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 )
6 * This file is part of Libav.
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.
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.
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
25 * AAC definitions and structures
26 * @author Oded Shimon ( ods15 ods15 dyndns org )
27 * @author Maxim Gavrilov ( maxim.gavrilov gmail com )
33 #include "libavutil/float_dsp.h"
37 #include "mpeg4audio.h"
39 #include "fmtconvert.h"
43 #define MAX_CHANNELS 64
44 #define MAX_ELEM_ID 16
46 #define TNS_MAX_ORDER 20
47 #define MAX_LTP_LONG_SFB 40
49 enum RawDataBlockType {
60 enum ExtensionPayloadID {
64 EXT_DYNAMIC_RANGE = 0xb,
66 EXT_SBR_DATA_CRC = 0xe,
77 ZERO_BT = 0, ///< Scalefactors and spectral data are all zero.
78 FIRST_PAIR_BT = 5, ///< This and later band types encode two values (rather than four) with one code word.
79 ESC_BT = 11, ///< Spectral data are coded with an escape sequence.
80 NOISE_BT = 13, ///< Spectral data are scaled white noise not coded in the bitstream.
81 INTENSITY_BT2 = 14, ///< Scalefactor data are intensity stereo positions.
82 INTENSITY_BT = 15, ///< Scalefactor data are intensity stereo positions.
85 #define IS_CODEBOOK_UNSIGNED(x) ((x - 1) & 10)
87 enum ChannelPosition {
89 AAC_CHANNEL_FRONT = 1,
97 * The point during decoding at which channel coupling is applied.
101 BETWEEN_TNS_AND_IMDCT,
106 * Output configuration status
109 OC_NONE, ///< Output unconfigured
110 OC_TRIAL_PCE, ///< Output configuration under trial specified by an inband PCE
111 OC_TRIAL_FRAME, ///< Output configuration under trial specified by a frame header
112 OC_GLOBAL_HDR, ///< Output configuration set in a global header but not yet locked
113 OC_LOCKED, ///< Output configuration locked in place
116 typedef struct OutputConfiguration {
117 MPEG4AudioConfig m4ac;
118 uint8_t layout_map[MAX_ELEM_ID*4][3];
121 uint64_t channel_layout;
122 enum OCStatus status;
123 } OutputConfiguration;
128 typedef struct PredictorState {
137 #define MAX_PREDICTORS 672
139 #define SCALE_DIV_512 36 ///< scalefactor difference that corresponds to scale difference in 512 times
140 #define SCALE_ONE_POS 140 ///< scalefactor index that corresponds to scale=1.0
141 #define SCALE_MAX_POS 255 ///< scalefactor index maximum value
142 #define SCALE_MAX_DIFF 60 ///< maximum scalefactor difference allowed by standard
143 #define SCALE_DIFF_ZERO 60 ///< codebook index corresponding to zero scalefactor indices difference
144 #define POW_SF2_ZERO 200 ///< ff_aac_pow2sf_tab index corresponding to pow(2, 0);
147 * Long Term Prediction
149 typedef struct LongTermPrediction {
153 int8_t used[MAX_LTP_LONG_SFB];
154 } LongTermPrediction;
157 * Individual Channel Stream
159 typedef struct IndividualChannelStream {
160 uint8_t max_sfb; ///< number of scalefactor bands per group
161 enum WindowSequence window_sequence[2];
162 uint8_t use_kb_window[2]; ///< If set, use Kaiser-Bessel window, otherwise use a sinus window.
163 int num_window_groups;
164 uint8_t group_len[8];
165 LongTermPrediction ltp;
166 const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
167 const uint8_t *swb_sizes; ///< table of scalefactor band sizes for a particular window
168 int num_swb; ///< number of scalefactor window bands
171 int predictor_present;
172 int predictor_initialized;
173 int predictor_reset_group;
174 uint8_t prediction_used[41];
175 } IndividualChannelStream;
178 * Temporal Noise Shaping
180 typedef struct TemporalNoiseShaping {
186 float coef[8][4][TNS_MAX_ORDER];
187 } TemporalNoiseShaping;
190 * Dynamic Range Control - decoded from the bitstream but not processed further.
192 typedef struct DynamicRangeControl {
193 int pce_instance_tag; ///< Indicates with which program the DRC info is associated.
194 int dyn_rng_sgn[17]; ///< DRC sign information; 0 - positive, 1 - negative
195 int dyn_rng_ctl[17]; ///< DRC magnitude information
196 int exclude_mask[MAX_CHANNELS]; ///< Channels to be excluded from DRC processing.
197 int band_incr; ///< Number of DRC bands greater than 1 having DRC info.
198 int interpolation_scheme; ///< Indicates the interpolation scheme used in the SBR QMF domain.
199 int band_top[17]; ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
200 int prog_ref_level; /**< A reference level for the long-term program audio level for all
203 } DynamicRangeControl;
205 typedef struct Pulse {
213 * coupling parameters
215 typedef struct ChannelCoupling {
216 enum CouplingPoint coupling_point; ///< The point during decoding at which coupling is applied.
217 int num_coupled; ///< number of target elements
218 enum RawDataBlockType type[8]; ///< Type of channel element to be coupled - SCE or CPE.
219 int id_select[8]; ///< element id
220 int ch_select[8]; /**< [0] shared list of gains; [1] list of gains for right channel;
221 * [2] list of gains for left channel; [3] lists of gains for both channels
227 * Single Channel Element - used for both SCE and LFE elements.
229 typedef struct SingleChannelElement {
230 IndividualChannelStream ics;
231 TemporalNoiseShaping tns;
233 enum BandType band_type[128]; ///< band types
234 int band_type_run_end[120]; ///< band type run end points
235 float sf[120]; ///< scalefactors
236 int sf_idx[128]; ///< scalefactor indices (used by encoder)
237 uint8_t zeroes[128]; ///< band is not coded (used by encoder)
238 DECLARE_ALIGNED(32, float, coeffs)[1024]; ///< coefficients for IMDCT
239 DECLARE_ALIGNED(32, float, saved)[1024]; ///< overlap
240 DECLARE_ALIGNED(32, float, ret)[2048]; ///< PCM output
241 DECLARE_ALIGNED(16, float, ltp_state)[3072]; ///< time signal for LTP
242 PredictorState predictor_state[MAX_PREDICTORS];
243 } SingleChannelElement;
246 * channel element - generic struct for SCE/CPE/CCE/LFE
248 typedef struct ChannelElement {
250 int common_window; ///< Set if channels share a common 'IndividualChannelStream' in bitstream.
251 int ms_mode; ///< Signals mid/side stereo flags coding mode (used by encoder)
252 uint8_t ms_mask[128]; ///< Set if mid/side stereo is used for each scalefactor window band
254 SingleChannelElement ch[2];
256 ChannelCoupling coup;
257 SpectralBandReplication sbr;
263 typedef struct AACContext {
264 AVCodecContext *avctx;
267 int is_saved; ///< Set if elements have stored overlap from previous frame.
268 DynamicRangeControl che_drc;
271 * @name Channel element related data
274 ChannelElement *che[4][MAX_ELEM_ID];
275 ChannelElement *tag_che_map[4][MAX_ELEM_ID];
280 * @name temporary aligned temporary buffers
281 * (We do not want to have these on the stack.)
284 DECLARE_ALIGNED(32, float, buf_mdct)[1024];
288 * @name Computed / set up during initialization
292 FFTContext mdct_small;
295 FmtConvertContext fmt_conv;
296 AVFloatDSPContext fdsp;
301 * @name Members used for output interleaving
304 float *output_data[MAX_CHANNELS]; ///< Points to each element's 'ret' buffer (PCM output).
307 DECLARE_ALIGNED(32, float, temp)[128];
309 OutputConfiguration oc[2];
312 #endif /* AVCODEC_AAC_H */