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 superframe_size; ///< current superframe size, in bytes 80 int frame_size; ///< current frame size, in bytes 81 int bit_rate; ///< stream bit rate, in bits-per-second 82 int sample_rate; ///< sample frequency, in Hz 83 int num_blocks; ///< number of audio blocks 84 int bitstream_id; ///< bitstream id (bsid) 85 int bitstream_mode; ///< bitstream mode (bsmod) 86 int channel_mode; ///< channel mode (acmod) 87 int lfe_on; ///< lfe channel in use 88 int dialog_normalization[2]; ///< dialog level in dBFS (dialnorm) 89 int compression_exists[2]; ///< compression field is valid for frame (compre) 90 int compression_gain[2]; ///< gain to apply for heavy compression (compr) 91 int channel_map; ///< custom channel map (chanmap) 92 int preferred_downmix; ///< Preferred 2-channel downmix mode (dmixmod) 93 int center_mix_level; ///< Center mix level index 94 int center_mix_level_ltrt; ///< Center mix level index for Lt/Rt (ltrtcmixlev) 95 int surround_mix_level; ///< Surround mix level index 96 int surround_mix_level_ltrt; ///< Surround mix level index for Lt/Rt (ltrtsurmixlev) 97 int lfe_mix_level_exists; ///< indicates if lfemixlevcod is specified (lfemixlevcode) 98 int lfe_mix_level; ///< LFE mix level index (lfemixlevcod) 99 int eac3; ///< indicates if current frame is E-AC-3 100 int eac3_frame_dependent_found; ///< bitstream has E-AC-3 dependent frame(s) 101 int eac3_subsbtreamid_found; ///< bitstream has E-AC-3 additional substream(s) 102 int dolby_surround_mode; ///< dolby surround mode (dsurmod) 103 int dolby_surround_ex_mode; ///< dolby surround ex mode (dsurexmod) 104 int dolby_headphone_mode; ///< dolby headphone mode (dheadphonmod) 105 ///@} 106 107 int preferred_stereo_downmix; 108 float ltrt_center_mix_level; 109 float ltrt_surround_mix_level; 110 float loro_center_mix_level; 111 float loro_surround_mix_level; 112 int target_level; ///< target level in dBFS 113 float level_gain[2]; 114 115 ///@name Frame syntax parameters 116 int snr_offset_strategy; ///< SNR offset strategy (snroffststr) 117 int block_switch_syntax; ///< block switch syntax enabled (blkswe) 118 int dither_flag_syntax; ///< dither flag syntax enabled (dithflage) 119 int bit_allocation_syntax; ///< bit allocation model syntax enabled (bamode) 120 int fast_gain_syntax; ///< fast gain codes enabled (frmfgaincode) 121 int dba_syntax; ///< delta bit allocation syntax enabled (dbaflde) 122 int skip_syntax; ///< skip field syntax enabled (skipflde) 123 ///@} 124 125 ///@name Standard coupling 126 int cpl_in_use[AC3_MAX_BLOCKS]; ///< coupling in use (cplinu) 127 int cpl_strategy_exists[AC3_MAX_BLOCKS];///< coupling strategy exists (cplstre) 128 int channel_in_cpl[AC3_MAX_CHANNELS]; ///< channel in coupling (chincpl) 129 int phase_flags_in_use; ///< phase flags in use (phsflginu) 130 int phase_flags[AC3_MAX_CPL_BANDS]; ///< phase flags (phsflg) 131 int num_cpl_bands; ///< number of coupling bands (ncplbnd) 132 uint8_t cpl_band_struct[AC3_MAX_CPL_BANDS]; 133 uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band 134 int firstchincpl; ///< first channel in coupling 135 int first_cpl_coords[AC3_MAX_CHANNELS]; ///< first coupling coordinates states (firstcplcos) 136 int cpl_coords[AC3_MAX_CHANNELS][AC3_MAX_CPL_BANDS]; ///< coupling coordinates (cplco) 137 ///@} 138 139 ///@name Spectral extension 140 ///@{ 141 int spx_in_use; ///< spectral extension in use (spxinu) 142 uint8_t channel_uses_spx[AC3_MAX_CHANNELS]; ///< channel uses spectral extension (chinspx) 143 int8_t spx_atten_code[AC3_MAX_CHANNELS]; ///< spx attenuation code (spxattencod) 144 int spx_src_start_freq; ///< spx start frequency bin 145 int spx_dst_end_freq; ///< spx end frequency bin 146 int spx_dst_start_freq; ///< spx starting frequency bin for copying (copystartmant) 147 ///< the copy region ends at the start of the spx region. 148 int num_spx_bands; ///< number of spx bands (nspxbnds) 149 uint8_t spx_band_struct[SPX_MAX_BANDS]; 150 uint8_t spx_band_sizes[SPX_MAX_BANDS]; ///< number of bins in each spx band 151 uint8_t first_spx_coords[AC3_MAX_CHANNELS]; ///< first spx coordinates states (firstspxcos) 152 INTFLOAT spx_noise_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS]; ///< spx noise blending factor (nblendfact) 153 INTFLOAT spx_signal_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS];///< spx signal blending factor (sblendfact) 154 ///@} 155 156 ///@name Adaptive hybrid transform 157 int channel_uses_aht[AC3_MAX_CHANNELS]; ///< channel AHT in use (chahtinu) 158 int pre_mantissa[AC3_MAX_CHANNELS][AC3_MAX_COEFS][AC3_MAX_BLOCKS]; ///< pre-IDCT mantissas 159 ///@} 160 161 ///@name Channel 162 int fbw_channels; ///< number of full-bandwidth channels 163 int channels; ///< number of total channels 164 int lfe_ch; ///< index of LFE channel 165 SHORTFLOAT *downmix_coeffs[2]; ///< stereo downmix coefficients 166 int downmixed; ///< indicates if coeffs are currently downmixed 167 int output_mode; ///< output channel configuration 168 int prev_output_mode; ///< output channel configuration for previous frame 169 int out_channels; ///< number of output channels 170 int prev_bit_rate; ///< stream bit rate, in bits-per-second for previous frame 171 ///@} 172 173 ///@name Dynamic range 174 INTFLOAT dynamic_range[2]; ///< dynamic range 175 INTFLOAT drc_scale; ///< percentage of dynamic range compression to be applied 176 int heavy_compression; ///< apply heavy compression 177 INTFLOAT heavy_dynamic_range[2]; ///< heavy dynamic range compression 178 ///@} 179 180 ///@name Bandwidth 181 int start_freq[AC3_MAX_CHANNELS]; ///< start frequency bin (strtmant) 182 int end_freq[AC3_MAX_CHANNELS]; ///< end frequency bin (endmant) 183 ///@} 184 185 ///@name Consistent noise generation 186 int consistent_noise_generation; ///< seed noise generation with AC-3 frame on decode 187 ///@} 188 189 ///@name Rematrixing 190 int num_rematrixing_bands; ///< number of rematrixing bands (nrematbnd) 191 int rematrixing_flags[4]; ///< rematrixing flags (rematflg) 192 ///@} 193 194 ///@name Exponents 195 int num_exp_groups[AC3_MAX_CHANNELS]; ///< Number of exponent groups (nexpgrp) 196 int8_t dexps[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< decoded exponents 197 int exp_strategy[AC3_MAX_BLOCKS][AC3_MAX_CHANNELS]; ///< exponent strategies (expstr) 198 ///@} 199 200 ///@name Bit allocation 201 AC3BitAllocParameters bit_alloc_params; ///< bit allocation parameters 202 int first_cpl_leak; ///< first coupling leak state (firstcplleak) 203 int snr_offset[AC3_MAX_CHANNELS]; ///< signal-to-noise ratio offsets (snroffst) 204 int fast_gain[AC3_MAX_CHANNELS]; ///< fast gain values/SMR's (fgain) 205 uint8_t bap[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< bit allocation pointers 206 int16_t psd[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< scaled exponents 207 int16_t band_psd[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< interpolated exponents 208 int16_t mask[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< masking curve values 209 int dba_mode[AC3_MAX_CHANNELS]; ///< delta bit allocation mode 210 int dba_nsegs[AC3_MAX_CHANNELS]; ///< number of delta segments 211 uint8_t dba_offsets[AC3_MAX_CHANNELS][8]; ///< delta segment offsets 212 uint8_t dba_lengths[AC3_MAX_CHANNELS][8]; ///< delta segment lengths 213 uint8_t dba_values[AC3_MAX_CHANNELS][8]; ///< delta values for each segment 214 ///@} 215 216 ///@name Zero-mantissa dithering 217 int dither_flag[AC3_MAX_CHANNELS]; ///< dither flags (dithflg) 218 AVLFG dith_state; ///< for dither generation 219 ///@} 220 221 ///@name IMDCT 222 int block_switch[AC3_MAX_CHANNELS]; ///< block switch flags (blksw) 223 FFTContext imdct_512; ///< for 512 sample IMDCT 224 FFTContext imdct_256; ///< for 256 sample IMDCT 225 ///@} 226 227 ///@name Optimization 228 BswapDSPContext bdsp; 229 #if USE_FIXED 230 AVFixedDSPContext *fdsp; 231 #else 232 AVFloatDSPContext *fdsp; 233 #endif 234 AC3DSPContext ac3dsp; 235 FmtConvertContext fmt_conv; ///< optimized conversion functions 236 ///@} 237 238 SHORTFLOAT *outptr[AC3_MAX_CHANNELS]; 239 INTFLOAT *xcfptr[AC3_MAX_CHANNELS]; 240 INTFLOAT *dlyptr[AC3_MAX_CHANNELS]; 241 242 ///@name Aligned arrays 243 DECLARE_ALIGNED(16, int, fixed_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< fixed-point transform coefficients 244 DECLARE_ALIGNED(32, INTFLOAT, transform_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< transform coefficients 245 DECLARE_ALIGNED(32, INTFLOAT, delay)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< delay - added to the next block 246 DECLARE_ALIGNED(32, INTFLOAT, window)[AC3_BLOCK_SIZE]; ///< window coefficients 247 DECLARE_ALIGNED(32, INTFLOAT, tmp_output)[AC3_BLOCK_SIZE]; ///< temporary storage for output before windowing 248 DECLARE_ALIGNED(32, SHORTFLOAT, output)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< output after imdct transform and windowing 249 DECLARE_ALIGNED(32, uint8_t, input_buffer)[AC3_FRAME_BUFFER_SIZE + AV_INPUT_BUFFER_PADDING_SIZE]; ///< temp buffer to prevent overread 250 DECLARE_ALIGNED(32, SHORTFLOAT, output_buffer)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE * 6]; ///< final output buffer 251 ///@} 252 } AC3DecodeContext; 253 254 /** 255 * Parse the E-AC-3 frame header. 256 * This parses both the bit stream info and audio frame header. 257 */ 258 static int ff_eac3_parse_header(AC3DecodeContext *s); 259 260 /** 261 * Decode mantissas in a single channel for the entire frame. 262 * This is used when AHT mode is enabled. 263 */ 264 static void ff_eac3_decode_transform_coeffs_aht_ch(AC3DecodeContext *s, int ch); 265 266 /** 267 * Apply spectral extension to each channel by copying lower frequency 268 * coefficients to higher frequency bins and applying side information to 269 * approximate the original high frequency signal. 270 */ 271 static void ff_eac3_apply_spectral_extension(AC3DecodeContext *s); 272 273 #if (!USE_FIXED) 274 extern float ff_ac3_heavy_dynamic_range_tab[256]; 275 #endif 276 277 #endif /* AVCODEC_AC3DEC_H */ 278