2 * This file is part of FFmpeg.
4 * FFmpeg is free software; you can redistribute it and/or
5 * modify it under the terms of the GNU Lesser General Public
6 * License as published by the Free Software Foundation; either
7 * version 2.1 of the License, or (at your option) any later version.
9 * FFmpeg is distributed in the hope that it will be useful,
10 * but WITHOUT ANY WARRANTY; without even the implied warranty of
11 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
12 * Lesser General Public License for more details.
14 * You should have received a copy of the GNU Lesser General Public
15 * License along with FFmpeg; if not, write to the Free Software
16 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 * simple audio converter
23 * @example transcode_aac.c
24 * Convert an input audio file to AAC in an MP4 container using FFmpeg.
25 * @author Andreas Unterweger (dustsigns@gmail.com)
30 #include "libavformat/avformat.h"
31 #include "libavformat/avio.h"
33 #include "libavcodec/avcodec.h"
35 #include "libavutil/audio_fifo.h"
36 #include "libavutil/avassert.h"
37 #include "libavutil/avstring.h"
38 #include "libavutil/frame.h"
39 #include "libavutil/opt.h"
41 #include "libswresample/swresample.h"
43 /** The output bit rate in kbit/s */
44 #define OUTPUT_BIT_RATE 48000
45 /** The number of output channels */
46 #define OUTPUT_CHANNELS 2
47 /** The audio sample output format */
48 #define OUTPUT_SAMPLE_FORMAT AV_SAMPLE_FMT_S16
51 * Convert an error code into a text message.
52 * @param error Error code to be converted
53 * @return Corresponding error text (not thread-safe)
55 static char *const get_error_text(const int error)
57 static char error_buffer[255];
58 av_strerror(error, error_buffer, sizeof(error_buffer));
62 /** Open an input file and the required decoder. */
63 static int open_input_file(const char *filename,
64 AVFormatContext **input_format_context,
65 AVCodecContext **input_codec_context)
70 /** Open the input file to read from it. */
71 if ((error = avformat_open_input(input_format_context, filename, NULL,
73 fprintf(stderr, "Could not open input file '%s' (error '%s')\n",
74 filename, get_error_text(error));
75 *input_format_context = NULL;
79 /** Get information on the input file (number of streams etc.). */
80 if ((error = avformat_find_stream_info(*input_format_context, NULL)) < 0) {
81 fprintf(stderr, "Could not open find stream info (error '%s')\n",
82 get_error_text(error));
83 avformat_close_input(input_format_context);
87 /** Make sure that there is only one stream in the input file. */
88 if ((*input_format_context)->nb_streams != 1) {
89 fprintf(stderr, "Expected one audio input stream, but found %d\n",
90 (*input_format_context)->nb_streams);
91 avformat_close_input(input_format_context);
95 /** Find a decoder for the audio stream. */
96 if (!(input_codec = avcodec_find_decoder((*input_format_context)->streams[0]->codec->codec_id))) {
97 fprintf(stderr, "Could not find input codec\n");
98 avformat_close_input(input_format_context);
102 /** Open the decoder for the audio stream to use it later. */
103 if ((error = avcodec_open2((*input_format_context)->streams[0]->codec,
104 input_codec, NULL)) < 0) {
105 fprintf(stderr, "Could not open input codec (error '%s')\n",
106 get_error_text(error));
107 avformat_close_input(input_format_context);
111 /** Save the decoder context for easier access later. */
112 *input_codec_context = (*input_format_context)->streams[0]->codec;
118 * Open an output file and the required encoder.
119 * Also set some basic encoder parameters.
120 * Some of these parameters are based on the input file's parameters.
122 static int open_output_file(const char *filename,
123 AVCodecContext *input_codec_context,
124 AVFormatContext **output_format_context,
125 AVCodecContext **output_codec_context)
127 AVIOContext *output_io_context = NULL;
128 AVStream *stream = NULL;
129 AVCodec *output_codec = NULL;
132 /** Open the output file to write to it. */
133 if ((error = avio_open(&output_io_context, filename,
134 AVIO_FLAG_WRITE)) < 0) {
135 fprintf(stderr, "Could not open output file '%s' (error '%s')\n",
136 filename, get_error_text(error));
140 /** Create a new format context for the output container format. */
141 if (!(*output_format_context = avformat_alloc_context())) {
142 fprintf(stderr, "Could not allocate output format context\n");
143 return AVERROR(ENOMEM);
146 /** Associate the output file (pointer) with the container format context. */
147 (*output_format_context)->pb = output_io_context;
149 /** Guess the desired container format based on the file extension. */
150 if (!((*output_format_context)->oformat = av_guess_format(NULL, filename,
152 fprintf(stderr, "Could not find output file format\n");
156 av_strlcpy((*output_format_context)->filename, filename,
157 sizeof((*output_format_context)->filename));
159 /** Find the encoder to be used by its name. */
160 if (!(output_codec = avcodec_find_encoder(AV_CODEC_ID_AAC))) {
161 fprintf(stderr, "Could not find an AAC encoder.\n");
165 /** Create a new audio stream in the output file container. */
166 if (!(stream = avformat_new_stream(*output_format_context, output_codec))) {
167 fprintf(stderr, "Could not create new stream\n");
168 error = AVERROR(ENOMEM);
172 /** Save the encoder context for easiert access later. */
173 *output_codec_context = stream->codec;
176 * Set the basic encoder parameters.
177 * The input file's sample rate is used to avoid a sample rate conversion.
179 (*output_codec_context)->channels = OUTPUT_CHANNELS;
180 (*output_codec_context)->channel_layout = av_get_default_channel_layout(OUTPUT_CHANNELS);
181 (*output_codec_context)->sample_rate = input_codec_context->sample_rate;
182 (*output_codec_context)->sample_fmt = AV_SAMPLE_FMT_S16;
183 (*output_codec_context)->bit_rate = OUTPUT_BIT_RATE;
186 * Some container formats (like MP4) require global headers to be present
187 * Mark the encoder so that it behaves accordingly.
189 if ((*output_format_context)->oformat->flags & AVFMT_GLOBALHEADER)
190 (*output_codec_context)->flags |= CODEC_FLAG_GLOBAL_HEADER;
192 /** Open the encoder for the audio stream to use it later. */
193 if ((error = avcodec_open2(*output_codec_context, output_codec, NULL)) < 0) {
194 fprintf(stderr, "Could not open output codec (error '%s')\n",
195 get_error_text(error));
202 avio_close((*output_format_context)->pb);
203 avformat_free_context(*output_format_context);
204 *output_format_context = NULL;
205 return error < 0 ? error : AVERROR_EXIT;
208 /** Initialize one data packet for reading or writing. */
209 static void init_packet(AVPacket *packet)
211 av_init_packet(packet);
212 /** Set the packet data and size so that it is recognized as being empty. */
217 /** Initialize one audio frame for reading from the input file */
218 static int init_input_frame(AVFrame **frame)
220 if (!(*frame = av_frame_alloc())) {
221 fprintf(stderr, "Could not allocate input frame\n");
222 return AVERROR(ENOMEM);
228 * Initialize the audio resampler based on the input and output codec settings.
229 * If the input and output sample formats differ, a conversion is required
230 * libswresample takes care of this, but requires initialization.
232 static int init_resampler(AVCodecContext *input_codec_context,
233 AVCodecContext *output_codec_context,
234 SwrContext **resample_context)
239 * Create a resampler context for the conversion.
240 * Set the conversion parameters.
241 * Default channel layouts based on the number of channels
242 * are assumed for simplicity (they are sometimes not detected
243 * properly by the demuxer and/or decoder).
245 *resample_context = swr_alloc_set_opts(NULL,
246 av_get_default_channel_layout(output_codec_context->channels),
247 output_codec_context->sample_fmt,
248 output_codec_context->sample_rate,
249 av_get_default_channel_layout(input_codec_context->channels),
250 input_codec_context->sample_fmt,
251 input_codec_context->sample_rate,
253 if (!*resample_context) {
254 fprintf(stderr, "Could not allocate resample context\n");
255 return AVERROR(ENOMEM);
258 * Perform a sanity check so that the number of converted samples is
259 * not greater than the number of samples to be converted.
260 * If the sample rates differ, this case has to be handled differently
262 av_assert0(output_codec_context->sample_rate == input_codec_context->sample_rate);
264 /** Open the resampler with the specified parameters. */
265 if ((error = swr_init(*resample_context)) < 0) {
266 fprintf(stderr, "Could not open resample context\n");
267 swr_free(resample_context);
273 /** Initialize a FIFO buffer for the audio samples to be encoded. */
274 static int init_fifo(AVAudioFifo **fifo)
276 /** Create the FIFO buffer based on the specified output sample format. */
277 if (!(*fifo = av_audio_fifo_alloc(OUTPUT_SAMPLE_FORMAT, OUTPUT_CHANNELS, 1))) {
278 fprintf(stderr, "Could not allocate FIFO\n");
279 return AVERROR(ENOMEM);
284 /** Write the header of the output file container. */
285 static int write_output_file_header(AVFormatContext *output_format_context)
288 if ((error = avformat_write_header(output_format_context, NULL)) < 0) {
289 fprintf(stderr, "Could not write output file header (error '%s')\n",
290 get_error_text(error));
296 /** Decode one audio frame from the input file. */
297 static int decode_audio_frame(AVFrame *frame,
298 AVFormatContext *input_format_context,
299 AVCodecContext *input_codec_context,
300 int *data_present, int *finished)
302 /** Packet used for temporary storage. */
303 AVPacket input_packet;
305 init_packet(&input_packet);
307 /** Read one audio frame from the input file into a temporary packet. */
308 if ((error = av_read_frame(input_format_context, &input_packet)) < 0) {
309 /** If we are the the end of the file, flush the decoder below. */
310 if (error == AVERROR_EOF)
313 fprintf(stderr, "Could not read frame (error '%s')\n",
314 get_error_text(error));
320 * Decode the audio frame stored in the temporary packet.
321 * The input audio stream decoder is used to do this.
322 * If we are at the end of the file, pass an empty packet to the decoder
325 if ((error = avcodec_decode_audio4(input_codec_context, frame,
326 data_present, &input_packet)) < 0) {
327 fprintf(stderr, "Could not decode frame (error '%s')\n",
328 get_error_text(error));
329 av_free_packet(&input_packet);
334 * If the decoder has not been flushed completely, we are not finished,
335 * so that this function has to be called again.
337 if (*finished && *data_present)
339 av_free_packet(&input_packet);
344 * Initialize a temporary storage for the specified number of audio samples.
345 * The conversion requires temporary storage due to the different format.
346 * The number of audio samples to be allocated is specified in frame_size.
348 static int init_converted_samples(uint8_t ***converted_input_samples,
349 AVCodecContext *output_codec_context,
355 * Allocate as many pointers as there are audio channels.
356 * Each pointer will later point to the audio samples of the corresponding
357 * channels (although it may be NULL for interleaved formats).
359 if (!(*converted_input_samples = calloc(output_codec_context->channels,
360 sizeof(**converted_input_samples)))) {
361 fprintf(stderr, "Could not allocate converted input sample pointers\n");
362 return AVERROR(ENOMEM);
366 * Allocate memory for the samples of all channels in one consecutive
367 * block for convenience.
369 if ((error = av_samples_alloc(*converted_input_samples, NULL,
370 output_codec_context->channels,
372 output_codec_context->sample_fmt, 0)) < 0) {
374 "Could not allocate converted input samples (error '%s')\n",
375 get_error_text(error));
376 av_freep(&(*converted_input_samples)[0]);
377 free(*converted_input_samples);
384 * Convert the input audio samples into the output sample format.
385 * The conversion happens on a per-frame basis, the size of which is specified
388 static int convert_samples(const uint8_t **input_data,
389 uint8_t **converted_data, const int frame_size,
390 SwrContext *resample_context)
394 /** Convert the samples using the resampler. */
395 if ((error = swr_convert(resample_context,
396 converted_data, frame_size,
397 input_data , frame_size)) < 0) {
398 fprintf(stderr, "Could not convert input samples (error '%s')\n",
399 get_error_text(error));
406 /** Add converted input audio samples to the FIFO buffer for later processing. */
407 static int add_samples_to_fifo(AVAudioFifo *fifo,
408 uint8_t **converted_input_samples,
409 const int frame_size)
414 * Make the FIFO as large as it needs to be to hold both,
415 * the old and the new samples.
417 if ((error = av_audio_fifo_realloc(fifo, av_audio_fifo_size(fifo) + frame_size)) < 0) {
418 fprintf(stderr, "Could not reallocate FIFO\n");
422 /** Store the new samples in the FIFO buffer. */
423 if (av_audio_fifo_write(fifo, (void **)converted_input_samples,
424 frame_size) < frame_size) {
425 fprintf(stderr, "Could not write data to FIFO\n");
432 * Read one audio frame from the input file, decodes, converts and stores
433 * it in the FIFO buffer.
435 static int read_decode_convert_and_store(AVAudioFifo *fifo,
436 AVFormatContext *input_format_context,
437 AVCodecContext *input_codec_context,
438 AVCodecContext *output_codec_context,
439 SwrContext *resampler_context,
442 /** Temporary storage of the input samples of the frame read from the file. */
443 AVFrame *input_frame = NULL;
444 /** Temporary storage for the converted input samples. */
445 uint8_t **converted_input_samples = NULL;
447 int ret = AVERROR_EXIT;
449 /** Initialize temporary storage for one input frame. */
450 if (init_input_frame(&input_frame))
452 /** Decode one frame worth of audio samples. */
453 if (decode_audio_frame(input_frame, input_format_context,
454 input_codec_context, &data_present, finished))
457 * If we are at the end of the file and there are no more samples
458 * in the decoder which are delayed, we are actually finished.
459 * This must not be treated as an error.
461 if (*finished && !data_present) {
465 /** If there is decoded data, convert and store it */
467 /** Initialize the temporary storage for the converted input samples. */
468 if (init_converted_samples(&converted_input_samples, output_codec_context,
469 input_frame->nb_samples))
473 * Convert the input samples to the desired output sample format.
474 * This requires a temporary storage provided by converted_input_samples.
476 if (convert_samples((const uint8_t**)input_frame->extended_data, converted_input_samples,
477 input_frame->nb_samples, resampler_context))
480 /** Add the converted input samples to the FIFO buffer for later processing. */
481 if (add_samples_to_fifo(fifo, converted_input_samples,
482 input_frame->nb_samples))
489 if (converted_input_samples) {
490 av_freep(&converted_input_samples[0]);
491 free(converted_input_samples);
493 av_frame_free(&input_frame);
499 * Initialize one input frame for writing to the output file.
500 * The frame will be exactly frame_size samples large.
502 static int init_output_frame(AVFrame **frame,
503 AVCodecContext *output_codec_context,
508 /** Create a new frame to store the audio samples. */
509 if (!(*frame = av_frame_alloc())) {
510 fprintf(stderr, "Could not allocate output frame\n");
515 * Set the frame's parameters, especially its size and format.
516 * av_frame_get_buffer needs this to allocate memory for the
517 * audio samples of the frame.
518 * Default channel layouts based on the number of channels
519 * are assumed for simplicity.
521 (*frame)->nb_samples = frame_size;
522 (*frame)->channel_layout = output_codec_context->channel_layout;
523 (*frame)->format = output_codec_context->sample_fmt;
524 (*frame)->sample_rate = output_codec_context->sample_rate;
527 * Allocate the samples of the created frame. This call will make
528 * sure that the audio frame can hold as many samples as specified.
530 if ((error = av_frame_get_buffer(*frame, 0)) < 0) {
531 fprintf(stderr, "Could allocate output frame samples (error '%s')\n",
532 get_error_text(error));
533 av_frame_free(frame);
540 /** Encode one frame worth of audio to the output file. */
541 static int encode_audio_frame(AVFrame *frame,
542 AVFormatContext *output_format_context,
543 AVCodecContext *output_codec_context,
546 /** Packet used for temporary storage. */
547 AVPacket output_packet;
549 init_packet(&output_packet);
552 * Encode the audio frame and store it in the temporary packet.
553 * The output audio stream encoder is used to do this.
555 if ((error = avcodec_encode_audio2(output_codec_context, &output_packet,
556 frame, data_present)) < 0) {
557 fprintf(stderr, "Could not encode frame (error '%s')\n",
558 get_error_text(error));
559 av_free_packet(&output_packet);
563 /** Write one audio frame from the temporary packet to the output file. */
565 if ((error = av_write_frame(output_format_context, &output_packet)) < 0) {
566 fprintf(stderr, "Could not write frame (error '%s')\n",
567 get_error_text(error));
568 av_free_packet(&output_packet);
572 av_free_packet(&output_packet);
579 * Load one audio frame from the FIFO buffer, encode and write it to the
582 static int load_encode_and_write(AVAudioFifo *fifo,
583 AVFormatContext *output_format_context,
584 AVCodecContext *output_codec_context)
586 /** Temporary storage of the output samples of the frame written to the file. */
587 AVFrame *output_frame;
589 * Use the maximum number of possible samples per frame.
590 * If there is less than the maximum possible frame size in the FIFO
591 * buffer use this number. Otherwise, use the maximum possible frame size
593 const int frame_size = FFMIN(av_audio_fifo_size(fifo),
594 output_codec_context->frame_size);
597 /** Initialize temporary storage for one output frame. */
598 if (init_output_frame(&output_frame, output_codec_context, frame_size))
602 * Read as many samples from the FIFO buffer as required to fill the frame.
603 * The samples are stored in the frame temporarily.
605 if (av_audio_fifo_read(fifo, (void **)output_frame->data, frame_size) < frame_size) {
606 fprintf(stderr, "Could not read data from FIFO\n");
607 av_frame_free(&output_frame);
611 /** Encode one frame worth of audio samples. */
612 if (encode_audio_frame(output_frame, output_format_context,
613 output_codec_context, &data_written)) {
614 av_frame_free(&output_frame);
617 av_frame_free(&output_frame);
621 /** Write the trailer of the output file container. */
622 static int write_output_file_trailer(AVFormatContext *output_format_context)
625 if ((error = av_write_trailer(output_format_context)) < 0) {
626 fprintf(stderr, "Could not write output file trailer (error '%s')\n",
627 get_error_text(error));
633 /** Convert an audio file to an AAC file in an MP4 container. */
634 int main(int argc, char **argv)
636 AVFormatContext *input_format_context = NULL, *output_format_context = NULL;
637 AVCodecContext *input_codec_context = NULL, *output_codec_context = NULL;
638 SwrContext *resample_context = NULL;
639 AVAudioFifo *fifo = NULL;
640 int ret = AVERROR_EXIT;
643 fprintf(stderr, "Usage: %s <input file> <output file>\n", argv[0]);
647 /** Register all codecs and formats so that they can be used. */
649 /** Open the input file for reading. */
650 if (open_input_file(argv[1], &input_format_context,
651 &input_codec_context))
653 /** Open the output file for writing. */
654 if (open_output_file(argv[2], input_codec_context,
655 &output_format_context, &output_codec_context))
657 /** Initialize the resampler to be able to convert audio sample formats. */
658 if (init_resampler(input_codec_context, output_codec_context,
661 /** Initialize the FIFO buffer to store audio samples to be encoded. */
662 if (init_fifo(&fifo))
664 /** Write the header of the output file container. */
665 if (write_output_file_header(output_format_context))
669 * Loop as long as we have input samples to read or output samples
670 * to write; abort as soon as we have neither.
673 /** Use the encoder's desired frame size for processing. */
674 const int output_frame_size = output_codec_context->frame_size;
678 * Make sure that there is one frame worth of samples in the FIFO
679 * buffer so that the encoder can do its work.
680 * Since the decoder's and the encoder's frame size may differ, we
681 * need to FIFO buffer to store as many frames worth of input samples
682 * that they make up at least one frame worth of output samples.
684 while (av_audio_fifo_size(fifo) < output_frame_size) {
686 * Decode one frame worth of audio samples, convert it to the
687 * output sample format and put it into the FIFO buffer.
689 if (read_decode_convert_and_store(fifo, input_format_context,
691 output_codec_context,
692 resample_context, &finished))
696 * If we are at the end of the input file, we continue
697 * encoding the remaining audio samples to the output file.
704 * If we have enough samples for the encoder, we encode them.
705 * At the end of the file, we pass the remaining samples to
708 while (av_audio_fifo_size(fifo) >= output_frame_size ||
709 (finished && av_audio_fifo_size(fifo) > 0))
711 * Take one frame worth of audio samples from the FIFO buffer,
712 * encode it and write it to the output file.
714 if (load_encode_and_write(fifo, output_format_context,
715 output_codec_context))
719 * If we are at the end of the input file and have encoded
720 * all remaining samples, we can exit this loop and finish.
724 /** Flush the encoder as it may have delayed frames. */
726 if (encode_audio_frame(NULL, output_format_context,
727 output_codec_context, &data_written))
729 } while (data_written);
734 /** Write the trailer of the output file container. */
735 if (write_output_file_trailer(output_format_context))
741 av_audio_fifo_free(fifo);
742 swr_free(&resample_context);
743 if (output_codec_context)
744 avcodec_close(output_codec_context);
745 if (output_format_context) {
746 avio_close(output_format_context->pb);
747 avformat_free_context(output_format_context);
749 if (input_codec_context)
750 avcodec_close(input_codec_context);
751 if (input_format_context)
752 avformat_close_input(&input_format_context);