2 * DCA compatible decoder
3 * Copyright (C) 2004 Gildas Bazin
4 * Copyright (C) 2004 Benjamin Zores
5 * Copyright (C) 2006 Benjamin Larsson
6 * Copyright (C) 2007 Konstantin Shishkov
8 * This file is part of FFmpeg.
10 * FFmpeg is free software; you can redistribute it and/or
11 * modify it under the terms of the GNU Lesser General Public
12 * License as published by the Free Software Foundation; either
13 * version 2.1 of the License, or (at your option) any later version.
15 * FFmpeg is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
18 * Lesser General Public License for more details.
20 * You should have received a copy of the GNU Lesser General Public
21 * License along with FFmpeg; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
30 #include "libavutil/float_dsp.h"
31 #include "libavutil/internal.h"
35 #include "fmtconvert.h"
38 #define DCA_PRIM_CHANNELS_MAX (7)
39 #define DCA_ABITS_MAX (32) /* Should be 28 */
40 #define DCA_SUBSUBFRAMES_MAX (4)
41 #define DCA_SUBFRAMES_MAX (16)
42 #define DCA_BLOCKS_MAX (16)
43 #define DCA_LFE_MAX (3)
44 #define DCA_CHSETS_MAX (4)
45 #define DCA_CHSET_CHANS_MAX (8)
47 #define DCA_PRIM_CHANNELS_MAX (7)
48 #define DCA_ABITS_MAX (32) /* Should be 28 */
49 #define DCA_SUBSUBFRAMES_MAX (4)
50 #define DCA_SUBFRAMES_MAX (16)
51 #define DCA_BLOCKS_MAX (16)
52 #define DCA_LFE_MAX (3)
53 #define DCA_XLL_FBANDS_MAX (4)
54 #define DCA_XLL_SEGMENTS_MAX (16)
55 #define DCA_XLL_CHSETS_MAX (16)
56 #define DCA_XLL_CHANNELS_MAX (16)
57 #define DCA_XLL_AORDER_MAX (15)
59 /* Arbitrary limit; not sure what the maximum really is, but much larger. */
60 #define DCA_XLL_DMIX_NCOEFFS_MAX (18)
62 #define DCA_MAX_FRAME_SIZE 16384
63 #define DCA_MAX_EXSS_HEADER_SIZE 4096
65 #define DCA_BUFFER_PADDING_SIZE 1024
67 enum DCAExtensionMask {
68 DCA_EXT_CORE = 0x001, ///< core in core substream
69 DCA_EXT_XXCH = 0x002, ///< XXCh channels extension in core substream
70 DCA_EXT_X96 = 0x004, ///< 96/24 extension in core substream
71 DCA_EXT_XCH = 0x008, ///< XCh channel extension in core substream
72 DCA_EXT_EXSS_CORE = 0x010, ///< core in ExSS (extension substream)
73 DCA_EXT_EXSS_XBR = 0x020, ///< extended bitrate extension in ExSS
74 DCA_EXT_EXSS_XXCH = 0x040, ///< XXCh channels extension in ExSS
75 DCA_EXT_EXSS_X96 = 0x080, ///< 96/24 extension in ExSS
76 DCA_EXT_EXSS_LBR = 0x100, ///< low bitrate component in ExSS
77 DCA_EXT_EXSS_XLL = 0x200, ///< lossless extension in ExSS
80 typedef struct XllChSetSubHeader {
81 int channels; ///< number of channels in channel set, at most 16
82 int residual_encode; ///< residual channel encoding
83 int bit_resolution; ///< input sample bit-width
84 int bit_width; ///< original input sample bit-width
85 int sampling_frequency; ///< sampling frequency
86 int samp_freq_interp; ///< sampling frequency interpolation multiplier
87 int replacement_set; ///< replacement channel set group
88 int active_replace_set; ///< current channel set is active channel set
90 int downmix_coeff_code_embedded;
93 int hier_chset; ///< hierarchical channel set
95 int downmix_coeffs[DCA_XLL_DMIX_NCOEFFS_MAX];
98 int mapping_coeffs_present;
101 /* m_nOrigChanOrder */
102 uint8_t orig_chan_order[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
103 uint8_t orig_chan_order_inv[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
104 /* Coefficients for channel pairs (at most 8), m_anPWChPairsCoeffs */
105 int8_t pw_ch_pairs_coeffs[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX/2];
106 /* m_nCurrHighestLPCOrder */
107 uint8_t adapt_order_max[DCA_XLL_FBANDS_MAX];
108 /* m_pnAdaptPredOrder */
109 uint8_t adapt_order[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
110 /* m_pnFixedPredOrder */
111 uint8_t fixed_order[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
112 /* m_pnLPCReflCoeffsQInd, unsigned version */
113 uint8_t lpc_refl_coeffs_q_ind[DCA_XLL_FBANDS_MAX]
114 [DCA_XLL_CHANNELS_MAX][DCA_XLL_AORDER_MAX];
116 int lsb_fsize[DCA_XLL_FBANDS_MAX];
117 int8_t scalable_lsbs[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
118 int8_t bit_width_adj_per_ch[DCA_XLL_FBANDS_MAX][DCA_XLL_CHANNELS_MAX];
121 typedef struct XllNavi {
122 GetBitContext gb; // Context for parsing the data segments
123 unsigned band_size[DCA_XLL_FBANDS_MAX];
124 unsigned segment_size[DCA_XLL_FBANDS_MAX][DCA_XLL_SEGMENTS_MAX];
125 unsigned chset_size[DCA_XLL_FBANDS_MAX][DCA_XLL_SEGMENTS_MAX][DCA_XLL_CHSETS_MAX];
128 typedef struct QMF64_table {
129 float dct4_coeff[32][32];
130 float dct2_coeff[32][32];
135 /* Primary audio coding header */
136 typedef struct DCAAudioHeader {
137 int subband_activity[DCA_PRIM_CHANNELS_MAX]; ///< subband activity count
138 int vq_start_subband[DCA_PRIM_CHANNELS_MAX]; ///< high frequency vq start subband
139 int joint_intensity[DCA_PRIM_CHANNELS_MAX]; ///< joint intensity coding index
140 int transient_huffman[DCA_PRIM_CHANNELS_MAX]; ///< transient mode code book
141 int scalefactor_huffman[DCA_PRIM_CHANNELS_MAX]; ///< scale factor code book
142 int bitalloc_huffman[DCA_PRIM_CHANNELS_MAX]; ///< bit allocation quantizer select
143 int quant_index_huffman[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< quantization index codebook select
144 float scalefactor_adj[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< scale factor adjustment
146 int subframes; ///< number of subframes
147 int total_channels; ///< number of channels including extensions
148 int prim_channels; ///< number of primary audio channels
151 typedef struct DCAChan {
152 DECLARE_ALIGNED(32, float, subband_samples)[DCA_BLOCKS_MAX][DCA_SUBBANDS][8];
154 /* Subband samples history (for ADPCM) */
155 DECLARE_ALIGNED(16, float, subband_samples_hist)[DCA_SUBBANDS][4];
158 /* Half size is sufficient for core decoding, but for 96 kHz data
159 * we need QMF with 64 subbands and 1024 samples. */
160 DECLARE_ALIGNED(32, float, subband_fir_hist)[1024];
161 DECLARE_ALIGNED(32, float, subband_fir_noidea)[64];
163 /* Primary audio coding side information */
164 int prediction_mode[DCA_SUBBANDS]; ///< prediction mode (ADPCM used or not)
165 int prediction_vq[DCA_SUBBANDS]; ///< prediction VQ coefs
166 int bitalloc[DCA_SUBBANDS]; ///< bit allocation index
167 int transition_mode[DCA_SUBBANDS]; ///< transition mode (transients)
168 int32_t scale_factor[DCA_SUBBANDS][2];///< scale factors (2 if transient)
169 int joint_huff; ///< joint subband scale factors codebook
170 int joint_scale_factor[DCA_SUBBANDS]; ///< joint subband scale factors
172 int32_t high_freq_vq[DCA_SUBBANDS]; ///< VQ encoded high frequency subbands
176 typedef struct DCAContext {
177 const AVClass *class; ///< class for AVOptions
178 AVCodecContext *avctx;
180 int frame_type; ///< type of the current frame
181 int samples_deficit; ///< deficit sample count
182 int crc_present; ///< crc is present in the bitstream
183 int sample_blocks; ///< number of PCM sample blocks
184 int frame_size; ///< primary frame byte size
185 int amode; ///< audio channels arrangement
186 int sample_rate; ///< audio sampling rate
187 int bit_rate; ///< transmission bit rate
188 int bit_rate_index; ///< transmission bit rate index
190 int dynrange; ///< embedded dynamic range flag
191 int timestamp; ///< embedded time stamp flag
192 int aux_data; ///< auxiliary data flag
193 int hdcd; ///< source material is mastered in HDCD
194 int ext_descr; ///< extension audio descriptor flag
195 int ext_coding; ///< extended coding flag
196 int aspf; ///< audio sync word insertion flag
197 int lfe; ///< low frequency effects flag
198 int predictor_history; ///< predictor history flag
199 int header_crc; ///< header crc check bytes
200 int multirate_inter; ///< multirate interpolator switch
201 int version; ///< encoder software revision
202 int copy_history; ///< copy history
203 int source_pcm_res; ///< source pcm resolution
204 int front_sum; ///< front sum/difference flag
205 int surround_sum; ///< surround sum/difference flag
206 int dialog_norm; ///< dialog normalisation parameter
208 /* Primary audio coding header */
209 DCAAudioHeader audio_header;
211 /* Primary audio coding side information */
212 int subsubframes[DCA_SUBFRAMES_MAX]; ///< number of subsubframes
213 int partial_samples[DCA_SUBFRAMES_MAX]; ///< partial subsubframe samples count
214 float downmix_coef[DCA_PRIM_CHANNELS_MAX + 1][2]; ///< stereo downmix coefficients
215 int dynrange_coef; ///< dynamic range coefficient
217 /* Core substream's embedded downmix coefficients (cf. ETSI TS 102 114 V1.4.1)
218 * Input: primary audio channels (incl. LFE if present)
219 * Output: downmix audio channels (up to 4, no LFE) */
220 uint8_t core_downmix; ///< embedded downmix coefficients available
221 uint8_t core_downmix_amode; ///< audio channel arrangement of embedded downmix
222 uint16_t core_downmix_codes[DCA_PRIM_CHANNELS_MAX + 1][4]; ///< embedded downmix coefficients (9-bit codes)
225 float lfe_data[2 * DCA_LFE_MAX * (DCA_BLOCKS_MAX + 4)]; ///< Low frequency effect data
226 int lfe_scale_factor;
228 /* Subband samples history (for ADPCM) */
229 DECLARE_ALIGNED(32, float, raXin)[32];
231 DCAChan dca_chan[DCA_PRIM_CHANNELS_MAX];
233 int output; ///< type of output
235 float *samples_chanptr[DCA_PRIM_CHANNELS_MAX + 1];
236 float *extra_channels[DCA_PRIM_CHANNELS_MAX + 1];
237 uint8_t *extra_channels_buffer;
238 unsigned int extra_channels_buffer_size;
240 uint8_t dca_buffer[DCA_MAX_FRAME_SIZE + DCA_MAX_EXSS_HEADER_SIZE + DCA_BUFFER_PADDING_SIZE];
241 int dca_buffer_size; ///< how much data is in the dca_buffer
243 const int8_t *channel_order_tab; ///< channel reordering table, lfe and non lfe
245 /* Current position in DCA frame */
246 int current_subframe;
247 int current_subsubframe;
249 int core_ext_mask; ///< present extensions in the core substream
250 int exss_ext_mask; ///< Non-core extensions
252 /* XCh extension information */
253 int xch_present; ///< XCh extension present and valid
254 int xch_base_channel; ///< index of first (only) channel containing XCH data
255 int xch_disable; ///< whether the XCh extension should be decoded or not
257 /* XXCH extension information */
259 int xxch_nbits_spk_mask;
260 uint32_t xxch_core_spkmask;
261 uint32_t xxch_spk_masks[4]; /* speaker masks, last element is core mask */
262 int xxch_chset_nch[4];
263 float xxch_dmix_sf[DCA_CHSETS_MAX];
265 uint32_t xxch_dmix_embedded; /* lower layer has mix pre-embedded, per chset */
266 float xxch_dmix_coeff[DCA_PRIM_CHANNELS_MAX][32]; /* worst case sizing */
268 int8_t xxch_order_tab[32];
271 /* XLL extension information */
273 int xll_nch_sets; ///< number of channel sets per frame
274 int xll_channels; ///< total number of channels (in all channel sets)
275 int xll_residual_channels; ///< number of residual channels
276 int xll_segments; ///< number of segments per frame
277 int xll_log_smpl_in_seg; ///< supposedly this is "nBits4SamplLoci"
278 int xll_smpl_in_seg; ///< samples in segment per one frequency band for the first channel set
279 int xll_bits4seg_size; ///< number of bits used to read segment size
280 int xll_banddata_crc; ///< presence of CRC16 within each frequency band
281 int xll_scalable_lsb;
282 int xll_bits4ch_mask; ///< channel position mask
283 int xll_fixed_lsb_width;
284 XllChSetSubHeader xll_chsets[DCA_XLL_CHSETS_MAX];
287 unsigned int xll_sample_buf_size;
289 /* ExSS header parser */
290 int static_fields; ///< static fields present
291 int mix_metadata; ///< mixing metadata present
292 int num_mix_configs; ///< number of mix out configurations
293 int mix_config_num_ch[4]; ///< number of channels in each mix out configuration
296 int one2one_map_chtospkr;
298 int debug_flag; ///< used for suppressing repeated error messages output
299 AVFloatDSPContext *fdsp;
301 SynthFilterContext synth;
302 DCADSPContext dcadsp;
303 QMF64_table *qmf64_table;
304 FmtConvertContext fmt_conv;
307 extern av_export const uint32_t avpriv_dca_sample_rates[16];
310 * Convert bitstream to one representation based on sync marker
312 int avpriv_dca_convert_bitstream(const uint8_t *src, int src_size, uint8_t *dst,
315 int ff_dca_xbr_parse_frame(DCAContext *s);
316 int ff_dca_xxch_decode_frame(DCAContext *s);
318 void ff_dca_exss_parse_header(DCAContext *s);
320 int ff_dca_xll_decode_header(DCAContext *s);
321 int ff_dca_xll_decode_navi(DCAContext *s, int asset_end);
322 int ff_dca_xll_decode_audio(DCAContext *s, AVFrame *frame);
324 #endif /* AVCODEC_DCA_H */