path: root/src/decoder/mp4ff_decoder_plugin.c

/*
 * Copyright (C) 2003-2010 The Music Player Daemon Project
 * http://www.musicpd.org
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program 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 General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License along
 * with this program; if not, write to the Free Software Foundation, Inc.,
 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.
 */

#include "config.h"
#include "decoder_api.h"
#include "audio_check.h"
#include "tag_table.h"

#include <glib.h>

#include <mp4ff.h>
#include <faad.h>

#include <assert.h>
#include <stdlib.h>
#include <unistd.h>

#undef G_LOG_DOMAIN
#define G_LOG_DOMAIN "mp4ff"

/* all code here is either based on or copied from FAAD2's frontend code */

struct mp4ff_input_stream {
	mp4ff_callback_t callback;

	struct decoder *decoder;
	struct input_stream *input_stream;
};

static int
mp4_get_aac_track(mp4ff_t * infile, faacDecHandle decoder,
		  uint32_t *sample_rate, unsigned char *channels_r)
{
#ifdef HAVE_FAAD_LONG
	/* neaacdec.h declares all arguments as "unsigned long", but
	   internally expects uint32_t pointers.  To avoid gcc
	   warnings, use this workaround. */
	unsigned long *sample_rate_r = (unsigned long*)sample_rate;
#else
	uint32_t *sample_rate_r = sample_rate;
#endif
	int i, rc;
	int num_tracks = mp4ff_total_tracks(infile);

	for (i = 0; i < num_tracks; i++) {
		unsigned char *buff = NULL;
		unsigned int buff_size = 0;

		if (mp4ff_get_track_type(infile, i) != 1)
			/* not an audio track */
			continue;

		if (decoder == NULL)
			/* have don't have a decoder to initialize -
			   we're done now, because we found an audio
			   track */
			return i;

		mp4ff_get_decoder_config(infile, i, &buff, &buff_size);
		if (buff == NULL)
			continue;

		rc = faacDecInit2(decoder, buff, buff_size,
				  sample_rate_r, channels_r);
		free(buff);

		if (rc >= 0)
			/* found a valid AAC track */
			return i;
	}

	/* can't decode this */
	return -1;
}

static uint32_t
mp4_read(void *user_data, void *buffer, uint32_t length)
{
	struct mp4ff_input_stream *mis = user_data;

	return decoder_read(mis->decoder, mis->input_stream, buffer, length);
}

static uint32_t
mp4_seek(void *user_data, uint64_t position)
{
	struct mp4ff_input_stream *mis = user_data;

	return input_stream_seek(mis->input_stream, position, SEEK_SET, NULL)
		? 0 : -1;
}

static const mp4ff_callback_t mpd_mp4ff_callback = {
	.read = mp4_read,
	.seek = mp4_seek,
};

static faacDecHandle
mp4_faad_new(mp4ff_t *mp4fh, int *track_r, struct audio_format *audio_format)
{
	faacDecHandle decoder;
	faacDecConfigurationPtr config;
	int track;
	uint32_t sample_rate;
	unsigned char channels;
	GError *error = NULL;

	decoder = faacDecOpen();

	config = faacDecGetCurrentConfiguration(decoder);
	config->outputFormat = FAAD_FMT_16BIT;
#ifdef HAVE_FAACDECCONFIGURATION_DOWNMATRIX
	config->downMatrix = 1;
#endif
#ifdef HAVE_FAACDECCONFIGURATION_DONTUPSAMPLEIMPLICITSBR
	config->dontUpSampleImplicitSBR = 0;
#endif
	faacDecSetConfiguration(decoder, config);

	track = mp4_get_aac_track(mp4fh, decoder, &sample_rate, &channels);
	if (track < 0) {
		g_warning("No AAC track found");
		faacDecClose(decoder);
		return NULL;
	}

	if (!audio_format_init_checked(audio_format, sample_rate,
				       SAMPLE_FORMAT_S16, channels,
				       &error)) {
		g_warning("%s", error->message);
		g_error_free(error);
		faacDecClose(decoder);
		return NULL;
	}

	*track_r = track;

	return decoder;
}

static void
mp4_decode(struct decoder *mpd_decoder, struct input_stream *input_stream)
{
	struct mp4ff_input_stream mis = {
		.callback = mpd_mp4ff_callback,
		.decoder = mpd_decoder,
		.input_stream = input_stream,
	};
	mp4ff_t *mp4fh;
	int32_t track;
	float file_time, total_time;
	int32_t scale;
	faacDecHandle decoder;
	struct audio_format audio_format;
	faacDecFrameInfo frame_info;
	unsigned char *mp4_buffer;
	unsigned int mp4_buffer_size;
	long sample_id;
	long num_samples;
	long dur;
	unsigned int sample_count;
	char *sample_buffer;
	size_t sample_buffer_length;
	unsigned int initial = 1;
	float *seek_table;
	long seek_table_end = -1;
	bool seek_position_found = false;
	long offset;
	uint16_t bit_rate = 0;
	bool seeking = false;
	double seek_where = 0;
	enum decoder_command cmd = DECODE_COMMAND_NONE;

	mis.callback.user_data = &mis;

	mp4fh = mp4ff_open_read(&mis.callback);
	if (!mp4fh) {
		g_warning("Input does not appear to be a mp4 stream.\n");
		return;
	}

	decoder = mp4_faad_new(mp4fh, &track, &audio_format);
	if (decoder == NULL) {
		mp4ff_close(mp4fh);
		return;
	}

	file_time = mp4ff_get_track_duration_use_offsets(mp4fh, track);
	scale = mp4ff_time_scale(mp4fh, track);

	if (scale < 0) {
		g_warning("Error getting audio format of mp4 AAC track.\n");
		faacDecClose(decoder);
		mp4ff_close(mp4fh);
		return;
	}
	total_time = ((float)file_time) / scale;

	num_samples = mp4ff_num_samples(mp4fh, track);
	if (num_samples > (long)(G_MAXINT / sizeof(float))) {
		 g_warning("Integer overflow.\n");
		 faacDecClose(decoder);
		 mp4ff_close(mp4fh);
		 return;
	}

	file_time = 0.0;

	seek_table = input_stream->seekable
		? g_malloc(sizeof(float) * num_samples)
		: NULL;

	decoder_initialized(mpd_decoder, &audio_format,
			    input_stream->seekable,
			    total_time);

	for (sample_id = 0;
	     sample_id < num_samples && cmd != DECODE_COMMAND_STOP;
	     sample_id++) {
		if (cmd == DECODE_COMMAND_SEEK) {
			assert(seek_table != NULL);

			seeking = true;
			seek_where = decoder_seek_where(mpd_decoder);
		}

		if (seeking && seek_table_end > 1 &&
		    seek_table[seek_table_end] >= seek_where) {
			int i = 2;

			assert(seek_table != NULL);

			while (seek_table[i] < seek_where)
				i++;
			sample_id = i - 1;
			file_time = seek_table[sample_id];
		}

		dur = mp4ff_get_sample_duration(mp4fh, track, sample_id);
		offset = mp4ff_get_sample_offset(mp4fh, track, sample_id);

		if (seek_table != NULL && sample_id > seek_table_end) {
			seek_table[sample_id] = file_time;
			seek_table_end = sample_id;
		}

		if (sample_id == 0)
			dur = 0;
		if (offset > dur)
			dur = 0;
		else
			dur -= offset;
		file_time += ((float)dur) / scale;

		if (seeking && file_time >= seek_where)
			seek_position_found = true;

		if (seeking && seek_position_found) {
			seek_position_found = false;
			seeking = 0;
			decoder_command_finished(mpd_decoder);
		}

		if (seeking)
			continue;

		if (mp4ff_read_sample(mp4fh, track, sample_id, &mp4_buffer,
				      &mp4_buffer_size) == 0)
			break;

#ifdef HAVE_FAAD_BUFLEN_FUNCS
		sample_buffer = faacDecDecode(decoder, &frame_info, mp4_buffer,
					      mp4_buffer_size);
#else
		sample_buffer = faacDecDecode(decoder, &frame_info, mp4_buffer);
#endif

		free(mp4_buffer);

		if (frame_info.error > 0) {
			g_warning("faad2 error: %s\n",
				  faacDecGetErrorMessage(frame_info.error));
			break;
		}

		if (frame_info.channels != audio_format.channels) {
			g_warning("channel count changed from %u to %u",
				  audio_format.channels, frame_info.channels);
			break;
		}

#ifdef HAVE_FAACDECFRAMEINFO_SAMPLERATE
		if (frame_info.samplerate != audio_format.sample_rate) {
			g_warning("sample rate changed from %u to %lu",
				  audio_format.sample_rate,
				  (unsigned long)frame_info.samplerate);
			break;
		}
#endif

		if (audio_format.channels * (unsigned long)(dur + offset) > frame_info.samples) {
			dur = frame_info.samples / audio_format.channels;
			offset = 0;
		}

		sample_count = (unsigned long)(dur * audio_format.channels);

		if (sample_count > 0) {
			initial = 0;
			bit_rate = frame_info.bytesconsumed * 8.0 *
			    frame_info.channels * scale /
			    frame_info.samples / 1000 + 0.5;
		}

		sample_buffer_length = sample_count * 2;

		sample_buffer += offset * audio_format.channels * 2;

		cmd = decoder_data(mpd_decoder, input_stream,
				   sample_buffer, sample_buffer_length,
				   bit_rate);
	}

	g_free(seek_table);
	faacDecClose(decoder);
	mp4ff_close(mp4fh);
}

static const char *const mp4ff_tag_names[TAG_NUM_OF_ITEM_TYPES] = {
	[TAG_COMPOSER] = "writer",
	[TAG_PERFORMER] = "band",
};

static enum tag_type
mp4ff_tag_name_parse(const char *name)
{
	enum tag_type type = tag_table_lookup(mp4ff_tag_names, name);
	if (type == TAG_NUM_OF_ITEM_TYPES)
		type = tag_name_parse_i(name);

	return type;
}

static struct tag *
mp4_stream_tag(struct input_stream *is)
{
	struct tag *ret = NULL;
	struct mp4ff_input_stream mis = {
		.callback = mpd_mp4ff_callback,
		.decoder = NULL,
		.input_stream = is,
	};
	mp4ff_t *mp4fh;
	int32_t track;
	int32_t file_time;
	int32_t scale;
	int i;

	mis.callback.user_data = &mis;

	mp4fh = mp4ff_open_read(&mis.callback);
	if (mp4fh == NULL)
		return NULL;

	track = mp4_get_aac_track(mp4fh, NULL, NULL, NULL);
	if (track < 0) {
		mp4ff_close(mp4fh);
		return NULL;
	}

	ret = tag_new();
	file_time = mp4ff_get_track_duration_use_offsets(mp4fh, track);
	scale = mp4ff_time_scale(mp4fh, track);
	if (scale < 0) {
		mp4ff_close(mp4fh);
		tag_free(ret);
		return NULL;
	}
	ret->time = ((float)file_time) / scale + 0.5;

	for (i = 0; i < mp4ff_meta_get_num_items(mp4fh); i++) {
		char *item;
		char *value;

		mp4ff_meta_get_by_index(mp4fh, i, &item, &value);

		enum tag_type type = mp4ff_tag_name_parse(item);
		if (type != TAG_NUM_OF_ITEM_TYPES)
			tag_add_item(ret, type, value);

		free(item);
		free(value);
	}

	mp4ff_close(mp4fh);

	return ret;
}

static const char *const mp4_suffixes[] = { "m4a", "mp4", NULL };
static const char *const mp4_mime_types[] = { "audio/mp4", "audio/m4a", NULL };

const struct decoder_plugin mp4ff_decoder_plugin = {
	.name = "mp4ff",
	.stream_decode = mp4_decode,
	.stream_tag = mp4_stream_tag,
	.suffixes = mp4_suffixes,
	.mime_types = mp4_mime_types,
};