1 /*
2 * Copyright (c) 2013-2018 Andreas Unterweger
3 *
4 * This file is part of FFmpeg.
5 *
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
10 *
11 * FFmpeg 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 GNU
14 * Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19 */
20
21 /**
22 * @file
23 * Simple audio converter
24 *
25 * @example transcode_aac.c
26 * Convert an input audio file to AAC in an MP4 container using FFmpeg.
27 * Formats other than MP4 are supported based on the output file extension.
28 * @author Andreas Unterweger (dustsigns@gmail.com)
29 */
30
31 #include <stdio.h>
32
33 #include "libavformat/avformat.h"
34 #include "libavformat/avio.h"
35
36 #include "libavcodec/avcodec.h"
37
38 #include "libavutil/audio_fifo.h"
39 #include "libavutil/avassert.h"
40 #include "libavutil/avstring.h"
41 #include "libavutil/frame.h"
42 #include "libavutil/opt.h"
43
44 #include "libswresample/swresample.h"
45
46 /* The output bit rate in bit/s */
47 #define OUTPUT_BIT_RATE 96000
48 /* The number of output channels */
49 #define OUTPUT_CHANNELS 2
50
51 /**
52 * Open an input file and the required decoder.
53 * @param filename File to be opened
54 * @param[out] input_format_context Format context of opened file
55 * @param[out] input_codec_context Codec context of opened file
56 * @return Error code (0 if successful)
57 */
open_input_file(const char * filename,AVFormatContext ** input_format_context,AVCodecContext ** input_codec_context)58 static int open_input_file(const char *filename,
59 AVFormatContext **input_format_context,
60 AVCodecContext **input_codec_context)
61 {
62 AVCodecContext *avctx;
63 AVCodec *input_codec;
64 int error;
65
66 /* Open the input file to read from it. */
67 if ((error = avformat_open_input(input_format_context, filename, NULL,
68 NULL)) < 0) {
69 fprintf(stderr, "Could not open input file '%s' (error '%s')\n",
70 filename, av_err2str(error));
71 *input_format_context = NULL;
72 return error;
73 }
74
75 /* Get information on the input file (number of streams etc.). */
76 if ((error = avformat_find_stream_info(*input_format_context, NULL)) < 0) {
77 fprintf(stderr, "Could not open find stream info (error '%s')\n",
78 av_err2str(error));
79 avformat_close_input(input_format_context);
80 return error;
81 }
82
83 /* Make sure that there is only one stream in the input file. */
84 if ((*input_format_context)->nb_streams != 1) {
85 fprintf(stderr, "Expected one audio input stream, but found %d\n",
86 (*input_format_context)->nb_streams);
87 avformat_close_input(input_format_context);
88 return AVERROR_EXIT;
89 }
90
91 /* Find a decoder for the audio stream. */
92 if (!(input_codec = avcodec_find_decoder((*input_format_context)->streams[0]->codecpar->codec_id))) {
93 fprintf(stderr, "Could not find input codec\n");
94 avformat_close_input(input_format_context);
95 return AVERROR_EXIT;
96 }
97
98 /* Allocate a new decoding context. */
99 avctx = avcodec_alloc_context3(input_codec);
100 if (!avctx) {
101 fprintf(stderr, "Could not allocate a decoding context\n");
102 avformat_close_input(input_format_context);
103 return AVERROR(ENOMEM);
104 }
105
106 /* Initialize the stream parameters with demuxer information. */
107 error = avcodec_parameters_to_context(avctx, (*input_format_context)->streams[0]->codecpar);
108 if (error < 0) {
109 avformat_close_input(input_format_context);
110 avcodec_free_context(&avctx);
111 return error;
112 }
113
114 /* Open the decoder for the audio stream to use it later. */
115 if ((error = avcodec_open2(avctx, input_codec, NULL)) < 0) {
116 fprintf(stderr, "Could not open input codec (error '%s')\n",
117 av_err2str(error));
118 avcodec_free_context(&avctx);
119 avformat_close_input(input_format_context);
120 return error;
121 }
122
123 /* Save the decoder context for easier access later. */
124 *input_codec_context = avctx;
125
126 return 0;
127 }
128
129 /**
130 * Open an output file and the required encoder.
131 * Also set some basic encoder parameters.
132 * Some of these parameters are based on the input file's parameters.
133 * @param filename File to be opened
134 * @param input_codec_context Codec context of input file
135 * @param[out] output_format_context Format context of output file
136 * @param[out] output_codec_context Codec context of output file
137 * @return Error code (0 if successful)
138 */
open_output_file(const char * filename,AVCodecContext * input_codec_context,AVFormatContext ** output_format_context,AVCodecContext ** output_codec_context)139 static int open_output_file(const char *filename,
140 AVCodecContext *input_codec_context,
141 AVFormatContext **output_format_context,
142 AVCodecContext **output_codec_context)
143 {
144 AVCodecContext *avctx = NULL;
145 AVIOContext *output_io_context = NULL;
146 AVStream *stream = NULL;
147 AVCodec *output_codec = NULL;
148 int error;
149
150 /* Open the output file to write to it. */
151 if ((error = avio_open(&output_io_context, filename,
152 AVIO_FLAG_WRITE)) < 0) {
153 fprintf(stderr, "Could not open output file '%s' (error '%s')\n",
154 filename, av_err2str(error));
155 return error;
156 }
157
158 /* Create a new format context for the output container format. */
159 if (!(*output_format_context = avformat_alloc_context())) {
160 fprintf(stderr, "Could not allocate output format context\n");
161 return AVERROR(ENOMEM);
162 }
163
164 /* Associate the output file (pointer) with the container format context. */
165 (*output_format_context)->pb = output_io_context;
166
167 /* Guess the desired container format based on the file extension. */
168 if (!((*output_format_context)->oformat = av_guess_format(NULL, filename,
169 NULL))) {
170 fprintf(stderr, "Could not find output file format\n");
171 goto cleanup;
172 }
173
174 if (!((*output_format_context)->url = av_strdup(filename))) {
175 fprintf(stderr, "Could not allocate url.\n");
176 error = AVERROR(ENOMEM);
177 goto cleanup;
178 }
179
180 /* Find the encoder to be used by its name. */
181 if (!(output_codec = avcodec_find_encoder(AV_CODEC_ID_AAC))) {
182 fprintf(stderr, "Could not find an AAC encoder.\n");
183 goto cleanup;
184 }
185
186 /* Create a new audio stream in the output file container. */
187 if (!(stream = avformat_new_stream(*output_format_context, NULL))) {
188 fprintf(stderr, "Could not create new stream\n");
189 error = AVERROR(ENOMEM);
190 goto cleanup;
191 }
192
193 avctx = avcodec_alloc_context3(output_codec);
194 if (!avctx) {
195 fprintf(stderr, "Could not allocate an encoding context\n");
196 error = AVERROR(ENOMEM);
197 goto cleanup;
198 }
199
200 /* Set the basic encoder parameters.
201 * The input file's sample rate is used to avoid a sample rate conversion. */
202 avctx->channels = OUTPUT_CHANNELS;
203 avctx->channel_layout = av_get_default_channel_layout(OUTPUT_CHANNELS);
204 avctx->sample_rate = input_codec_context->sample_rate;
205 avctx->sample_fmt = output_codec->sample_fmts[0];
206 avctx->bit_rate = OUTPUT_BIT_RATE;
207
208 /* Allow the use of the experimental AAC encoder. */
209 avctx->strict_std_compliance = FF_COMPLIANCE_EXPERIMENTAL;
210
211 /* Set the sample rate for the container. */
212 stream->time_base.den = input_codec_context->sample_rate;
213 stream->time_base.num = 1;
214
215 /* Some container formats (like MP4) require global headers to be present.
216 * Mark the encoder so that it behaves accordingly. */
217 if ((*output_format_context)->oformat->flags & AVFMT_GLOBALHEADER)
218 avctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
219
220 /* Open the encoder for the audio stream to use it later. */
221 if ((error = avcodec_open2(avctx, output_codec, NULL)) < 0) {
222 fprintf(stderr, "Could not open output codec (error '%s')\n",
223 av_err2str(error));
224 goto cleanup;
225 }
226
227 error = avcodec_parameters_from_context(stream->codecpar, avctx);
228 if (error < 0) {
229 fprintf(stderr, "Could not initialize stream parameters\n");
230 goto cleanup;
231 }
232
233 /* Save the encoder context for easier access later. */
234 *output_codec_context = avctx;
235
236 return 0;
237
238 cleanup:
239 avcodec_free_context(&avctx);
240 avio_closep(&(*output_format_context)->pb);
241 avformat_free_context(*output_format_context);
242 *output_format_context = NULL;
243 return error < 0 ? error : AVERROR_EXIT;
244 }
245
246 /**
247 * Initialize one data packet for reading or writing.
248 * @param packet Packet to be initialized
249 */
init_packet(AVPacket * packet)250 static void init_packet(AVPacket *packet)
251 {
252 av_init_packet(packet);
253 /* Set the packet data and size so that it is recognized as being empty. */
254 packet->data = NULL;
255 packet->size = 0;
256 }
257
258 /**
259 * Initialize one audio frame for reading from the input file.
260 * @param[out] frame Frame to be initialized
261 * @return Error code (0 if successful)
262 */
init_input_frame(AVFrame ** frame)263 static int init_input_frame(AVFrame **frame)
264 {
265 if (!(*frame = av_frame_alloc())) {
266 fprintf(stderr, "Could not allocate input frame\n");
267 return AVERROR(ENOMEM);
268 }
269 return 0;
270 }
271
272 /**
273 * Initialize the audio resampler based on the input and output codec settings.
274 * If the input and output sample formats differ, a conversion is required
275 * libswresample takes care of this, but requires initialization.
276 * @param input_codec_context Codec context of the input file
277 * @param output_codec_context Codec context of the output file
278 * @param[out] resample_context Resample context for the required conversion
279 * @return Error code (0 if successful)
280 */
init_resampler(AVCodecContext * input_codec_context,AVCodecContext * output_codec_context,SwrContext ** resample_context)281 static int init_resampler(AVCodecContext *input_codec_context,
282 AVCodecContext *output_codec_context,
283 SwrContext **resample_context)
284 {
285 int error;
286
287 /*
288 * Create a resampler context for the conversion.
289 * Set the conversion parameters.
290 * Default channel layouts based on the number of channels
291 * are assumed for simplicity (they are sometimes not detected
292 * properly by the demuxer and/or decoder).
293 */
294 *resample_context = swr_alloc_set_opts(NULL,
295 av_get_default_channel_layout(output_codec_context->channels),
296 output_codec_context->sample_fmt,
297 output_codec_context->sample_rate,
298 av_get_default_channel_layout(input_codec_context->channels),
299 input_codec_context->sample_fmt,
300 input_codec_context->sample_rate,
301 0, NULL);
302 if (!*resample_context) {
303 fprintf(stderr, "Could not allocate resample context\n");
304 return AVERROR(ENOMEM);
305 }
306 /*
307 * Perform a sanity check so that the number of converted samples is
308 * not greater than the number of samples to be converted.
309 * If the sample rates differ, this case has to be handled differently
310 */
311 av_assert0(output_codec_context->sample_rate == input_codec_context->sample_rate);
312
313 /* Open the resampler with the specified parameters. */
314 if ((error = swr_init(*resample_context)) < 0) {
315 fprintf(stderr, "Could not open resample context\n");
316 swr_free(resample_context);
317 return error;
318 }
319 return 0;
320 }
321
322 /**
323 * Initialize a FIFO buffer for the audio samples to be encoded.
324 * @param[out] fifo Sample buffer
325 * @param output_codec_context Codec context of the output file
326 * @return Error code (0 if successful)
327 */
init_fifo(AVAudioFifo ** fifo,AVCodecContext * output_codec_context)328 static int init_fifo(AVAudioFifo **fifo, AVCodecContext *output_codec_context)
329 {
330 /* Create the FIFO buffer based on the specified output sample format. */
331 if (!(*fifo = av_audio_fifo_alloc(output_codec_context->sample_fmt,
332 output_codec_context->channels, 1))) {
333 fprintf(stderr, "Could not allocate FIFO\n");
334 return AVERROR(ENOMEM);
335 }
336 return 0;
337 }
338
339 /**
340 * Write the header of the output file container.
341 * @param output_format_context Format context of the output file
342 * @return Error code (0 if successful)
343 */
write_output_file_header(AVFormatContext * output_format_context)344 static int write_output_file_header(AVFormatContext *output_format_context)
345 {
346 int error;
347 if ((error = avformat_write_header(output_format_context, NULL)) < 0) {
348 fprintf(stderr, "Could not write output file header (error '%s')\n",
349 av_err2str(error));
350 return error;
351 }
352 return 0;
353 }
354
355 /**
356 * Decode one audio frame from the input file.
357 * @param frame Audio frame to be decoded
358 * @param input_format_context Format context of the input file
359 * @param input_codec_context Codec context of the input file
360 * @param[out] data_present Indicates whether data has been decoded
361 * @param[out] finished Indicates whether the end of file has
362 * been reached and all data has been
363 * decoded. If this flag is false, there
364 * is more data to be decoded, i.e., this
365 * function has to be called again.
366 * @return Error code (0 if successful)
367 */
decode_audio_frame(AVFrame * frame,AVFormatContext * input_format_context,AVCodecContext * input_codec_context,int * data_present,int * finished)368 static int decode_audio_frame(AVFrame *frame,
369 AVFormatContext *input_format_context,
370 AVCodecContext *input_codec_context,
371 int *data_present, int *finished)
372 {
373 /* Packet used for temporary storage. */
374 AVPacket input_packet;
375 int error;
376 init_packet(&input_packet);
377
378 /* Read one audio frame from the input file into a temporary packet. */
379 if ((error = av_read_frame(input_format_context, &input_packet)) < 0) {
380 /* If we are at the end of the file, flush the decoder below. */
381 if (error == AVERROR_EOF)
382 *finished = 1;
383 else {
384 fprintf(stderr, "Could not read frame (error '%s')\n",
385 av_err2str(error));
386 return error;
387 }
388 }
389
390 /* Send the audio frame stored in the temporary packet to the decoder.
391 * The input audio stream decoder is used to do this. */
392 if ((error = avcodec_send_packet(input_codec_context, &input_packet)) < 0) {
393 fprintf(stderr, "Could not send packet for decoding (error '%s')\n",
394 av_err2str(error));
395 return error;
396 }
397
398 /* Receive one frame from the decoder. */
399 error = avcodec_receive_frame(input_codec_context, frame);
400 /* If the decoder asks for more data to be able to decode a frame,
401 * return indicating that no data is present. */
402 if (error == AVERROR(EAGAIN)) {
403 error = 0;
404 goto cleanup;
405 /* If the end of the input file is reached, stop decoding. */
406 } else if (error == AVERROR_EOF) {
407 *finished = 1;
408 error = 0;
409 goto cleanup;
410 } else if (error < 0) {
411 fprintf(stderr, "Could not decode frame (error '%s')\n",
412 av_err2str(error));
413 goto cleanup;
414 /* Default case: Return decoded data. */
415 } else {
416 *data_present = 1;
417 goto cleanup;
418 }
419
420 cleanup:
421 av_packet_unref(&input_packet);
422 return error;
423 }
424
425 /**
426 * Initialize a temporary storage for the specified number of audio samples.
427 * The conversion requires temporary storage due to the different format.
428 * The number of audio samples to be allocated is specified in frame_size.
429 * @param[out] converted_input_samples Array of converted samples. The
430 * dimensions are reference, channel
431 * (for multi-channel audio), sample.
432 * @param output_codec_context Codec context of the output file
433 * @param frame_size Number of samples to be converted in
434 * each round
435 * @return Error code (0 if successful)
436 */
init_converted_samples(uint8_t *** converted_input_samples,AVCodecContext * output_codec_context,int frame_size)437 static int init_converted_samples(uint8_t ***converted_input_samples,
438 AVCodecContext *output_codec_context,
439 int frame_size)
440 {
441 int error;
442
443 /* Allocate as many pointers as there are audio channels.
444 * Each pointer will later point to the audio samples of the corresponding
445 * channels (although it may be NULL for interleaved formats).
446 */
447 if (!(*converted_input_samples = calloc(output_codec_context->channels,
448 sizeof(**converted_input_samples)))) {
449 fprintf(stderr, "Could not allocate converted input sample pointers\n");
450 return AVERROR(ENOMEM);
451 }
452
453 /* Allocate memory for the samples of all channels in one consecutive
454 * block for convenience. */
455 if ((error = av_samples_alloc(*converted_input_samples, NULL,
456 output_codec_context->channels,
457 frame_size,
458 output_codec_context->sample_fmt, 0)) < 0) {
459 fprintf(stderr,
460 "Could not allocate converted input samples (error '%s')\n",
461 av_err2str(error));
462 av_freep(&(*converted_input_samples)[0]);
463 free(*converted_input_samples);
464 return error;
465 }
466 return 0;
467 }
468
469 /**
470 * Convert the input audio samples into the output sample format.
471 * The conversion happens on a per-frame basis, the size of which is
472 * specified by frame_size.
473 * @param input_data Samples to be decoded. The dimensions are
474 * channel (for multi-channel audio), sample.
475 * @param[out] converted_data Converted samples. The dimensions are channel
476 * (for multi-channel audio), sample.
477 * @param frame_size Number of samples to be converted
478 * @param resample_context Resample context for the conversion
479 * @return Error code (0 if successful)
480 */
convert_samples(const uint8_t ** input_data,uint8_t ** converted_data,const int frame_size,SwrContext * resample_context)481 static int convert_samples(const uint8_t **input_data,
482 uint8_t **converted_data, const int frame_size,
483 SwrContext *resample_context)
484 {
485 int error;
486
487 /* Convert the samples using the resampler. */
488 if ((error = swr_convert(resample_context,
489 converted_data, frame_size,
490 input_data , frame_size)) < 0) {
491 fprintf(stderr, "Could not convert input samples (error '%s')\n",
492 av_err2str(error));
493 return error;
494 }
495
496 return 0;
497 }
498
499 /**
500 * Add converted input audio samples to the FIFO buffer for later processing.
501 * @param fifo Buffer to add the samples to
502 * @param converted_input_samples Samples to be added. The dimensions are channel
503 * (for multi-channel audio), sample.
504 * @param frame_size Number of samples to be converted
505 * @return Error code (0 if successful)
506 */
add_samples_to_fifo(AVAudioFifo * fifo,uint8_t ** converted_input_samples,const int frame_size)507 static int add_samples_to_fifo(AVAudioFifo *fifo,
508 uint8_t **converted_input_samples,
509 const int frame_size)
510 {
511 int error;
512
513 /* Make the FIFO as large as it needs to be to hold both,
514 * the old and the new samples. */
515 if ((error = av_audio_fifo_realloc(fifo, av_audio_fifo_size(fifo) + frame_size)) < 0) {
516 fprintf(stderr, "Could not reallocate FIFO\n");
517 return error;
518 }
519
520 /* Store the new samples in the FIFO buffer. */
521 if (av_audio_fifo_write(fifo, (void **)converted_input_samples,
522 frame_size) < frame_size) {
523 fprintf(stderr, "Could not write data to FIFO\n");
524 return AVERROR_EXIT;
525 }
526 return 0;
527 }
528
529 /**
530 * Read one audio frame from the input file, decode, convert and store
531 * it in the FIFO buffer.
532 * @param fifo Buffer used for temporary storage
533 * @param input_format_context Format context of the input file
534 * @param input_codec_context Codec context of the input file
535 * @param output_codec_context Codec context of the output file
536 * @param resampler_context Resample context for the conversion
537 * @param[out] finished Indicates whether the end of file has
538 * been reached and all data has been
539 * decoded. If this flag is false,
540 * there is more data to be decoded,
541 * i.e., this function has to be called
542 * again.
543 * @return Error code (0 if successful)
544 */
read_decode_convert_and_store(AVAudioFifo * fifo,AVFormatContext * input_format_context,AVCodecContext * input_codec_context,AVCodecContext * output_codec_context,SwrContext * resampler_context,int * finished)545 static int read_decode_convert_and_store(AVAudioFifo *fifo,
546 AVFormatContext *input_format_context,
547 AVCodecContext *input_codec_context,
548 AVCodecContext *output_codec_context,
549 SwrContext *resampler_context,
550 int *finished)
551 {
552 /* Temporary storage of the input samples of the frame read from the file. */
553 AVFrame *input_frame = NULL;
554 /* Temporary storage for the converted input samples. */
555 uint8_t **converted_input_samples = NULL;
556 int data_present = 0;
557 int ret = AVERROR_EXIT;
558
559 /* Initialize temporary storage for one input frame. */
560 if (init_input_frame(&input_frame))
561 goto cleanup;
562 /* Decode one frame worth of audio samples. */
563 if (decode_audio_frame(input_frame, input_format_context,
564 input_codec_context, &data_present, finished))
565 goto cleanup;
566 /* If we are at the end of the file and there are no more samples
567 * in the decoder which are delayed, we are actually finished.
568 * This must not be treated as an error. */
569 if (*finished) {
570 ret = 0;
571 goto cleanup;
572 }
573 /* If there is decoded data, convert and store it. */
574 if (data_present) {
575 /* Initialize the temporary storage for the converted input samples. */
576 if (init_converted_samples(&converted_input_samples, output_codec_context,
577 input_frame->nb_samples))
578 goto cleanup;
579
580 /* Convert the input samples to the desired output sample format.
581 * This requires a temporary storage provided by converted_input_samples. */
582 if (convert_samples((const uint8_t**)input_frame->extended_data, converted_input_samples,
583 input_frame->nb_samples, resampler_context))
584 goto cleanup;
585
586 /* Add the converted input samples to the FIFO buffer for later processing. */
587 if (add_samples_to_fifo(fifo, converted_input_samples,
588 input_frame->nb_samples))
589 goto cleanup;
590 ret = 0;
591 }
592 ret = 0;
593
594 cleanup:
595 if (converted_input_samples) {
596 av_freep(&converted_input_samples[0]);
597 free(converted_input_samples);
598 }
599 av_frame_free(&input_frame);
600
601 return ret;
602 }
603
604 /**
605 * Initialize one input frame for writing to the output file.
606 * The frame will be exactly frame_size samples large.
607 * @param[out] frame Frame to be initialized
608 * @param output_codec_context Codec context of the output file
609 * @param frame_size Size of the frame
610 * @return Error code (0 if successful)
611 */
init_output_frame(AVFrame ** frame,AVCodecContext * output_codec_context,int frame_size)612 static int init_output_frame(AVFrame **frame,
613 AVCodecContext *output_codec_context,
614 int frame_size)
615 {
616 int error;
617
618 /* Create a new frame to store the audio samples. */
619 if (!(*frame = av_frame_alloc())) {
620 fprintf(stderr, "Could not allocate output frame\n");
621 return AVERROR_EXIT;
622 }
623
624 /* Set the frame's parameters, especially its size and format.
625 * av_frame_get_buffer needs this to allocate memory for the
626 * audio samples of the frame.
627 * Default channel layouts based on the number of channels
628 * are assumed for simplicity. */
629 (*frame)->nb_samples = frame_size;
630 (*frame)->channel_layout = output_codec_context->channel_layout;
631 (*frame)->format = output_codec_context->sample_fmt;
632 (*frame)->sample_rate = output_codec_context->sample_rate;
633
634 /* Allocate the samples of the created frame. This call will make
635 * sure that the audio frame can hold as many samples as specified. */
636 if ((error = av_frame_get_buffer(*frame, 0)) < 0) {
637 fprintf(stderr, "Could not allocate output frame samples (error '%s')\n",
638 av_err2str(error));
639 av_frame_free(frame);
640 return error;
641 }
642
643 return 0;
644 }
645
646 /* Global timestamp for the audio frames. */
647 static int64_t pts = 0;
648
649 /**
650 * Encode one frame worth of audio to the output file.
651 * @param frame Samples to be encoded
652 * @param output_format_context Format context of the output file
653 * @param output_codec_context Codec context of the output file
654 * @param[out] data_present Indicates whether data has been
655 * encoded
656 * @return Error code (0 if successful)
657 */
encode_audio_frame(AVFrame * frame,AVFormatContext * output_format_context,AVCodecContext * output_codec_context,int * data_present)658 static int encode_audio_frame(AVFrame *frame,
659 AVFormatContext *output_format_context,
660 AVCodecContext *output_codec_context,
661 int *data_present)
662 {
663 /* Packet used for temporary storage. */
664 AVPacket output_packet;
665 int error;
666 init_packet(&output_packet);
667
668 /* Set a timestamp based on the sample rate for the container. */
669 if (frame) {
670 frame->pts = pts;
671 pts += frame->nb_samples;
672 }
673
674 /* Send the audio frame stored in the temporary packet to the encoder.
675 * The output audio stream encoder is used to do this. */
676 error = avcodec_send_frame(output_codec_context, frame);
677 /* The encoder signals that it has nothing more to encode. */
678 if (error == AVERROR_EOF) {
679 error = 0;
680 goto cleanup;
681 } else if (error < 0) {
682 fprintf(stderr, "Could not send packet for encoding (error '%s')\n",
683 av_err2str(error));
684 return error;
685 }
686
687 /* Receive one encoded frame from the encoder. */
688 error = avcodec_receive_packet(output_codec_context, &output_packet);
689 /* If the encoder asks for more data to be able to provide an
690 * encoded frame, return indicating that no data is present. */
691 if (error == AVERROR(EAGAIN)) {
692 error = 0;
693 goto cleanup;
694 /* If the last frame has been encoded, stop encoding. */
695 } else if (error == AVERROR_EOF) {
696 error = 0;
697 goto cleanup;
698 } else if (error < 0) {
699 fprintf(stderr, "Could not encode frame (error '%s')\n",
700 av_err2str(error));
701 goto cleanup;
702 /* Default case: Return encoded data. */
703 } else {
704 *data_present = 1;
705 }
706
707 /* Write one audio frame from the temporary packet to the output file. */
708 if (*data_present &&
709 (error = av_write_frame(output_format_context, &output_packet)) < 0) {
710 fprintf(stderr, "Could not write frame (error '%s')\n",
711 av_err2str(error));
712 goto cleanup;
713 }
714
715 cleanup:
716 av_packet_unref(&output_packet);
717 return error;
718 }
719
720 /**
721 * Load one audio frame from the FIFO buffer, encode and write it to the
722 * output file.
723 * @param fifo Buffer used for temporary storage
724 * @param output_format_context Format context of the output file
725 * @param output_codec_context Codec context of the output file
726 * @return Error code (0 if successful)
727 */
load_encode_and_write(AVAudioFifo * fifo,AVFormatContext * output_format_context,AVCodecContext * output_codec_context)728 static int load_encode_and_write(AVAudioFifo *fifo,
729 AVFormatContext *output_format_context,
730 AVCodecContext *output_codec_context)
731 {
732 /* Temporary storage of the output samples of the frame written to the file. */
733 AVFrame *output_frame;
734 /* Use the maximum number of possible samples per frame.
735 * If there is less than the maximum possible frame size in the FIFO
736 * buffer use this number. Otherwise, use the maximum possible frame size. */
737 const int frame_size = FFMIN(av_audio_fifo_size(fifo),
738 output_codec_context->frame_size);
739 int data_written;
740
741 /* Initialize temporary storage for one output frame. */
742 if (init_output_frame(&output_frame, output_codec_context, frame_size))
743 return AVERROR_EXIT;
744
745 /* Read as many samples from the FIFO buffer as required to fill the frame.
746 * The samples are stored in the frame temporarily. */
747 if (av_audio_fifo_read(fifo, (void **)output_frame->data, frame_size) < frame_size) {
748 fprintf(stderr, "Could not read data from FIFO\n");
749 av_frame_free(&output_frame);
750 return AVERROR_EXIT;
751 }
752
753 /* Encode one frame worth of audio samples. */
754 if (encode_audio_frame(output_frame, output_format_context,
755 output_codec_context, &data_written)) {
756 av_frame_free(&output_frame);
757 return AVERROR_EXIT;
758 }
759 av_frame_free(&output_frame);
760 return 0;
761 }
762
763 /**
764 * Write the trailer of the output file container.
765 * @param output_format_context Format context of the output file
766 * @return Error code (0 if successful)
767 */
write_output_file_trailer(AVFormatContext * output_format_context)768 static int write_output_file_trailer(AVFormatContext *output_format_context)
769 {
770 int error;
771 if ((error = av_write_trailer(output_format_context)) < 0) {
772 fprintf(stderr, "Could not write output file trailer (error '%s')\n",
773 av_err2str(error));
774 return error;
775 }
776 return 0;
777 }
778
main(int argc,char ** argv)779 int main(int argc, char **argv)
780 {
781 AVFormatContext *input_format_context = NULL, *output_format_context = NULL;
782 AVCodecContext *input_codec_context = NULL, *output_codec_context = NULL;
783 SwrContext *resample_context = NULL;
784 AVAudioFifo *fifo = NULL;
785 int ret = AVERROR_EXIT;
786
787 if (argc != 3) {
788 fprintf(stderr, "Usage: %s <input file> <output file>\n", argv[0]);
789 exit(1);
790 }
791
792 /* Open the input file for reading. */
793 if (open_input_file(argv[1], &input_format_context,
794 &input_codec_context))
795 goto cleanup;
796 /* Open the output file for writing. */
797 if (open_output_file(argv[2], input_codec_context,
798 &output_format_context, &output_codec_context))
799 goto cleanup;
800 /* Initialize the resampler to be able to convert audio sample formats. */
801 if (init_resampler(input_codec_context, output_codec_context,
802 &resample_context))
803 goto cleanup;
804 /* Initialize the FIFO buffer to store audio samples to be encoded. */
805 if (init_fifo(&fifo, output_codec_context))
806 goto cleanup;
807 /* Write the header of the output file container. */
808 if (write_output_file_header(output_format_context))
809 goto cleanup;
810
811 /* Loop as long as we have input samples to read or output samples
812 * to write; abort as soon as we have neither. */
813 while (1) {
814 /* Use the encoder's desired frame size for processing. */
815 const int output_frame_size = output_codec_context->frame_size;
816 int finished = 0;
817
818 /* Make sure that there is one frame worth of samples in the FIFO
819 * buffer so that the encoder can do its work.
820 * Since the decoder's and the encoder's frame size may differ, we
821 * need to FIFO buffer to store as many frames worth of input samples
822 * that they make up at least one frame worth of output samples. */
823 while (av_audio_fifo_size(fifo) < output_frame_size) {
824 /* Decode one frame worth of audio samples, convert it to the
825 * output sample format and put it into the FIFO buffer. */
826 if (read_decode_convert_and_store(fifo, input_format_context,
827 input_codec_context,
828 output_codec_context,
829 resample_context, &finished))
830 goto cleanup;
831
832 /* If we are at the end of the input file, we continue
833 * encoding the remaining audio samples to the output file. */
834 if (finished)
835 break;
836 }
837
838 /* If we have enough samples for the encoder, we encode them.
839 * At the end of the file, we pass the remaining samples to
840 * the encoder. */
841 while (av_audio_fifo_size(fifo) >= output_frame_size ||
842 (finished && av_audio_fifo_size(fifo) > 0))
843 /* Take one frame worth of audio samples from the FIFO buffer,
844 * encode it and write it to the output file. */
845 if (load_encode_and_write(fifo, output_format_context,
846 output_codec_context))
847 goto cleanup;
848
849 /* If we are at the end of the input file and have encoded
850 * all remaining samples, we can exit this loop and finish. */
851 if (finished) {
852 int data_written;
853 /* Flush the encoder as it may have delayed frames. */
854 do {
855 data_written = 0;
856 if (encode_audio_frame(NULL, output_format_context,
857 output_codec_context, &data_written))
858 goto cleanup;
859 } while (data_written);
860 break;
861 }
862 }
863
864 /* Write the trailer of the output file container. */
865 if (write_output_file_trailer(output_format_context))
866 goto cleanup;
867 ret = 0;
868
869 cleanup:
870 if (fifo)
871 av_audio_fifo_free(fifo);
872 swr_free(&resample_context);
873 if (output_codec_context)
874 avcodec_free_context(&output_codec_context);
875 if (output_format_context) {
876 avio_closep(&output_format_context->pb);
877 avformat_free_context(output_format_context);
878 }
879 if (input_codec_context)
880 avcodec_free_context(&input_codec_context);
881 if (input_format_context)
882 avformat_close_input(&input_format_context);
883
884 return ret;
885 }
886