2 * Common code between the AC-3 and E-AC-3 decoders
3 * Copyright (c) 2007 Bartlomiej Wolowiec <bartek.wolowiec@gmail.com>
5 * This file is part of FFmpeg.
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.
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.
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
24 * Common code between the AC-3 and E-AC-3 decoders.
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.
36 * +-+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+-+
37 * |1| |12| | [12|12|12|12] | | | | | | | | | | | | |3|
38 * +-+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+--+-+
39 * ~~~ ~~~~ ~~~~~~~~~~~~~ ~~~
41 * | | | 3 unused frequency bins--+
43 * | | +--1 band containing 4 subbands
45 * | +--1 subband of 12 frequency bins
50 #ifndef AVCODEC_AC3DEC_H
51 #define AVCODEC_AC3DEC_H
53 #include "libavutil/float_dsp.h"
54 #include "libavutil/fixed_dsp.h"
55 #include "libavutil/lfg.h"
61 #include "fmtconvert.h"
63 #define AC3_OUTPUT_LFEON 8
65 #define SPX_MAX_BANDS 17
67 /** Large enough for maximum possible frame size when the specification limit is ignored */
68 #define AC3_FRAME_BUFFER_SIZE 32768
70 typedef struct AC3DecodeContext {
71 AVClass *class; ///< class for AVOptions
72 AVCodecContext *avctx; ///< parent context
73 GetBitContext gbc; ///< bitstream reader
75 ///@name Bit stream information
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 channel_map; ///< custom channel map
88 int preferred_downmix; ///< Preferred 2-channel downmix mode (dmixmod)
89 int center_mix_level; ///< Center mix level index
90 int center_mix_level_ltrt; ///< Center mix level index for Lt/Rt (ltrtcmixlev)
91 int surround_mix_level; ///< Surround mix level index
92 int surround_mix_level_ltrt; ///< Surround mix level index for Lt/Rt (ltrtsurmixlev)
93 int lfe_mix_level_exists; ///< indicates if lfemixlevcod is specified (lfemixlevcode)
94 int lfe_mix_level; ///< LFE mix level index (lfemixlevcod)
95 int eac3; ///< indicates if current frame is E-AC-3
96 int dolby_surround_mode; ///< dolby surround mode (dsurmod)
97 int dolby_surround_ex_mode; ///< dolby surround ex mode (dsurexmod)
98 int dolby_headphone_mode; ///< dolby headphone mode (dheadphonmod)
101 int preferred_stereo_downmix;
102 float ltrt_center_mix_level;
103 float ltrt_surround_mix_level;
104 float loro_center_mix_level;
105 float loro_surround_mix_level;
107 ///@name Frame syntax parameters
108 int snr_offset_strategy; ///< SNR offset strategy (snroffststr)
109 int block_switch_syntax; ///< block switch syntax enabled (blkswe)
110 int dither_flag_syntax; ///< dither flag syntax enabled (dithflage)
111 int bit_allocation_syntax; ///< bit allocation model syntax enabled (bamode)
112 int fast_gain_syntax; ///< fast gain codes enabled (frmfgaincode)
113 int dba_syntax; ///< delta bit allocation syntax enabled (dbaflde)
114 int skip_syntax; ///< skip field syntax enabled (skipflde)
117 ///@name Standard coupling
118 int cpl_in_use[AC3_MAX_BLOCKS]; ///< coupling in use (cplinu)
119 int cpl_strategy_exists[AC3_MAX_BLOCKS];///< coupling strategy exists (cplstre)
120 int channel_in_cpl[AC3_MAX_CHANNELS]; ///< channel in coupling (chincpl)
121 int phase_flags_in_use; ///< phase flags in use (phsflginu)
122 int phase_flags[AC3_MAX_CPL_BANDS]; ///< phase flags (phsflg)
123 int num_cpl_bands; ///< number of coupling bands (ncplbnd)
124 uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band
125 int firstchincpl; ///< first channel in coupling
126 int first_cpl_coords[AC3_MAX_CHANNELS]; ///< first coupling coordinates states (firstcplcos)
127 int cpl_coords[AC3_MAX_CHANNELS][AC3_MAX_CPL_BANDS]; ///< coupling coordinates (cplco)
130 ///@name Spectral extension
132 int spx_in_use; ///< spectral extension in use (spxinu)
133 uint8_t channel_uses_spx[AC3_MAX_CHANNELS]; ///< channel uses spectral extension (chinspx)
134 int8_t spx_atten_code[AC3_MAX_CHANNELS]; ///< spx attenuation code (spxattencod)
135 int spx_src_start_freq; ///< spx start frequency bin
136 int spx_dst_end_freq; ///< spx end frequency bin
137 int spx_dst_start_freq; ///< spx starting frequency bin for copying (copystartmant)
138 ///< the copy region ends at the start of the spx region.
139 int num_spx_bands; ///< number of spx bands (nspxbnds)
140 uint8_t spx_band_sizes[SPX_MAX_BANDS]; ///< number of bins in each spx band
141 uint8_t first_spx_coords[AC3_MAX_CHANNELS]; ///< first spx coordinates states (firstspxcos)
142 INTFLOAT spx_noise_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS]; ///< spx noise blending factor (nblendfact)
143 INTFLOAT spx_signal_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS];///< spx signal blending factor (sblendfact)
146 ///@name Adaptive hybrid transform
147 int channel_uses_aht[AC3_MAX_CHANNELS]; ///< channel AHT in use (chahtinu)
148 int pre_mantissa[AC3_MAX_CHANNELS][AC3_MAX_COEFS][AC3_MAX_BLOCKS]; ///< pre-IDCT mantissas
152 int fbw_channels; ///< number of full-bandwidth channels
153 int channels; ///< number of total channels
154 int lfe_ch; ///< index of LFE channel
155 SHORTFLOAT downmix_coeffs[AC3_MAX_CHANNELS][2]; ///< stereo downmix coefficients
156 int downmixed; ///< indicates if coeffs are currently downmixed
157 int output_mode; ///< output channel configuration
158 int out_channels; ///< number of output channels
161 ///@name Dynamic range
162 INTFLOAT dynamic_range[2]; ///< dynamic range
163 INTFLOAT drc_scale; ///< percentage of dynamic range compression to be applied
167 int start_freq[AC3_MAX_CHANNELS]; ///< start frequency bin (strtmant)
168 int end_freq[AC3_MAX_CHANNELS]; ///< end frequency bin (endmant)
172 int num_rematrixing_bands; ///< number of rematrixing bands (nrematbnd)
173 int rematrixing_flags[4]; ///< rematrixing flags (rematflg)
177 int num_exp_groups[AC3_MAX_CHANNELS]; ///< Number of exponent groups (nexpgrp)
178 int8_t dexps[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< decoded exponents
179 int exp_strategy[AC3_MAX_BLOCKS][AC3_MAX_CHANNELS]; ///< exponent strategies (expstr)
182 ///@name Bit allocation
183 AC3BitAllocParameters bit_alloc_params; ///< bit allocation parameters
184 int first_cpl_leak; ///< first coupling leak state (firstcplleak)
185 int snr_offset[AC3_MAX_CHANNELS]; ///< signal-to-noise ratio offsets (snroffst)
186 int fast_gain[AC3_MAX_CHANNELS]; ///< fast gain values/SMR's (fgain)
187 uint8_t bap[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< bit allocation pointers
188 int16_t psd[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< scaled exponents
189 int16_t band_psd[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< interpolated exponents
190 int16_t mask[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< masking curve values
191 int dba_mode[AC3_MAX_CHANNELS]; ///< delta bit allocation mode
192 int dba_nsegs[AC3_MAX_CHANNELS]; ///< number of delta segments
193 uint8_t dba_offsets[AC3_MAX_CHANNELS][8]; ///< delta segment offsets
194 uint8_t dba_lengths[AC3_MAX_CHANNELS][8]; ///< delta segment lengths
195 uint8_t dba_values[AC3_MAX_CHANNELS][8]; ///< delta values for each segment
198 ///@name Zero-mantissa dithering
199 int dither_flag[AC3_MAX_CHANNELS]; ///< dither flags (dithflg)
200 AVLFG dith_state; ///< for dither generation
204 int block_switch[AC3_MAX_CHANNELS]; ///< block switch flags (blksw)
205 FFTContext imdct_512; ///< for 512 sample IMDCT
206 FFTContext imdct_256; ///< for 256 sample IMDCT
209 ///@name Optimization
210 BswapDSPContext bdsp;
212 AVFixedDSPContext *fdsp;
214 AVFloatDSPContext fdsp;
216 AC3DSPContext ac3dsp;
217 FmtConvertContext fmt_conv; ///< optimized conversion functions
220 SHORTFLOAT *outptr[AC3_MAX_CHANNELS];
221 INTFLOAT *xcfptr[AC3_MAX_CHANNELS];
222 INTFLOAT *dlyptr[AC3_MAX_CHANNELS];
224 ///@name Aligned arrays
225 DECLARE_ALIGNED(16, int, fixed_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< fixed-point transform coefficients
226 DECLARE_ALIGNED(32, INTFLOAT, transform_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< transform coefficients
227 DECLARE_ALIGNED(32, INTFLOAT, delay)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< delay - added to the next block
228 DECLARE_ALIGNED(32, INTFLOAT, window)[AC3_BLOCK_SIZE]; ///< window coefficients
229 DECLARE_ALIGNED(32, INTFLOAT, tmp_output)[AC3_BLOCK_SIZE]; ///< temporary storage for output before windowing
230 DECLARE_ALIGNED(32, SHORTFLOAT, output)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< output after imdct transform and windowing
231 DECLARE_ALIGNED(32, uint8_t, input_buffer)[AC3_FRAME_BUFFER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE]; ///< temp buffer to prevent overread
236 * Parse the E-AC-3 frame header.
237 * This parses both the bit stream info and audio frame header.
239 int ff_eac3_parse_header(AC3DecodeContext *s);
242 * Decode mantissas in a single channel for the entire frame.
243 * This is used when AHT mode is enabled.
245 void ff_eac3_decode_transform_coeffs_aht_ch(AC3DecodeContext *s, int ch);
248 * Apply spectral extension to each channel by copying lower frequency
249 * coefficients to higher frequency bins and applying side information to
250 * approximate the original high frequency signal.
252 void ff_eac3_apply_spectral_extension(AC3DecodeContext *s);
254 #endif /* AVCODEC_AC3DEC_H */