2 * This file is part of Libav.
4 * Libav 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 * Libav 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 Libav; 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 Libav.
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/avstring.h"
37 #include "libavutil/frame.h"
38 #include "libavutil/opt.h"
40 #include "libavresample/avresample.h"
42 /** The output bit rate in kbit/s */
43 #define OUTPUT_BIT_RATE 96000
44 /** The number of output channels */
45 #define OUTPUT_CHANNELS 2
48 * Convert an error code into a text message.
49 * @param error Error code to be converted
50 * @return Corresponding error text (not thread-safe)
52 static char *const get_error_text(const int error)
54 static char error_buffer[255];
55 av_strerror(error, error_buffer, sizeof(error_buffer));
59 /** Open an input file and the required decoder. */
60 static int open_input_file(const char *filename,
61 AVFormatContext **input_format_context,
62 AVCodecContext **input_codec_context)
67 /** Open the input file to read from it. */
68 if ((error = avformat_open_input(input_format_context, filename, NULL,
70 fprintf(stderr, "Could not open input file '%s' (error '%s')\n",
71 filename, get_error_text(error));
72 *input_format_context = NULL;
76 /** Get information on the input file (number of streams etc.). */
77 if ((error = avformat_find_stream_info(*input_format_context, NULL)) < 0) {
78 fprintf(stderr, "Could not open find stream info (error '%s')\n",
79 get_error_text(error));
80 avformat_close_input(input_format_context);
84 /** Make sure that there is only one stream in the input file. */
85 if ((*input_format_context)->nb_streams != 1) {
86 fprintf(stderr, "Expected one audio input stream, but found %d\n",
87 (*input_format_context)->nb_streams);
88 avformat_close_input(input_format_context);
92 /** Find a decoder for the audio stream. */
93 if (!(input_codec = avcodec_find_decoder((*input_format_context)->streams[0]->codec->codec_id))) {
94 fprintf(stderr, "Could not find input codec\n");
95 avformat_close_input(input_format_context);
99 /** Open the decoder for the audio stream to use it later. */
100 if ((error = avcodec_open2((*input_format_context)->streams[0]->codec,
101 input_codec, NULL)) < 0) {
102 fprintf(stderr, "Could not open input codec (error '%s')\n",
103 get_error_text(error));
104 avformat_close_input(input_format_context);
108 /** Save the decoder context for easier access later. */
109 *input_codec_context = (*input_format_context)->streams[0]->codec;
115 * Open an output file and the required encoder.
116 * Also set some basic encoder parameters.
117 * Some of these parameters are based on the input file's parameters.
119 static int open_output_file(const char *filename,
120 AVCodecContext *input_codec_context,
121 AVFormatContext **output_format_context,
122 AVCodecContext **output_codec_context)
124 AVIOContext *output_io_context = NULL;
125 AVStream *stream = NULL;
126 AVCodec *output_codec = NULL;
129 /** Open the output file to write to it. */
130 if ((error = avio_open(&output_io_context, filename,
131 AVIO_FLAG_WRITE)) < 0) {
132 fprintf(stderr, "Could not open output file '%s' (error '%s')\n",
133 filename, get_error_text(error));
137 /** Create a new format context for the output container format. */
138 if (!(*output_format_context = avformat_alloc_context())) {
139 fprintf(stderr, "Could not allocate output format context\n");
140 return AVERROR(ENOMEM);
143 /** Associate the output file (pointer) with the container format context. */
144 (*output_format_context)->pb = output_io_context;
146 /** Guess the desired container format based on the file extension. */
147 if (!((*output_format_context)->oformat = av_guess_format(NULL, filename,
149 fprintf(stderr, "Could not find output file format\n");
153 av_strlcpy((*output_format_context)->filename, filename,
154 sizeof((*output_format_context)->filename));
156 /** Find the encoder to be used by its name. */
157 if (!(output_codec = avcodec_find_encoder(AV_CODEC_ID_AAC))) {
158 fprintf(stderr, "Could not find an AAC encoder.\n");
162 /** Create a new audio stream in the output file container. */
163 if (!(stream = avformat_new_stream(*output_format_context, output_codec))) {
164 fprintf(stderr, "Could not create new stream\n");
165 error = AVERROR(ENOMEM);
169 /** Save the encoder context for easier access later. */
170 *output_codec_context = stream->codec;
173 * Set the basic encoder parameters.
174 * The input file's sample rate is used to avoid a sample rate conversion.
176 (*output_codec_context)->channels = OUTPUT_CHANNELS;
177 (*output_codec_context)->channel_layout = av_get_default_channel_layout(OUTPUT_CHANNELS);
178 (*output_codec_context)->sample_rate = input_codec_context->sample_rate;
179 (*output_codec_context)->sample_fmt = output_codec->sample_fmts[0];
180 (*output_codec_context)->bit_rate = OUTPUT_BIT_RATE;
182 /** Allow the use of the experimental AAC encoder */
183 (*output_codec_context)->strict_std_compliance = FF_COMPLIANCE_EXPERIMENTAL;
185 /** Set the sample rate for the container. */
186 stream->time_base.den = input_codec_context->sample_rate;
187 stream->time_base.num = 1;
190 * Some container formats (like MP4) require global headers to be present
191 * Mark the encoder so that it behaves accordingly.
193 if ((*output_format_context)->oformat->flags & AVFMT_GLOBALHEADER)
194 (*output_codec_context)->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
196 /** Open the encoder for the audio stream to use it later. */
197 if ((error = avcodec_open2(*output_codec_context, output_codec, NULL)) < 0) {
198 fprintf(stderr, "Could not open output codec (error '%s')\n",
199 get_error_text(error));
206 avio_close((*output_format_context)->pb);
207 avformat_free_context(*output_format_context);
208 *output_format_context = NULL;
209 return error < 0 ? error : AVERROR_EXIT;
212 /** Initialize one data packet for reading or writing. */
213 static void init_packet(AVPacket *packet)
215 av_init_packet(packet);
216 /** Set the packet data and size so that it is recognized as being empty. */
221 /** Initialize one audio frame for reading from the input file */
222 static int init_input_frame(AVFrame **frame)
224 if (!(*frame = av_frame_alloc())) {
225 fprintf(stderr, "Could not allocate input frame\n");
226 return AVERROR(ENOMEM);
232 * Initialize the audio resampler based on the input and output codec settings.
233 * If the input and output sample formats differ, a conversion is required
234 * libavresample takes care of this, but requires initialization.
236 static int init_resampler(AVCodecContext *input_codec_context,
237 AVCodecContext *output_codec_context,
238 AVAudioResampleContext **resample_context)
241 * Only initialize the resampler if it is necessary, i.e.,
242 * if and only if the sample formats differ.
244 if (input_codec_context->sample_fmt != output_codec_context->sample_fmt ||
245 input_codec_context->channels != output_codec_context->channels) {
248 /** Create a resampler context for the conversion. */
249 if (!(*resample_context = avresample_alloc_context())) {
250 fprintf(stderr, "Could not allocate resample context\n");
251 return AVERROR(ENOMEM);
255 * Set the conversion parameters.
256 * Default channel layouts based on the number of channels
257 * are assumed for simplicity (they are sometimes not detected
258 * properly by the demuxer and/or decoder).
260 av_opt_set_int(*resample_context, "in_channel_layout",
261 av_get_default_channel_layout(input_codec_context->channels), 0);
262 av_opt_set_int(*resample_context, "out_channel_layout",
263 av_get_default_channel_layout(output_codec_context->channels), 0);
264 av_opt_set_int(*resample_context, "in_sample_rate",
265 input_codec_context->sample_rate, 0);
266 av_opt_set_int(*resample_context, "out_sample_rate",
267 output_codec_context->sample_rate, 0);
268 av_opt_set_int(*resample_context, "in_sample_fmt",
269 input_codec_context->sample_fmt, 0);
270 av_opt_set_int(*resample_context, "out_sample_fmt",
271 output_codec_context->sample_fmt, 0);
273 /** Open the resampler with the specified parameters. */
274 if ((error = avresample_open(*resample_context)) < 0) {
275 fprintf(stderr, "Could not open resample context\n");
276 avresample_free(resample_context);
283 /** Initialize a FIFO buffer for the audio samples to be encoded. */
284 static int init_fifo(AVAudioFifo **fifo, AVCodecContext *output_codec_context)
286 /** Create the FIFO buffer based on the specified output sample format. */
287 if (!(*fifo = av_audio_fifo_alloc(output_codec_context->sample_fmt,
288 output_codec_context->channels, 1))) {
289 fprintf(stderr, "Could not allocate FIFO\n");
290 return AVERROR(ENOMEM);
295 /** Write the header of the output file container. */
296 static int write_output_file_header(AVFormatContext *output_format_context)
299 if ((error = avformat_write_header(output_format_context, NULL)) < 0) {
300 fprintf(stderr, "Could not write output file header (error '%s')\n",
301 get_error_text(error));
307 /** Decode one audio frame from the input file. */
308 static int decode_audio_frame(AVFrame *frame,
309 AVFormatContext *input_format_context,
310 AVCodecContext *input_codec_context,
311 int *data_present, int *finished)
313 /** Packet used for temporary storage. */
314 AVPacket input_packet;
316 init_packet(&input_packet);
318 /** Read one audio frame from the input file into a temporary packet. */
319 if ((error = av_read_frame(input_format_context, &input_packet)) < 0) {
320 /** If we are the the end of the file, flush the decoder below. */
321 if (error == AVERROR_EOF)
324 fprintf(stderr, "Could not read frame (error '%s')\n",
325 get_error_text(error));
331 * Decode the audio frame stored in the temporary packet.
332 * The input audio stream decoder is used to do this.
333 * If we are at the end of the file, pass an empty packet to the decoder
336 if ((error = avcodec_decode_audio4(input_codec_context, frame,
337 data_present, &input_packet)) < 0) {
338 fprintf(stderr, "Could not decode frame (error '%s')\n",
339 get_error_text(error));
340 av_free_packet(&input_packet);
345 * If the decoder has not been flushed completely, we are not finished,
346 * so that this function has to be called again.
348 if (*finished && *data_present)
350 av_free_packet(&input_packet);
355 * Initialize a temporary storage for the specified number of audio samples.
356 * The conversion requires temporary storage due to the different format.
357 * The number of audio samples to be allocated is specified in frame_size.
359 static int init_converted_samples(uint8_t ***converted_input_samples,
360 AVCodecContext *output_codec_context,
366 * Allocate as many pointers as there are audio channels.
367 * Each pointer will later point to the audio samples of the corresponding
368 * channels (although it may be NULL for interleaved formats).
370 if (!(*converted_input_samples = calloc(output_codec_context->channels,
371 sizeof(**converted_input_samples)))) {
372 fprintf(stderr, "Could not allocate converted input sample pointers\n");
373 return AVERROR(ENOMEM);
377 * Allocate memory for the samples of all channels in one consecutive
378 * block for convenience.
380 if ((error = av_samples_alloc(*converted_input_samples, NULL,
381 output_codec_context->channels,
383 output_codec_context->sample_fmt, 0)) < 0) {
385 "Could not allocate converted input samples (error '%s')\n",
386 get_error_text(error));
387 av_freep(&(*converted_input_samples)[0]);
388 free(*converted_input_samples);
395 * Convert the input audio samples into the output sample format.
396 * The conversion happens on a per-frame basis, the size of which is specified
399 static int convert_samples(uint8_t **input_data,
400 uint8_t **converted_data, const int frame_size,
401 AVAudioResampleContext *resample_context)
405 /** Convert the samples using the resampler. */
406 if ((error = avresample_convert(resample_context, converted_data, 0,
407 frame_size, input_data, 0, frame_size)) < 0) {
408 fprintf(stderr, "Could not convert input samples (error '%s')\n",
409 get_error_text(error));
414 * Perform a sanity check so that the number of converted samples is
415 * not greater than the number of samples to be converted.
416 * If the sample rates differ, this case has to be handled differently
418 if (avresample_available(resample_context)) {
419 fprintf(stderr, "Converted samples left over\n");
426 /** Add converted input audio samples to the FIFO buffer for later processing. */
427 static int add_samples_to_fifo(AVAudioFifo *fifo,
428 uint8_t **converted_input_samples,
429 const int frame_size)
434 * Make the FIFO as large as it needs to be to hold both,
435 * the old and the new samples.
437 if ((error = av_audio_fifo_realloc(fifo, av_audio_fifo_size(fifo) + frame_size)) < 0) {
438 fprintf(stderr, "Could not reallocate FIFO\n");
442 /** Store the new samples in the FIFO buffer. */
443 if (av_audio_fifo_write(fifo, (void **)converted_input_samples,
444 frame_size) < frame_size) {
445 fprintf(stderr, "Could not write data to FIFO\n");
452 * Read one audio frame from the input file, decodes, converts and stores
453 * it in the FIFO buffer.
455 static int read_decode_convert_and_store(AVAudioFifo *fifo,
456 AVFormatContext *input_format_context,
457 AVCodecContext *input_codec_context,
458 AVCodecContext *output_codec_context,
459 AVAudioResampleContext *resampler_context,
462 /** Temporary storage of the input samples of the frame read from the file. */
463 AVFrame *input_frame = NULL;
464 /** Temporary storage for the converted input samples. */
465 uint8_t **converted_input_samples = NULL;
467 int ret = AVERROR_EXIT;
469 /** Initialize temporary storage for one input frame. */
470 if (init_input_frame(&input_frame))
472 /** Decode one frame worth of audio samples. */
473 if (decode_audio_frame(input_frame, input_format_context,
474 input_codec_context, &data_present, finished))
477 * If we are at the end of the file and there are no more samples
478 * in the decoder which are delayed, we are actually finished.
479 * This must not be treated as an error.
481 if (*finished && !data_present) {
485 /** If there is decoded data, convert and store it */
487 /** Initialize the temporary storage for the converted input samples. */
488 if (init_converted_samples(&converted_input_samples, output_codec_context,
489 input_frame->nb_samples))
493 * Convert the input samples to the desired output sample format.
494 * This requires a temporary storage provided by converted_input_samples.
496 if (convert_samples(input_frame->extended_data, converted_input_samples,
497 input_frame->nb_samples, resampler_context))
500 /** Add the converted input samples to the FIFO buffer for later processing. */
501 if (add_samples_to_fifo(fifo, converted_input_samples,
502 input_frame->nb_samples))
509 if (converted_input_samples) {
510 av_freep(&converted_input_samples[0]);
511 free(converted_input_samples);
513 av_frame_free(&input_frame);
519 * Initialize one input frame for writing to the output file.
520 * The frame will be exactly frame_size samples large.
522 static int init_output_frame(AVFrame **frame,
523 AVCodecContext *output_codec_context,
528 /** Create a new frame to store the audio samples. */
529 if (!(*frame = av_frame_alloc())) {
530 fprintf(stderr, "Could not allocate output frame\n");
535 * Set the frame's parameters, especially its size and format.
536 * av_frame_get_buffer needs this to allocate memory for the
537 * audio samples of the frame.
538 * Default channel layouts based on the number of channels
539 * are assumed for simplicity.
541 (*frame)->nb_samples = frame_size;
542 (*frame)->channel_layout = output_codec_context->channel_layout;
543 (*frame)->format = output_codec_context->sample_fmt;
544 (*frame)->sample_rate = output_codec_context->sample_rate;
547 * Allocate the samples of the created frame. This call will make
548 * sure that the audio frame can hold as many samples as specified.
550 if ((error = av_frame_get_buffer(*frame, 0)) < 0) {
551 fprintf(stderr, "Could allocate output frame samples (error '%s')\n",
552 get_error_text(error));
553 av_frame_free(frame);
560 /** Global timestamp for the audio frames */
561 static int64_t pts = 0;
563 /** Encode one frame worth of audio to the output file. */
564 static int encode_audio_frame(AVFrame *frame,
565 AVFormatContext *output_format_context,
566 AVCodecContext *output_codec_context,
569 /** Packet used for temporary storage. */
570 AVPacket output_packet;
572 init_packet(&output_packet);
574 /** Set a timestamp based on the sample rate for the container. */
577 pts += frame->nb_samples;
581 * Encode the audio frame and store it in the temporary packet.
582 * The output audio stream encoder is used to do this.
584 if ((error = avcodec_encode_audio2(output_codec_context, &output_packet,
585 frame, data_present)) < 0) {
586 fprintf(stderr, "Could not encode frame (error '%s')\n",
587 get_error_text(error));
588 av_free_packet(&output_packet);
592 /** Write one audio frame from the temporary packet to the output file. */
594 if ((error = av_write_frame(output_format_context, &output_packet)) < 0) {
595 fprintf(stderr, "Could not write frame (error '%s')\n",
596 get_error_text(error));
597 av_free_packet(&output_packet);
601 av_free_packet(&output_packet);
608 * Load one audio frame from the FIFO buffer, encode and write it to the
611 static int load_encode_and_write(AVAudioFifo *fifo,
612 AVFormatContext *output_format_context,
613 AVCodecContext *output_codec_context)
615 /** Temporary storage of the output samples of the frame written to the file. */
616 AVFrame *output_frame;
618 * Use the maximum number of possible samples per frame.
619 * If there is less than the maximum possible frame size in the FIFO
620 * buffer use this number. Otherwise, use the maximum possible frame size
622 const int frame_size = FFMIN(av_audio_fifo_size(fifo),
623 output_codec_context->frame_size);
626 /** Initialize temporary storage for one output frame. */
627 if (init_output_frame(&output_frame, output_codec_context, frame_size))
631 * Read as many samples from the FIFO buffer as required to fill the frame.
632 * The samples are stored in the frame temporarily.
634 if (av_audio_fifo_read(fifo, (void **)output_frame->data, frame_size) < frame_size) {
635 fprintf(stderr, "Could not read data from FIFO\n");
636 av_frame_free(&output_frame);
640 /** Encode one frame worth of audio samples. */
641 if (encode_audio_frame(output_frame, output_format_context,
642 output_codec_context, &data_written)) {
643 av_frame_free(&output_frame);
646 av_frame_free(&output_frame);
650 /** Write the trailer of the output file container. */
651 static int write_output_file_trailer(AVFormatContext *output_format_context)
654 if ((error = av_write_trailer(output_format_context)) < 0) {
655 fprintf(stderr, "Could not write output file trailer (error '%s')\n",
656 get_error_text(error));
662 /** Convert an audio file to an AAC file in an MP4 container. */
663 int main(int argc, char **argv)
665 AVFormatContext *input_format_context = NULL, *output_format_context = NULL;
666 AVCodecContext *input_codec_context = NULL, *output_codec_context = NULL;
667 AVAudioResampleContext *resample_context = NULL;
668 AVAudioFifo *fifo = NULL;
669 int ret = AVERROR_EXIT;
672 fprintf(stderr, "Usage: %s <input file> <output file>\n", argv[0]);
676 /** Register all codecs and formats so that they can be used. */
678 /** Open the input file for reading. */
679 if (open_input_file(argv[1], &input_format_context,
680 &input_codec_context))
682 /** Open the output file for writing. */
683 if (open_output_file(argv[2], input_codec_context,
684 &output_format_context, &output_codec_context))
686 /** Initialize the resampler to be able to convert audio sample formats. */
687 if (init_resampler(input_codec_context, output_codec_context,
690 /** Initialize the FIFO buffer to store audio samples to be encoded. */
691 if (init_fifo(&fifo, output_codec_context))
693 /** Write the header of the output file container. */
694 if (write_output_file_header(output_format_context))
698 * Loop as long as we have input samples to read or output samples
699 * to write; abort as soon as we have neither.
702 /** Use the encoder's desired frame size for processing. */
703 const int output_frame_size = output_codec_context->frame_size;
707 * Make sure that there is one frame worth of samples in the FIFO
708 * buffer so that the encoder can do its work.
709 * Since the decoder's and the encoder's frame size may differ, we
710 * need to FIFO buffer to store as many frames worth of input samples
711 * that they make up at least one frame worth of output samples.
713 while (av_audio_fifo_size(fifo) < output_frame_size) {
715 * Decode one frame worth of audio samples, convert it to the
716 * output sample format and put it into the FIFO buffer.
718 if (read_decode_convert_and_store(fifo, input_format_context,
720 output_codec_context,
721 resample_context, &finished))
725 * If we are at the end of the input file, we continue
726 * encoding the remaining audio samples to the output file.
733 * If we have enough samples for the encoder, we encode them.
734 * At the end of the file, we pass the remaining samples to
737 while (av_audio_fifo_size(fifo) >= output_frame_size ||
738 (finished && av_audio_fifo_size(fifo) > 0))
740 * Take one frame worth of audio samples from the FIFO buffer,
741 * encode it and write it to the output file.
743 if (load_encode_and_write(fifo, output_format_context,
744 output_codec_context))
748 * If we are at the end of the input file and have encoded
749 * all remaining samples, we can exit this loop and finish.
753 /** Flush the encoder as it may have delayed frames. */
755 if (encode_audio_frame(NULL, output_format_context,
756 output_codec_context, &data_written))
758 } while (data_written);
763 /** Write the trailer of the output file container. */
764 if (write_output_file_trailer(output_format_context))
770 av_audio_fifo_free(fifo);
771 if (resample_context) {
772 avresample_close(resample_context);
773 avresample_free(&resample_context);
775 if (output_codec_context)
776 avcodec_close(output_codec_context);
777 if (output_format_context) {
778 avio_close(output_format_context->pb);
779 avformat_free_context(output_format_context);
781 if (input_codec_context)
782 avcodec_close(input_codec_context);
783 if (input_format_context)
784 avformat_close_input(&input_format_context);