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 FFmpeg.
7 *
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.
12 *
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.
17 *
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
21 */
22
23/**
24 * @file libavcodec/aac.h
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 "libavutil/internal.h"
34#include "avcodec.h"
35#include "dsputil.h"
36#include "mpeg4audio.h"
37
38#include <stdint.h>
39
40#define AAC_INIT_VLC_STATIC(num, size) \
41    INIT_VLC_STATIC(&vlc_spectral[num], 6, ff_aac_spectral_sizes[num], \
42         ff_aac_spectral_bits[num], sizeof( ff_aac_spectral_bits[num][0]), sizeof( ff_aac_spectral_bits[num][0]), \
43        ff_aac_spectral_codes[num], sizeof(ff_aac_spectral_codes[num][0]), sizeof(ff_aac_spectral_codes[num][0]), \
44        size);
45
46#define MAX_CHANNELS 64
47#define MAX_ELEM_ID 16
48
49#define TNS_MAX_ORDER 20
50
51enum AudioObjectType {
52    AOT_NULL,
53                               // Support?                Name
54    AOT_AAC_MAIN,              ///< Y                       Main
55    AOT_AAC_LC,                ///< Y                       Low Complexity
56    AOT_AAC_SSR,               ///< N (code in SoC repo)    Scalable Sample Rate
57    AOT_AAC_LTP,               ///< N (code in SoC repo)    Long Term Prediction
58    AOT_SBR,                   ///< N (in progress)         Spectral Band Replication
59    AOT_AAC_SCALABLE,          ///< N                       Scalable
60    AOT_TWINVQ,                ///< N                       Twin Vector Quantizer
61    AOT_CELP,                  ///< N                       Code Excited Linear Prediction
62    AOT_HVXC,                  ///< N                       Harmonic Vector eXcitation Coding
63    AOT_TTSI             = 12, ///< N                       Text-To-Speech Interface
64    AOT_MAINSYNTH,             ///< N                       Main Synthesis
65    AOT_WAVESYNTH,             ///< N                       Wavetable Synthesis
66    AOT_MIDI,                  ///< N                       General MIDI
67    AOT_SAFX,                  ///< N                       Algorithmic Synthesis and Audio Effects
68    AOT_ER_AAC_LC,             ///< N                       Error Resilient Low Complexity
69    AOT_ER_AAC_LTP       = 19, ///< N                       Error Resilient Long Term Prediction
70    AOT_ER_AAC_SCALABLE,       ///< N                       Error Resilient Scalable
71    AOT_ER_TWINVQ,             ///< N                       Error Resilient Twin Vector Quantizer
72    AOT_ER_BSAC,               ///< N                       Error Resilient Bit-Sliced Arithmetic Coding
73    AOT_ER_AAC_LD,             ///< N                       Error Resilient Low Delay
74    AOT_ER_CELP,               ///< N                       Error Resilient Code Excited Linear Prediction
75    AOT_ER_HVXC,               ///< N                       Error Resilient Harmonic Vector eXcitation Coding
76    AOT_ER_HILN,               ///< N                       Error Resilient Harmonic and Individual Lines plus Noise
77    AOT_ER_PARAM,              ///< N                       Error Resilient Parametric
78    AOT_SSC,                   ///< N                       SinuSoidal Coding
79};
80
81enum RawDataBlockType {
82    TYPE_SCE,
83    TYPE_CPE,
84    TYPE_CCE,
85    TYPE_LFE,
86    TYPE_DSE,
87    TYPE_PCE,
88    TYPE_FIL,
89    TYPE_END,
90};
91
92enum ExtensionPayloadID {
93    EXT_FILL,
94    EXT_FILL_DATA,
95    EXT_DATA_ELEMENT,
96    EXT_DYNAMIC_RANGE = 0xb,
97    EXT_SBR_DATA      = 0xd,
98    EXT_SBR_DATA_CRC  = 0xe,
99};
100
101enum WindowSequence {
102    ONLY_LONG_SEQUENCE,
103    LONG_START_SEQUENCE,
104    EIGHT_SHORT_SEQUENCE,
105    LONG_STOP_SEQUENCE,
106};
107
108enum BandType {
109    ZERO_BT        = 0,     ///< Scalefactors and spectral data are all zero.
110    FIRST_PAIR_BT  = 5,     ///< This and later band types encode two values (rather than four) with one code word.
111    ESC_BT         = 11,    ///< Spectral data are coded with an escape sequence.
112    NOISE_BT       = 13,    ///< Spectral data are scaled white noise not coded in the bitstream.
113    INTENSITY_BT2  = 14,    ///< Scalefactor data are intensity stereo positions.
114    INTENSITY_BT   = 15,    ///< Scalefactor data are intensity stereo positions.
115};
116
117#define IS_CODEBOOK_UNSIGNED(x) ((x - 1) & 10)
118
119enum ChannelPosition {
120    AAC_CHANNEL_FRONT = 1,
121    AAC_CHANNEL_SIDE  = 2,
122    AAC_CHANNEL_BACK  = 3,
123    AAC_CHANNEL_LFE   = 4,
124    AAC_CHANNEL_CC    = 5,
125};
126
127/**
128 * The point during decoding at which channel coupling is applied.
129 */
130enum CouplingPoint {
131    BEFORE_TNS,
132    BETWEEN_TNS_AND_IMDCT,
133    AFTER_IMDCT = 3,
134};
135
136/**
137 * Predictor State
138 */
139typedef struct {
140    float cor0;
141    float cor1;
142    float var0;
143    float var1;
144    float r0;
145    float r1;
146} PredictorState;
147
148#define MAX_PREDICTORS 672
149
150/**
151 * Individual Channel Stream
152 */
153typedef struct {
154    uint8_t max_sfb;            ///< number of scalefactor bands per group
155    enum WindowSequence window_sequence[2];
156    uint8_t use_kb_window[2];   ///< If set, use Kaiser-Bessel window, otherwise use a sinus window.
157    int num_window_groups;
158    uint8_t group_len[8];
159    const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
160    int num_swb;                ///< number of scalefactor window bands
161    int num_windows;
162    int tns_max_bands;
163    int predictor_present;
164    int predictor_initialized;
165    int predictor_reset_group;
166    uint8_t prediction_used[41];
167} IndividualChannelStream;
168
169/**
170 * Temporal Noise Shaping
171 */
172typedef struct {
173    int present;
174    int n_filt[8];
175    int length[8][4];
176    int direction[8][4];
177    int order[8][4];
178    float coef[8][4][TNS_MAX_ORDER];
179} TemporalNoiseShaping;
180
181/**
182 * Dynamic Range Control - decoded from the bitstream but not processed further.
183 */
184typedef struct {
185    int pce_instance_tag;                           ///< Indicates with which program the DRC info is associated.
186    int dyn_rng_sgn[17];                            ///< DRC sign information; 0 - positive, 1 - negative
187    int dyn_rng_ctl[17];                            ///< DRC magnitude information
188    int exclude_mask[MAX_CHANNELS];                 ///< Channels to be excluded from DRC processing.
189    int band_incr;                                  ///< Number of DRC bands greater than 1 having DRC info.
190    int interpolation_scheme;                       ///< Indicates the interpolation scheme used in the SBR QMF domain.
191    int band_top[17];                               ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
192    int prog_ref_level;                             /**< A reference level for the long-term program audio level for all
193                                                     *   channels combined.
194                                                     */
195} DynamicRangeControl;
196
197typedef struct {
198    int num_pulse;
199    int pos[4];
200    int amp[4];
201} Pulse;
202
203/**
204 * coupling parameters
205 */
206typedef struct {
207    enum CouplingPoint coupling_point;  ///< The point during decoding at which coupling is applied.
208    int num_coupled;       ///< number of target elements
209    enum RawDataBlockType type[8];   ///< Type of channel element to be coupled - SCE or CPE.
210    int id_select[8];      ///< element id
211    int ch_select[8];      /**< [0] shared list of gains; [1] list of gains for right channel;
212                            *   [2] list of gains for left channel; [3] lists of gains for both channels
213                            */
214    float gain[16][120];
215} ChannelCoupling;
216
217/**
218 * Single Channel Element - used for both SCE and LFE elements.
219 */
220typedef struct {
221    IndividualChannelStream ics;
222    TemporalNoiseShaping tns;
223    enum BandType band_type[120];             ///< band types
224    int band_type_run_end[120];               ///< band type run end points
225    float sf[120];                            ///< scalefactors
226    DECLARE_ALIGNED_16(float, coeffs[1024]);  ///< coefficients for IMDCT
227    DECLARE_ALIGNED_16(float, saved[512]);    ///< overlap
228    DECLARE_ALIGNED_16(float, ret[1024]);     ///< PCM output
229    PredictorState predictor_state[MAX_PREDICTORS];
230} SingleChannelElement;
231
232/**
233 * channel element - generic struct for SCE/CPE/CCE/LFE
234 */
235typedef struct {
236    // CPE specific
237    uint8_t ms_mask[120];     ///< Set if mid/side stereo is used for each scalefactor window band
238    // shared
239    SingleChannelElement ch[2];
240    // CCE specific
241    ChannelCoupling coup;
242} ChannelElement;
243
244/**
245 * main AAC context
246 */
247typedef struct {
248    AVCodecContext * avccontext;
249
250    MPEG4AudioConfig m4ac;
251
252    int is_saved;                 ///< Set if elements have stored overlap from previous frame.
253    DynamicRangeControl che_drc;
254
255    /**
256     * @defgroup elements Channel element related data.
257     * @{
258     */
259    enum ChannelPosition che_pos[4][MAX_ELEM_ID]; /**< channel element channel mapping with the
260                                                   *   first index as the first 4 raw data block types
261                                                   */
262    ChannelElement * che[4][MAX_ELEM_ID];
263    ChannelElement * tag_che_map[4][MAX_ELEM_ID];
264    int tags_mapped;
265    /** @} */
266
267    /**
268     * @defgroup temporary aligned temporary buffers (We do not want to have these on the stack.)
269     * @{
270     */
271    DECLARE_ALIGNED_16(float, buf_mdct[1024]);
272    /** @} */
273
274    /**
275     * @defgroup tables   Computed / set up during initialization.
276     * @{
277     */
278    MDCTContext mdct;
279    MDCTContext mdct_small;
280    DSPContext dsp;
281    int random_state;
282    /** @} */
283
284    /**
285     * @defgroup output   Members used for output interleaving.
286     * @{
287     */
288    float *output_data[MAX_CHANNELS];                 ///< Points to each element's 'ret' buffer (PCM output).
289    float add_bias;                                   ///< offset for dsp.float_to_int16
290    float sf_scale;                                   ///< Pre-scale for correct IMDCT and dsp.float_to_int16.
291    int sf_offset;                                    ///< offset into pow2sf_tab as appropriate for dsp.float_to_int16
292    /** @} */
293
294    DECLARE_ALIGNED(16, float, temp[128]);
295} AACContext;
296
297#endif /* AVCODEC_AAC_H */
298