| 1 | /* |
| 2 | * Common code between the AC-3 and E-AC-3 decoders |
| 3 | * Copyright (c) 2007 Bartlomiej Wolowiec <bartek.wolowiec@gmail.com> |
| 4 | * |
| 5 | * This file is part of FFmpeg. |
| 6 | * |
| 7 | * FFmpeg is free software; you can redistribute it and/or |
| 8 | * modify it under the terms of the GNU Lesser General Public |
| 9 | * 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 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 15 | * Lesser General Public License for more details. |
| 16 | * |
| 17 | * 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 | |
| 22 | /** |
| 23 | * @file |
| 24 | * Common code between the AC-3 and E-AC-3 decoders. |
| 25 | * |
| 26 | * Summary of MDCT Coefficient Grouping: |
| 27 | * The individual MDCT coefficient indices are often referred to in the |
| 28 | * (E-)AC-3 specification as frequency bins. These bins are grouped together |
| 29 | * into subbands of 12 coefficients each. The subbands are grouped together |
| 30 | * into bands as defined in the bitstream by the band structures, which |
| 31 | * determine the number of bands and the size of each band. The full spectrum |
| 32 | * of 256 frequency bins is divided into 1 DC bin + 21 subbands = 253 bins. |
| 33 | * This system of grouping coefficients is used for channel bandwidth, stereo |
| 34 | * rematrixing, channel coupling, enhanced coupling, and spectral extension. |
| 35 | * |
| 36 | * +-+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+-+ |
| 37 | * |1| |12| | [12|12|12|12] | | | | | | | | | | | | |3| |
| 38 | * +-+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+-+ |
| 39 | * ~~~ ~~~~ ~~~~~~~~~~~~~ ~~~ |
| 40 | * | | | | |
| 41 | * | | | 3 unused frequency bins--+ |
| 42 | * | | | |
| 43 | * | | +--1 band containing 4 subbands |
| 44 | * | | |
| 45 | * | +--1 subband of 12 frequency bins |
| 46 | * | |
| 47 | * +--DC frequency bin |
| 48 | */ |
| 49 | |
| 50 | #ifndef AVCODEC_AC3DEC_H |
| 51 | #define AVCODEC_AC3DEC_H |
| 52 | |
| 53 | #include "libavutil/float_dsp.h" |
| 54 | #include "libavutil/fixed_dsp.h" |
| 55 | #include "libavutil/lfg.h" |
| 56 | #include "ac3.h" |
| 57 | #include "ac3dsp.h" |
| 58 | #include "bswapdsp.h" |
| 59 | #include "get_bits.h" |
| 60 | #include "fft.h" |
| 61 | #include "fmtconvert.h" |
| 62 | |
| 63 | #define AC3_OUTPUT_LFEON 8 |
| 64 | |
| 65 | #define SPX_MAX_BANDS 17 |
| 66 | |
| 67 | /** Large enough for maximum possible frame size when the specification limit is ignored */ |
| 68 | #define AC3_FRAME_BUFFER_SIZE 32768 |
| 69 | |
| 70 | typedef struct AC3DecodeContext { |
| 71 | AVClass *class; ///< class for AVOptions |
| 72 | AVCodecContext *avctx; ///< parent context |
| 73 | GetBitContext gbc; ///< bitstream reader |
| 74 | |
| 75 | ///@name Bit stream information |
| 76 | ///@{ |
| 77 | int frame_type; ///< frame type (strmtyp) |
| 78 | int substreamid; ///< substream identification |
| 79 | int frame_size; ///< current frame size, in bytes |
| 80 | int bit_rate; ///< stream bit rate, in bits-per-second |
| 81 | int sample_rate; ///< sample frequency, in Hz |
| 82 | int num_blocks; ///< number of audio blocks |
| 83 | int bitstream_id; ///< bitstream id (bsid) |
| 84 | int bitstream_mode; ///< bitstream mode (bsmod) |
| 85 | int channel_mode; ///< channel mode (acmod) |
| 86 | int lfe_on; ///< lfe channel in use |
| 87 | int dialog_normalization[2]; ///< dialog level in dBFS (dialnorm) |
| 88 | int compression_exists[2]; ///< compression field is valid for frame (compre) |
| 89 | int compression_gain[2]; ///< gain to apply for heavy compression (compr) |
| 90 | int channel_map; ///< custom channel map |
| 91 | int preferred_downmix; ///< Preferred 2-channel downmix mode (dmixmod) |
| 92 | int center_mix_level; ///< Center mix level index |
| 93 | int center_mix_level_ltrt; ///< Center mix level index for Lt/Rt (ltrtcmixlev) |
| 94 | int surround_mix_level; ///< Surround mix level index |
| 95 | int surround_mix_level_ltrt; ///< Surround mix level index for Lt/Rt (ltrtsurmixlev) |
| 96 | int lfe_mix_level_exists; ///< indicates if lfemixlevcod is specified (lfemixlevcode) |
| 97 | int lfe_mix_level; ///< LFE mix level index (lfemixlevcod) |
| 98 | int eac3; ///< indicates if current frame is E-AC-3 |
| 99 | int dolby_surround_mode; ///< dolby surround mode (dsurmod) |
| 100 | int dolby_surround_ex_mode; ///< dolby surround ex mode (dsurexmod) |
| 101 | int dolby_headphone_mode; ///< dolby headphone mode (dheadphonmod) |
| 102 | ///@} |
| 103 | |
| 104 | int preferred_stereo_downmix; |
| 105 | float ltrt_center_mix_level; |
| 106 | float ltrt_surround_mix_level; |
| 107 | float loro_center_mix_level; |
| 108 | float loro_surround_mix_level; |
| 109 | int target_level; ///< target level in dBFS |
| 110 | float level_gain[2]; |
| 111 | |
| 112 | ///@name Frame syntax parameters |
| 113 | int snr_offset_strategy; ///< SNR offset strategy (snroffststr) |
| 114 | int block_switch_syntax; ///< block switch syntax enabled (blkswe) |
| 115 | int dither_flag_syntax; ///< dither flag syntax enabled (dithflage) |
| 116 | int bit_allocation_syntax; ///< bit allocation model syntax enabled (bamode) |
| 117 | int fast_gain_syntax; ///< fast gain codes enabled (frmfgaincode) |
| 118 | int dba_syntax; ///< delta bit allocation syntax enabled (dbaflde) |
| 119 | int skip_syntax; ///< skip field syntax enabled (skipflde) |
| 120 | ///@} |
| 121 | |
| 122 | ///@name Standard coupling |
| 123 | int cpl_in_use[AC3_MAX_BLOCKS]; ///< coupling in use (cplinu) |
| 124 | int cpl_strategy_exists[AC3_MAX_BLOCKS];///< coupling strategy exists (cplstre) |
| 125 | int channel_in_cpl[AC3_MAX_CHANNELS]; ///< channel in coupling (chincpl) |
| 126 | int phase_flags_in_use; ///< phase flags in use (phsflginu) |
| 127 | int phase_flags[AC3_MAX_CPL_BANDS]; ///< phase flags (phsflg) |
| 128 | int num_cpl_bands; ///< number of coupling bands (ncplbnd) |
| 129 | uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band |
| 130 | int firstchincpl; ///< first channel in coupling |
| 131 | int first_cpl_coords[AC3_MAX_CHANNELS]; ///< first coupling coordinates states (firstcplcos) |
| 132 | int cpl_coords[AC3_MAX_CHANNELS][AC3_MAX_CPL_BANDS]; ///< coupling coordinates (cplco) |
| 133 | ///@} |
| 134 | |
| 135 | ///@name Spectral extension |
| 136 | ///@{ |
| 137 | int spx_in_use; ///< spectral extension in use (spxinu) |
| 138 | uint8_t channel_uses_spx[AC3_MAX_CHANNELS]; ///< channel uses spectral extension (chinspx) |
| 139 | int8_t spx_atten_code[AC3_MAX_CHANNELS]; ///< spx attenuation code (spxattencod) |
| 140 | int spx_src_start_freq; ///< spx start frequency bin |
| 141 | int spx_dst_end_freq; ///< spx end frequency bin |
| 142 | int spx_dst_start_freq; ///< spx starting frequency bin for copying (copystartmant) |
| 143 | ///< the copy region ends at the start of the spx region. |
| 144 | int num_spx_bands; ///< number of spx bands (nspxbnds) |
| 145 | uint8_t spx_band_sizes[SPX_MAX_BANDS]; ///< number of bins in each spx band |
| 146 | uint8_t first_spx_coords[AC3_MAX_CHANNELS]; ///< first spx coordinates states (firstspxcos) |
| 147 | INTFLOAT spx_noise_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS]; ///< spx noise blending factor (nblendfact) |
| 148 | INTFLOAT spx_signal_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS];///< spx signal blending factor (sblendfact) |
| 149 | ///@} |
| 150 | |
| 151 | ///@name Adaptive hybrid transform |
| 152 | int channel_uses_aht[AC3_MAX_CHANNELS]; ///< channel AHT in use (chahtinu) |
| 153 | int pre_mantissa[AC3_MAX_CHANNELS][AC3_MAX_COEFS][AC3_MAX_BLOCKS]; ///< pre-IDCT mantissas |
| 154 | ///@} |
| 155 | |
| 156 | ///@name Channel |
| 157 | int fbw_channels; ///< number of full-bandwidth channels |
| 158 | int channels; ///< number of total channels |
| 159 | int lfe_ch; ///< index of LFE channel |
| 160 | SHORTFLOAT downmix_coeffs[AC3_MAX_CHANNELS][2]; ///< stereo downmix coefficients |
| 161 | int downmixed; ///< indicates if coeffs are currently downmixed |
| 162 | int output_mode; ///< output channel configuration |
| 163 | int out_channels; ///< number of output channels |
| 164 | ///@} |
| 165 | |
| 166 | ///@name Dynamic range |
| 167 | INTFLOAT dynamic_range[2]; ///< dynamic range |
| 168 | INTFLOAT drc_scale; ///< percentage of dynamic range compression to be applied |
| 169 | int heavy_compression; ///< apply heavy compression |
| 170 | INTFLOAT heavy_dynamic_range[2]; ///< heavy dynamic range compression |
| 171 | ///@} |
| 172 | |
| 173 | ///@name Bandwidth |
| 174 | int start_freq[AC3_MAX_CHANNELS]; ///< start frequency bin (strtmant) |
| 175 | int end_freq[AC3_MAX_CHANNELS]; ///< end frequency bin (endmant) |
| 176 | ///@} |
| 177 | |
| 178 | ///@name Rematrixing |
| 179 | int num_rematrixing_bands; ///< number of rematrixing bands (nrematbnd) |
| 180 | int rematrixing_flags[4]; ///< rematrixing flags (rematflg) |
| 181 | ///@} |
| 182 | |
| 183 | ///@name Exponents |
| 184 | int num_exp_groups[AC3_MAX_CHANNELS]; ///< Number of exponent groups (nexpgrp) |
| 185 | int8_t dexps[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< decoded exponents |
| 186 | int exp_strategy[AC3_MAX_BLOCKS][AC3_MAX_CHANNELS]; ///< exponent strategies (expstr) |
| 187 | ///@} |
| 188 | |
| 189 | ///@name Bit allocation |
| 190 | AC3BitAllocParameters bit_alloc_params; ///< bit allocation parameters |
| 191 | int first_cpl_leak; ///< first coupling leak state (firstcplleak) |
| 192 | int snr_offset[AC3_MAX_CHANNELS]; ///< signal-to-noise ratio offsets (snroffst) |
| 193 | int fast_gain[AC3_MAX_CHANNELS]; ///< fast gain values/SMR's (fgain) |
| 194 | uint8_t bap[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< bit allocation pointers |
| 195 | int16_t psd[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< scaled exponents |
| 196 | int16_t band_psd[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< interpolated exponents |
| 197 | int16_t mask[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< masking curve values |
| 198 | int dba_mode[AC3_MAX_CHANNELS]; ///< delta bit allocation mode |
| 199 | int dba_nsegs[AC3_MAX_CHANNELS]; ///< number of delta segments |
| 200 | uint8_t dba_offsets[AC3_MAX_CHANNELS][8]; ///< delta segment offsets |
| 201 | uint8_t dba_lengths[AC3_MAX_CHANNELS][8]; ///< delta segment lengths |
| 202 | uint8_t dba_values[AC3_MAX_CHANNELS][8]; ///< delta values for each segment |
| 203 | ///@} |
| 204 | |
| 205 | ///@name Zero-mantissa dithering |
| 206 | int dither_flag[AC3_MAX_CHANNELS]; ///< dither flags (dithflg) |
| 207 | AVLFG dith_state; ///< for dither generation |
| 208 | ///@} |
| 209 | |
| 210 | ///@name IMDCT |
| 211 | int block_switch[AC3_MAX_CHANNELS]; ///< block switch flags (blksw) |
| 212 | FFTContext imdct_512; ///< for 512 sample IMDCT |
| 213 | FFTContext imdct_256; ///< for 256 sample IMDCT |
| 214 | ///@} |
| 215 | |
| 216 | ///@name Optimization |
| 217 | BswapDSPContext bdsp; |
| 218 | #if USE_FIXED |
| 219 | AVFixedDSPContext *fdsp; |
| 220 | #else |
| 221 | AVFloatDSPContext fdsp; |
| 222 | #endif |
| 223 | AC3DSPContext ac3dsp; |
| 224 | FmtConvertContext fmt_conv; ///< optimized conversion functions |
| 225 | ///@} |
| 226 | |
| 227 | SHORTFLOAT *outptr[AC3_MAX_CHANNELS]; |
| 228 | INTFLOAT *xcfptr[AC3_MAX_CHANNELS]; |
| 229 | INTFLOAT *dlyptr[AC3_MAX_CHANNELS]; |
| 230 | |
| 231 | ///@name Aligned arrays |
| 232 | DECLARE_ALIGNED(16, int, fixed_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< fixed-point transform coefficients |
| 233 | DECLARE_ALIGNED(32, INTFLOAT, transform_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< transform coefficients |
| 234 | DECLARE_ALIGNED(32, INTFLOAT, delay)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< delay - added to the next block |
| 235 | DECLARE_ALIGNED(32, INTFLOAT, window)[AC3_BLOCK_SIZE]; ///< window coefficients |
| 236 | DECLARE_ALIGNED(32, INTFLOAT, tmp_output)[AC3_BLOCK_SIZE]; ///< temporary storage for output before windowing |
| 237 | DECLARE_ALIGNED(32, SHORTFLOAT, output)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< output after imdct transform and windowing |
| 238 | DECLARE_ALIGNED(32, uint8_t, input_buffer)[AC3_FRAME_BUFFER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE]; ///< temp buffer to prevent overread |
| 239 | ///@} |
| 240 | } AC3DecodeContext; |
| 241 | |
| 242 | /** |
| 243 | * Parse the E-AC-3 frame header. |
| 244 | * This parses both the bit stream info and audio frame header. |
| 245 | */ |
| 246 | int ff_eac3_parse_header(AC3DecodeContext *s); |
| 247 | |
| 248 | /** |
| 249 | * Decode mantissas in a single channel for the entire frame. |
| 250 | * This is used when AHT mode is enabled. |
| 251 | */ |
| 252 | void ff_eac3_decode_transform_coeffs_aht_ch(AC3DecodeContext *s, int ch); |
| 253 | |
| 254 | /** |
| 255 | * Apply spectral extension to each channel by copying lower frequency |
| 256 | * coefficients to higher frequency bins and applying side information to |
| 257 | * approximate the original high frequency signal. |
| 258 | */ |
| 259 | void ff_eac3_apply_spectral_extension(AC3DecodeContext *s); |
| 260 | |
| 261 | #endif /* AVCODEC_AC3DEC_H */ |