1/* 2 * AAC definitions and structures 3 * Copyright (c) 2005-2006 Oded Shimon ( ods15 ods15 dyndns org ) 4 * Copyright (c) 2006-2007 Maxim Gavrilov ( maxim.gavrilov gmail com ) 5 * 6 * This file is part of Libav. 7 * 8 * Libav is free software; you can redistribute it and/or 9 * modify it under the terms of the GNU Lesser General Public 10 * License as published by the Free Software Foundation; either 11 * version 2.1 of the License, or (at your option) any later version. 12 * 13 * Libav is distributed in the hope that it will be useful, 14 * but WITHOUT ANY WARRANTY; without even the implied warranty of 15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU 16 * Lesser General Public License for more details. 17 * 18 * You should have received a copy of the GNU Lesser General Public 19 * License along with Libav; if not, write to the Free Software 20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA 21 */ 22 23/** 24 * @file 25 * AAC definitions and structures 26 * @author Oded Shimon ( ods15 ods15 dyndns org ) 27 * @author Maxim Gavrilov ( maxim.gavrilov gmail com ) 28 */ 29 30#ifndef AVCODEC_AAC_H 31#define AVCODEC_AAC_H 32 33#include "avcodec.h" 34#include "dsputil.h" 35#include "fft.h" 36#include "mpeg4audio.h" 37#include "sbr.h" 38#include "fmtconvert.h" 39 40#include <stdint.h> 41 42#define MAX_CHANNELS 64 43#define MAX_ELEM_ID 16 44 45#define TNS_MAX_ORDER 20 46#define MAX_LTP_LONG_SFB 40 47 48enum RawDataBlockType { 49 TYPE_SCE, 50 TYPE_CPE, 51 TYPE_CCE, 52 TYPE_LFE, 53 TYPE_DSE, 54 TYPE_PCE, 55 TYPE_FIL, 56 TYPE_END, 57}; 58 59enum ExtensionPayloadID { 60 EXT_FILL, 61 EXT_FILL_DATA, 62 EXT_DATA_ELEMENT, 63 EXT_DYNAMIC_RANGE = 0xb, 64 EXT_SBR_DATA = 0xd, 65 EXT_SBR_DATA_CRC = 0xe, 66}; 67 68enum WindowSequence { 69 ONLY_LONG_SEQUENCE, 70 LONG_START_SEQUENCE, 71 EIGHT_SHORT_SEQUENCE, 72 LONG_STOP_SEQUENCE, 73}; 74 75enum BandType { 76 ZERO_BT = 0, ///< Scalefactors and spectral data are all zero. 77 FIRST_PAIR_BT = 5, ///< This and later band types encode two values (rather than four) with one code word. 78 ESC_BT = 11, ///< Spectral data are coded with an escape sequence. 79 NOISE_BT = 13, ///< Spectral data are scaled white noise not coded in the bitstream. 80 INTENSITY_BT2 = 14, ///< Scalefactor data are intensity stereo positions. 81 INTENSITY_BT = 15, ///< Scalefactor data are intensity stereo positions. 82}; 83 84#define IS_CODEBOOK_UNSIGNED(x) ((x - 1) & 10) 85 86enum ChannelPosition { 87 AAC_CHANNEL_OFF = 0, 88 AAC_CHANNEL_FRONT = 1, 89 AAC_CHANNEL_SIDE = 2, 90 AAC_CHANNEL_BACK = 3, 91 AAC_CHANNEL_LFE = 4, 92 AAC_CHANNEL_CC = 5, 93}; 94 95/** 96 * The point during decoding at which channel coupling is applied. 97 */ 98enum CouplingPoint { 99 BEFORE_TNS, 100 BETWEEN_TNS_AND_IMDCT, 101 AFTER_IMDCT = 3, 102}; 103 104/** 105 * Output configuration status 106 */ 107enum OCStatus { 108 OC_NONE, ///< Output unconfigured 109 OC_TRIAL_PCE, ///< Output configuration under trial specified by an inband PCE 110 OC_TRIAL_FRAME, ///< Output configuration under trial specified by a frame header 111 OC_GLOBAL_HDR, ///< Output configuration set in a global header but not yet locked 112 OC_LOCKED, ///< Output configuration locked in place 113}; 114 115/** 116 * Predictor State 117 */ 118typedef struct { 119 float cor0; 120 float cor1; 121 float var0; 122 float var1; 123 float r0; 124 float r1; 125} PredictorState; 126 127#define MAX_PREDICTORS 672 128 129#define SCALE_DIV_512 36 ///< scalefactor difference that corresponds to scale difference in 512 times 130#define SCALE_ONE_POS 140 ///< scalefactor index that corresponds to scale=1.0 131#define SCALE_MAX_POS 255 ///< scalefactor index maximum value 132#define SCALE_MAX_DIFF 60 ///< maximum scalefactor difference allowed by standard 133#define SCALE_DIFF_ZERO 60 ///< codebook index corresponding to zero scalefactor indices difference 134#define POW_SF2_ZERO 200 ///< ff_aac_pow2sf_tab index corresponding to pow(2, 0); 135 136/** 137 * Long Term Prediction 138 */ 139typedef struct { 140 int8_t present; 141 int16_t lag; 142 float coef; 143 int8_t used[MAX_LTP_LONG_SFB]; 144} LongTermPrediction; 145 146/** 147 * Individual Channel Stream 148 */ 149typedef struct { 150 uint8_t max_sfb; ///< number of scalefactor bands per group 151 enum WindowSequence window_sequence[2]; 152 uint8_t use_kb_window[2]; ///< If set, use Kaiser-Bessel window, otherwise use a sinus window. 153 int num_window_groups; 154 uint8_t group_len[8]; 155 LongTermPrediction ltp; 156 const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window 157 const uint8_t *swb_sizes; ///< table of scalefactor band sizes for a particular window 158 int num_swb; ///< number of scalefactor window bands 159 int num_windows; 160 int tns_max_bands; 161 int predictor_present; 162 int predictor_initialized; 163 int predictor_reset_group; 164 uint8_t prediction_used[41]; 165} IndividualChannelStream; 166 167/** 168 * Temporal Noise Shaping 169 */ 170typedef struct { 171 int present; 172 int n_filt[8]; 173 int length[8][4]; 174 int direction[8][4]; 175 int order[8][4]; 176 float coef[8][4][TNS_MAX_ORDER]; 177} TemporalNoiseShaping; 178 179/** 180 * Dynamic Range Control - decoded from the bitstream but not processed further. 181 */ 182typedef struct { 183 int pce_instance_tag; ///< Indicates with which program the DRC info is associated. 184 int dyn_rng_sgn[17]; ///< DRC sign information; 0 - positive, 1 - negative 185 int dyn_rng_ctl[17]; ///< DRC magnitude information 186 int exclude_mask[MAX_CHANNELS]; ///< Channels to be excluded from DRC processing. 187 int band_incr; ///< Number of DRC bands greater than 1 having DRC info. 188 int interpolation_scheme; ///< Indicates the interpolation scheme used in the SBR QMF domain. 189 int band_top[17]; ///< Indicates the top of the i-th DRC band in units of 4 spectral lines. 190 int prog_ref_level; /**< A reference level for the long-term program audio level for all 191 * channels combined. 192 */ 193} DynamicRangeControl; 194 195typedef struct { 196 int num_pulse; 197 int start; 198 int pos[4]; 199 int amp[4]; 200} Pulse; 201 202/** 203 * coupling parameters 204 */ 205typedef struct { 206 enum CouplingPoint coupling_point; ///< The point during decoding at which coupling is applied. 207 int num_coupled; ///< number of target elements 208 enum RawDataBlockType type[8]; ///< Type of channel element to be coupled - SCE or CPE. 209 int id_select[8]; ///< element id 210 int ch_select[8]; /**< [0] shared list of gains; [1] list of gains for right channel; 211 * [2] list of gains for left channel; [3] lists of gains for both channels 212 */ 213 float gain[16][120]; 214} ChannelCoupling; 215 216/** 217 * Single Channel Element - used for both SCE and LFE elements. 218 */ 219typedef struct { 220 IndividualChannelStream ics; 221 TemporalNoiseShaping tns; 222 Pulse pulse; 223 enum BandType band_type[128]; ///< band types 224 int band_type_run_end[120]; ///< band type run end points 225 float sf[120]; ///< scalefactors 226 int sf_idx[128]; ///< scalefactor indices (used by encoder) 227 uint8_t zeroes[128]; ///< band is not coded (used by encoder) 228 DECLARE_ALIGNED(32, float, coeffs)[1024]; ///< coefficients for IMDCT 229 DECLARE_ALIGNED(32, float, saved)[1024]; ///< overlap 230 DECLARE_ALIGNED(32, float, ret)[2048]; ///< PCM output 231 DECLARE_ALIGNED(16, float, ltp_state)[3072]; ///< time signal for LTP 232 PredictorState predictor_state[MAX_PREDICTORS]; 233} SingleChannelElement; 234 235/** 236 * channel element - generic struct for SCE/CPE/CCE/LFE 237 */ 238typedef struct { 239 // CPE specific 240 int common_window; ///< Set if channels share a common 'IndividualChannelStream' in bitstream. 241 int ms_mode; ///< Signals mid/side stereo flags coding mode (used by encoder) 242 uint8_t ms_mask[128]; ///< Set if mid/side stereo is used for each scalefactor window band 243 // shared 244 SingleChannelElement ch[2]; 245 // CCE specific 246 ChannelCoupling coup; 247 SpectralBandReplication sbr; 248} ChannelElement; 249 250/** 251 * main AAC context 252 */ 253typedef struct { 254 AVCodecContext *avctx; 255 AVFrame frame; 256 257 MPEG4AudioConfig m4ac; 258 259 int is_saved; ///< Set if elements have stored overlap from previous frame. 260 DynamicRangeControl che_drc; 261 262 /** 263 * @name Channel element related data 264 * @{ 265 */ 266 enum ChannelPosition che_pos[4][MAX_ELEM_ID]; /**< channel element channel mapping with the 267 * first index as the first 4 raw data block types 268 */ 269 ChannelElement *che[4][MAX_ELEM_ID]; 270 ChannelElement *tag_che_map[4][MAX_ELEM_ID]; 271 int tags_mapped; 272 /** @} */ 273 274 /** 275 * @name temporary aligned temporary buffers 276 * (We do not want to have these on the stack.) 277 * @{ 278 */ 279 DECLARE_ALIGNED(32, float, buf_mdct)[1024]; 280 /** @} */ 281 282 /** 283 * @name Computed / set up during initialization 284 * @{ 285 */ 286 FFTContext mdct; 287 FFTContext mdct_small; 288 FFTContext mdct_ltp; 289 DSPContext dsp; 290 FmtConvertContext fmt_conv; 291 int random_state; 292 /** @} */ 293 294 /** 295 * @name Members used for output interleaving 296 * @{ 297 */ 298 float *output_data[MAX_CHANNELS]; ///< Points to each element's 'ret' buffer (PCM output). 299 /** @} */ 300 301 DECLARE_ALIGNED(32, float, temp)[128]; 302 303 enum OCStatus output_configured; 304} AACContext; 305 306#endif /* AVCODEC_AAC_H */ 307