path: root/libavcodec/dca_lbr.h (plain)
blob: e6ca8057e9d0739fcf74b69c91828cf5ff472d34

1	/*
2	* Copyright (C) 2016 foo86
3	*
4	* This file is part of FFmpeg.
5	*
6	* FFmpeg is free software; you can redistribute it and/or
7	* modify it under the terms of the GNU Lesser General Public
8	* License as published by the Free Software Foundation; either
9	* version 2.1 of the License, or (at your option) any later version.
10	*
11	* FFmpeg is distributed in the hope that it will be useful,
12	* but WITHOUT ANY WARRANTY; without even the implied warranty of
13	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14	* Lesser General Public License for more details.
15	*
16	* You should have received a copy of the GNU Lesser General Public
17	* License along with FFmpeg; if not, write to the Free Software
18	* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
19	*/
20
21	#ifndef AVCODEC_DCA_LBR_H
22	#define AVCODEC_DCA_LBR_H
23
24	#include "libavutil/common.h"
25	#include "libavutil/float_dsp.h"
26	#include "libavutil/mem.h"
27
28	#include "avcodec.h"
29	#include "internal.h"
30	#include "get_bits.h"
31	#include "dca.h"
32	#include "dca_exss.h"
33	#include "dcadsp.h"
34	#include "fft.h"
35
36	#define DCA_LBR_CHANNELS 6
37	#define DCA_LBR_CHANNELS_TOTAL 32
38	#define DCA_LBR_SUBBANDS 32
39	#define DCA_LBR_TONES 512
40
41	#define DCA_LBR_TIME_SAMPLES 128
42	#define DCA_LBR_TIME_HISTORY 8
43
44	typedef struct DCALbrTone {
45	uint8_t x_freq; ///< Spectral line offset
46	uint8_t f_delt; ///< Difference between original and center frequency
47	uint8_t ph_rot; ///< Phase rotation
48	uint8_t pad; ///< Padding field
49	uint8_t amp[DCA_LBR_CHANNELS]; ///< Per-channel amplitude
50	uint8_t phs[DCA_LBR_CHANNELS]; ///< Per-channel phase
51	} DCALbrTone;
52
53	typedef struct DCALbrDecoder {
54	AVCodecContext *avctx;
55	GetBitContext gb;
56
57	int sample_rate; ///< Sample rate of LBR audio
58	int ch_mask; ///< LBR speaker mask
59	int flags; ///< Flags for LBR decoder initialization
60	int bit_rate_orig; ///< Original bit rate
61	int bit_rate_scaled; ///< Scaled bit rate
62
63	int nchannels; ///< Number of fullband channels to decode
64	int nchannels_total; ///< Total number of fullband channels
65	int freq_range; ///< Frequency range of LBR audio
66	int band_limit; ///< Band limit factor
67	int limited_rate; ///< Band limited sample rate
68	int limited_range; ///< Band limited frequency range
69	int res_profile; ///< Resolution profile
70	int nsubbands; ///< Number of encoded subbands
71	int g3_avg_only_start_sb; ///< Subband index where grid 3 scale factors end
72	int min_mono_subband; ///< Subband index where mono encoding starts
73	int max_mono_subband; ///< Subband index where mono encoding ends
74
75	int framenum; ///< Lower 5 bits of current frame number
76	int lbr_rand; ///< Seed for subband randomization
77	int warned; ///< Flags for warning suppression
78
79	uint8_t quant_levels[DCA_LBR_CHANNELS / 2][DCA_LBR_SUBBANDS]; ///< Quantization levels
80	uint8_t sb_indices[DCA_LBR_SUBBANDS]; ///< Subband reordering indices
81
82	uint8_t sec_ch_sbms[DCA_LBR_CHANNELS / 2][DCA_LBR_SUBBANDS]; ///< Right channel inversion or mid/side decoding flags
83	uint8_t sec_ch_lrms[DCA_LBR_CHANNELS / 2][DCA_LBR_SUBBANDS]; ///< Flags indicating if left/right channel are swapped
84	uint32_t ch_pres[DCA_LBR_CHANNELS]; ///< Subband allocation flags
85
86	uint8_t grid_1_scf[DCA_LBR_CHANNELS][12][8]; ///< Grid 1 scale factors
87	uint8_t grid_2_scf[DCA_LBR_CHANNELS][3][64]; ///< Grid 2 scale factors
88
89	int8_t grid_3_avg[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS - 4]; ///< Grid 3 average values
90	int8_t grid_3_scf[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS - 4][8]; ///< Grid 3 scale factors
91	uint32_t grid_3_pres[DCA_LBR_CHANNELS]; ///< Grid 3 scale factors presence flags
92
93	uint8_t high_res_scf[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS][8]; ///< High-frequency resolution scale factors
94
95	uint8_t part_stereo[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS / 4][5]; ///< Partial stereo coefficients
96	uint8_t part_stereo_pres; ///< Partial stereo coefficients presence flags
97
98	float lpc_coeff[2][DCA_LBR_CHANNELS][3][2][8]; ///< Predictor coefficients
99
100	float sb_scf[DCA_LBR_SUBBANDS]; ///< Subband randomization scale factors
101
102	float *time_samples[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS]; ///< Time samples
103
104	float *ts_buffer; ///< Time sample buffer base
105	unsigned int ts_size; ///< Time sample buffer size
106
107	DECLARE_ALIGNED(32, float, history)[DCA_LBR_CHANNELS][DCA_LBR_SUBBANDS * 4]; ///< IMDCT history
108	DECLARE_ALIGNED(32, float, window)[DCA_LBR_SUBBANDS * 4]; ///< Long window for IMDCT
109
110	DECLARE_ALIGNED(32, float, lfe_data)[64]; ///< Decimated LFE samples
111	DECLARE_ALIGNED(32, float, lfe_history)[5][2]; ///< LFE IIR filter history
112	float lfe_scale; ///< Scale factor of LFE samples before IIR filter
113
114	uint8_t tonal_scf[6]; ///< Tonal scale factors
115	uint16_t tonal_bounds[5][32][2]; ///< Per-group per-subframe start/end positions of tones
116	DCALbrTone tones[DCA_LBR_TONES]; ///< Circular buffer of tones
117	int ntones; ///< Circular buffer head position
118
119	FFTContext imdct;
120	AVFloatDSPContext *fdsp;
121	DCADSPContext *dcadsp;
122	} DCALbrDecoder;
123
124	int ff_dca_lbr_parse(DCALbrDecoder *s, uint8_t *data, DCAExssAsset *asset);
125	int ff_dca_lbr_filter_frame(DCALbrDecoder *s, AVFrame *frame);
126	av_cold void ff_dca_lbr_flush(DCALbrDecoder *s);
127	av_cold int ff_dca_lbr_init(DCALbrDecoder *s);
128	av_cold void ff_dca_lbr_close(DCALbrDecoder *s);
129
130	#endif
131