Statistics
| Branch: | Revision:

ffmpeg / libavcodec / aac.h @ e22910b2

History | View | Annotate | Download (9.87 KB)

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
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

    
48
enum 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

    
59
enum 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

    
68
enum WindowSequence {
69
    ONLY_LONG_SEQUENCE,
70
    LONG_START_SEQUENCE,
71
    EIGHT_SHORT_SEQUENCE,
72
    LONG_STOP_SEQUENCE,
73
};
74

    
75
enum 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

    
86
enum ChannelPosition {
87
    AAC_CHANNEL_FRONT = 1,
88
    AAC_CHANNEL_SIDE  = 2,
89
    AAC_CHANNEL_BACK  = 3,
90
    AAC_CHANNEL_LFE   = 4,
91
    AAC_CHANNEL_CC    = 5,
92
};
93

    
94
/**
95
 * The point during decoding at which channel coupling is applied.
96
 */
97
enum CouplingPoint {
98
    BEFORE_TNS,
99
    BETWEEN_TNS_AND_IMDCT,
100
    AFTER_IMDCT = 3,
101
};
102

    
103
/**
104
 * Output configuration status
105
 */
106
enum OCStatus {
107
    OC_NONE,        //< Output unconfigured
108
    OC_TRIAL_PCE,   //< Output configuration under trial specified by an inband PCE
109
    OC_TRIAL_FRAME, //< Output configuration under trial specified by a frame header
110
    OC_GLOBAL_HDR,  //< Output configuration set in a global header but not yet locked
111
    OC_LOCKED,      //< Output configuration locked in place
112
};
113

    
114
/**
115
 * Predictor State
116
 */
117
typedef struct {
118
    float cor0;
119
    float cor1;
120
    float var0;
121
    float var1;
122
    float r0;
123
    float r1;
124
} PredictorState;
125

    
126
#define MAX_PREDICTORS 672
127

    
128
#define SCALE_DIV_512    36    ///< scalefactor difference that corresponds to scale difference in 512 times
129
#define SCALE_ONE_POS   140    ///< scalefactor index that corresponds to scale=1.0
130
#define SCALE_MAX_POS   255    ///< scalefactor index maximum value
131
#define SCALE_MAX_DIFF   60    ///< maximum scalefactor difference allowed by standard
132
#define SCALE_DIFF_ZERO  60    ///< codebook index corresponding to zero scalefactor indices difference
133

    
134
/**
135
 * Long Term Prediction
136
 */
137
typedef struct {
138
    int8_t present;
139
    int16_t lag;
140
    float coef;
141
    int8_t used[MAX_LTP_LONG_SFB];
142
} LongTermPrediction;
143

    
144
/**
145
 * Individual Channel Stream
146
 */
147
typedef struct {
148
    uint8_t max_sfb;            ///< number of scalefactor bands per group
149
    enum WindowSequence window_sequence[2];
150
    uint8_t use_kb_window[2];   ///< If set, use Kaiser-Bessel window, otherwise use a sinus window.
151
    int num_window_groups;
152
    uint8_t group_len[8];
153
    LongTermPrediction ltp;
154
    const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
155
    const uint8_t *swb_sizes;   ///< table of scalefactor band sizes for a particular window
156
    int num_swb;                ///< number of scalefactor window bands
157
    int num_windows;
158
    int tns_max_bands;
159
    int predictor_present;
160
    int predictor_initialized;
161
    int predictor_reset_group;
162
    uint8_t prediction_used[41];
163
} IndividualChannelStream;
164

    
165
/**
166
 * Temporal Noise Shaping
167
 */
168
typedef struct {
169
    int present;
170
    int n_filt[8];
171
    int length[8][4];
172
    int direction[8][4];
173
    int order[8][4];
174
    float coef[8][4][TNS_MAX_ORDER];
175
} TemporalNoiseShaping;
176

    
177
/**
178
 * Dynamic Range Control - decoded from the bitstream but not processed further.
179
 */
180
typedef struct {
181
    int pce_instance_tag;                           ///< Indicates with which program the DRC info is associated.
182
    int dyn_rng_sgn[17];                            ///< DRC sign information; 0 - positive, 1 - negative
183
    int dyn_rng_ctl[17];                            ///< DRC magnitude information
184
    int exclude_mask[MAX_CHANNELS];                 ///< Channels to be excluded from DRC processing.
185
    int band_incr;                                  ///< Number of DRC bands greater than 1 having DRC info.
186
    int interpolation_scheme;                       ///< Indicates the interpolation scheme used in the SBR QMF domain.
187
    int band_top[17];                               ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
188
    int prog_ref_level;                             /**< A reference level for the long-term program audio level for all
189
                                                     *   channels combined.
190
                                                     */
191
} DynamicRangeControl;
192

    
193
typedef struct {
194
    int num_pulse;
195
    int start;
196
    int pos[4];
197
    int amp[4];
198
} Pulse;
199

    
200
/**
201
 * coupling parameters
202
 */
203
typedef struct {
204
    enum CouplingPoint coupling_point;  ///< The point during decoding at which coupling is applied.
205
    int num_coupled;       ///< number of target elements
206
    enum RawDataBlockType type[8];   ///< Type of channel element to be coupled - SCE or CPE.
207
    int id_select[8];      ///< element id
208
    int ch_select[8];      /**< [0] shared list of gains; [1] list of gains for right channel;
209
                            *   [2] list of gains for left channel; [3] lists of gains for both channels
210
                            */
211
    float gain[16][120];
212
} ChannelCoupling;
213

    
214
/**
215
 * Single Channel Element - used for both SCE and LFE elements.
216
 */
217
typedef struct {
218
    IndividualChannelStream ics;
219
    TemporalNoiseShaping tns;
220
    Pulse pulse;
221
    enum BandType band_type[128];                   ///< band types
222
    int band_type_run_end[120];                     ///< band type run end points
223
    float sf[120];                                  ///< scalefactors
224
    int sf_idx[128];                                ///< scalefactor indices (used by encoder)
225
    uint8_t zeroes[128];                            ///< band is not coded (used by encoder)
226
    DECLARE_ALIGNED(16, float,   coeffs)[1024];     ///< coefficients for IMDCT
227
    DECLARE_ALIGNED(16, float,   saved)[1024];      ///< overlap
228
    DECLARE_ALIGNED(16, float,   ret)[2048];        ///< PCM output
229
    DECLARE_ALIGNED(16, int16_t, ltp_state)[3072];  ///< time signal for LTP
230
    PredictorState predictor_state[MAX_PREDICTORS];
231
} SingleChannelElement;
232

    
233
/**
234
 * channel element - generic struct for SCE/CPE/CCE/LFE
235
 */
236
typedef struct {
237
    // CPE specific
238
    int common_window;        ///< Set if channels share a common 'IndividualChannelStream' in bitstream.
239
    int     ms_mode;          ///< Signals mid/side stereo flags coding mode (used by encoder)
240
    uint8_t ms_mask[128];     ///< Set if mid/side stereo is used for each scalefactor window band
241
    // shared
242
    SingleChannelElement ch[2];
243
    // CCE specific
244
    ChannelCoupling coup;
245
    SpectralBandReplication sbr;
246
} ChannelElement;
247

    
248
/**
249
 * main AAC context
250
 */
251
typedef struct {
252
    AVCodecContext *avctx;
253

    
254
    MPEG4AudioConfig m4ac;
255

    
256
    int is_saved;                 ///< Set if elements have stored overlap from previous frame.
257
    DynamicRangeControl che_drc;
258

    
259
    /**
260
     * @defgroup elements Channel element related data.
261
     * @{
262
     */
263
    enum ChannelPosition che_pos[4][MAX_ELEM_ID]; /**< channel element channel mapping with the
264
                                                   *   first index as the first 4 raw data block types
265
                                                   */
266
    ChannelElement          *che[4][MAX_ELEM_ID];
267
    ChannelElement  *tag_che_map[4][MAX_ELEM_ID];
268
    int tags_mapped;
269
    /** @} */
270

    
271
    /**
272
     * @defgroup temporary aligned temporary buffers (We do not want to have these on the stack.)
273
     * @{
274
     */
275
    DECLARE_ALIGNED(16, float, buf_mdct)[1024];
276
    /** @} */
277

    
278
    /**
279
     * @defgroup tables   Computed / set up during initialization.
280
     * @{
281
     */
282
    FFTContext mdct;
283
    FFTContext mdct_small;
284
    FFTContext mdct_ltp;
285
    DSPContext dsp;
286
    FmtConvertContext fmt_conv;
287
    int random_state;
288
    /** @} */
289

    
290
    /**
291
     * @defgroup output   Members used for output interleaving.
292
     * @{
293
     */
294
    float *output_data[MAX_CHANNELS];                 ///< Points to each element's 'ret' buffer (PCM output).
295
    float sf_scale;                                   ///< Pre-scale for correct IMDCT and dsp.float_to_int16.
296
    int sf_offset;                                    ///< offset into pow2sf_tab as appropriate for dsp.float_to_int16
297
    /** @} */
298

    
299
    DECLARE_ALIGNED(16, float, temp)[128];
300

    
301
    enum OCStatus output_configured;
302
} AACContext;
303

    
304
#endif /* AVCODEC_AAC_H */