ac3.h 6.59 KB
Newer Older
1
/*
2
 * Common code between the AC-3 encoder and decoder
3
 * Copyright (c) 2000, 2001, 2002 Fabrice Bellard
4
 *
5 6 7
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
8 9
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
10
 * version 2.1 of the License, or (at your option) any later version.
11
 *
12
 * FFmpeg is distributed in the hope that it will be useful,
13 14 15 16 17
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
18
 * License along with FFmpeg; if not, write to the Free Software
19
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20 21
 */

Michael Niedermayer's avatar
Michael Niedermayer committed
22
/**
23
 * @file libavcodec/ac3.h
24
 * Common code between the AC-3 encoder and decoder.
Michael Niedermayer's avatar
Michael Niedermayer committed
25 26
 */

27 28
#ifndef AVCODEC_AC3_H
#define AVCODEC_AC3_H
29

30 31
#include "ac3tab.h"

32 33 34
#define AC3_MAX_CODED_FRAME_SIZE 3840 /* in bytes */
#define AC3_MAX_CHANNELS 6 /* including LFE channel */

35
#define NB_BLOCKS 6 /* number of PCM blocks inside an AC-3 frame */
36 37 38 39 40 41 42 43 44 45
#define AC3_FRAME_SIZE (NB_BLOCKS * 256)

/* exponent encoding strategy */
#define EXP_REUSE 0
#define EXP_NEW   1

#define EXP_D15   1
#define EXP_D25   2
#define EXP_D45   3

46 47 48 49 50 51 52 53 54 55
/** Delta bit allocation strategy */
typedef enum {
    DBA_REUSE = 0,
    DBA_NEW,
    DBA_NONE,
    DBA_RESERVED
} AC3DeltaStrategy;

/** Channel mode (audio coding mode) */
typedef enum {
56 57 58 59 60 61 62 63
    AC3_CHMODE_DUALMONO = 0,
    AC3_CHMODE_MONO,
    AC3_CHMODE_STEREO,
    AC3_CHMODE_3F,
    AC3_CHMODE_2F1R,
    AC3_CHMODE_3F1R,
    AC3_CHMODE_2F2R,
    AC3_CHMODE_3F2R
64 65
} AC3ChannelMode;

66
typedef struct AC3BitAllocParameters {
67 68 69 70
    int sr_code;
    int sr_shift;
    int slow_gain, slow_decay, fast_decay, db_per_bit, floor;
    int cpl_fast_leak, cpl_slow_leak;
71 72
} AC3BitAllocParameters;

73 74 75 76 77 78 79 80 81 82
/**
 * @struct AC3HeaderInfo
 * Coded AC-3 header values up to the lfeon element, plus derived values.
 */
typedef struct {
    /** @defgroup coded Coded elements
     * @{
     */
    uint16_t sync_word;
    uint16_t crc1;
83
    uint8_t sr_code;
84 85 86
    uint8_t bitstream_id;
    uint8_t channel_mode;
    uint8_t lfe_on;
87
    uint8_t frame_type;
88
    int substreamid;                        ///< substream identification
89 90 91
    int center_mix_level;                   ///< Center mix level index
    int surround_mix_level;                 ///< Surround mix level index
    uint16_t channel_map;
92
    int num_blocks;                         ///< number of audio blocks
93 94 95 96 97
    /** @} */

    /** @defgroup derived Derived values
     * @{
     */
98
    uint8_t sr_shift;
99 100 101 102 103 104 105
    uint16_t sample_rate;
    uint32_t bit_rate;
    uint8_t channels;
    uint16_t frame_size;
    /** @} */
} AC3HeaderInfo;

106
typedef enum {
107 108 109 110 111
    EAC3_FRAME_TYPE_INDEPENDENT = 0,
    EAC3_FRAME_TYPE_DEPENDENT,
    EAC3_FRAME_TYPE_AC3_CONVERT,
    EAC3_FRAME_TYPE_RESERVED
} EAC3FrameType;
112

Måns Rullgård's avatar
Måns Rullgård committed
113
void ac3_common_init(void);
114 115 116 117 118 119 120 121 122 123 124 125 126

/**
 * Calculates the log power-spectral density of the input signal.
 * This gives a rough estimate of signal power in the frequency domain by using
 * the spectral envelope (exponents).  The psd is also separately grouped
 * into critical bands for use in the calculating the masking curve.
 * 128 units in psd = -6 dB.  The dbknee parameter in AC3BitAllocParameters
 * determines the reference level.
 *
 * @param[in]  exp        frequency coefficient exponents
 * @param[in]  start      starting bin location
 * @param[in]  end        ending bin location
 * @param[out] psd        signal power for each frequency bin
127
 * @param[out] band_psd   signal power for each critical band
128 129
 */
void ff_ac3_bit_alloc_calc_psd(int8_t *exp, int start, int end, int16_t *psd,
130
                               int16_t *band_psd);
131 132 133 134 135 136 137 138 139

/**
 * Calculates the masking curve.
 * First, the excitation is calculated using parameters in \p s and the signal
 * power in each critical band.  The excitation is compared with a predefined
 * hearing threshold table to produce the masking curve.  If delta bit
 * allocation information is provided, it is used for adjusting the masking
 * curve, usually to give a closer match to a better psychoacoustic model.
 *
140
 * @param[in]  s            adjustable bit allocation parameters
141
 * @param[in]  band_psd     signal power for each critical band
142 143 144 145
 * @param[in]  start        starting bin location
 * @param[in]  end          ending bin location
 * @param[in]  fast_gain    fast gain (estimated signal-to-mask ratio)
 * @param[in]  is_lfe       whether or not the channel being processed is the LFE
146 147 148 149 150
 * @param[in]  dba_mode     delta bit allocation mode (none, reuse, or new)
 * @param[in]  dba_nsegs    number of delta segments
 * @param[in]  dba_offsets  location offsets for each segment
 * @param[in]  dba_lengths  length of each segment
 * @param[in]  dba_values   delta bit allocation for each segment
151
 * @param[out] mask         calculated masking curve
152
 * @return returns 0 for success, non-zero for error
153
 */
154
int ff_ac3_bit_alloc_calc_mask(AC3BitAllocParameters *s, int16_t *band_psd,
155 156 157 158
                               int start, int end, int fast_gain, int is_lfe,
                               int dba_mode, int dba_nsegs, uint8_t *dba_offsets,
                               uint8_t *dba_lengths, uint8_t *dba_values,
                               int16_t *mask);
159 160 161 162 163 164 165 166 167 168 169

/**
 * Calculates bit allocation pointers.
 * The SNR is the difference between the masking curve and the signal.  AC-3
 * uses this value for each frequency bin to allocate bits.  The \p snroffset
 * parameter is a global adjustment to the SNR for all bins.
 *
 * @param[in]  mask       masking curve
 * @param[in]  psd        signal power for each frequency bin
 * @param[in]  start      starting bin location
 * @param[in]  end        ending bin location
170
 * @param[in]  snr_offset SNR adjustment
171
 * @param[in]  floor      noise floor
172
 * @param[in]  bap_tab    look-up table for bit allocation pointers
173 174 175
 * @param[out] bap        bit allocation pointers
 */
void ff_ac3_bit_alloc_calc_bap(int16_t *mask, int16_t *psd, int start, int end,
176 177
                               int snr_offset, int floor,
                               const uint8_t *bap_tab, uint8_t *bap);
178

Måns Rullgård's avatar
Måns Rullgård committed
179 180
void ac3_parametric_bit_allocation(AC3BitAllocParameters *s, uint8_t *bap,
                                   int8_t *exp, int start, int end,
181 182 183 184
                                   int snr_offset, int fast_gain, int is_lfe,
                                   int dba_mode, int dba_nsegs,
                                   uint8_t *dba_offsets, uint8_t *dba_lengths,
                                   uint8_t *dba_values);
185

186
#endif /* AVCODEC_AC3_H */