ac3.h 6.09 KB
Newer Older
1 2 3 4
/*
 * Common code between AC3 encoder and decoder
 * Copyright (c) 2000, 2001, 2002 Fabrice Bellard.
 *
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 24 25 26
/**
 * @file ac3.h
 * Common code between AC3 encoder and decoder.
 */

27 28
#ifndef FFMPEG_AC3_H
#define FFMPEG_AC3_H
29

30 31
#include "ac3tab.h"

32 33 34 35 36 37 38 39 40 41 42 43 44 45
#define AC3_MAX_CODED_FRAME_SIZE 3840 /* in bytes */
#define AC3_MAX_CHANNELS 6 /* including LFE channel */

#define NB_BLOCKS 6 /* number of PCM blocks inside an AC3 frame */
#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 89 90 91 92
    /** @} */

    /** @defgroup derived Derived values
     * @{
     */
93
    uint8_t sr_shift;
94 95 96 97 98 99 100
    uint16_t sample_rate;
    uint32_t bit_rate;
    uint8_t channels;
    uint16_t frame_size;
    /** @} */
} AC3HeaderInfo;

101
typedef enum {
102 103 104 105 106
    EAC3_FRAME_TYPE_INDEPENDENT = 0,
    EAC3_FRAME_TYPE_DEPENDENT,
    EAC3_FRAME_TYPE_AC3_CONVERT,
    EAC3_FRAME_TYPE_RESERVED
} EAC3FrameType;
107

Måns Rullgård's avatar
Måns Rullgård committed
108
void ac3_common_init(void);
109 110 111 112 113 114 115 116 117 118 119 120 121

/**
 * 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
122
 * @param[out] band_psd   signal power for each critical band
123 124
 */
void ff_ac3_bit_alloc_calc_psd(int8_t *exp, int start, int end, int16_t *psd,
125
                               int16_t *band_psd);
126 127 128 129 130 131 132 133 134

/**
 * 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.
 *
135
 * @param[in]  s            adjustable bit allocation parameters
136
 * @param[in]  band_psd     signal power for each critical band
137 138 139 140
 * @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
141 142 143 144 145
 * @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
146
 * @param[out] mask         calculated masking curve
147
 */
148 149 150 151
void ff_ac3_bit_alloc_calc_mask(AC3BitAllocParameters *s, int16_t *band_psd,
                                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,
152 153 154 155 156 157 158 159 160 161 162 163
                                int16_t *mask);

/**
 * 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
164
 * @param[in]  snr_offset SNR adjustment
165 166 167 168
 * @param[in]  floor      noise floor
 * @param[out] bap        bit allocation pointers
 */
void ff_ac3_bit_alloc_calc_bap(int16_t *mask, int16_t *psd, int start, int end,
169
                               int snr_offset, int floor, uint8_t *bap);
170

Måns Rullgård's avatar
Måns Rullgård committed
171 172
void ac3_parametric_bit_allocation(AC3BitAllocParameters *s, uint8_t *bap,
                                   int8_t *exp, int start, int end,
173 174 175 176
                                   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);
177

178
#endif /* FFMPEG_AC3_H */