2 * Copyright (C) 2016 foo86
4 * This file is part of FFmpeg.
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.
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.
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
21 #ifndef AVCODEC_DCA_CORE_H
22 #define AVCODEC_DCA_CORE_H
24 #include "libavutil/common.h"
25 #include "libavutil/float_dsp.h"
26 #include "libavutil/fixed_dsp.h"
27 #include "libavutil/mem.h"
39 #include "synth_filter.h"
41 #define DCA_CHANNELS 7
42 #define DCA_SUBBANDS 32
43 #define DCA_SUBBANDS_X96 64
44 #define DCA_SUBFRAMES 16
45 #define DCA_SUBBAND_SAMPLES 8
46 #define DCA_PCMBLOCK_SAMPLES 32
47 #define DCA_LFE_HISTORY 8
48 #define DCA_ABITS_MAX 26
50 #define DCA_CORE_CHANNELS_MAX 6
51 #define DCA_DMIX_CHANNELS_MAX 4
52 #define DCA_XXCH_CHANNELS_MAX 2
53 #define DCA_EXSS_CHANNELS_MAX 8
54 #define DCA_EXSS_CHSETS_MAX 4
56 #define DCA_FILTER_MODE_X96 0x01
57 #define DCA_FILTER_MODE_FIXED 0x02
59 enum DCACoreAudioMode {
60 DCA_AMODE_MONO, // Mode 0: A (mono)
61 DCA_AMODE_MONO_DUAL, // Mode 1: A + B (dual mono)
62 DCA_AMODE_STEREO, // Mode 2: L + R (stereo)
63 DCA_AMODE_STEREO_SUMDIFF, // Mode 3: (L+R) + (L-R) (sum-diff)
64 DCA_AMODE_STEREO_TOTAL, // Mode 4: LT + RT (left and right total)
65 DCA_AMODE_3F, // Mode 5: C + L + R
66 DCA_AMODE_2F1R, // Mode 6: L + R + S
67 DCA_AMODE_3F1R, // Mode 7: C + L + R + S
68 DCA_AMODE_2F2R, // Mode 8: L + R + SL + SR
69 DCA_AMODE_3F2R, // Mode 9: C + L + R + SL + SR
74 enum DCACoreExtAudioType {
75 DCA_EXT_AUDIO_XCH = 0,
76 DCA_EXT_AUDIO_X96 = 2,
77 DCA_EXT_AUDIO_XXCH = 6
87 typedef struct DCADSPData {
90 DECLARE_ALIGNED(32, float, hist1)[1024];
91 DECLARE_ALIGNED(32, float, hist2)[64];
94 DECLARE_ALIGNED(32, int32_t, hist1)[1024];
95 DECLARE_ALIGNED(32, int32_t, hist2)[64];
101 typedef struct DCACoreDecoder {
102 AVCodecContext *avctx;
107 int crc_present; ///< CRC present flag
108 int npcmblocks; ///< Number of PCM sample blocks
109 int frame_size; ///< Primary frame byte size
110 int audio_mode; ///< Audio channel arrangement
111 int sample_rate; ///< Core audio sampling frequency
112 int bit_rate; ///< Transmission bit rate
113 int drc_present; ///< Embedded dynamic range flag
114 int ts_present; ///< Embedded time stamp flag
115 int aux_present; ///< Auxiliary data flag
116 int ext_audio_type; ///< Extension audio descriptor flag
117 int ext_audio_present; ///< Extended coding flag
118 int sync_ssf; ///< Audio sync word insertion flag
119 int lfe_present; ///< Low frequency effects flag
120 int predictor_history; ///< Predictor history flag switch
121 int filter_perfect; ///< Multirate interpolator switch
122 int source_pcm_res; ///< Source PCM resolution
123 int es_format; ///< Extended surround (ES) mastering flag
124 int sumdiff_front; ///< Front sum/difference flag
125 int sumdiff_surround; ///< Surround sum/difference flag
127 // Primary audio coding header
128 int nsubframes; ///< Number of subframes
129 int nchannels; ///< Number of primary audio channels (incl. extension channels)
130 int ch_mask; ///< Speaker layout mask (incl. LFE and extension channels)
131 int8_t nsubbands[DCA_CHANNELS]; ///< Subband activity count
132 int8_t subband_vq_start[DCA_CHANNELS]; ///< High frequency VQ start subband
133 int8_t joint_intensity_index[DCA_CHANNELS]; ///< Joint intensity coding index
134 int8_t transition_mode_sel[DCA_CHANNELS]; ///< Transient mode code book
135 int8_t scale_factor_sel[DCA_CHANNELS]; ///< Scale factor code book
136 int8_t bit_allocation_sel[DCA_CHANNELS]; ///< Bit allocation quantizer select
137 int8_t quant_index_sel[DCA_CHANNELS][DCA_CODE_BOOKS]; ///< Quantization index codebook select
138 int32_t scale_factor_adj[DCA_CHANNELS][DCA_CODE_BOOKS]; ///< Scale factor adjustment
140 // Primary audio coding side information
141 int8_t nsubsubframes[DCA_SUBFRAMES]; ///< Subsubframe count for each subframe
142 int8_t prediction_mode[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Prediction mode
143 int16_t prediction_vq_index[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Prediction coefficients VQ address
144 int8_t bit_allocation[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Bit allocation index
145 int8_t transition_mode[DCA_SUBFRAMES][DCA_CHANNELS][DCA_SUBBANDS]; ///< Transition mode
146 int32_t scale_factors[DCA_CHANNELS][DCA_SUBBANDS][2]; ///< Scale factors (2x for transients and X96)
147 int8_t joint_scale_sel[DCA_CHANNELS]; ///< Joint subband codebook select
148 int32_t joint_scale_factors[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< Scale factors for joint subband coding
151 int prim_dmix_embedded; ///< Auxiliary dynamic downmix flag
152 int prim_dmix_type; ///< Auxiliary primary channel downmix type
153 int prim_dmix_coeff[DCA_DMIX_CHANNELS_MAX * DCA_CORE_CHANNELS_MAX]; ///< Dynamic downmix code coefficients
156 int ext_audio_mask; ///< Bit mask of fully decoded core extensions
158 // XCH extension data
159 int xch_pos; ///< Bit position of XCH frame in core substream
161 // XXCH extension data
162 int xxch_crc_present; ///< CRC presence flag for XXCH channel set header
163 int xxch_mask_nbits; ///< Number of bits for loudspeaker mask
164 int xxch_core_mask; ///< Core loudspeaker activity mask
165 int xxch_spkr_mask; ///< Loudspeaker layout mask
166 int xxch_dmix_embedded; ///< Downmix already performed by encoder
167 int xxch_dmix_scale_inv; ///< Downmix scale factor
168 int xxch_dmix_mask[DCA_XXCH_CHANNELS_MAX]; ///< Downmix channel mapping mask
169 int xxch_dmix_coeff[DCA_XXCH_CHANNELS_MAX * DCA_CORE_CHANNELS_MAX]; ///< Downmix coefficients
170 int xxch_pos; ///< Bit position of XXCH frame in core substream
172 // X96 extension data
173 int x96_rev_no; ///< X96 revision number
174 int x96_crc_present; ///< CRC presence flag for X96 channel set header
175 int x96_nchannels; ///< Number of primary channels in X96 extension
176 int x96_high_res; ///< X96 high resolution flag
177 int x96_subband_start; ///< First encoded subband in X96 extension
178 int x96_rand; ///< Random seed for generating samples for unallocated X96 subbands
179 int x96_pos; ///< Bit position of X96 frame in core substream
182 unsigned int x96_subband_size;
183 int32_t *x96_subband_buffer; ///< X96 subband sample buffer base
184 int32_t *x96_subband_samples[DCA_CHANNELS][DCA_SUBBANDS_X96]; ///< X96 subband samples
186 unsigned int subband_size;
187 int32_t *subband_buffer; ///< Subband sample buffer base
188 int32_t *subband_samples[DCA_CHANNELS][DCA_SUBBANDS]; ///< Subband samples
189 int32_t *lfe_samples; ///< Decimated LFE samples
192 DCADSPData dcadsp_data[DCA_CHANNELS]; ///< FIR history buffers
193 DCADSPContext *dcadsp;
194 DCADCTContext dcadct;
196 SynthFilterContext synth;
197 AVFloatDSPContext *float_dsp;
198 AVFixedDSPContext *fixed_dsp;
201 unsigned int output_size;
202 void *output_buffer; ///< PCM output buffer base
203 int32_t *output_samples[DCA_SPEAKER_COUNT]; ///< PCM output for fixed point mode
204 int32_t output_history_lfe_fixed; ///< LFE PCM history for X96 filter
205 float output_history_lfe_float; ///< LFE PCM history for X96 filter
207 int ch_remap[DCA_SPEAKER_COUNT]; ///< Channel to speaker map
208 int request_mask; ///< Requested channel layout (for stereo downmix)
210 int npcmsamples; ///< Number of PCM samples per channel
211 int output_rate; ///< Output sample rate (1x or 2x header rate)
213 int filter_mode; ///< Previous filtering mode for detecting changes
216 static inline int ff_dca_core_map_spkr(DCACoreDecoder *core, int spkr)
218 if (core->ch_mask & (1U << spkr))
220 if (spkr == DCA_SPEAKER_Lss && (core->ch_mask & DCA_SPEAKER_MASK_Ls))
221 return DCA_SPEAKER_Ls;
222 if (spkr == DCA_SPEAKER_Rss && (core->ch_mask & DCA_SPEAKER_MASK_Rs))
223 return DCA_SPEAKER_Rs;
227 static inline void ff_dca_core_dequantize(int32_t *output, const int32_t *input,
228 int32_t step_size, int32_t scale, int residual, int len)
230 // Account for quantizer step size
231 int64_t step_scale = (int64_t)step_size * scale;
234 // Limit scale factor resolution to 22 bits
235 if (step_scale > (1 << 23)) {
236 shift = av_log2(step_scale >> 23) + 1;
237 step_scale >>= shift;
242 for (n = 0; n < len; n++)
243 output[n] += clip23(norm__(input[n] * step_scale, 22 - shift));
245 for (n = 0; n < len; n++)
246 output[n] = clip23(norm__(input[n] * step_scale, 22 - shift));
250 int ff_dca_core_parse(DCACoreDecoder *s, uint8_t *data, int size);
251 int ff_dca_core_parse_exss(DCACoreDecoder *s, uint8_t *data, DCAExssAsset *asset);
252 int ff_dca_core_filter_fixed(DCACoreDecoder *s, int x96_synth);
253 int ff_dca_core_filter_frame(DCACoreDecoder *s, AVFrame *frame);
254 av_cold void ff_dca_core_flush(DCACoreDecoder *s);
255 av_cold int ff_dca_core_init(DCACoreDecoder *s);
256 av_cold void ff_dca_core_close(DCACoreDecoder *s);