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 FFmpeg.
8 * FFmpeg 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 * FFmpeg 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 FFmpeg; 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"
42 #define MAX_CHANNELS 64
43 #define MAX_ELEM_ID 16
45 #define TNS_MAX_ORDER 20
46 #define MAX_LTP_LONG_SFB 40
48 enum RawDataBlockType {
59 enum ExtensionPayloadID {
63 EXT_DYNAMIC_RANGE = 0xb,
65 EXT_SBR_DATA_CRC = 0xe,
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 RESERVED_BT = 12, ///< Band types following are encoded differently from others.
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 (out of phase).
82 INTENSITY_BT = 15, ///< Scalefactor data are intensity stereo positions (in phase).
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
145 #define NOISE_PRE 256 ///< preamble for NOISE_BT, put in bitstream with the first noise band
146 #define NOISE_PRE_BITS 9 ///< length of preamble
147 #define NOISE_OFFSET 90 ///< subtracted from global gain, used as offset for the preamble
150 * Long Term Prediction
152 typedef struct LongTermPrediction {
156 int8_t used[MAX_LTP_LONG_SFB];
157 } LongTermPrediction;
160 * Individual Channel Stream
162 typedef struct IndividualChannelStream {
163 uint8_t max_sfb; ///< number of scalefactor bands per group
164 enum WindowSequence window_sequence[2];
165 uint8_t use_kb_window[2]; ///< If set, use Kaiser-Bessel window, otherwise use a sine window.
166 int num_window_groups;
167 uint8_t group_len[8];
168 LongTermPrediction ltp;
169 const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
170 const uint8_t *swb_sizes; ///< table of scalefactor band sizes for a particular window
171 int num_swb; ///< number of scalefactor window bands
174 int predictor_present;
175 int predictor_initialized;
176 int predictor_reset_group;
177 uint8_t prediction_used[41];
178 } IndividualChannelStream;
181 * Temporal Noise Shaping
183 typedef struct TemporalNoiseShaping {
189 float coef[8][4][TNS_MAX_ORDER];
190 } TemporalNoiseShaping;
193 * Dynamic Range Control - decoded from the bitstream but not processed further.
195 typedef struct DynamicRangeControl {
196 int pce_instance_tag; ///< Indicates with which program the DRC info is associated.
197 int dyn_rng_sgn[17]; ///< DRC sign information; 0 - positive, 1 - negative
198 int dyn_rng_ctl[17]; ///< DRC magnitude information
199 int exclude_mask[MAX_CHANNELS]; ///< Channels to be excluded from DRC processing.
200 int band_incr; ///< Number of DRC bands greater than 1 having DRC info.
201 int interpolation_scheme; ///< Indicates the interpolation scheme used in the SBR QMF domain.
202 int band_top[17]; ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
203 int prog_ref_level; /**< A reference level for the long-term program audio level for all
206 } DynamicRangeControl;
208 typedef struct Pulse {
216 * coupling parameters
218 typedef struct ChannelCoupling {
219 enum CouplingPoint coupling_point; ///< The point during decoding at which coupling is applied.
220 int num_coupled; ///< number of target elements
221 enum RawDataBlockType type[8]; ///< Type of channel element to be coupled - SCE or CPE.
222 int id_select[8]; ///< element id
223 int ch_select[8]; /**< [0] shared list of gains; [1] list of gains for right channel;
224 * [2] list of gains for left channel; [3] lists of gains for both channels
230 * Single Channel Element - used for both SCE and LFE elements.
232 typedef struct SingleChannelElement {
233 IndividualChannelStream ics;
234 TemporalNoiseShaping tns;
236 enum BandType band_type[128]; ///< band types
237 int band_type_run_end[120]; ///< band type run end points
238 float sf[120]; ///< scalefactors
239 int sf_idx[128]; ///< scalefactor indices (used by encoder)
240 uint8_t zeroes[128]; ///< band is not coded (used by encoder)
241 float is_ener[128]; ///< Intensity stereo pos (used by encoder)
242 float pns_ener[128]; ///< Noise energy values (used by encoder)
243 DECLARE_ALIGNED(32, float, pcoeffs)[1024]; ///< coefficients for IMDCT, pristine
244 DECLARE_ALIGNED(32, float, coeffs)[1024]; ///< coefficients for IMDCT, maybe processed
245 DECLARE_ALIGNED(32, float, saved)[1536]; ///< overlap
246 DECLARE_ALIGNED(32, float, ret_buf)[2048]; ///< PCM output buffer
247 DECLARE_ALIGNED(16, float, ltp_state)[3072]; ///< time signal for LTP
248 PredictorState predictor_state[MAX_PREDICTORS];
249 float *ret; ///< PCM output
250 } SingleChannelElement;
253 * channel element - generic struct for SCE/CPE/CCE/LFE
255 typedef struct ChannelElement {
258 int common_window; ///< Set if channels share a common 'IndividualChannelStream' in bitstream.
259 int ms_mode; ///< Signals mid/side stereo flags coding mode (used by encoder)
260 uint8_t is_mode; ///< Set if any bands have been encoded using intensity stereo (used by encoder)
261 uint8_t ms_mask[128]; ///< Set if mid/side stereo is used for each scalefactor window band
262 uint8_t is_mask[128]; ///< Set if intensity stereo is used (used by encoder)
264 SingleChannelElement ch[2];
266 ChannelCoupling coup;
267 SpectralBandReplication sbr;
275 AVCodecContext *avctx;
278 int is_saved; ///< Set if elements have stored overlap from previous frame.
279 DynamicRangeControl che_drc;
282 * @name Channel element related data
285 ChannelElement *che[4][MAX_ELEM_ID];
286 ChannelElement *tag_che_map[4][MAX_ELEM_ID];
288 int warned_remapping_once;
292 * @name temporary aligned temporary buffers
293 * (We do not want to have these on the stack.)
296 DECLARE_ALIGNED(32, float, buf_mdct)[1024];
300 * @name Computed / set up during initialization
304 FFTContext mdct_small;
307 IMDCT15Context *mdct480;
308 AVFloatDSPContext *fdsp;
313 * @name Members used for output
316 SingleChannelElement *output_element[MAX_CHANNELS]; ///< Points to each SingleChannelElement
321 * @name Japanese DTV specific extension
324 int force_dmono_mode;///< 0->not dmono, 1->use first channel, 2->use second channel
325 int dmono_mode; ///< 0->not dmono, 1->use first channel, 2->use second channel
328 DECLARE_ALIGNED(32, float, temp)[128];
330 OutputConfiguration oc[2];
331 int warned_num_aac_frames;
333 /* aacdec functions pointers */
334 void (*imdct_and_windowing)(AACContext *ac, SingleChannelElement *sce);
335 void (*apply_ltp)(AACContext *ac, SingleChannelElement *sce);
336 void (*apply_tns)(float coef[1024], TemporalNoiseShaping *tns,
337 IndividualChannelStream *ics, int decode);
338 void (*windowing_and_mdct_ltp)(AACContext *ac, float *out,
339 float *in, IndividualChannelStream *ics);
340 void (*update_ltp)(AACContext *ac, SingleChannelElement *sce);
344 void ff_aacdec_init_mips(AACContext *c);
346 #endif /* AVCODEC_AAC_H */