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 96000
45 /** The number of output channels */
46 #define OUTPUT_CHANNELS 2
49 * Convert an error code into a text message.
50 * @param error Error code to be converted
51 * @return Corresponding error text (not thread-safe)
53 static const char *get_error_text(const int error)
55 static char error_buffer[255];
56 av_strerror(error, error_buffer, sizeof(error_buffer));
60 /** Open an input file and the required decoder. */
61 static int open_input_file(const char *filename,
62 AVFormatContext **input_format_context,
63 AVCodecContext **input_codec_context)
65 AVCodecContext *avctx;
69 /** Open the input file to read from it. */
70 if ((error = avformat_open_input(input_format_context, filename, NULL,
72 fprintf(stderr, "Could not open input file '%s' (error '%s')\n",
73 filename, get_error_text(error));
74 *input_format_context = NULL;
78 /** Get information on the input file (number of streams etc.). */
79 if ((error = avformat_find_stream_info(*input_format_context, NULL)) < 0) {
80 fprintf(stderr, "Could not open find stream info (error '%s')\n",
81 get_error_text(error));
82 avformat_close_input(input_format_context);
86 /** Make sure that there is only one stream in the input file. */
87 if ((*input_format_context)->nb_streams != 1) {
88 fprintf(stderr, "Expected one audio input stream, but found %d\n",
89 (*input_format_context)->nb_streams);
90 avformat_close_input(input_format_context);
94 /** Find a decoder for the audio stream. */
95 if (!(input_codec = avcodec_find_decoder((*input_format_context)->streams[0]->codecpar->codec_id))) {
96 fprintf(stderr, "Could not find input codec\n");
97 avformat_close_input(input_format_context);
101 /** allocate a new decoding context */
102 avctx = avcodec_alloc_context3(input_codec);
104 fprintf(stderr, "Could not allocate a decoding context\n");
105 avformat_close_input(input_format_context);
106 return AVERROR(ENOMEM);
109 /** initialize the stream parameters with demuxer information */
110 error = avcodec_parameters_to_context(avctx, (*input_format_context)->streams[0]->codecpar);
112 avformat_close_input(input_format_context);
113 avcodec_free_context(&avctx);
117 /** Open the decoder for the audio stream to use it later. */
118 if ((error = avcodec_open2(avctx, input_codec, NULL)) < 0) {
119 fprintf(stderr, "Could not open input codec (error '%s')\n",
120 get_error_text(error));
121 avcodec_free_context(&avctx);
122 avformat_close_input(input_format_context);
126 /** Save the decoder context for easier access later. */
127 *input_codec_context = avctx;
133 * Open an output file and the required encoder.
134 * Also set some basic encoder parameters.
135 * Some of these parameters are based on the input file's parameters.
137 static int open_output_file(const char *filename,
138 AVCodecContext *input_codec_context,
139 AVFormatContext **output_format_context,
140 AVCodecContext **output_codec_context)
142 AVCodecContext *avctx = NULL;
143 AVIOContext *output_io_context = NULL;
144 AVStream *stream = NULL;
145 AVCodec *output_codec = NULL;
148 /** Open the output file to write to it. */
149 if ((error = avio_open(&output_io_context, filename,
150 AVIO_FLAG_WRITE)) < 0) {
151 fprintf(stderr, "Could not open output file '%s' (error '%s')\n",
152 filename, get_error_text(error));
156 /** Create a new format context for the output container format. */
157 if (!(*output_format_context = avformat_alloc_context())) {
158 fprintf(stderr, "Could not allocate output format context\n");
159 return AVERROR(ENOMEM);
162 /** Associate the output file (pointer) with the container format context. */
163 (*output_format_context)->pb = output_io_context;
165 /** Guess the desired container format based on the file extension. */
166 if (!((*output_format_context)->oformat = av_guess_format(NULL, filename,
168 fprintf(stderr, "Could not find output file format\n");
172 av_strlcpy((*output_format_context)->filename, filename,
173 sizeof((*output_format_context)->filename));
175 /** Find the encoder to be used by its name. */
176 if (!(output_codec = avcodec_find_encoder(AV_CODEC_ID_AAC))) {
177 fprintf(stderr, "Could not find an AAC encoder.\n");
181 /** Create a new audio stream in the output file container. */
182 if (!(stream = avformat_new_stream(*output_format_context, NULL))) {
183 fprintf(stderr, "Could not create new stream\n");
184 error = AVERROR(ENOMEM);
188 avctx = avcodec_alloc_context3(output_codec);
190 fprintf(stderr, "Could not allocate an encoding context\n");
191 error = AVERROR(ENOMEM);
196 * Set the basic encoder parameters.
197 * The input file's sample rate is used to avoid a sample rate conversion.
199 avctx->channels = OUTPUT_CHANNELS;
200 avctx->channel_layout = av_get_default_channel_layout(OUTPUT_CHANNELS);
201 avctx->sample_rate = input_codec_context->sample_rate;
202 avctx->sample_fmt = output_codec->sample_fmts[0];
203 avctx->bit_rate = OUTPUT_BIT_RATE;
205 /** Allow the use of the experimental AAC encoder */
206 avctx->strict_std_compliance = FF_COMPLIANCE_EXPERIMENTAL;
208 /** Set the sample rate for the container. */
209 stream->time_base.den = input_codec_context->sample_rate;
210 stream->time_base.num = 1;
213 * Some container formats (like MP4) require global headers to be present
214 * Mark the encoder so that it behaves accordingly.
216 if ((*output_format_context)->oformat->flags & AVFMT_GLOBALHEADER)
217 avctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
219 /** Open the encoder for the audio stream to use it later. */
220 if ((error = avcodec_open2(avctx, output_codec, NULL)) < 0) {
221 fprintf(stderr, "Could not open output codec (error '%s')\n",
222 get_error_text(error));
226 error = avcodec_parameters_from_context(stream->codecpar, avctx);
228 fprintf(stderr, "Could not initialize stream parameters\n");
232 /** Save the encoder context for easier access later. */
233 *output_codec_context = avctx;
238 avcodec_free_context(&avctx);
239 avio_closep(&(*output_format_context)->pb);
240 avformat_free_context(*output_format_context);
241 *output_format_context = NULL;
242 return error < 0 ? error : AVERROR_EXIT;
245 /** Initialize one data packet for reading or writing. */
246 static void init_packet(AVPacket *packet)
248 av_init_packet(packet);
249 /** Set the packet data and size so that it is recognized as being empty. */
254 /** Initialize one audio frame for reading from the input file */
255 static int init_input_frame(AVFrame **frame)
257 if (!(*frame = av_frame_alloc())) {
258 fprintf(stderr, "Could not allocate input frame\n");
259 return AVERROR(ENOMEM);
265 * Initialize the audio resampler based on the input and output codec settings.
266 * If the input and output sample formats differ, a conversion is required
267 * libswresample takes care of this, but requires initialization.
269 static int init_resampler(AVCodecContext *input_codec_context,
270 AVCodecContext *output_codec_context,
271 SwrContext **resample_context)
276 * Create a resampler context for the conversion.
277 * Set the conversion parameters.
278 * Default channel layouts based on the number of channels
279 * are assumed for simplicity (they are sometimes not detected
280 * properly by the demuxer and/or decoder).
282 *resample_context = swr_alloc_set_opts(NULL,
283 av_get_default_channel_layout(output_codec_context->channels),
284 output_codec_context->sample_fmt,
285 output_codec_context->sample_rate,
286 av_get_default_channel_layout(input_codec_context->channels),
287 input_codec_context->sample_fmt,
288 input_codec_context->sample_rate,
290 if (!*resample_context) {
291 fprintf(stderr, "Could not allocate resample context\n");
292 return AVERROR(ENOMEM);
295 * Perform a sanity check so that the number of converted samples is
296 * not greater than the number of samples to be converted.
297 * If the sample rates differ, this case has to be handled differently
299 av_assert0(output_codec_context->sample_rate == input_codec_context->sample_rate);
301 /** Open the resampler with the specified parameters. */
302 if ((error = swr_init(*resample_context)) < 0) {
303 fprintf(stderr, "Could not open resample context\n");
304 swr_free(resample_context);
310 /** Initialize a FIFO buffer for the audio samples to be encoded. */
311 static int init_fifo(AVAudioFifo **fifo, AVCodecContext *output_codec_context)
313 /** Create the FIFO buffer based on the specified output sample format. */
314 if (!(*fifo = av_audio_fifo_alloc(output_codec_context->sample_fmt,
315 output_codec_context->channels, 1))) {
316 fprintf(stderr, "Could not allocate FIFO\n");
317 return AVERROR(ENOMEM);
322 /** Write the header of the output file container. */
323 static int write_output_file_header(AVFormatContext *output_format_context)
326 if ((error = avformat_write_header(output_format_context, NULL)) < 0) {
327 fprintf(stderr, "Could not write output file header (error '%s')\n",
328 get_error_text(error));
334 /** Decode one audio frame from the input file. */
335 static int decode_audio_frame(AVFrame *frame,
336 AVFormatContext *input_format_context,
337 AVCodecContext *input_codec_context,
338 int *data_present, int *finished)
340 /** Packet used for temporary storage. */
341 AVPacket input_packet;
343 init_packet(&input_packet);
345 /** Read one audio frame from the input file into a temporary packet. */
346 if ((error = av_read_frame(input_format_context, &input_packet)) < 0) {
347 /** If we are at the end of the file, flush the decoder below. */
348 if (error == AVERROR_EOF)
351 fprintf(stderr, "Could not read frame (error '%s')\n",
352 get_error_text(error));
358 * Decode the audio frame stored in the temporary packet.
359 * The input audio stream decoder is used to do this.
360 * If we are at the end of the file, pass an empty packet to the decoder
363 if ((error = avcodec_decode_audio4(input_codec_context, frame,
364 data_present, &input_packet)) < 0) {
365 fprintf(stderr, "Could not decode frame (error '%s')\n",
366 get_error_text(error));
367 av_packet_unref(&input_packet);
372 * If the decoder has not been flushed completely, we are not finished,
373 * so that this function has to be called again.
375 if (*finished && *data_present)
377 av_packet_unref(&input_packet);
382 * Initialize a temporary storage for the specified number of audio samples.
383 * The conversion requires temporary storage due to the different format.
384 * The number of audio samples to be allocated is specified in frame_size.
386 static int init_converted_samples(uint8_t ***converted_input_samples,
387 AVCodecContext *output_codec_context,
393 * Allocate as many pointers as there are audio channels.
394 * Each pointer will later point to the audio samples of the corresponding
395 * channels (although it may be NULL for interleaved formats).
397 if (!(*converted_input_samples = calloc(output_codec_context->channels,
398 sizeof(**converted_input_samples)))) {
399 fprintf(stderr, "Could not allocate converted input sample pointers\n");
400 return AVERROR(ENOMEM);
404 * Allocate memory for the samples of all channels in one consecutive
405 * block for convenience.
407 if ((error = av_samples_alloc(*converted_input_samples, NULL,
408 output_codec_context->channels,
410 output_codec_context->sample_fmt, 0)) < 0) {
412 "Could not allocate converted input samples (error '%s')\n",
413 get_error_text(error));
414 av_freep(&(*converted_input_samples)[0]);
415 free(*converted_input_samples);
422 * Convert the input audio samples into the output sample format.
423 * The conversion happens on a per-frame basis, the size of which is specified
426 static int convert_samples(const uint8_t **input_data,
427 uint8_t **converted_data, const int frame_size,
428 SwrContext *resample_context)
432 /** Convert the samples using the resampler. */
433 if ((error = swr_convert(resample_context,
434 converted_data, frame_size,
435 input_data , frame_size)) < 0) {
436 fprintf(stderr, "Could not convert input samples (error '%s')\n",
437 get_error_text(error));
444 /** Add converted input audio samples to the FIFO buffer for later processing. */
445 static int add_samples_to_fifo(AVAudioFifo *fifo,
446 uint8_t **converted_input_samples,
447 const int frame_size)
452 * Make the FIFO as large as it needs to be to hold both,
453 * the old and the new samples.
455 if ((error = av_audio_fifo_realloc(fifo, av_audio_fifo_size(fifo) + frame_size)) < 0) {
456 fprintf(stderr, "Could not reallocate FIFO\n");
460 /** Store the new samples in the FIFO buffer. */
461 if (av_audio_fifo_write(fifo, (void **)converted_input_samples,
462 frame_size) < frame_size) {
463 fprintf(stderr, "Could not write data to FIFO\n");
470 * Read one audio frame from the input file, decodes, converts and stores
471 * it in the FIFO buffer.
473 static int read_decode_convert_and_store(AVAudioFifo *fifo,
474 AVFormatContext *input_format_context,
475 AVCodecContext *input_codec_context,
476 AVCodecContext *output_codec_context,
477 SwrContext *resampler_context,
480 /** Temporary storage of the input samples of the frame read from the file. */
481 AVFrame *input_frame = NULL;
482 /** Temporary storage for the converted input samples. */
483 uint8_t **converted_input_samples = NULL;
485 int ret = AVERROR_EXIT;
487 /** Initialize temporary storage for one input frame. */
488 if (init_input_frame(&input_frame))
490 /** Decode one frame worth of audio samples. */
491 if (decode_audio_frame(input_frame, input_format_context,
492 input_codec_context, &data_present, finished))
495 * If we are at the end of the file and there are no more samples
496 * in the decoder which are delayed, we are actually finished.
497 * This must not be treated as an error.
499 if (*finished && !data_present) {
503 /** If there is decoded data, convert and store it */
505 /** Initialize the temporary storage for the converted input samples. */
506 if (init_converted_samples(&converted_input_samples, output_codec_context,
507 input_frame->nb_samples))
511 * Convert the input samples to the desired output sample format.
512 * This requires a temporary storage provided by converted_input_samples.
514 if (convert_samples((const uint8_t**)input_frame->extended_data, converted_input_samples,
515 input_frame->nb_samples, resampler_context))
518 /** Add the converted input samples to the FIFO buffer for later processing. */
519 if (add_samples_to_fifo(fifo, converted_input_samples,
520 input_frame->nb_samples))
527 if (converted_input_samples) {
528 av_freep(&converted_input_samples[0]);
529 free(converted_input_samples);
531 av_frame_free(&input_frame);
537 * Initialize one input frame for writing to the output file.
538 * The frame will be exactly frame_size samples large.
540 static int init_output_frame(AVFrame **frame,
541 AVCodecContext *output_codec_context,
546 /** Create a new frame to store the audio samples. */
547 if (!(*frame = av_frame_alloc())) {
548 fprintf(stderr, "Could not allocate output frame\n");
553 * Set the frame's parameters, especially its size and format.
554 * av_frame_get_buffer needs this to allocate memory for the
555 * audio samples of the frame.
556 * Default channel layouts based on the number of channels
557 * are assumed for simplicity.
559 (*frame)->nb_samples = frame_size;
560 (*frame)->channel_layout = output_codec_context->channel_layout;
561 (*frame)->format = output_codec_context->sample_fmt;
562 (*frame)->sample_rate = output_codec_context->sample_rate;
565 * Allocate the samples of the created frame. This call will make
566 * sure that the audio frame can hold as many samples as specified.
568 if ((error = av_frame_get_buffer(*frame, 0)) < 0) {
569 fprintf(stderr, "Could not allocate output frame samples (error '%s')\n",
570 get_error_text(error));
571 av_frame_free(frame);
578 /** Global timestamp for the audio frames */
579 static int64_t pts = 0;
581 /** Encode one frame worth of audio to the output file. */
582 static int encode_audio_frame(AVFrame *frame,
583 AVFormatContext *output_format_context,
584 AVCodecContext *output_codec_context,
587 /** Packet used for temporary storage. */
588 AVPacket output_packet;
590 init_packet(&output_packet);
592 /** Set a timestamp based on the sample rate for the container. */
595 pts += frame->nb_samples;
599 * Encode the audio frame and store it in the temporary packet.
600 * The output audio stream encoder is used to do this.
602 if ((error = avcodec_encode_audio2(output_codec_context, &output_packet,
603 frame, data_present)) < 0) {
604 fprintf(stderr, "Could not encode frame (error '%s')\n",
605 get_error_text(error));
606 av_packet_unref(&output_packet);
610 /** Write one audio frame from the temporary packet to the output file. */
612 if ((error = av_write_frame(output_format_context, &output_packet)) < 0) {
613 fprintf(stderr, "Could not write frame (error '%s')\n",
614 get_error_text(error));
615 av_packet_unref(&output_packet);
619 av_packet_unref(&output_packet);
626 * Load one audio frame from the FIFO buffer, encode and write it to the
629 static int load_encode_and_write(AVAudioFifo *fifo,
630 AVFormatContext *output_format_context,
631 AVCodecContext *output_codec_context)
633 /** Temporary storage of the output samples of the frame written to the file. */
634 AVFrame *output_frame;
636 * Use the maximum number of possible samples per frame.
637 * If there is less than the maximum possible frame size in the FIFO
638 * buffer use this number. Otherwise, use the maximum possible frame size
640 const int frame_size = FFMIN(av_audio_fifo_size(fifo),
641 output_codec_context->frame_size);
644 /** Initialize temporary storage for one output frame. */
645 if (init_output_frame(&output_frame, output_codec_context, frame_size))
649 * Read as many samples from the FIFO buffer as required to fill the frame.
650 * The samples are stored in the frame temporarily.
652 if (av_audio_fifo_read(fifo, (void **)output_frame->data, frame_size) < frame_size) {
653 fprintf(stderr, "Could not read data from FIFO\n");
654 av_frame_free(&output_frame);
658 /** Encode one frame worth of audio samples. */
659 if (encode_audio_frame(output_frame, output_format_context,
660 output_codec_context, &data_written)) {
661 av_frame_free(&output_frame);
664 av_frame_free(&output_frame);
668 /** Write the trailer of the output file container. */
669 static int write_output_file_trailer(AVFormatContext *output_format_context)
672 if ((error = av_write_trailer(output_format_context)) < 0) {
673 fprintf(stderr, "Could not write output file trailer (error '%s')\n",
674 get_error_text(error));
680 /** Convert an audio file to an AAC file in an MP4 container. */
681 int main(int argc, char **argv)
683 AVFormatContext *input_format_context = NULL, *output_format_context = NULL;
684 AVCodecContext *input_codec_context = NULL, *output_codec_context = NULL;
685 SwrContext *resample_context = NULL;
686 AVAudioFifo *fifo = NULL;
687 int ret = AVERROR_EXIT;
690 fprintf(stderr, "Usage: %s <input file> <output file>\n", argv[0]);
694 /** Register all codecs and formats so that they can be used. */
696 /** Open the input file for reading. */
697 if (open_input_file(argv[1], &input_format_context,
698 &input_codec_context))
700 /** Open the output file for writing. */
701 if (open_output_file(argv[2], input_codec_context,
702 &output_format_context, &output_codec_context))
704 /** Initialize the resampler to be able to convert audio sample formats. */
705 if (init_resampler(input_codec_context, output_codec_context,
708 /** Initialize the FIFO buffer to store audio samples to be encoded. */
709 if (init_fifo(&fifo, output_codec_context))
711 /** Write the header of the output file container. */
712 if (write_output_file_header(output_format_context))
716 * Loop as long as we have input samples to read or output samples
717 * to write; abort as soon as we have neither.
720 /** Use the encoder's desired frame size for processing. */
721 const int output_frame_size = output_codec_context->frame_size;
725 * Make sure that there is one frame worth of samples in the FIFO
726 * buffer so that the encoder can do its work.
727 * Since the decoder's and the encoder's frame size may differ, we
728 * need to FIFO buffer to store as many frames worth of input samples
729 * that they make up at least one frame worth of output samples.
731 while (av_audio_fifo_size(fifo) < output_frame_size) {
733 * Decode one frame worth of audio samples, convert it to the
734 * output sample format and put it into the FIFO buffer.
736 if (read_decode_convert_and_store(fifo, input_format_context,
738 output_codec_context,
739 resample_context, &finished))
743 * If we are at the end of the input file, we continue
744 * encoding the remaining audio samples to the output file.
751 * If we have enough samples for the encoder, we encode them.
752 * At the end of the file, we pass the remaining samples to
755 while (av_audio_fifo_size(fifo) >= output_frame_size ||
756 (finished && av_audio_fifo_size(fifo) > 0))
758 * Take one frame worth of audio samples from the FIFO buffer,
759 * encode it and write it to the output file.
761 if (load_encode_and_write(fifo, output_format_context,
762 output_codec_context))
766 * If we are at the end of the input file and have encoded
767 * all remaining samples, we can exit this loop and finish.
771 /** Flush the encoder as it may have delayed frames. */
773 if (encode_audio_frame(NULL, output_format_context,
774 output_codec_context, &data_written))
776 } while (data_written);
781 /** Write the trailer of the output file container. */
782 if (write_output_file_trailer(output_format_context))
788 av_audio_fifo_free(fifo);
789 swr_free(&resample_context);
790 if (output_codec_context)
791 avcodec_free_context(&output_codec_context);
792 if (output_format_context) {
793 avio_closep(&output_format_context->pb);
794 avformat_free_context(output_format_context);
796 if (input_codec_context)
797 avcodec_free_context(&input_codec_context);
798 if (input_format_context)
799 avformat_close_input(&input_format_context);