You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

791 lines
28 KiB

5 years ago
/*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
/**
* @file
* simple audio converter
*
* @example transcode_aac.c
* Convert an input audio file to AAC in an MP4 container using FFmpeg.
* @author Andreas Unterweger (dustsigns@gmail.com)
*/
#include <stdio.h>
#include "libavformat/avformat.h"
#include "libavformat/avio.h"
#include "libavcodec/avcodec.h"
#include "libavutil/audio_fifo.h"
#include "libavutil/avassert.h"
#include "libavutil/avstring.h"
#include "libavutil/frame.h"
#include "libavutil/opt.h"
#include "libswresample/swresample.h"
/** The output bit rate in kbit/s */
#define OUTPUT_BIT_RATE 96000
/** The number of output channels */
#define OUTPUT_CHANNELS 2
/** Open an input file and the required decoder. */
static int open_input_file(const char *filename,
AVFormatContext **input_format_context,
AVCodecContext **input_codec_context)
{
AVCodecContext *avctx;
AVCodec *input_codec;
int error;
/** Open the input file to read from it. */
if ((error = avformat_open_input(input_format_context, filename, NULL,
NULL)) < 0) {
fprintf(stderr, "Could not open input file '%s' (error '%s')\n",
filename, av_err2str(error));
*input_format_context = NULL;
return error;
}
/** Get information on the input file (number of streams etc.). */
if ((error = avformat_find_stream_info(*input_format_context, NULL)) < 0) {
fprintf(stderr, "Could not open find stream info (error '%s')\n",
av_err2str(error));
avformat_close_input(input_format_context);
return error;
}
/** Make sure that there is only one stream in the input file. */
if ((*input_format_context)->nb_streams != 1) {
fprintf(stderr, "Expected one audio input stream, but found %d\n",
(*input_format_context)->nb_streams);
avformat_close_input(input_format_context);
return AVERROR_EXIT;
}
/** Find a decoder for the audio stream. */
if (!(input_codec = avcodec_find_decoder((*input_format_context)->streams[0]->codecpar->codec_id))) {
fprintf(stderr, "Could not find input codec\n");
avformat_close_input(input_format_context);
return AVERROR_EXIT;
}
/** allocate a new decoding context */
avctx = avcodec_alloc_context3(input_codec);
if (!avctx) {
fprintf(stderr, "Could not allocate a decoding context\n");
avformat_close_input(input_format_context);
return AVERROR(ENOMEM);
}
/** initialize the stream parameters with demuxer information */
error = avcodec_parameters_to_context(avctx, (*input_format_context)->streams[0]->codecpar);
if (error < 0) {
avformat_close_input(input_format_context);
avcodec_free_context(&avctx);
return error;
}
/** Open the decoder for the audio stream to use it later. */
if ((error = avcodec_open2(avctx, input_codec, NULL)) < 0) {
fprintf(stderr, "Could not open input codec (error '%s')\n",
av_err2str(error));
avcodec_free_context(&avctx);
avformat_close_input(input_format_context);
return error;
}
/** Save the decoder context for easier access later. */
*input_codec_context = avctx;
return 0;
}
/**
* Open an output file and the required encoder.
* Also set some basic encoder parameters.
* Some of these parameters are based on the input file's parameters.
*/
static int open_output_file(const char *filename,
AVCodecContext *input_codec_context,
AVFormatContext **output_format_context,
AVCodecContext **output_codec_context)
{
AVCodecContext *avctx = NULL;
AVIOContext *output_io_context = NULL;
AVStream *stream = NULL;
AVCodec *output_codec = NULL;
int error;
/** Open the output file to write to it. */
if ((error = avio_open(&output_io_context, filename,
AVIO_FLAG_WRITE)) < 0) {
fprintf(stderr, "Could not open output file '%s' (error '%s')\n",
filename, av_err2str(error));
return error;
}
/** Create a new format context for the output container format. */
if (!(*output_format_context = avformat_alloc_context())) {
fprintf(stderr, "Could not allocate output format context\n");
return AVERROR(ENOMEM);
}
/** Associate the output file (pointer) with the container format context. */
(*output_format_context)->pb = output_io_context;
/** Guess the desired container format based on the file extension. */
if (!((*output_format_context)->oformat = av_guess_format(NULL, filename,
NULL))) {
fprintf(stderr, "Could not find output file format\n");
goto cleanup;
}
av_strlcpy((*output_format_context)->filename, filename,
sizeof((*output_format_context)->filename));
/** Find the encoder to be used by its name. */
if (!(output_codec = avcodec_find_encoder(AV_CODEC_ID_AAC))) {
fprintf(stderr, "Could not find an AAC encoder.\n");
goto cleanup;
}
/** Create a new audio stream in the output file container. */
if (!(stream = avformat_new_stream(*output_format_context, NULL))) {
fprintf(stderr, "Could not create new stream\n");
error = AVERROR(ENOMEM);
goto cleanup;
}
avctx = avcodec_alloc_context3(output_codec);
if (!avctx) {
fprintf(stderr, "Could not allocate an encoding context\n");
error = AVERROR(ENOMEM);
goto cleanup;
}
/**
* Set the basic encoder parameters.
* The input file's sample rate is used to avoid a sample rate conversion.
*/
avctx->channels = OUTPUT_CHANNELS;
avctx->channel_layout = av_get_default_channel_layout(OUTPUT_CHANNELS);
avctx->sample_rate = input_codec_context->sample_rate;
avctx->sample_fmt = output_codec->sample_fmts[0];
avctx->bit_rate = OUTPUT_BIT_RATE;
/** Allow the use of the experimental AAC encoder */
avctx->strict_std_compliance = FF_COMPLIANCE_EXPERIMENTAL;
/** Set the sample rate for the container. */
stream->time_base.den = input_codec_context->sample_rate;
stream->time_base.num = 1;
/**
* Some container formats (like MP4) require global headers to be present
* Mark the encoder so that it behaves accordingly.
*/
if ((*output_format_context)->oformat->flags & AVFMT_GLOBALHEADER)
avctx->flags |= AV_CODEC_FLAG_GLOBAL_HEADER;
/** Open the encoder for the audio stream to use it later. */
if ((error = avcodec_open2(avctx, output_codec, NULL)) < 0) {
fprintf(stderr, "Could not open output codec (error '%s')\n",
av_err2str(error));
goto cleanup;
}
error = avcodec_parameters_from_context(stream->codecpar, avctx);
if (error < 0) {
fprintf(stderr, "Could not initialize stream parameters\n");
goto cleanup;
}
/** Save the encoder context for easier access later. */
*output_codec_context = avctx;
return 0;
cleanup:
avcodec_free_context(&avctx);
avio_closep(&(*output_format_context)->pb);
avformat_free_context(*output_format_context);
*output_format_context = NULL;
return error < 0 ? error : AVERROR_EXIT;
}
/** Initialize one data packet for reading or writing. */
static void init_packet(AVPacket *packet)
{
av_init_packet(packet);
/** Set the packet data and size so that it is recognized as being empty. */
packet->data = NULL;
packet->size = 0;
}
/** Initialize one audio frame for reading from the input file */
static int init_input_frame(AVFrame **frame)
{
if (!(*frame = av_frame_alloc())) {
fprintf(stderr, "Could not allocate input frame\n");
return AVERROR(ENOMEM);
}
return 0;
}
/**
* Initialize the audio resampler based on the input and output codec settings.
* If the input and output sample formats differ, a conversion is required
* libswresample takes care of this, but requires initialization.
*/
static int init_resampler(AVCodecContext *input_codec_context,
AVCodecContext *output_codec_context,
SwrContext **resample_context)
{
int error;
/**
* Create a resampler context for the conversion.
* Set the conversion parameters.
* Default channel layouts based on the number of channels
* are assumed for simplicity (they are sometimes not detected
* properly by the demuxer and/or decoder).
*/
*resample_context = swr_alloc_set_opts(NULL,
av_get_default_channel_layout(output_codec_context->channels),
output_codec_context->sample_fmt,
output_codec_context->sample_rate,
av_get_default_channel_layout(input_codec_context->channels),
input_codec_context->sample_fmt,
input_codec_context->sample_rate,
0, NULL);
if (!*resample_context) {
fprintf(stderr, "Could not allocate resample context\n");
return AVERROR(ENOMEM);
}
/**
* Perform a sanity check so that the number of converted samples is
* not greater than the number of samples to be converted.
* If the sample rates differ, this case has to be handled differently
*/
av_assert0(output_codec_context->sample_rate == input_codec_context->sample_rate);
/** Open the resampler with the specified parameters. */
if ((error = swr_init(*resample_context)) < 0) {
fprintf(stderr, "Could not open resample context\n");
swr_free(resample_context);
return error;
}
return 0;
}
/** Initialize a FIFO buffer for the audio samples to be encoded. */
static int init_fifo(AVAudioFifo **fifo, AVCodecContext *output_codec_context)
{
/** Create the FIFO buffer based on the specified output sample format. */
if (!(*fifo = av_audio_fifo_alloc(output_codec_context->sample_fmt,
output_codec_context->channels, 1))) {
fprintf(stderr, "Could not allocate FIFO\n");
return AVERROR(ENOMEM);
}
return 0;
}
/** Write the header of the output file container. */
static int write_output_file_header(AVFormatContext *output_format_context)
{
int error;
if ((error = avformat_write_header(output_format_context, NULL)) < 0) {
fprintf(stderr, "Could not write output file header (error '%s')\n",
av_err2str(error));
return error;
}
return 0;
}
/** Decode one audio frame from the input file. */
static int decode_audio_frame(AVFrame *frame,
AVFormatContext *input_format_context,
AVCodecContext *input_codec_context,
int *data_present, int *finished)
{
/** Packet used for temporary storage. */
AVPacket input_packet;
int error;
init_packet(&input_packet);
/** Read one audio frame from the input file into a temporary packet. */
if ((error = av_read_frame(input_format_context, &input_packet)) < 0) {
/** If we are at the end of the file, flush the decoder below. */
if (error == AVERROR_EOF)
*finished = 1;
else {
fprintf(stderr, "Could not read frame (error '%s')\n",
av_err2str(error));
return error;
}
}
/**
* Decode the audio frame stored in the temporary packet.
* The input audio stream decoder is used to do this.
* If we are at the end of the file, pass an empty packet to the decoder
* to flush it.
*/
if ((error = avcodec_decode_audio4(input_codec_context, frame,
data_present, &input_packet)) < 0) {
fprintf(stderr, "Could not decode frame (error '%s')\n",
av_err2str(error));
av_packet_unref(&input_packet);
return error;
}
/**
* If the decoder has not been flushed completely, we are not finished,
* so that this function has to be called again.
*/
if (*finished && *data_present)
*finished = 0;
av_packet_unref(&input_packet);
return 0;
}
/**
* Initialize a temporary storage for the specified number of audio samples.
* The conversion requires temporary storage due to the different format.
* The number of audio samples to be allocated is specified in frame_size.
*/
static int init_converted_samples(uint8_t ***converted_input_samples,
AVCodecContext *output_codec_context,
int frame_size)
{
int error;
/**
* Allocate as many pointers as there are audio channels.
* Each pointer will later point to the audio samples of the corresponding
* channels (although it may be NULL for interleaved formats).
*/
if (!(*converted_input_samples = calloc(output_codec_context->channels,
sizeof(**converted_input_samples)))) {
fprintf(stderr, "Could not allocate converted input sample pointers\n");
return AVERROR(ENOMEM);
}
/**
* Allocate memory for the samples of all channels in one consecutive
* block for convenience.
*/
if ((error = av_samples_alloc(*converted_input_samples, NULL,
output_codec_context->channels,
frame_size,
output_codec_context->sample_fmt, 0)) < 0) {
fprintf(stderr,
"Could not allocate converted input samples (error '%s')\n",
av_err2str(error));
av_freep(&(*converted_input_samples)[0]);
free(*converted_input_samples);
return error;
}
return 0;
}
/**
* Convert the input audio samples into the output sample format.
* The conversion happens on a per-frame basis, the size of which is specified
* by frame_size.
*/
static int convert_samples(const uint8_t **input_data,
uint8_t **converted_data, const int frame_size,
SwrContext *resample_context)
{
int error;
/** Convert the samples using the resampler. */
if ((error = swr_convert(resample_context,
converted_data, frame_size,
input_data , frame_size)) < 0) {
fprintf(stderr, "Could not convert input samples (error '%s')\n",
av_err2str(error));
return error;
}
return 0;
}
/** Add converted input audio samples to the FIFO buffer for later processing. */
static int add_samples_to_fifo(AVAudioFifo *fifo,
uint8_t **converted_input_samples,
const int frame_size)
{
int error;
/**
* Make the FIFO as large as it needs to be to hold both,
* the old and the new samples.
*/
if ((error = av_audio_fifo_realloc(fifo, av_audio_fifo_size(fifo) + frame_size)) < 0) {
fprintf(stderr, "Could not reallocate FIFO\n");
return error;
}
/** Store the new samples in the FIFO buffer. */
if (av_audio_fifo_write(fifo, (void **)converted_input_samples,
frame_size) < frame_size) {
fprintf(stderr, "Could not write data to FIFO\n");
return AVERROR_EXIT;
}
return 0;
}
/**
* Read one audio frame from the input file, decodes, converts and stores
* it in the FIFO buffer.
*/
static int read_decode_convert_and_store(AVAudioFifo *fifo,
AVFormatContext *input_format_context,
AVCodecContext *input_codec_context,
AVCodecContext *output_codec_context,
SwrContext *resampler_context,
int *finished)
{
/** Temporary storage of the input samples of the frame read from the file. */
AVFrame *input_frame = NULL;
/** Temporary storage for the converted input samples. */
uint8_t **converted_input_samples = NULL;
int data_present;
int ret = AVERROR_EXIT;
/** Initialize temporary storage for one input frame. */
if (init_input_frame(&input_frame))
goto cleanup;
/** Decode one frame worth of audio samples. */
if (decode_audio_frame(input_frame, input_format_context,
input_codec_context, &data_present, finished))
goto cleanup;
/**
* If we are at the end of the file and there are no more samples
* in the decoder which are delayed, we are actually finished.
* This must not be treated as an error.
*/
if (*finished && !data_present) {
ret = 0;
goto cleanup;
}
/** If there is decoded data, convert and store it */
if (data_present) {
/** Initialize the temporary storage for the converted input samples. */
if (init_converted_samples(&converted_input_samples, output_codec_context,
input_frame->nb_samples))
goto cleanup;
/**
* Convert the input samples to the desired output sample format.
* This requires a temporary storage provided by converted_input_samples.
*/
if (convert_samples((const uint8_t**)input_frame->extended_data, converted_input_samples,
input_frame->nb_samples, resampler_context))
goto cleanup;
/** Add the converted input samples to the FIFO buffer for later processing. */
if (add_samples_to_fifo(fifo, converted_input_samples,
input_frame->nb_samples))
goto cleanup;
ret = 0;
}
ret = 0;
cleanup:
if (converted_input_samples) {
av_freep(&converted_input_samples[0]);
free(converted_input_samples);
}
av_frame_free(&input_frame);
return ret;
}
/**
* Initialize one input frame for writing to the output file.
* The frame will be exactly frame_size samples large.
*/
static int init_output_frame(AVFrame **frame,
AVCodecContext *output_codec_context,
int frame_size)
{
int error;
/** Create a new frame to store the audio samples. */
if (!(*frame = av_frame_alloc())) {
fprintf(stderr, "Could not allocate output frame\n");
return AVERROR_EXIT;
}
/**
* Set the frame's parameters, especially its size and format.
* av_frame_get_buffer needs this to allocate memory for the
* audio samples of the frame.
* Default channel layouts based on the number of channels
* are assumed for simplicity.
*/
(*frame)->nb_samples = frame_size;
(*frame)->channel_layout = output_codec_context->channel_layout;
(*frame)->format = output_codec_context->sample_fmt;
(*frame)->sample_rate = output_codec_context->sample_rate;
/**
* Allocate the samples of the created frame. This call will make
* sure that the audio frame can hold as many samples as specified.
*/
if ((error = av_frame_get_buffer(*frame, 0)) < 0) {
fprintf(stderr, "Could not allocate output frame samples (error '%s')\n",
av_err2str(error));
av_frame_free(frame);
return error;
}
return 0;
}
/** Global timestamp for the audio frames */
static int64_t pts = 0;
/** Encode one frame worth of audio to the output file. */
static int encode_audio_frame(AVFrame *frame,
AVFormatContext *output_format_context,
AVCodecContext *output_codec_context,
int *data_present)
{
/** Packet used for temporary storage. */
AVPacket output_packet;
int error;
init_packet(&output_packet);
/** Set a timestamp based on the sample rate for the container. */
if (frame) {
frame->pts = pts;
pts += frame->nb_samples;
}
/**
* Encode the audio frame and store it in the temporary packet.
* The output audio stream encoder is used to do this.
*/
if ((error = avcodec_encode_audio2(output_codec_context, &output_packet,
frame, data_present)) < 0) {
fprintf(stderr, "Could not encode frame (error '%s')\n",
av_err2str(error));
av_packet_unref(&output_packet);
return error;
}
/** Write one audio frame from the temporary packet to the output file. */
if (*data_present) {
if ((error = av_write_frame(output_format_context, &output_packet)) < 0) {
fprintf(stderr, "Could not write frame (error '%s')\n",
av_err2str(error));
av_packet_unref(&output_packet);
return error;
}
av_packet_unref(&output_packet);
}
return 0;
}
/**
* Load one audio frame from the FIFO buffer, encode and write it to the
* output file.
*/
static int load_encode_and_write(AVAudioFifo *fifo,
AVFormatContext *output_format_context,
AVCodecContext *output_codec_context)
{
/** Temporary storage of the output samples of the frame written to the file. */
AVFrame *output_frame;
/**
* Use the maximum number of possible samples per frame.
* If there is less than the maximum possible frame size in the FIFO
* buffer use this number. Otherwise, use the maximum possible frame size
*/
const int frame_size = FFMIN(av_audio_fifo_size(fifo),
output_codec_context->frame_size);
int data_written;
/** Initialize temporary storage for one output frame. */
if (init_output_frame(&output_frame, output_codec_context, frame_size))
return AVERROR_EXIT;
/**
* Read as many samples from the FIFO buffer as required to fill the frame.
* The samples are stored in the frame temporarily.
*/
if (av_audio_fifo_read(fifo, (void **)output_frame->data, frame_size) < frame_size) {
fprintf(stderr, "Could not read data from FIFO\n");
av_frame_free(&output_frame);
return AVERROR_EXIT;
}
/** Encode one frame worth of audio samples. */
if (encode_audio_frame(output_frame, output_format_context,
output_codec_context, &data_written)) {
av_frame_free(&output_frame);
return AVERROR_EXIT;
}
av_frame_free(&output_frame);
return 0;
}
/** Write the trailer of the output file container. */
static int write_output_file_trailer(AVFormatContext *output_format_context)
{
int error;
if ((error = av_write_trailer(output_format_context)) < 0) {
fprintf(stderr, "Could not write output file trailer (error '%s')\n",
av_err2str(error));
return error;
}
return 0;
}
/** Convert an audio file to an AAC file in an MP4 container. */
int main(int argc, char **argv)
{
AVFormatContext *input_format_context = NULL, *output_format_context = NULL;
AVCodecContext *input_codec_context = NULL, *output_codec_context = NULL;
SwrContext *resample_context = NULL;
AVAudioFifo *fifo = NULL;
int ret = AVERROR_EXIT;
if (argc < 3) {
fprintf(stderr, "Usage: %s <input file> <output file>\n", argv[0]);
exit(1);
}
/** Register all codecs and formats so that they can be used. */
av_register_all();
/** Open the input file for reading. */
if (open_input_file(argv[1], &input_format_context,
&input_codec_context))
goto cleanup;
/** Open the output file for writing. */
if (open_output_file(argv[2], input_codec_context,
&output_format_context, &output_codec_context))
goto cleanup;
/** Initialize the resampler to be able to convert audio sample formats. */
if (init_resampler(input_codec_context, output_codec_context,
&resample_context))
goto cleanup;
/** Initialize the FIFO buffer to store audio samples to be encoded. */
if (init_fifo(&fifo, output_codec_context))
goto cleanup;
/** Write the header of the output file container. */
if (write_output_file_header(output_format_context))
goto cleanup;
/**
* Loop as long as we have input samples to read or output samples
* to write; abort as soon as we have neither.
*/
while (1) {
/** Use the encoder's desired frame size for processing. */
const int output_frame_size = output_codec_context->frame_size;
int finished = 0;
/**
* Make sure that there is one frame worth of samples in the FIFO
* buffer so that the encoder can do its work.
* Since the decoder's and the encoder's frame size may differ, we
* need to FIFO buffer to store as many frames worth of input samples
* that they make up at least one frame worth of output samples.
*/
while (av_audio_fifo_size(fifo) < output_frame_size) {
/**
* Decode one frame worth of audio samples, convert it to the
* output sample format and put it into the FIFO buffer.
*/
if (read_decode_convert_and_store(fifo, input_format_context,
input_codec_context,
output_codec_context,
resample_context, &finished))
goto cleanup;
/**
* If we are at the end of the input file, we continue
* encoding the remaining audio samples to the output file.
*/
if (finished)
break;
}
/**
* If we have enough samples for the encoder, we encode them.
* At the end of the file, we pass the remaining samples to
* the encoder.
*/
while (av_audio_fifo_size(fifo) >= output_frame_size ||
(finished && av_audio_fifo_size(fifo) > 0))
/**
* Take one frame worth of audio samples from the FIFO buffer,
* encode it and write it to the output file.
*/
if (load_encode_and_write(fifo, output_format_context,
output_codec_context))
goto cleanup;
/**
* If we are at the end of the input file and have encoded
* all remaining samples, we can exit this loop and finish.
*/
if (finished) {
int data_written;
/** Flush the encoder as it may have delayed frames. */
do {
if (encode_audio_frame(NULL, output_format_context,
output_codec_context, &data_written))
goto cleanup;
} while (data_written);
break;
}
}
/** Write the trailer of the output file container. */
if (write_output_file_trailer(output_format_context))
goto cleanup;
ret = 0;
cleanup:
if (fifo)
av_audio_fifo_free(fifo);
swr_free(&resample_context);
if (output_codec_context)
avcodec_free_context(&output_codec_context);
if (output_format_context) {
avio_closep(&output_format_context->pb);
avformat_free_context(output_format_context);
}
if (input_codec_context)
avcodec_free_context(&input_codec_context);
if (input_format_context)
avformat_close_input(&input_format_context);
return ret;
}