1
0
mirror of https://github.com/FFmpeg/FFmpeg.git synced 2024-11-21 10:55:51 +02:00
FFmpeg/libavcodec/aac/aacdec_latm.h
Lynne eee5fa0808
aacdec: add a decoder for AAC USAC (xHE-AAC)
This commit adds a decoder for the frequency-domain part of USAC.

What works:
 - Mono
 - Stereo (no prediction)
 - Stereo (mid/side coding)
 - Stereo (complex prediction)

What's left:
 - SBR
 - Speech coding

Known issues:
 - Desync with certain sequences
 - Preroll crossover missing (shouldn't matter, bitrate adaptation only)
2024-06-02 18:34:45 +02:00

353 lines
12 KiB
C

/*
* AAC decoder
* Copyright (c) 2005-2006 Oded Shimon ( ods15 ods15 dyndns org )
* Copyright (c) 2006-2007 Maxim Gavrilov ( maxim.gavrilov gmail com )
* Copyright (c) 2008-2013 Alex Converse <alex.converse@gmail.com>
*
* AAC LATM decoder
* Copyright (c) 2008-2010 Paul Kendall <paul@kcbbs.gen.nz>
* Copyright (c) 2010 Janne Grunau <janne-libav@jannau.net>
*
* AAC decoder fixed-point implementation
* Copyright (c) 2013
* MIPS Technologies, Inc., California.
*
* 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
*/
#ifndef AVCODEC_AAC_AACDEC_LATM_H
#define AVCODEC_AAC_AACDEC_LATM_H
#define LOAS_SYNC_WORD 0x2b7 ///< 11 bits LOAS sync word
struct LATMContext {
AACDecContext aac_ctx; ///< containing AACContext
int initialized; ///< initialized after a valid extradata was seen
// parser data
int audio_mux_version_A; ///< LATM syntax version
int frame_length_type; ///< 0/1 variable/fixed frame length
int frame_length; ///< frame length for fixed frame length
};
static inline uint32_t latm_get_value(GetBitContext *b)
{
int length = get_bits(b, 2);
return get_bits_long(b, (length+1)*8);
}
static int latm_decode_audio_specific_config(struct LATMContext *latmctx,
GetBitContext *gb, int asclen)
{
AACDecContext *ac = &latmctx->aac_ctx;
AVCodecContext *avctx = ac->avctx;
OutputConfiguration oc = { 0 };
MPEG4AudioConfig *m4ac = &oc.m4ac;
GetBitContext gbc;
int config_start_bit = get_bits_count(gb);
int sync_extension = 0;
int bits_consumed, esize, i;
if (asclen > 0) {
sync_extension = 1;
asclen = FFMIN(asclen, get_bits_left(gb));
init_get_bits(&gbc, gb->buffer, config_start_bit + asclen);
skip_bits_long(&gbc, config_start_bit);
} else if (asclen == 0) {
gbc = *gb;
} else {
return AVERROR_INVALIDDATA;
}
if (get_bits_left(gb) <= 0)
return AVERROR_INVALIDDATA;
bits_consumed = decode_audio_specific_config_gb(NULL, avctx, &oc,
&gbc, config_start_bit,
sync_extension);
if (bits_consumed < config_start_bit)
return AVERROR_INVALIDDATA;
bits_consumed -= config_start_bit;
if (asclen == 0)
asclen = bits_consumed;
if (!latmctx->initialized ||
ac->oc[1].m4ac.sample_rate != m4ac->sample_rate ||
ac->oc[1].m4ac.chan_config != m4ac->chan_config) {
if (latmctx->initialized) {
av_log(avctx, AV_LOG_INFO, "audio config changed (sample_rate=%d, chan_config=%d)\n",
m4ac->sample_rate, m4ac->chan_config);
} else {
av_log(avctx, AV_LOG_DEBUG, "initializing latmctx\n");
}
latmctx->initialized = 0;
esize = (asclen + 7) / 8;
if (avctx->extradata_size < esize) {
av_free(avctx->extradata);
avctx->extradata = av_malloc(esize + AV_INPUT_BUFFER_PADDING_SIZE);
if (!avctx->extradata)
return AVERROR(ENOMEM);
}
avctx->extradata_size = esize;
gbc = *gb;
for (i = 0; i < esize; i++) {
avctx->extradata[i] = get_bits(&gbc, 8);
}
memset(avctx->extradata+esize, 0, AV_INPUT_BUFFER_PADDING_SIZE);
}
skip_bits_long(gb, asclen);
return 0;
}
static int read_stream_mux_config(struct LATMContext *latmctx,
GetBitContext *gb)
{
int ret, audio_mux_version = get_bits(gb, 1);
latmctx->audio_mux_version_A = 0;
if (audio_mux_version)
latmctx->audio_mux_version_A = get_bits(gb, 1);
if (!latmctx->audio_mux_version_A) {
if (audio_mux_version)
latm_get_value(gb); // taraFullness
skip_bits(gb, 1); // allStreamSameTimeFraming
skip_bits(gb, 6); // numSubFrames
// numPrograms
if (get_bits(gb, 4)) { // numPrograms
avpriv_request_sample(latmctx->aac_ctx.avctx, "Multiple programs");
return AVERROR_PATCHWELCOME;
}
// for each program (which there is only one in DVB)
// for each layer (which there is only one in DVB)
if (get_bits(gb, 3)) { // numLayer
avpriv_request_sample(latmctx->aac_ctx.avctx, "Multiple layers");
return AVERROR_PATCHWELCOME;
}
// for all but first stream: use_same_config = get_bits(gb, 1);
if (!audio_mux_version) {
if ((ret = latm_decode_audio_specific_config(latmctx, gb, 0)) < 0)
return ret;
} else {
int ascLen = latm_get_value(gb);
if ((ret = latm_decode_audio_specific_config(latmctx, gb, ascLen)) < 0)
return ret;
}
latmctx->frame_length_type = get_bits(gb, 3);
switch (latmctx->frame_length_type) {
case 0:
skip_bits(gb, 8); // latmBufferFullness
break;
case 1:
latmctx->frame_length = get_bits(gb, 9);
break;
case 3:
case 4:
case 5:
skip_bits(gb, 6); // CELP frame length table index
break;
case 6:
case 7:
skip_bits(gb, 1); // HVXC frame length table index
break;
}
if (get_bits(gb, 1)) { // other data
if (audio_mux_version) {
latm_get_value(gb); // other_data_bits
} else {
int esc;
do {
if (get_bits_left(gb) < 9)
return AVERROR_INVALIDDATA;
esc = get_bits(gb, 1);
skip_bits(gb, 8);
} while (esc);
}
}
if (get_bits(gb, 1)) // crc present
skip_bits(gb, 8); // config_crc
}
return 0;
}
static int read_payload_length_info(struct LATMContext *ctx, GetBitContext *gb)
{
uint8_t tmp;
if (ctx->frame_length_type == 0) {
int mux_slot_length = 0;
do {
if (get_bits_left(gb) < 8)
return AVERROR_INVALIDDATA;
tmp = get_bits(gb, 8);
mux_slot_length += tmp;
} while (tmp == 255);
return mux_slot_length;
} else if (ctx->frame_length_type == 1) {
return ctx->frame_length;
} else if (ctx->frame_length_type == 3 ||
ctx->frame_length_type == 5 ||
ctx->frame_length_type == 7) {
skip_bits(gb, 2); // mux_slot_length_coded
}
return 0;
}
static int read_audio_mux_element(struct LATMContext *latmctx,
GetBitContext *gb)
{
int err;
uint8_t use_same_mux = get_bits(gb, 1);
if (!use_same_mux) {
if ((err = read_stream_mux_config(latmctx, gb)) < 0)
return err;
} else if (!latmctx->aac_ctx.avctx->extradata) {
av_log(latmctx->aac_ctx.avctx, AV_LOG_DEBUG,
"no decoder config found\n");
return 1;
}
if (latmctx->audio_mux_version_A == 0) {
int mux_slot_length_bytes = read_payload_length_info(latmctx, gb);
if (mux_slot_length_bytes < 0 || mux_slot_length_bytes * 8LL > get_bits_left(gb)) {
av_log(latmctx->aac_ctx.avctx, AV_LOG_ERROR, "incomplete frame\n");
return AVERROR_INVALIDDATA;
} else if (mux_slot_length_bytes * 8 + 256 < get_bits_left(gb)) {
av_log(latmctx->aac_ctx.avctx, AV_LOG_ERROR,
"frame length mismatch %d << %d\n",
mux_slot_length_bytes * 8, get_bits_left(gb));
return AVERROR_INVALIDDATA;
}
}
return 0;
}
static int latm_decode_frame(AVCodecContext *avctx, AVFrame *out,
int *got_frame_ptr, AVPacket *avpkt)
{
struct LATMContext *latmctx = avctx->priv_data;
int muxlength, err;
GetBitContext gb;
if ((err = init_get_bits8(&gb, avpkt->data, avpkt->size)) < 0)
return err;
// check for LOAS sync word
if (get_bits(&gb, 11) != LOAS_SYNC_WORD)
return AVERROR_INVALIDDATA;
muxlength = get_bits(&gb, 13) + 3;
// not enough data, the parser should have sorted this out
if (muxlength > avpkt->size)
return AVERROR_INVALIDDATA;
if ((err = read_audio_mux_element(latmctx, &gb)))
return (err < 0) ? err : avpkt->size;
if (!latmctx->initialized) {
if (!avctx->extradata) {
*got_frame_ptr = 0;
return avpkt->size;
} else {
push_output_configuration(&latmctx->aac_ctx);
if ((err = decode_audio_specific_config(
&latmctx->aac_ctx, avctx, &latmctx->aac_ctx.oc[1],
avctx->extradata, avctx->extradata_size*8LL, 1)) < 0) {
pop_output_configuration(&latmctx->aac_ctx);
return err;
}
latmctx->initialized = 1;
}
}
if (show_bits(&gb, 12) == 0xfff) {
av_log(latmctx->aac_ctx.avctx, AV_LOG_ERROR,
"ADTS header detected, probably as result of configuration "
"misparsing\n");
return AVERROR_INVALIDDATA;
}
switch (latmctx->aac_ctx.oc[1].m4ac.object_type) {
case AOT_ER_AAC_LC:
case AOT_ER_AAC_LTP:
case AOT_ER_AAC_LD:
case AOT_ER_AAC_ELD:
err = aac_decode_er_frame(avctx, out, got_frame_ptr, &gb);
break;
default:
err = aac_decode_frame_int(avctx, out, got_frame_ptr, &gb, avpkt);
}
if (err < 0)
return err;
return muxlength;
}
static av_cold int latm_decode_init(AVCodecContext *avctx)
{
struct LATMContext *latmctx = avctx->priv_data;
int ret = ff_aac_decode_init_float(avctx);
if (avctx->extradata_size > 0)
latmctx->initialized = !ret;
return ret;
}
/*
Note: This decoder filter is intended to decode LATM streams transferred
in MPEG transport streams which only contain one program.
To do a more complex LATM demuxing a separate LATM demuxer should be used.
*/
const FFCodec ff_aac_latm_decoder = {
.p.name = "aac_latm",
CODEC_LONG_NAME("AAC LATM (Advanced Audio Coding LATM syntax)"),
.p.type = AVMEDIA_TYPE_AUDIO,
.p.id = AV_CODEC_ID_AAC_LATM,
.priv_data_size = sizeof(struct LATMContext),
.init = latm_decode_init,
.close = decode_close,
FF_CODEC_DECODE_CB(latm_decode_frame),
.p.sample_fmts = (const enum AVSampleFormat[]) {
AV_SAMPLE_FMT_FLTP, AV_SAMPLE_FMT_NONE
},
.p.capabilities = AV_CODEC_CAP_CHANNEL_CONF | AV_CODEC_CAP_DR1,
.caps_internal = FF_CODEC_CAP_INIT_CLEANUP,
.p.ch_layouts = ff_aac_ch_layout,
.flush = flush,
.p.profiles = NULL_IF_CONFIG_SMALL(ff_aac_profiles),
};
#endif /* AVCODEC_AAC_AACDEC_LATM_H */