref av_free_packet
56 #include "aubio_priv.h"
59 #include "source_avcodec.h"
61 #if LIBAVCODEC_VERSION_INT < AV_VERSION_INT(56, 56, 0)
62 #define AUBIO_AVCODEC_MAX_BUFFER_SIZE FF_MIN_BUFFER_SIZE
64 #define AUBIO_AVCODEC_MAX_BUFFER_SIZE AV_INPUT_BUFFER_MIN_SIZE
67 struct _aubio_source_avcodec_t {
72 // some data about the file
74 uint_t input_samplerate;
75 uint_t input_channels;
78 AVFormatContext *avFormatCtx;
79 AVCodecContext *avCodecCtx;
82 #ifdef HAVE_AVRESAMPLE
83 AVAudioResampleContext *avr;
84 #elif defined(HAVE_SWRESAMPLE)
90 sint_t selected_stream;
95 // hack to create or re-create the context the first time _do or _do_multi is called
96 void aubio_source_avcodec_reset_resampler(aubio_source_avcodec_t * s, uint_t multi);
97 void aubio_source_avcodec_readframe(aubio_source_avcodec_t *s, uint_t * read_samples);
99 uint_t aubio_source_avcodec_has_network_url(aubio_source_avcodec_t *s);
101 uint_t aubio_source_avcodec_has_network_url(aubio_source_avcodec_t *s) {
102 char proto[20], authorization[256], hostname[128], uripath[256];
103 int proto_size = 20, authorization_size = 256, hostname_size = 128,
104 *port_ptr = 0, path_size = 256;
105 av_url_split(proto, proto_size, authorization, authorization_size, hostname,
106 hostname_size, port_ptr, uripath, path_size, s->path);
114 aubio_source_avcodec_t * new_aubio_source_avcodec(const char_t * path, uint_t samplerate, uint_t hop_size) {
115 aubio_source_avcodec_t * s = AUBIO_NEW(aubio_source_avcodec_t);
116 AVFormatContext *avFormatCtx = s->avFormatCtx;
117 AVCodecContext *avCodecCtx = s->avCodecCtx;
118 AVFrame *avFrame = s->avFrame;
119 sint_t selected_stream = -1;
120 #if FF_API_LAVF_AVCTX
121 AVCodecParameters *codecpar;
127 AUBIO_ERR("source_avcodec: Aborted opening null path\n");
130 if ((sint_t)samplerate < 0) {
131 AUBIO_ERR("source_avcodec: Can not open %s with samplerate %d\n", path, samplerate);
134 if ((sint_t)hop_size <= 0) {
135 AUBIO_ERR("source_avcodec: Can not open %s with hop_size %d\n", path, hop_size);
139 s->hop_size = hop_size;
142 if (s->path) AUBIO_FREE(s->path);
143 s->path = AUBIO_ARRAY(char_t, strnlen(path, PATH_MAX) + 1);
144 strncpy(s->path, path, strnlen(path, PATH_MAX) + 1);
146 #if LIBAVFORMAT_VERSION_INT < AV_VERSION_INT(58,0,0)
147 // register all formats and codecs
151 if (aubio_source_avcodec_has_network_url(s)) {
152 avformat_network_init();
155 // try opening the file and get some info about it
157 if ( (err = avformat_open_input(&avFormatCtx, s->path, NULL, NULL) ) < 0 ) {
159 av_strerror (err, errorstr, sizeof(errorstr));
160 AUBIO_ERR("source_avcodec: Failed opening %s (%s)\n", s->path, errorstr);
164 // try to make sure max_analyze_duration is big enough for most songs
165 #if FFMPEG_LIBAVFORMAT_MAX_DUR2
166 avFormatCtx->max_analyze_duration2 *= 100;
168 avFormatCtx->max_analyze_duration *= 100;
171 // retrieve stream information
172 if ( (err = avformat_find_stream_info(avFormatCtx, NULL)) < 0 ) {
174 av_strerror (err, errorstr, sizeof(errorstr));
175 AUBIO_ERR("source_avcodec: Could not find stream information " "for %s (%s)\n", s->path,
180 // dump information about file onto standard error
181 //av_dump_format(avFormatCtx, 0, s->path, 0);
183 // look for the first audio stream
184 for (i = 0; i < avFormatCtx->nb_streams; i++) {
185 #if FF_API_LAVF_AVCTX
186 if (avFormatCtx->streams[i]->codecpar->codec_type == AVMEDIA_TYPE_AUDIO) {
188 if (avFormatCtx->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO) {
190 if (selected_stream == -1) {
193 AUBIO_WRN("source_avcodec: More than one audio stream in %s, "
194 "taking the first one\n", s->path);
198 if (selected_stream == -1) {
199 AUBIO_ERR("source_avcodec: No audio stream in %s\n", s->path);
202 //AUBIO_DBG("Taking stream %d in file %s\n", selected_stream, s->path);
203 s->selected_stream = selected_stream;
205 #if FF_API_LAVF_AVCTX
206 codecpar = avFormatCtx->streams[selected_stream]->codecpar;
207 if (codecpar == NULL) {
208 AUBIO_ERR("source_avcodec: Could not find decoder for %s", s->path);
211 codec = avcodec_find_decoder(codecpar->codec_id);
213 /* Allocate a codec context for the decoder */
214 avCodecCtx = avcodec_alloc_context3(codec);
216 AUBIO_ERR("source_avcodec: Failed to allocate the %s codec context for path %s\n",
217 av_get_media_type_string(AVMEDIA_TYPE_AUDIO), s->path);
221 avCodecCtx = avFormatCtx->streams[selected_stream]->codec;
222 codec = avcodec_find_decoder(avCodecCtx->codec_id);
225 AUBIO_ERR("source_avcodec: Could not find decoder for %s", s->path);
229 #if FF_API_LAVF_AVCTX
230 /* Copy codec parameters from input stream to output codec context */
231 if ((err = avcodec_parameters_to_context(avCodecCtx, codecpar)) < 0) {
232 AUBIO_ERR("source_avcodec: Failed to copy %s codec parameters to decoder context for %s\n",
233 av_get_media_type_string(AVMEDIA_TYPE_AUDIO), s->path);
238 if ( ( err = avcodec_open2(avCodecCtx, codec, NULL) ) < 0) {
240 av_strerror (err, errorstr, sizeof(errorstr));
241 AUBIO_ERR("source_avcodec: Could not load codec for %s (%s)\n", s->path, errorstr);
245 /* get input specs */
246 s->input_samplerate = avCodecCtx->sample_rate;
247 s->input_channels = avCodecCtx->channels;
248 //AUBIO_DBG("input_samplerate: %d\n", s->input_samplerate);
249 //AUBIO_DBG("input_channels: %d\n", s->input_channels);
251 if (samplerate == 0) {
252 s->samplerate = s->input_samplerate;
254 s->samplerate = samplerate;
257 if (s->samplerate > s->input_samplerate) {
258 AUBIO_WRN("source_avcodec: upsampling %s from %d to %d\n", s->path,
259 s->input_samplerate, s->samplerate);
262 avFrame = av_frame_alloc();
264 AUBIO_ERR("source_avcodec: Could not allocate frame for (%s)\n", s->path);
267 /* allocate output for avr */
268 s->output = (smpl_t *)av_malloc(AUBIO_AVCODEC_MAX_BUFFER_SIZE * sizeof(smpl_t));
273 s->avFormatCtx = avFormatCtx;
274 s->avCodecCtx = avCodecCtx;
275 s->avFrame = avFrame;
277 // default to mono output
278 aubio_source_avcodec_reset_resampler(s, 0);
280 if (s->avr == NULL) goto beach;
285 //av_log_set_level(AV_LOG_QUIET);
290 //AUBIO_ERR("can not read %s at samplerate %dHz with a hop_size of %d\n",
291 // s->path, s->samplerate, s->hop_size);
292 del_aubio_source_avcodec(s);
296 void aubio_source_avcodec_reset_resampler(aubio_source_avcodec_t * s, uint_t multi) {
297 // create or reset resampler to/from mono/multi-channel
298 if ( (multi != s->multi) || (s->avr == NULL) ) {
300 int64_t input_layout = av_get_default_channel_layout(s->input_channels);
301 uint_t output_channels = multi ? s->input_channels : 1;
302 int64_t output_layout = av_get_default_channel_layout(output_channels);
303 #ifdef HAVE_AVRESAMPLE
304 AVAudioResampleContext *avr = avresample_alloc_context();
305 AVAudioResampleContext *oldavr = s->avr;
306 #elif defined(HAVE_SWRESAMPLE)
307 SwrContext *avr = swr_alloc();
308 SwrContext *oldavr = s->avr;
309 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
311 av_opt_set_int(avr, "in_channel_layout", input_layout, 0);
312 av_opt_set_int(avr, "out_channel_layout", output_layout, 0);
313 av_opt_set_int(avr, "in_sample_rate", s->input_samplerate, 0);
314 av_opt_set_int(avr, "out_sample_rate", s->samplerate, 0);
315 av_opt_set_int(avr, "in_sample_fmt", s->avCodecCtx->sample_fmt, 0);
316 #if HAVE_AUBIO_DOUBLE
317 av_opt_set_int(avr, "out_sample_fmt", AV_SAMPLE_FMT_DBL, 0);
319 av_opt_set_int(avr, "out_sample_fmt", AV_SAMPLE_FMT_FLT, 0);
322 //av_opt_set_int(avr, "out_sample_fmt", AV_SAMPLE_FMT_FLTP, 0);
323 #ifdef HAVE_AVRESAMPLE
324 if ( ( err = avresample_open(avr) ) < 0)
325 #elif defined(HAVE_SWRESAMPLE)
326 if ( ( err = swr_init(avr) ) < 0)
327 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
330 av_strerror (err, errorstr, sizeof(errorstr));
331 AUBIO_ERR("source_avcodec: Could not open resampling context for %s (%s)\n",
336 if (oldavr != NULL) {
337 #ifdef HAVE_AVRESAMPLE
338 avresample_close( oldavr );
339 #elif defined(HAVE_SWRESAMPLE)
340 swr_close ( oldavr );
341 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
349 void aubio_source_avcodec_readframe(aubio_source_avcodec_t *s, uint_t * read_samples) {
350 AVFormatContext *avFormatCtx = s->avFormatCtx;
351 AVCodecContext *avCodecCtx = s->avCodecCtx;
352 AVFrame *avFrame = s->avFrame;
353 AVPacket avPacket = s->avPacket;
354 #ifdef HAVE_AVRESAMPLE
355 AVAudioResampleContext *avr = s->avr;
356 #elif defined(HAVE_SWRESAMPLE)
357 SwrContext *avr = s->avr;
358 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
360 #ifdef HAVE_AVRESAMPLE
362 int in_samples = avFrame->nb_samples;
363 int out_linesize = 0;
364 int max_out_samples = AUBIO_AVCODEC_MAX_BUFFER_SIZE;
366 #elif defined(HAVE_SWRESAMPLE)
367 int in_samples = avFrame->nb_samples;
368 int max_out_samples = AUBIO_AVCODEC_MAX_BUFFER_SIZE / avCodecCtx->channels;
370 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
371 smpl_t *output = s->output;
372 #ifndef FF_API_LAVF_AVCTX
377 av_init_packet (&avPacket);
382 int err = av_read_frame (avFormatCtx, &avPacket);
383 if (err == AVERROR_EOF) {
389 av_strerror (err, errorstr, sizeof(errorstr));
390 AUBIO_ERR("source_avcodec: could not read frame in %s (%s)\n", s->path, errorstr);
394 } while (avPacket.stream_index != s->selected_stream);
396 #if FF_API_LAVF_AVCTX
397 ret = avcodec_send_packet(avCodecCtx, &avPacket);
398 if (ret < 0 && ret != AVERROR_EOF) {
399 AUBIO_ERR("source_avcodec: error when sending packet for %s\n", s->path);
402 ret = avcodec_receive_frame(avCodecCtx, avFrame);
407 if (ret == AVERROR(EAGAIN)) {
408 //AUBIO_WRN("source_avcodec: output is not available right now - user must try to send new input\n");
410 } else if (ret == AVERROR_EOF) {
411 AUBIO_WRN("source_avcodec: the decoder has been fully flushed, and there will be no more output frames\n");
413 AUBIO_ERR("source_avcodec: decoding errors on %s\n", s->path);
418 len = avcodec_decode_audio4(avCodecCtx, avFrame, &got_frame, &avPacket);
421 AUBIO_ERR("source_avcodec: error while decoding %s\n", s->path);
425 if (got_frame == 0) {
426 AUBIO_WRN("source_avcodec: did not get a frame when reading %s\n", s->path);
430 #if LIBAVUTIL_VERSION_MAJOR > 52
431 if (avFrame->channels != (sint_t)s->input_channels) {
432 AUBIO_WRN ("source_avcodec: trying to read from %d channel(s),"
433 "but configured for %d; is '%s' corrupt?\n", avFrame->channels,
434 s->input_channels, s->path);
439 #ifdef HAVE_AVRESAMPLE
441 av_samples_get_buffer_size(&in_linesize, avCodecCtx->channels,
442 avFrame->nb_samples, avCodecCtx->sample_fmt, 1);
443 in_samples = avFrame->nb_samples;
445 max_out_samples = AUBIO_AVCODEC_MAX_BUFFER_SIZE;
446 out_samples = avresample_convert ( avr,
447 (uint8_t **)&output, out_linesize, max_out_samples,
448 (uint8_t **)avFrame->data, in_linesize, in_samples);
449 #elif defined(HAVE_SWRESAMPLE)
450 in_samples = avFrame->nb_samples;
451 max_out_samples = AUBIO_AVCODEC_MAX_BUFFER_SIZE / avCodecCtx->channels;
452 out_samples = swr_convert( avr,
453 (uint8_t **)&output, max_out_samples,
454 (const uint8_t **)avFrame->data, in_samples);
455 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
456 if (out_samples <= 0) {
457 AUBIO_WRN("source_avcodec: no sample found while converting frame (%s)\n", s->path);
461 *read_samples = out_samples;
464 s->avFormatCtx = avFormatCtx;
465 s->avCodecCtx = avCodecCtx;
466 s->avFrame = avFrame;
467 #if defined(HAVE_AVRESAMPLE) || defined(HAVE_SWRESAMPLE)
469 #endif /* HAVE_AVRESAMPLE || HAVE_SWRESAMPLE */
472 av_packet_unref(&avPacket);
475 void aubio_source_avcodec_do(aubio_source_avcodec_t * s, fvec_t * read_data, uint_t * read){
478 uint_t total_wrote = 0;
480 if (s->multi == 1) aubio_source_avcodec_reset_resampler(s, 0);
481 while (total_wrote < s->hop_size) {
482 end = MIN(s->read_samples - s->read_index, s->hop_size - total_wrote);
483 for (i = 0; i < end; i++) {
484 read_data->data[i + total_wrote] = s->output[i + s->read_index];
487 if (total_wrote < s->hop_size) {
488 uint_t avcodec_read = 0;
489 aubio_source_avcodec_readframe(s, &avcodec_read);
490 s->read_samples = avcodec_read;
496 s->read_index += end;
499 if (total_wrote < s->hop_size) {
500 for (i = total_wrote; i < s->hop_size; i++) {
501 read_data->data[i] = 0.;
507 void aubio_source_avcodec_do_multi(aubio_source_avcodec_t * s, fmat_t * read_data, uint_t * read){
510 uint_t total_wrote = 0;
512 if (s->multi == 0) aubio_source_avcodec_reset_resampler(s, 1);
513 while (total_wrote < s->hop_size) {
514 end = MIN(s->read_samples - s->read_index, s->hop_size - total_wrote);
515 for (j = 0; j < read_data->height; j++) {
516 for (i = 0; i < end; i++) {
517 read_data->data[j][i + total_wrote] =
518 s->output[(i + s->read_index) * s->input_channels + j];
522 if (total_wrote < s->hop_size) {
523 uint_t avcodec_read = 0;
524 aubio_source_avcodec_readframe(s, &avcodec_read);
525 s->read_samples = avcodec_read;
531 s->read_index += end;
534 if (total_wrote < s->hop_size) {
535 for (j = 0; j < read_data->height; j++) {
536 for (i = total_wrote; i < s->hop_size; i++) {
537 read_data->data[j][i] = 0.;
544 uint_t aubio_source_avcodec_get_samplerate(const aubio_source_avcodec_t * s) {
545 return s->samplerate;
548 uint_t aubio_source_avcodec_get_channels(const aubio_source_avcodec_t * s) {
549 return s->input_channels;
552 uint_t aubio_source_avcodec_seek (aubio_source_avcodec_t * s, uint_t pos) {
553 int64_t resampled_pos = (uint_t)ROUND(pos * (s->input_samplerate * 1. / s->samplerate));
554 int64_t min_ts = MAX(resampled_pos - 2000, 0);
555 int64_t max_ts = MIN(resampled_pos + 2000, INT64_MAX);
556 int seek_flags = AVSEEK_FLAG_FRAME | AVSEEK_FLAG_ANY;
557 int ret = AUBIO_FAIL;
558 if (s->avFormatCtx != NULL && s->avr != NULL) {
561 AUBIO_ERR("source_avcodec: failed seeking in %s (file not opened?)", s->path);
564 if ((sint_t)pos < 0) {
565 AUBIO_ERR("source_avcodec: could not seek %s at %d (seeking position"
566 " should be >= 0)\n", s->path, pos);
569 ret = avformat_seek_file(s->avFormatCtx, s->selected_stream,
570 min_ts, resampled_pos, max_ts, seek_flags);
572 AUBIO_ERR("source_avcodec: failed seeking to %d in file %s", pos, s->path);
578 #ifdef HAVE_AVRESAMPLE
579 // reset the AVAudioResampleContext
580 avresample_close(s->avr);
581 avresample_open(s->avr);
582 #elif defined(HAVE_SWRESAMPLE)
589 uint_t aubio_source_avcodec_get_duration (aubio_source_avcodec_t * s) {
590 if (s && &(s->avFormatCtx) != NULL) {
591 int64_t duration = s->avFormatCtx->duration;
592 return s->samplerate * ((uint_t)duration / 1e6 );
597 uint_t aubio_source_avcodec_close(aubio_source_avcodec_t * s) {
598 if (s->avr != NULL) {
599 #ifdef HAVE_AVRESAMPLE
600 avresample_close( s->avr );
601 #elif defined(HAVE_SWRESAMPLE)
602 swr_close ( s->avr );
607 if (s->avCodecCtx != NULL) {
608 #ifndef HAVE_AUBIO_LIBAVCODEC_DEPRECATED
609 avcodec_free_context( &s->avCodecCtx );
611 avcodec_close ( s->avCodecCtx );
614 s->avCodecCtx = NULL;
615 if (s->avFormatCtx != NULL) {
616 avformat_close_input(&s->avFormatCtx);
617 s->avFormatCtx = NULL;
619 av_packet_unref(&s->avPacket);
623 void del_aubio_source_avcodec(aubio_source_avcodec_t * s){
625 aubio_source_avcodec_close(s);
626 if (s->output != NULL) {
630 if (s->avFrame != NULL) {
631 av_frame_free( &(s->avFrame) );
641 #endif /* HAVE_LIBAV */