Statistics
| Branch: | Revision:

ffmpeg / libavcodec / aac.h @ 78e65cd7

History | View | Annotate | Download (9.4 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 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 "avcodec.h"
34
#include "dsputil.h"
35
#include "mpeg4audio.h"
36

    
37
#include <stdint.h>
38

    
39
#define AAC_INIT_VLC_STATIC(num, size) \
40
    INIT_VLC_STATIC(&vlc_spectral[num], 6, ff_aac_spectral_sizes[num], \
41
         ff_aac_spectral_bits[num], sizeof( ff_aac_spectral_bits[num][0]), sizeof( ff_aac_spectral_bits[num][0]), \
42
        ff_aac_spectral_codes[num], sizeof(ff_aac_spectral_codes[num][0]), sizeof(ff_aac_spectral_codes[num][0]), \
43
        size);
44

    
45
#define MAX_CHANNELS 64
46
#define MAX_ELEM_ID 16
47

    
48
#define TNS_MAX_ORDER 20
49

    
50
enum RawDataBlockType {
51
    TYPE_SCE,
52
    TYPE_CPE,
53
    TYPE_CCE,
54
    TYPE_LFE,
55
    TYPE_DSE,
56
    TYPE_PCE,
57
    TYPE_FIL,
58
    TYPE_END,
59
};
60

    
61
enum ExtensionPayloadID {
62
    EXT_FILL,
63
    EXT_FILL_DATA,
64
    EXT_DATA_ELEMENT,
65
    EXT_DYNAMIC_RANGE = 0xb,
66
    EXT_SBR_DATA      = 0xd,
67
    EXT_SBR_DATA_CRC  = 0xe,
68
};
69

    
70
enum WindowSequence {
71
    ONLY_LONG_SEQUENCE,
72
    LONG_START_SEQUENCE,
73
    EIGHT_SHORT_SEQUENCE,
74
    LONG_STOP_SEQUENCE,
75
};
76

    
77
enum BandType {
78
    ZERO_BT        = 0,     ///< Scalefactors and spectral data are all zero.
79
    FIRST_PAIR_BT  = 5,     ///< This and later band types encode two values (rather than four) with one code word.
80
    ESC_BT         = 11,    ///< Spectral data are coded with an escape sequence.
81
    NOISE_BT       = 13,    ///< Spectral data are scaled white noise not coded in the bitstream.
82
    INTENSITY_BT2  = 14,    ///< Scalefactor data are intensity stereo positions.
83
    INTENSITY_BT   = 15,    ///< Scalefactor data are intensity stereo positions.
84
};
85

    
86
#define IS_CODEBOOK_UNSIGNED(x) ((x - 1) & 10)
87

    
88
enum ChannelPosition {
89
    AAC_CHANNEL_FRONT = 1,
90
    AAC_CHANNEL_SIDE  = 2,
91
    AAC_CHANNEL_BACK  = 3,
92
    AAC_CHANNEL_LFE   = 4,
93
    AAC_CHANNEL_CC    = 5,
94
};
95

    
96
/**
97
 * The point during decoding at which channel coupling is applied.
98
 */
99
enum CouplingPoint {
100
    BEFORE_TNS,
101
    BETWEEN_TNS_AND_IMDCT,
102
    AFTER_IMDCT = 3,
103
};
104

    
105
/**
106
 * Predictor State
107
 */
108
typedef struct {
109
    float cor0;
110
    float cor1;
111
    float var0;
112
    float var1;
113
    float r0;
114
    float r1;
115
} PredictorState;
116

    
117
#define MAX_PREDICTORS 672
118

    
119
#define SCALE_DIV_512    36    ///< scalefactor difference that corresponds to scale difference in 512 times
120
#define SCALE_ONE_POS   140    ///< scalefactor index that corresponds to scale=1.0
121
#define SCALE_MAX_POS   255    ///< scalefactor index maximum value
122
#define SCALE_MAX_DIFF   60    ///< maximum scalefactor difference allowed by standard
123
#define SCALE_DIFF_ZERO  60    ///< codebook index corresponding to zero scalefactor indices difference
124

    
125
/**
126
 * Individual Channel Stream
127
 */
128
typedef struct {
129
    uint8_t max_sfb;            ///< number of scalefactor bands per group
130
    enum WindowSequence window_sequence[2];
131
    uint8_t use_kb_window[2];   ///< If set, use Kaiser-Bessel window, otherwise use a sinus window.
132
    int num_window_groups;
133
    uint8_t group_len[8];
134
    const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
135
    const uint8_t *swb_sizes;   ///< table of scalefactor band sizes for a particular window
136
    int num_swb;                ///< number of scalefactor window bands
137
    int num_windows;
138
    int tns_max_bands;
139
    int predictor_present;
140
    int predictor_initialized;
141
    int predictor_reset_group;
142
    uint8_t prediction_used[41];
143
} IndividualChannelStream;
144

    
145
/**
146
 * Temporal Noise Shaping
147
 */
148
typedef struct {
149
    int present;
150
    int n_filt[8];
151
    int length[8][4];
152
    int direction[8][4];
153
    int order[8][4];
154
    float coef[8][4][TNS_MAX_ORDER];
155
} TemporalNoiseShaping;
156

    
157
/**
158
 * Dynamic Range Control - decoded from the bitstream but not processed further.
159
 */
160
typedef struct {
161
    int pce_instance_tag;                           ///< Indicates with which program the DRC info is associated.
162
    int dyn_rng_sgn[17];                            ///< DRC sign information; 0 - positive, 1 - negative
163
    int dyn_rng_ctl[17];                            ///< DRC magnitude information
164
    int exclude_mask[MAX_CHANNELS];                 ///< Channels to be excluded from DRC processing.
165
    int band_incr;                                  ///< Number of DRC bands greater than 1 having DRC info.
166
    int interpolation_scheme;                       ///< Indicates the interpolation scheme used in the SBR QMF domain.
167
    int band_top[17];                               ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
168
    int prog_ref_level;                             /**< A reference level for the long-term program audio level for all
169
                                                     *   channels combined.
170
                                                     */
171
} DynamicRangeControl;
172

    
173
typedef struct {
174
    int num_pulse;
175
    int start;
176
    int pos[4];
177
    int amp[4];
178
} Pulse;
179

    
180
/**
181
 * coupling parameters
182
 */
183
typedef struct {
184
    enum CouplingPoint coupling_point;  ///< The point during decoding at which coupling is applied.
185
    int num_coupled;       ///< number of target elements
186
    enum RawDataBlockType type[8];   ///< Type of channel element to be coupled - SCE or CPE.
187
    int id_select[8];      ///< element id
188
    int ch_select[8];      /**< [0] shared list of gains; [1] list of gains for right channel;
189
                            *   [2] list of gains for left channel; [3] lists of gains for both channels
190
                            */
191
    float gain[16][120];
192
} ChannelCoupling;
193

    
194
/**
195
 * Single Channel Element - used for both SCE and LFE elements.
196
 */
197
typedef struct {
198
    IndividualChannelStream ics;
199
    TemporalNoiseShaping tns;
200
    Pulse pulse;
201
    enum BandType band_type[128];             ///< band types
202
    int band_type_run_end[120];               ///< band type run end points
203
    float sf[120];                            ///< scalefactors
204
    int sf_idx[128];                          ///< scalefactor indices (used by encoder)
205
    uint8_t zeroes[128];                      ///< band is not coded (used by encoder)
206
    DECLARE_ALIGNED_16(float, coeffs[1024]);  ///< coefficients for IMDCT
207
    DECLARE_ALIGNED_16(float, saved[1024]);   ///< overlap
208
    DECLARE_ALIGNED_16(float, ret[1024]);     ///< PCM output
209
    PredictorState predictor_state[MAX_PREDICTORS];
210
} SingleChannelElement;
211

    
212
/**
213
 * channel element - generic struct for SCE/CPE/CCE/LFE
214
 */
215
typedef struct {
216
    // CPE specific
217
    int common_window;        ///< Set if channels share a common 'IndividualChannelStream' in bitstream.
218
    int     ms_mode;          ///< Signals mid/side stereo flags coding mode (used by encoder)
219
    uint8_t ms_mask[128];     ///< Set if mid/side stereo is used for each scalefactor window band
220
    // shared
221
    SingleChannelElement ch[2];
222
    // CCE specific
223
    ChannelCoupling coup;
224
} ChannelElement;
225

    
226
/**
227
 * main AAC context
228
 */
229
typedef struct {
230
    AVCodecContext * avccontext;
231

    
232
    MPEG4AudioConfig m4ac;
233

    
234
    int is_saved;                 ///< Set if elements have stored overlap from previous frame.
235
    DynamicRangeControl che_drc;
236

    
237
    /**
238
     * @defgroup elements Channel element related data.
239
     * @{
240
     */
241
    enum ChannelPosition che_pos[4][MAX_ELEM_ID]; /**< channel element channel mapping with the
242
                                                   *   first index as the first 4 raw data block types
243
                                                   */
244
    ChannelElement * che[4][MAX_ELEM_ID];
245
    ChannelElement * tag_che_map[4][MAX_ELEM_ID];
246
    int tags_mapped;
247
    /** @} */
248

    
249
    /**
250
     * @defgroup temporary aligned temporary buffers (We do not want to have these on the stack.)
251
     * @{
252
     */
253
    DECLARE_ALIGNED_16(float, buf_mdct[1024]);
254
    /** @} */
255

    
256
    /**
257
     * @defgroup tables   Computed / set up during initialization.
258
     * @{
259
     */
260
    MDCTContext mdct;
261
    MDCTContext mdct_small;
262
    DSPContext dsp;
263
    int random_state;
264
    /** @} */
265

    
266
    /**
267
     * @defgroup output   Members used for output interleaving.
268
     * @{
269
     */
270
    float *output_data[MAX_CHANNELS];                 ///< Points to each element's 'ret' buffer (PCM output).
271
    float add_bias;                                   ///< offset for dsp.float_to_int16
272
    float sf_scale;                                   ///< Pre-scale for correct IMDCT and dsp.float_to_int16.
273
    int sf_offset;                                    ///< offset into pow2sf_tab as appropriate for dsp.float_to_int16
274
    /** @} */
275

    
276
    DECLARE_ALIGNED(16, float, temp[128]);
277
} AACContext;
278

    
279
#endif /* AVCODEC_AAC_H */