Statistics
| Branch: | Revision:

ffmpeg / libavcodec / aac.h @ 848a5815

History | View | Annotate | Download (9.79 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 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 FFMPEG_AAC_H
31
#define FFMPEG_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
enum AudioObjectType {
49
    AOT_NULL,
50
                               // Support?                Name
51
    AOT_AAC_MAIN,              ///< Y                       Main
52
    AOT_AAC_LC,                ///< Y                       Low Complexity
53
    AOT_AAC_SSR,               ///< N (code in SoC repo)    Scalable Sample Rate
54
    AOT_AAC_LTP,               ///< N (code in SoC repo)    Long Term Prediction
55
    AOT_SBR,                   ///< N (in progress)         Spectral Band Replication
56
    AOT_AAC_SCALABLE,          ///< N                       Scalable
57
    AOT_TWINVQ,                ///< N                       Twin Vector Quantizer
58
    AOT_CELP,                  ///< N                       Code Excited Linear Prediction
59
    AOT_HVXC,                  ///< N                       Harmonic Vector eXcitation Coding
60
    AOT_TTSI             = 12, ///< N                       Text-To-Speech Interface
61
    AOT_MAINSYNTH,             ///< N                       Main Synthesis
62
    AOT_WAVESYNTH,             ///< N                       Wavetable Synthesis
63
    AOT_MIDI,                  ///< N                       General MIDI
64
    AOT_SAFX,                  ///< N                       Algorithmic Synthesis and Audio Effects
65
    AOT_ER_AAC_LC,             ///< N                       Error Resilient Low Complexity
66
    AOT_ER_AAC_LTP       = 19, ///< N                       Error Resilient Long Term Prediction
67
    AOT_ER_AAC_SCALABLE,       ///< N                       Error Resilient Scalable
68
    AOT_ER_TWINVQ,             ///< N                       Error Resilient Twin Vector Quantizer
69
    AOT_ER_BSAC,               ///< N                       Error Resilient Bit-Sliced Arithmetic Coding
70
    AOT_ER_AAC_LD,             ///< N                       Error Resilient Low Delay
71
    AOT_ER_CELP,               ///< N                       Error Resilient Code Excited Linear Prediction
72
    AOT_ER_HVXC,               ///< N                       Error Resilient Harmonic Vector eXcitation Coding
73
    AOT_ER_HILN,               ///< N                       Error Resilient Harmonic and Individual Lines plus Noise
74
    AOT_ER_PARAM,              ///< N                       Error Resilient Parametric
75
    AOT_SSC,                   ///< N                       SinuSoidal Coding
76
};
77

    
78
enum RawDataBlockType {
79
    TYPE_SCE,
80
    TYPE_CPE,
81
    TYPE_CCE,
82
    TYPE_LFE,
83
    TYPE_DSE,
84
    TYPE_PCE,
85
    TYPE_FIL,
86
    TYPE_END,
87
};
88

    
89
enum ExtensionPayloadID {
90
    EXT_FILL,
91
    EXT_FILL_DATA,
92
    EXT_DATA_ELEMENT,
93
    EXT_DYNAMIC_RANGE = 0xb,
94
    EXT_SBR_DATA      = 0xd,
95
    EXT_SBR_DATA_CRC  = 0xe,
96
};
97

    
98
enum WindowSequence {
99
    ONLY_LONG_SEQUENCE,
100
    LONG_START_SEQUENCE,
101
    EIGHT_SHORT_SEQUENCE,
102
    LONG_STOP_SEQUENCE,
103
};
104

    
105
enum BandType {
106
    ZERO_BT        = 0,     ///< Scalefactors and spectral data are all zero.
107
    FIRST_PAIR_BT  = 5,     ///< This and later band types encode two values (rather than four) with one code word.
108
    ESC_BT         = 11,    ///< Spectral data are coded with an escape sequence.
109
    NOISE_BT       = 13,    ///< Spectral data are scaled white noise not coded in the bitstream.
110
    INTENSITY_BT2  = 14,    ///< Scalefactor data are intensity stereo positions.
111
    INTENSITY_BT   = 15,    ///< Scalefactor data are intensity stereo positions.
112
};
113

    
114
#define IS_CODEBOOK_UNSIGNED(x) ((x - 1) & 10)
115

    
116
enum ChannelPosition {
117
    AAC_CHANNEL_FRONT = 1,
118
    AAC_CHANNEL_SIDE  = 2,
119
    AAC_CHANNEL_BACK  = 3,
120
    AAC_CHANNEL_LFE   = 4,
121
    AAC_CHANNEL_CC    = 5,
122
};
123

    
124
/**
125
 * The point during decoding at which channel coupling is applied.
126
 */
127
enum CouplingPoint {
128
    BEFORE_TNS,
129
    BETWEEN_TNS_AND_IMDCT,
130
    AFTER_IMDCT = 3,
131
};
132

    
133
/**
134
 * Individual Channel Stream
135
 */
136
typedef struct {
137
    uint8_t max_sfb;            ///< number of scalefactor bands per group
138
    enum WindowSequence window_sequence[2];
139
    uint8_t use_kb_window[2];   ///< If set, use Kaiser-Bessel window, otherwise use a sinus window.
140
    int num_window_groups;
141
    uint8_t group_len[8];
142
    const uint16_t *swb_offset; ///< table of offsets to the lowest spectral coefficient of a scalefactor band, sfb, for a particular window
143
    int num_swb;                ///< number of scalefactor window bands
144
    int num_windows;
145
    int tns_max_bands;
146
} IndividualChannelStream;
147

    
148
/**
149
 * Dynamic Range Control - decoded from the bitstream but not processed further.
150
 */
151
typedef struct {
152
    int pce_instance_tag;                           ///< Indicates with which program the DRC info is associated.
153
    int dyn_rng_sgn[17];                            ///< DRC sign information; 0 - positive, 1 - negative
154
    int dyn_rng_ctl[17];                            ///< DRC magnitude information
155
    int exclude_mask[MAX_CHANNELS];                 ///< Channels to be excluded from DRC processing.
156
    int band_incr;                                  ///< Number of DRC bands greater than 1 having DRC info.
157
    int interpolation_scheme;                       ///< Indicates the interpolation scheme used in the SBR QMF domain.
158
    int band_top[17];                               ///< Indicates the top of the i-th DRC band in units of 4 spectral lines.
159
    int prog_ref_level;                             /**< A reference level for the long-term program audio level for all
160
                                                     *   channels combined.
161
                                                     */
162
} DynamicRangeControl;
163

    
164
typedef struct {
165
    int num_pulse;
166
    int pos[4];
167
    int amp[4];
168
} Pulse;
169

    
170
/**
171
 * coupling parameters
172
 */
173
typedef struct {
174
    enum CouplingPoint coupling_point;  ///< The point during decoding at which coupling is applied.
175
    int num_coupled;       ///< number of target elements
176
    enum RawDataBlockType type[8];   ///< Type of channel element to be coupled - SCE or CPE.
177
    int id_select[8];      ///< element id
178
    int ch_select[8];      /**< [0] shared list of gains; [1] list of gains for left channel;
179
                            *   [2] list of gains for right channel; [3] lists of gains for both channels
180
                            */
181
    float gain[16][120];
182
} ChannelCoupling;
183

    
184
/**
185
 * Single Channel Element - used for both SCE and LFE elements.
186
 */
187
typedef struct {
188
    IndividualChannelStream ics;
189
    TemporalNoiseShaping tns;
190
    enum BandType band_type[120];             ///< band types
191
    int band_type_run_end[120];               ///< band type run end points
192
    float sf[120];                            ///< scalefactors
193
    DECLARE_ALIGNED_16(float, coeffs[1024]);  ///< coefficients for IMDCT
194
    DECLARE_ALIGNED_16(float, saved[1024]);   ///< overlap
195
    DECLARE_ALIGNED_16(float, ret[1024]);     ///< PCM output
196
} SingleChannelElement;
197

    
198
/**
199
 * channel element - generic struct for SCE/CPE/CCE/LFE
200
 */
201
typedef struct {
202
    // CPE specific
203
    uint8_t ms_mask[120];     ///< Set if mid/side stereo is used for each scalefactor window band
204
    // shared
205
    SingleChannelElement ch[2];
206
    // CCE specific
207
    ChannelCoupling coup;
208
} ChannelElement;
209

    
210
/**
211
 * main AAC context
212
 */
213
typedef struct {
214
    AVCodecContext * avccontext;
215

    
216
    MPEG4AudioConfig m4ac;
217

    
218
    int is_saved;                 ///< Set if elements have stored overlap from previous frame.
219
    DynamicRangeControl che_drc;
220

    
221
    /**
222
     * @defgroup elements
223
     * @{
224
     */
225
    enum ChannelPosition che_pos[4][MAX_ELEM_ID]; /**< channel element channel mapping with the
226
                                                   *   first index as the first 4 raw data block types
227
                                                   */
228
    ChannelElement * che[4][MAX_ELEM_ID];
229
    /** @} */
230

    
231
    /**
232
     * @defgroup tables   Computed / set up during initialization.
233
     * @{
234
     */
235
    MDCTContext mdct;
236
    MDCTContext mdct_small;
237
    DSPContext dsp;
238
    int random_state;
239
    /** @} */
240

    
241
    /**
242
     * @defgroup output   Members used for output interleaving.
243
     * @{
244
     */
245
    float *output_data[MAX_CHANNELS];                 ///< Points to each element's 'ret' buffer (PCM output).
246
    float add_bias;                                   ///< offset for dsp.float_to_int16
247
    float sf_scale;                                   ///< Pre-scale for correct IMDCT and dsp.float_to_int16.
248
    int sf_offset;                                    ///< offset into pow2sf_tab as appropriate for dsp.float_to_int16
249
    /** @} */
250

    
251
} AACContext;
252

    
253
#endif /* FFMPEG_AAC_H */