xref: /third_party/ffmpeg/libavcodec/ac3dec.h (revision cabdff1a)
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 "libavutil/mem_internal.h"
57
58#include "ac3.h"
59#include "ac3dsp.h"
60#include "bswapdsp.h"
61#include "get_bits.h"
62#include "fft.h"
63#include "fmtconvert.h"
64
65#define AC3_OUTPUT_LFEON  8
66
67#define SPX_MAX_BANDS    17
68
69/** Large enough for maximum possible frame size when the specification limit is ignored */
70#define AC3_FRAME_BUFFER_SIZE 32768
71
72typedef struct AC3DecodeContext {
73    AVClass        *class;                  ///< class for AVOptions
74    AVCodecContext *avctx;                  ///< parent context
75    GetBitContext gbc;                      ///< bitstream reader
76
77///@name Bit stream information
78///@{
79    int frame_type;                         ///< frame type                             (strmtyp)
80    int substreamid;                        ///< substream identification
81    int superframe_size;                    ///< current superframe size, in bytes
82    int frame_size;                         ///< current frame size, in bytes
83    int bit_rate;                           ///< stream bit rate, in bits-per-second
84    int sample_rate;                        ///< sample frequency, in Hz
85    int num_blocks;                         ///< number of audio blocks
86    int bitstream_id;                       ///< bitstream id                           (bsid)
87    int bitstream_mode;                     ///< bitstream mode                         (bsmod)
88    int channel_mode;                       ///< channel mode                           (acmod)
89    int lfe_on;                             ///< lfe channel in use
90    int dialog_normalization[2];            ///< dialog level in dBFS                   (dialnorm)
91    int compression_exists[2];              ///< compression field is valid for frame   (compre)
92    int compression_gain[2];                ///< gain to apply for heavy compression    (compr)
93    int channel_map;                        ///< custom channel map                     (chanmap)
94    int preferred_downmix;                  ///< Preferred 2-channel downmix mode       (dmixmod)
95    int center_mix_level;                   ///< Center mix level index
96    int center_mix_level_ltrt;              ///< Center mix level index for Lt/Rt       (ltrtcmixlev)
97    int surround_mix_level;                 ///< Surround mix level index
98    int surround_mix_level_ltrt;            ///< Surround mix level index for Lt/Rt     (ltrtsurmixlev)
99    int lfe_mix_level_exists;               ///< indicates if lfemixlevcod is specified (lfemixlevcode)
100    int lfe_mix_level;                      ///< LFE mix level index                    (lfemixlevcod)
101    int eac3;                               ///< indicates if current frame is E-AC-3
102    int eac3_frame_dependent_found;         ///< bitstream has E-AC-3 dependent frame(s)
103    int eac3_subsbtreamid_found;            ///< bitstream has E-AC-3 additional substream(s)
104    int dolby_surround_mode;                ///< dolby surround mode                    (dsurmod)
105    int dolby_surround_ex_mode;             ///< dolby surround ex mode                 (dsurexmod)
106    int dolby_headphone_mode;               ///< dolby headphone mode                   (dheadphonmod)
107///@}
108
109    int preferred_stereo_downmix;
110    float ltrt_center_mix_level;
111    float ltrt_surround_mix_level;
112    float loro_center_mix_level;
113    float loro_surround_mix_level;
114    int target_level;                       ///< target level in dBFS
115    float level_gain[2];
116
117///@name Frame syntax parameters
118    int snr_offset_strategy;                ///< SNR offset strategy                    (snroffststr)
119    int block_switch_syntax;                ///< block switch syntax enabled            (blkswe)
120    int dither_flag_syntax;                 ///< dither flag syntax enabled             (dithflage)
121    int bit_allocation_syntax;              ///< bit allocation model syntax enabled    (bamode)
122    int fast_gain_syntax;                   ///< fast gain codes enabled                (frmfgaincode)
123    int dba_syntax;                         ///< delta bit allocation syntax enabled    (dbaflde)
124    int skip_syntax;                        ///< skip field syntax enabled              (skipflde)
125 ///@}
126
127///@name Standard coupling
128    int cpl_in_use[AC3_MAX_BLOCKS];         ///< coupling in use                        (cplinu)
129    int cpl_strategy_exists[AC3_MAX_BLOCKS];///< coupling strategy exists               (cplstre)
130    int channel_in_cpl[AC3_MAX_CHANNELS];   ///< channel in coupling                    (chincpl)
131    int phase_flags_in_use;                 ///< phase flags in use                     (phsflginu)
132    int phase_flags[AC3_MAX_CPL_BANDS];     ///< phase flags                            (phsflg)
133    int num_cpl_bands;                      ///< number of coupling bands               (ncplbnd)
134    uint8_t cpl_band_struct[AC3_MAX_CPL_BANDS];
135    uint8_t cpl_band_sizes[AC3_MAX_CPL_BANDS]; ///< number of coeffs in each coupling band
136    int firstchincpl;                       ///< first channel in coupling
137    int first_cpl_coords[AC3_MAX_CHANNELS]; ///< first coupling coordinates states      (firstcplcos)
138    int cpl_coords[AC3_MAX_CHANNELS][AC3_MAX_CPL_BANDS]; ///< coupling coordinates      (cplco)
139///@}
140
141///@name Spectral extension
142///@{
143    int spx_in_use;                             ///< spectral extension in use              (spxinu)
144    uint8_t channel_uses_spx[AC3_MAX_CHANNELS]; ///< channel uses spectral extension        (chinspx)
145    int8_t spx_atten_code[AC3_MAX_CHANNELS];    ///< spx attenuation code                   (spxattencod)
146    int spx_src_start_freq;                     ///< spx start frequency bin
147    int spx_dst_end_freq;                       ///< spx end frequency bin
148    int spx_dst_start_freq;                     ///< spx starting frequency bin for copying (copystartmant)
149                                                ///< the copy region ends at the start of the spx region.
150    int num_spx_bands;                          ///< number of spx bands                    (nspxbnds)
151    uint8_t spx_band_struct[SPX_MAX_BANDS];
152    uint8_t spx_band_sizes[SPX_MAX_BANDS];      ///< number of bins in each spx band
153    uint8_t first_spx_coords[AC3_MAX_CHANNELS]; ///< first spx coordinates states           (firstspxcos)
154    INTFLOAT spx_noise_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS]; ///< spx noise blending factor  (nblendfact)
155    INTFLOAT spx_signal_blend[AC3_MAX_CHANNELS][SPX_MAX_BANDS];///< spx signal blending factor (sblendfact)
156///@}
157
158///@name Adaptive hybrid transform
159    int channel_uses_aht[AC3_MAX_CHANNELS];                         ///< channel AHT in use (chahtinu)
160    int pre_mantissa[AC3_MAX_CHANNELS][AC3_MAX_COEFS][AC3_MAX_BLOCKS];  ///< pre-IDCT mantissas
161///@}
162
163///@name Channel
164    int fbw_channels;                           ///< number of full-bandwidth channels
165    int channels;                               ///< number of total channels
166    int lfe_ch;                                 ///< index of LFE channel
167    SHORTFLOAT *downmix_coeffs[2];              ///< stereo downmix coefficients
168    int downmixed;                              ///< indicates if coeffs are currently downmixed
169    int output_mode;                            ///< output channel configuration
170    int prev_output_mode;                       ///< output channel configuration for previous frame
171    int out_channels;                           ///< number of output channels
172    int prev_bit_rate;                          ///< stream bit rate, in bits-per-second for previous frame
173///@}
174
175///@name Dynamic range
176    INTFLOAT dynamic_range[2];                 ///< dynamic range
177    INTFLOAT drc_scale;                        ///< percentage of dynamic range compression to be applied
178    int heavy_compression;                     ///< apply heavy compression
179    INTFLOAT heavy_dynamic_range[2];           ///< heavy dynamic range compression
180///@}
181
182///@name Bandwidth
183    int start_freq[AC3_MAX_CHANNELS];       ///< start frequency bin                    (strtmant)
184    int end_freq[AC3_MAX_CHANNELS];         ///< end frequency bin                      (endmant)
185///@}
186
187///@name Consistent noise generation
188    int consistent_noise_generation;        ///< seed noise generation with AC-3 frame on decode
189///@}
190
191///@name Rematrixing
192    int num_rematrixing_bands;              ///< number of rematrixing bands            (nrematbnd)
193    int rematrixing_flags[4];               ///< rematrixing flags                      (rematflg)
194///@}
195
196///@name Exponents
197    int num_exp_groups[AC3_MAX_CHANNELS];           ///< Number of exponent groups      (nexpgrp)
198    int8_t dexps[AC3_MAX_CHANNELS][AC3_MAX_COEFS];  ///< decoded exponents
199    int exp_strategy[AC3_MAX_BLOCKS][AC3_MAX_CHANNELS]; ///< exponent strategies        (expstr)
200///@}
201
202///@name Bit allocation
203    AC3BitAllocParameters bit_alloc_params;         ///< bit allocation parameters
204    int first_cpl_leak;                             ///< first coupling leak state      (firstcplleak)
205    int snr_offset[AC3_MAX_CHANNELS];               ///< signal-to-noise ratio offsets  (snroffst)
206    int fast_gain[AC3_MAX_CHANNELS];                ///< fast gain values/SMR's         (fgain)
207    uint8_t bap[AC3_MAX_CHANNELS][AC3_MAX_COEFS];   ///< bit allocation pointers
208    int16_t psd[AC3_MAX_CHANNELS][AC3_MAX_COEFS];   ///< scaled exponents
209    int16_t band_psd[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS]; ///< interpolated exponents
210    int16_t mask[AC3_MAX_CHANNELS][AC3_CRITICAL_BANDS];     ///< masking curve values
211    int dba_mode[AC3_MAX_CHANNELS];                 ///< delta bit allocation mode
212    int dba_nsegs[AC3_MAX_CHANNELS];                ///< number of delta segments
213    uint8_t dba_offsets[AC3_MAX_CHANNELS][8];       ///< delta segment offsets
214    uint8_t dba_lengths[AC3_MAX_CHANNELS][8];       ///< delta segment lengths
215    uint8_t dba_values[AC3_MAX_CHANNELS][8];        ///< delta values for each segment
216///@}
217
218///@name Zero-mantissa dithering
219    int dither_flag[AC3_MAX_CHANNELS];      ///< dither flags                           (dithflg)
220    AVLFG dith_state;                       ///< for dither generation
221///@}
222
223///@name IMDCT
224    int block_switch[AC3_MAX_CHANNELS];     ///< block switch flags                     (blksw)
225    FFTContext imdct_512;                   ///< for 512 sample IMDCT
226    FFTContext imdct_256;                   ///< for 256 sample IMDCT
227///@}
228
229///@name Optimization
230    BswapDSPContext bdsp;
231#if USE_FIXED
232    AVFixedDSPContext *fdsp;
233#else
234    AVFloatDSPContext *fdsp;
235#endif
236    AC3DSPContext ac3dsp;
237    FmtConvertContext fmt_conv;             ///< optimized conversion functions
238///@}
239
240    SHORTFLOAT *outptr[AC3_MAX_CHANNELS];
241    INTFLOAT *xcfptr[AC3_MAX_CHANNELS];
242    INTFLOAT *dlyptr[AC3_MAX_CHANNELS];
243
244///@name Aligned arrays
245    DECLARE_ALIGNED(16, int,   fixed_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS];       ///< fixed-point transform coefficients
246    DECLARE_ALIGNED(32, INTFLOAT, transform_coeffs)[AC3_MAX_CHANNELS][AC3_MAX_COEFS];   ///< transform coefficients
247    DECLARE_ALIGNED(32, INTFLOAT, delay)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE];         ///< delay - added to the next block
248    DECLARE_ALIGNED(32, INTFLOAT, window)[AC3_BLOCK_SIZE];                              ///< window coefficients
249    DECLARE_ALIGNED(32, INTFLOAT, tmp_output)[AC3_BLOCK_SIZE];                          ///< temporary storage for output before windowing
250    DECLARE_ALIGNED(32, SHORTFLOAT, output)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE];            ///< output after imdct transform and windowing
251    DECLARE_ALIGNED(32, uint8_t, input_buffer)[AC3_FRAME_BUFFER_SIZE + AV_INPUT_BUFFER_PADDING_SIZE]; ///< temp buffer to prevent overread
252    DECLARE_ALIGNED(32, SHORTFLOAT, output_buffer)[EAC3_MAX_CHANNELS][AC3_BLOCK_SIZE * 6];  ///< final output buffer
253///@}
254
255    AVChannelLayout downmix_layout;
256} AC3DecodeContext;
257
258/**
259 * Parse the E-AC-3 frame header.
260 * This parses both the bit stream info and audio frame header.
261 */
262static int ff_eac3_parse_header(AC3DecodeContext *s);
263
264/**
265 * Decode mantissas in a single channel for the entire frame.
266 * This is used when AHT mode is enabled.
267 */
268static void ff_eac3_decode_transform_coeffs_aht_ch(AC3DecodeContext *s, int ch);
269
270/**
271 * Apply spectral extension to each channel by copying lower frequency
272 * coefficients to higher frequency bins and applying side information to
273 * approximate the original high frequency signal.
274 */
275static void ff_eac3_apply_spectral_extension(AC3DecodeContext *s);
276
277#if (!USE_FIXED)
278extern float ff_ac3_heavy_dynamic_range_tab[256];
279#endif
280
281#endif /* AVCODEC_AC3DEC_H */
282