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/lfg.h"
59 #include "fmtconvert.h"
61 #define AC3_OUTPUT_LFEON 8
63 #define SPX_MAX_BANDS 17
65 /** Large enough for maximum possible frame size when the specification limit is ignored */
66 #define AC3_FRAME_BUFFER_SIZE 32768
69 AVClass *class; ///< class for AVOptions
70 AVCodecContext *avctx; ///< parent context
71 AVFrame frame; ///< AVFrame for decoded output
72 GetBitContext gbc; ///< bitstream reader
74 ///@name Bit stream information
76 int frame_type; ///< frame type (strmtyp)
77 int substreamid; ///< substream identification
78 int frame_size; ///< current frame size, in bytes
79 int bit_rate; ///< stream bit rate, in bits-per-second
80 int sample_rate; ///< sample frequency, in Hz
81 int num_blocks; ///< number of audio blocks
82 int bitstream_mode; ///< bitstream mode (bsmod)
83 int channel_mode; ///< channel mode (acmod)
84 int channel_layout; ///< channel layout
85 int lfe_on; ///< lfe channel in use
86 int channel_map; ///< custom channel map
87 int center_mix_level; ///< Center mix level index
88 int surround_mix_level; ///< Surround mix level index
89 int eac3; ///< indicates if current frame is E-AC-3
92 int preferred_stereo_downmix;
93 float ltrt_center_mix_level;
94 float ltrt_surround_mix_level;
95 float loro_center_mix_level;
96 float loro_surround_mix_level;
98 ///@name Frame syntax parameters
99 int snr_offset_strategy; ///< SNR offset strategy (snroffststr)
100 int block_switch_syntax; ///< block switch syntax enabled (blkswe)
101 int dither_flag_syntax; ///< dither flag syntax enabled (dithflage)
102 int bit_allocation_syntax; ///< bit allocation model syntax enabled (bamode)
103 int fast_gain_syntax; ///< fast gain codes enabled (frmfgaincode)
104 int dba_syntax; ///< delta bit allocation syntax enabled (dbaflde)
105 int skip_syntax; ///< skip field syntax enabled (skipflde)
108 ///@name Standard coupling
109 int cpl_in_use[AC3_MAX_BLOCKS]; ///< coupling in use (cplinu)
110 int cpl_strategy_exists[AC3_MAX_BLOCKS];///< coupling strategy exists (cplstre)
111 int channel_in_cpl[AC3_MAX_CHANNELS]; ///< channel in coupling (chincpl)
112 int phase_flags_in_use; ///< phase flags in use (phsflginu)
113 int phase_flags[AC3_MAX_CPL_BANDS]; ///< phase flags (phsflg)
114 int num_cpl_bands; ///< number of coupling bands (ncplbnd)
115 uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band
116 int firstchincpl; ///< first channel in coupling
117 int first_cpl_coords[AC3_MAX_CHANNELS]; ///< first coupling coordinates states (firstcplcos)
118 int cpl_coords[AC3_MAX_CHANNELS][AC3_MAX_CPL_BANDS]; ///< coupling coordinates (cplco)
121 ///@name Spectral extension
123 int spx_in_use; ///< spectral extension in use (spxinu)
124 uint8_t channel_uses_spx[AC3_MAX_CHANNELS]; ///< channel uses spectral extension (chinspx)
125 int8_t spx_atten_code[AC3_MAX_CHANNELS]; ///< spx attenuation code (spxattencod)
126 int spx_src_start_freq; ///< spx start frequency bin
127 int spx_dst_end_freq; ///< spx end frequency bin
128 int spx_dst_start_freq; ///< spx starting frequency bin for copying (copystartmant)
129 ///< the copy region ends at the start of the spx region.
130 int num_spx_bands; ///< number of spx bands (nspxbnds)
131 uint8_t spx_band_sizes[SPX_MAX_BANDS]; ///< number of bins in each spx band
132 uint8_t first_spx_coords[AC3_MAX_CHANNELS]; ///< first spx coordinates states (firstspxcos)
133 float spx_noise_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS]; ///< spx noise blending factor (nblendfact)
134 float spx_signal_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS];///< spx signal blending factor (sblendfact)
137 ///@name Adaptive hybrid transform
138 int channel_uses_aht[AC3_MAX_CHANNELS]; ///< channel AHT in use (chahtinu)
139 int pre_mantissa[AC3_MAX_CHANNELS][AC3_MAX_COEFS][AC3_MAX_BLOCKS]; ///< pre-IDCT mantissas
143 int fbw_channels; ///< number of full-bandwidth channels
144 int channels; ///< number of total channels
145 int lfe_ch; ///< index of LFE channel
146 float downmix_coeffs[AC3_MAX_CHANNELS][2]; ///< stereo downmix coefficients
147 int downmixed; ///< indicates if coeffs are currently downmixed
148 int output_mode; ///< output channel configuration
149 int out_channels; ///< number of output channels
152 ///@name Dynamic range
153 float dynamic_range[2]; ///< dynamic range
154 float drc_scale; ///< percentage of dynamic range compression to be applied
158 int start_freq[AC3_MAX_CHANNELS]; ///< start frequency bin (strtmant)
159 int end_freq[AC3_MAX_CHANNELS]; ///< end frequency bin (endmant)
163 int num_rematrixing_bands; ///< number of rematrixing bands (nrematbnd)
164 int rematrixing_flags[4]; ///< rematrixing flags (rematflg)
168 int num_exp_groups[AC3_MAX_CHANNELS]; ///< Number of exponent groups (nexpgrp)
169 int8_t dexps[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< decoded exponents
170 int exp_strategy[AC3_MAX_BLOCKS][AC3_MAX_CHANNELS]; ///< exponent strategies (expstr)
173 ///@name Bit allocation
174 AC3BitAllocParameters bit_alloc_params; ///< bit allocation parameters
175 int first_cpl_leak; ///< first coupling leak state (firstcplleak)
176 int snr_offset[AC3_MAX_CHANNELS]; ///< signal-to-noise ratio offsets (snroffst)
177 int fast_gain[AC3_MAX_CHANNELS]; ///< fast gain values/SMR's (fgain)
178 uint8_t bap[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< bit allocation pointers
179 int16_t psd[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< scaled exponents
180 int16_t band_psd[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< interpolated exponents
181 int16_t mask[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< masking curve values
182 int dba_mode[AC3_MAX_CHANNELS]; ///< delta bit allocation mode
183 int dba_nsegs[AC3_MAX_CHANNELS]; ///< number of delta segments
184 uint8_t dba_offsets[AC3_MAX_CHANNELS][8]; ///< delta segment offsets
185 uint8_t dba_lengths[AC3_MAX_CHANNELS][8]; ///< delta segment lengths
186 uint8_t dba_values[AC3_MAX_CHANNELS][8]; ///< delta values for each segment
189 ///@name Zero-mantissa dithering
190 int dither_flag[AC3_MAX_CHANNELS]; ///< dither flags (dithflg)
191 AVLFG dith_state; ///< for dither generation
195 int block_switch[AC3_MAX_CHANNELS]; ///< block switch flags (blksw)
196 FFTContext imdct_512; ///< for 512 sample IMDCT
197 FFTContext imdct_256; ///< for 256 sample IMDCT
200 ///@name Optimization
201 DSPContext dsp; ///< for optimization
202 AC3DSPContext ac3dsp;
203 FmtConvertContext fmt_conv; ///< optimized conversion functions
204 float mul_bias; ///< scaling for float_to_int16 conversion
207 ///@name Aligned arrays
208 DECLARE_ALIGNED(16, int, fixed_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< fixed-point transform coefficients
209 DECLARE_ALIGNED(32, float, transform_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS]; ///< transform coefficients
210 DECLARE_ALIGNED(32, float, delay)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< delay - added to the next block
211 DECLARE_ALIGNED(32, float, window)[AC3_BLOCK_SIZE]; ///< window coefficients
212 DECLARE_ALIGNED(32, float, tmp_output)[AC3_BLOCK_SIZE]; ///< temporary storage for output before windowing
213 DECLARE_ALIGNED(32, float, output)[AC3_MAX_CHANNELS][AC3_BLOCK_SIZE]; ///< output after imdct transform and windowing
214 DECLARE_ALIGNED(32, uint8_t, input_buffer)[AC3_FRAME_BUFFER_SIZE + FF_INPUT_BUFFER_PADDING_SIZE]; ///< temp buffer to prevent overread
219 * Parse the E-AC-3 frame header.
220 * This parses both the bit stream info and audio frame header.
222 int ff_eac3_parse_header(AC3DecodeContext *s);
225 * Decode mantissas in a single channel for the entire frame.
226 * This is used when AHT mode is enabled.
228 void ff_eac3_decode_transform_coeffs_aht_ch(AC3DecodeContext *s, int ch);
230 void ff_ac3_downmix_c(float (*samples)[256], float (*matrix)[2],
231 int out_ch, int in_ch, int len);
234 * Apply spectral extension to each channel by copying lower frequency
235 * coefficients to higher frequency bins and applying side information to
236 * approximate the original high frequency signal.
238 void ff_eac3_apply_spectral_extension(AC3DecodeContext *s);
240 #endif /* AVCODEC_AC3DEC_H */