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/>.
26 #include <libavcodec/avcodec.h>
27 #include <libavformat/avformat.h>
28 #include <libavresample/avresample.h>
29 #include <libavutil/opt.h>
32 #include "aubio_priv.h"
35 #include "source_avcodec.h"
37 #define AUBIO_AVCODEC_MAX_BUFFER_SIZE FF_MIN_BUFFER_SIZE
39 struct _aubio_source_avcodec_t {
44 // some data about the file
46 uint_t input_samplerate;
47 uint_t input_channels;
50 AVFormatContext *avFormatCtx;
51 AVCodecContext *avCodecCtx;
53 AVAudioResampleContext *avr;
57 sint_t selected_stream;
62 // hack to create or re-create the context the first time _do or _do_multi is called
63 void aubio_source_avcodec_reset_resampler(aubio_source_avcodec_t * s, uint_t multi);
64 void aubio_source_avcodec_readframe(aubio_source_avcodec_t *s, uint_t * read_samples);
66 aubio_source_avcodec_t * new_aubio_source_avcodec(char_t * path, uint_t samplerate, uint_t hop_size) {
67 aubio_source_avcodec_t * s = AUBIO_NEW(aubio_source_avcodec_t);
70 AUBIO_ERR("source_avcodec: Aborted opening null path\n");
73 if ((sint_t)samplerate < 0) {
74 AUBIO_ERR("source_avcodec: Can not open %s with samplerate %d\n", path, samplerate);
77 if ((sint_t)hop_size <= 0) {
78 AUBIO_ERR("source_avcodec: Can not open %s with hop_size %d\n", path, hop_size);
82 s->hop_size = hop_size;
86 // register all formats and codecs
90 //avformat_network_init();
92 // try opening the file and get some info about it
93 AVFormatContext *avFormatCtx = s->avFormatCtx;
95 if ( (err = avformat_open_input(&avFormatCtx, s->path, NULL, NULL) ) < 0 ) {
97 av_strerror (err, errorstr, sizeof(errorstr));
98 AUBIO_ERR("source_avcodec: Failed opening %s (%s)\n", s->path, errorstr);
102 // try to make sure max_analyze_duration is big enough for most songs
103 avFormatCtx->max_analyze_duration *= 100;
105 // retrieve stream information
106 if ( (err = avformat_find_stream_info(avFormatCtx, NULL)) < 0 ) {
108 av_strerror (err, errorstr, sizeof(errorstr));
109 AUBIO_ERR("source_avcodec: Could not find stream information " "for %s (%s)\n", s->path,
114 // dump information about file onto standard error
115 //av_dump_format(avFormatCtx, 0, s->path, 0);
117 // look for the first audio stream
119 sint_t selected_stream = -1;
120 for (i = 0; i < avFormatCtx->nb_streams; i++) {
121 if (avFormatCtx->streams[i]->codec->codec_type == AVMEDIA_TYPE_AUDIO) {
122 if (selected_stream == -1) {
125 AUBIO_WRN("source_avcodec: More than one audio stream in %s, "
126 "taking the first one\n", s->path);
130 if (selected_stream == -1) {
131 AUBIO_ERR("source_avcodec: No audio stream in %s\n", s->path);
134 //AUBIO_DBG("Taking stream %d in file %s\n", selected_stream, s->path);
135 s->selected_stream = selected_stream;
137 AVCodecContext *avCodecCtx = s->avCodecCtx;
138 avCodecCtx = avFormatCtx->streams[selected_stream]->codec;
139 AVCodec *codec = avcodec_find_decoder(avCodecCtx->codec_id);
141 AUBIO_ERR("source_avcodec: Could not find decoder for %s", s->path);
145 if ( ( err = avcodec_open2(avCodecCtx, codec, NULL) ) < 0) {
147 av_strerror (err, errorstr, sizeof(errorstr));
148 AUBIO_ERR("source_avcodec: Could not load codec for %s (%s)\n", s->path, errorstr);
152 /* get input specs */
153 s->input_samplerate = avCodecCtx->sample_rate;
154 s->input_channels = avCodecCtx->channels;
155 //AUBIO_DBG("input_samplerate: %d\n", s->input_samplerate);
156 //AUBIO_DBG("input_channels: %d\n", s->input_channels);
158 if (samplerate == 0) {
159 samplerate = s->input_samplerate;
160 //AUBIO_DBG("sampling rate set to 0, automagically adjusting to %d\n", samplerate);
162 s->samplerate = samplerate;
164 if (s->samplerate > s->input_samplerate) {
165 AUBIO_WRN("source_avcodec: upsampling %s from %d to %d\n", s->path,
166 s->input_samplerate, s->samplerate);
169 AVFrame *avFrame = s->avFrame;
170 avFrame = av_frame_alloc();
172 AUBIO_ERR("source_avcodec: Could not allocate frame for (%s)\n", s->path);
175 /* allocate output for avr */
176 s->output = (float *)av_malloc(AUBIO_AVCODEC_MAX_BUFFER_SIZE * sizeof(float));
181 s->avFormatCtx = avFormatCtx;
182 s->avCodecCtx = avCodecCtx;
183 s->avFrame = avFrame;
185 // default to mono output
186 aubio_source_avcodec_reset_resampler(s, 0);
191 //av_log_set_level(AV_LOG_QUIET);
196 //AUBIO_ERR("can not read %s at samplerate %dHz with a hop_size of %d\n",
197 // s->path, s->samplerate, s->hop_size);
198 del_aubio_source_avcodec(s);
202 void aubio_source_avcodec_reset_resampler(aubio_source_avcodec_t * s, uint_t multi) {
203 if ( (multi != s->multi) || (s->avr == NULL) ) {
204 int64_t input_layout = av_get_default_channel_layout(s->input_channels);
205 uint_t output_channels = multi ? s->input_channels : 1;
206 int64_t output_layout = av_get_default_channel_layout(output_channels);
207 if (s->avr != NULL) {
208 avresample_close( s->avr );
212 AVAudioResampleContext *avr = s->avr;
213 avr = avresample_alloc_context();
215 av_opt_set_int(avr, "in_channel_layout", input_layout, 0);
216 av_opt_set_int(avr, "out_channel_layout", output_layout, 0);
217 av_opt_set_int(avr, "in_sample_rate", s->input_samplerate, 0);
218 av_opt_set_int(avr, "out_sample_rate", s->samplerate, 0);
219 av_opt_set_int(avr, "in_sample_fmt", s->avCodecCtx->sample_fmt, 0);
220 av_opt_set_int(avr, "out_sample_fmt", AV_SAMPLE_FMT_FLT, 0);
222 if ( ( err = avresample_open(avr) ) < 0) {
224 av_strerror (err, errorstr, sizeof(errorstr));
225 AUBIO_ERR("source_avcodec: Could not open AVAudioResampleContext for %s (%s)\n",
235 void aubio_source_avcodec_readframe(aubio_source_avcodec_t *s, uint_t * read_samples) {
236 AVFormatContext *avFormatCtx = s->avFormatCtx;
237 AVCodecContext *avCodecCtx = s->avCodecCtx;
238 AVFrame *avFrame = s->avFrame;
240 av_init_packet (&avPacket);
241 AVAudioResampleContext *avr = s->avr;
242 float *output = s->output;
247 int err = av_read_frame (avFormatCtx, &avPacket);
248 if (err == AVERROR_EOF) {
254 av_strerror (err, errorstr, sizeof(errorstr));
255 AUBIO_ERR("Could not read frame in %s (%s)\n", s->path, errorstr);
258 } while (avPacket.stream_index != s->selected_stream);
261 int len = avcodec_decode_audio4(avCodecCtx, avFrame, &got_frame, &avPacket);
264 AUBIO_ERR("Error while decoding %s\n", s->path);
267 if (got_frame == 0) {
268 //AUBIO_ERR("Could not get frame for (%s)\n", s->path);
273 av_samples_get_buffer_size(&in_linesize, avCodecCtx->channels,
274 avFrame->nb_samples, avCodecCtx->sample_fmt, 1);
275 int in_samples = avFrame->nb_samples;
276 int out_linesize = 0;
277 int max_out_samples = AUBIO_AVCODEC_MAX_BUFFER_SIZE;
278 int out_samples = avresample_convert ( avr,
279 (uint8_t **)&output, out_linesize, max_out_samples,
280 (uint8_t **)avFrame->data, in_linesize, in_samples);
281 if (out_samples <= 0) {
282 //AUBIO_ERR("No sample found while converting frame (%s)\n", s->path);
286 *read_samples = out_samples;
289 s->avFormatCtx = avFormatCtx;
290 s->avCodecCtx = avCodecCtx;
291 s->avFrame = avFrame;
295 av_free_packet(&avPacket);
298 void aubio_source_avcodec_do(aubio_source_avcodec_t * s, fvec_t * read_data, uint_t * read){
299 if (s->multi == 1) aubio_source_avcodec_reset_resampler(s, 0);
302 uint_t total_wrote = 0;
303 while (total_wrote < s->hop_size) {
304 end = MIN(s->read_samples - s->read_index, s->hop_size - total_wrote);
305 for (i = 0; i < end; i++) {
306 read_data->data[i + total_wrote] = s->output[i + s->read_index];
309 if (total_wrote < s->hop_size) {
310 uint_t avcodec_read = 0;
311 aubio_source_avcodec_readframe(s, &avcodec_read);
312 s->read_samples = avcodec_read;
318 s->read_index += end;
321 if (total_wrote < s->hop_size) {
322 for (i = end; i < s->hop_size; i++) {
323 read_data->data[i] = 0.;
329 void aubio_source_avcodec_do_multi(aubio_source_avcodec_t * s, fmat_t * read_data, uint_t * read){
330 if (s->multi == 0) aubio_source_avcodec_reset_resampler(s, 1);
333 uint_t total_wrote = 0;
334 while (total_wrote < s->hop_size) {
335 end = MIN(s->read_samples - s->read_index, s->hop_size - total_wrote);
336 for (j = 0; j < read_data->height; j++) {
337 for (i = 0; i < end; i++) {
338 read_data->data[j][i + total_wrote] =
339 s->output[(i + s->read_index) * s->input_channels + j];
343 if (total_wrote < s->hop_size) {
344 uint_t avcodec_read = 0;
345 aubio_source_avcodec_readframe(s, &avcodec_read);
346 s->read_samples = avcodec_read;
352 s->read_index += end;
355 if (total_wrote < s->hop_size) {
356 for (j = 0; j < read_data->height; j++) {
357 for (i = end; i < s->hop_size; i++) {
358 read_data->data[j][i] = 0.;
365 uint_t aubio_source_avcodec_get_samplerate(aubio_source_avcodec_t * s) {
366 return s->samplerate;
369 uint_t aubio_source_avcodec_get_channels(aubio_source_avcodec_t * s) {
370 return s->input_channels;
373 uint_t aubio_source_avcodec_seek (aubio_source_avcodec_t * s, uint_t pos) {
374 int64_t resampled_pos = (uint_t)ROUND(pos * (s->input_samplerate * 1. / s->samplerate));
375 int64_t min_ts = MAX(resampled_pos - 2000, 0);
376 int64_t max_ts = MIN(resampled_pos + 2000, INT64_MAX);
377 int seek_flags = AVSEEK_FLAG_FRAME | AVSEEK_FLAG_ANY;
378 int ret = avformat_seek_file(s->avFormatCtx, s->selected_stream,
379 min_ts, resampled_pos, max_ts, seek_flags);
381 AUBIO_ERR("Failed seeking to %d in file %s", pos, s->path);
387 // reset the AVAudioResampleContext
388 avresample_close(s->avr);
389 avresample_open(s->avr);
393 uint_t aubio_source_avcodec_close(aubio_source_avcodec_t * s) {
394 if (s->avr != NULL) {
395 avresample_close( s->avr );
399 if (s->avCodecCtx != NULL) {
400 avcodec_close ( s->avCodecCtx );
402 s->avCodecCtx = NULL;
403 if (s->avFormatCtx != NULL) {
404 avformat_close_input ( &(s->avFormatCtx) );
406 s->avFormatCtx = NULL;
410 void del_aubio_source_avcodec(aubio_source_avcodec_t * s){
412 aubio_source_avcodec_close(s);
413 if (s->output != NULL) {
417 if (s->avFrame != NULL) {
418 av_frame_free( &(s->avFrame) );
424 #endif /* HAVE_LIBAV */