📄 dca.c
字号:
/* * DCA compatible decoder * Copyright (C) 2004 Gildas Bazin * Copyright (C) 2004 Benjamin Zores * Copyright (C) 2006 Benjamin Larsson * Copyright (C) 2007 Konstantin Shishkov * * This file is part of FFmpeg. * * FFmpeg is free software; you can redistribute it and/or * modify it under the terms of the GNU Lesser General Public * License as published by the Free Software Foundation; either * version 2.1 of the License, or (at your option) any later version. * * FFmpeg is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU * Lesser General Public License for more details. * * You should have received a copy of the GNU Lesser General Public * License along with FFmpeg; if not, write to the Free Software * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA *//** * @file dca.c */#include <math.h>#include <stddef.h>#include <stdio.h>#include "avcodec.h"#include "dsputil.h"#include "bitstream.h"#include "dcadata.h"#include "dcahuff.h"#include "dca.h"//#define TRACE#define DCA_PRIM_CHANNELS_MAX (5)#define DCA_SUBBANDS (32)#define DCA_ABITS_MAX (32) /* Should be 28 */#define DCA_SUBSUBFAMES_MAX (4)#define DCA_LFE_MAX (3)enum DCAMode { DCA_MONO = 0, DCA_CHANNEL, DCA_STEREO, DCA_STEREO_SUMDIFF, DCA_STEREO_TOTAL, DCA_3F, DCA_2F1R, DCA_3F1R, DCA_2F2R, DCA_3F2R, DCA_4F2R};#define DCA_DOLBY 101 /* FIXME */#define DCA_CHANNEL_BITS 6#define DCA_CHANNEL_MASK 0x3F#define DCA_LFE 0x80#define HEADER_SIZE 14#define CONVERT_BIAS 384#define DCA_MAX_FRAME_SIZE 16383/** Bit allocation */typedef struct { int offset; ///< code values offset int maxbits[8]; ///< max bits in VLC int wrap; ///< wrap for get_vlc2() VLC vlc[8]; ///< actual codes} BitAlloc;static BitAlloc dca_bitalloc_index; ///< indexes for samples VLC selectstatic BitAlloc dca_tmode; ///< transition mode VLCsstatic BitAlloc dca_scalefactor; ///< scalefactor VLCsstatic BitAlloc dca_smpl_bitalloc[11]; ///< samples VLCs/** Pre-calculated cosine modulation coefs for the QMF */static float cos_mod[544];static av_always_inline int get_bitalloc(GetBitContext *gb, BitAlloc *ba, int idx){ return get_vlc2(gb, ba->vlc[idx].table, ba->vlc[idx].bits, ba->wrap) + ba->offset;}typedef struct { AVCodecContext *avctx; /* Frame header */ int frame_type; ///< type of the current frame int samples_deficit; ///< deficit sample count int crc_present; ///< crc is present in the bitstream int sample_blocks; ///< number of PCM sample blocks int frame_size; ///< primary frame byte size int amode; ///< audio channels arrangement int sample_rate; ///< audio sampling rate int bit_rate; ///< transmission bit rate int downmix; ///< embedded downmix enabled int dynrange; ///< embedded dynamic range flag int timestamp; ///< embedded time stamp flag int aux_data; ///< auxiliary data flag int hdcd; ///< source material is mastered in HDCD int ext_descr; ///< extension audio descriptor flag int ext_coding; ///< extended coding flag int aspf; ///< audio sync word insertion flag int lfe; ///< low frequency effects flag int predictor_history; ///< predictor history flag int header_crc; ///< header crc check bytes int multirate_inter; ///< multirate interpolator switch int version; ///< encoder software revision int copy_history; ///< copy history int source_pcm_res; ///< source pcm resolution int front_sum; ///< front sum/difference flag int surround_sum; ///< surround sum/difference flag int dialog_norm; ///< dialog normalisation parameter /* Primary audio coding header */ int subframes; ///< number of subframes int prim_channels; ///< number of primary audio channels int subband_activity[DCA_PRIM_CHANNELS_MAX]; ///< subband activity count int vq_start_subband[DCA_PRIM_CHANNELS_MAX]; ///< high frequency vq start subband int joint_intensity[DCA_PRIM_CHANNELS_MAX]; ///< joint intensity coding index int transient_huffman[DCA_PRIM_CHANNELS_MAX]; ///< transient mode code book int scalefactor_huffman[DCA_PRIM_CHANNELS_MAX]; ///< scale factor code book int bitalloc_huffman[DCA_PRIM_CHANNELS_MAX]; ///< bit allocation quantizer select int quant_index_huffman[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< quantization index codebook select float scalefactor_adj[DCA_PRIM_CHANNELS_MAX][DCA_ABITS_MAX]; ///< scale factor adjustment /* Primary audio coding side information */ int subsubframes; ///< number of subsubframes int partial_samples; ///< partial subsubframe samples count int prediction_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction mode (ADPCM used or not) int prediction_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< prediction VQ coefs int bitalloc[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< bit allocation index int transition_mode[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< transition mode (transients) int scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][2]; ///< scale factors (2 if transient) int joint_huff[DCA_PRIM_CHANNELS_MAX]; ///< joint subband scale factors codebook int joint_scale_factor[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< joint subband scale factors int downmix_coef[DCA_PRIM_CHANNELS_MAX][2]; ///< stereo downmix coefficients int dynrange_coef; ///< dynamic range coefficient int high_freq_vq[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS]; ///< VQ encoded high frequency subbands float lfe_data[2 * DCA_SUBSUBFAMES_MAX * DCA_LFE_MAX * 2 /*history */ ]; ///< Low frequency effect data int lfe_scale_factor; /* Subband samples history (for ADPCM) */ float subband_samples_hist[DCA_PRIM_CHANNELS_MAX][DCA_SUBBANDS][4]; float subband_fir_hist[DCA_PRIM_CHANNELS_MAX][512]; float subband_fir_noidea[DCA_PRIM_CHANNELS_MAX][64]; int output; ///< type of output int bias; ///< output bias DECLARE_ALIGNED_16(float, samples[1536]); /* 6 * 256 = 1536, might only need 5 */ DECLARE_ALIGNED_16(int16_t, tsamples[1536]); uint8_t dca_buffer[DCA_MAX_FRAME_SIZE]; int dca_buffer_size; ///< how much data is in the dca_buffer GetBitContext gb; /* Current position in DCA frame */ int current_subframe; int current_subsubframe; int debug_flag; ///< used for suppressing repeated error messages output DSPContext dsp;} DCAContext;static void dca_init_vlcs(void){ static int vlcs_inited = 0; int i, j; if (vlcs_inited) return; dca_bitalloc_index.offset = 1; dca_bitalloc_index.wrap = 2; for (i = 0; i < 5; i++) init_vlc(&dca_bitalloc_index.vlc[i], bitalloc_12_vlc_bits[i], 12, bitalloc_12_bits[i], 1, 1, bitalloc_12_codes[i], 2, 2, 1); dca_scalefactor.offset = -64; dca_scalefactor.wrap = 2; for (i = 0; i < 5; i++) init_vlc(&dca_scalefactor.vlc[i], SCALES_VLC_BITS, 129, scales_bits[i], 1, 1, scales_codes[i], 2, 2, 1); dca_tmode.offset = 0; dca_tmode.wrap = 1; for (i = 0; i < 4; i++) init_vlc(&dca_tmode.vlc[i], tmode_vlc_bits[i], 4, tmode_bits[i], 1, 1, tmode_codes[i], 2, 2, 1); for(i = 0; i < 10; i++) for(j = 0; j < 7; j++){ if(!bitalloc_codes[i][j]) break; dca_smpl_bitalloc[i+1].offset = bitalloc_offsets[i]; dca_smpl_bitalloc[i+1].wrap = 1 + (j > 4); init_vlc(&dca_smpl_bitalloc[i+1].vlc[j], bitalloc_maxbits[i][j], bitalloc_sizes[i], bitalloc_bits[i][j], 1, 1, bitalloc_codes[i][j], 2, 2, 1); } vlcs_inited = 1;}static inline void get_array(GetBitContext *gb, int *dst, int len, int bits){ while(len--) *dst++ = get_bits(gb, bits);}static int dca_parse_frame_header(DCAContext * s){ int i, j; static const float adj_table[4] = { 1.0, 1.1250, 1.2500, 1.4375 }; static const int bitlen[11] = { 0, 1, 2, 2, 2, 2, 3, 3, 3, 3, 3 }; static const int thr[11] = { 0, 1, 3, 3, 3, 3, 7, 7, 7, 7, 7 }; s->bias = CONVERT_BIAS; init_get_bits(&s->gb, s->dca_buffer, s->dca_buffer_size * 8); /* Sync code */ get_bits(&s->gb, 32); /* Frame header */ s->frame_type = get_bits(&s->gb, 1); s->samples_deficit = get_bits(&s->gb, 5) + 1; s->crc_present = get_bits(&s->gb, 1); s->sample_blocks = get_bits(&s->gb, 7) + 1; s->frame_size = get_bits(&s->gb, 14) + 1; if (s->frame_size < 95) return -1; s->amode = get_bits(&s->gb, 6); s->sample_rate = dca_sample_rates[get_bits(&s->gb, 4)]; if (!s->sample_rate) return -1; s->bit_rate = dca_bit_rates[get_bits(&s->gb, 5)]; if (!s->bit_rate) return -1; s->downmix = get_bits(&s->gb, 1); s->dynrange = get_bits(&s->gb, 1); s->timestamp = get_bits(&s->gb, 1); s->aux_data = get_bits(&s->gb, 1); s->hdcd = get_bits(&s->gb, 1); s->ext_descr = get_bits(&s->gb, 3); s->ext_coding = get_bits(&s->gb, 1); s->aspf = get_bits(&s->gb, 1); s->lfe = get_bits(&s->gb, 2); s->predictor_history = get_bits(&s->gb, 1); /* TODO: check CRC */ if (s->crc_present) s->header_crc = get_bits(&s->gb, 16); s->multirate_inter = get_bits(&s->gb, 1); s->version = get_bits(&s->gb, 4); s->copy_history = get_bits(&s->gb, 2); s->source_pcm_res = get_bits(&s->gb, 3); s->front_sum = get_bits(&s->gb, 1); s->surround_sum = get_bits(&s->gb, 1); s->dialog_norm = get_bits(&s->gb, 4); /* FIXME: channels mixing levels */ s->output = s->amode; if(s->lfe) s->output |= DCA_LFE;#ifdef TRACE av_log(s->avctx, AV_LOG_DEBUG, "frame type: %i\n", s->frame_type); av_log(s->avctx, AV_LOG_DEBUG, "samples deficit: %i\n", s->samples_deficit); av_log(s->avctx, AV_LOG_DEBUG, "crc present: %i\n", s->crc_present); av_log(s->avctx, AV_LOG_DEBUG, "sample blocks: %i (%i samples)\n", s->sample_blocks, s->sample_blocks * 32); av_log(s->avctx, AV_LOG_DEBUG, "frame size: %i bytes\n", s->frame_size); av_log(s->avctx, AV_LOG_DEBUG, "amode: %i (%i channels)\n", s->amode, dca_channels[s->amode]); av_log(s->avctx, AV_LOG_DEBUG, "sample rate: %i (%i Hz)\n", s->sample_rate, dca_sample_rates[s->sample_rate]); av_log(s->avctx, AV_LOG_DEBUG, "bit rate: %i (%i bits/s)\n", s->bit_rate, dca_bit_rates[s->bit_rate]); av_log(s->avctx, AV_LOG_DEBUG, "downmix: %i\n", s->downmix); av_log(s->avctx, AV_LOG_DEBUG, "dynrange: %i\n", s->dynrange); av_log(s->avctx, AV_LOG_DEBUG, "timestamp: %i\n", s->timestamp); av_log(s->avctx, AV_LOG_DEBUG, "aux_data: %i\n", s->aux_data); av_log(s->avctx, AV_LOG_DEBUG, "hdcd: %i\n", s->hdcd); av_log(s->avctx, AV_LOG_DEBUG, "ext descr: %i\n", s->ext_descr); av_log(s->avctx, AV_LOG_DEBUG, "ext coding: %i\n", s->ext_coding); av_log(s->avctx, AV_LOG_DEBUG, "aspf: %i\n", s->aspf); av_log(s->avctx, AV_LOG_DEBUG, "lfe: %i\n", s->lfe); av_log(s->avctx, AV_LOG_DEBUG, "predictor history: %i\n", s->predictor_history); av_log(s->avctx, AV_LOG_DEBUG, "header crc: %i\n", s->header_crc); av_log(s->avctx, AV_LOG_DEBUG, "multirate inter: %i\n", s->multirate_inter); av_log(s->avctx, AV_LOG_DEBUG, "version number: %i\n", s->version); av_log(s->avctx, AV_LOG_DEBUG, "copy history: %i\n", s->copy_history); av_log(s->avctx, AV_LOG_DEBUG, "source pcm resolution: %i (%i bits/sample)\n", s->source_pcm_res, dca_bits_per_sample[s->source_pcm_res]); av_log(s->avctx, AV_LOG_DEBUG, "front sum: %i\n", s->front_sum); av_log(s->avctx, AV_LOG_DEBUG, "surround sum: %i\n", s->surround_sum); av_log(s->avctx, AV_LOG_DEBUG, "dialog norm: %i\n", s->dialog_norm); av_log(s->avctx, AV_LOG_DEBUG, "\n");#endif /* Primary audio coding header */ s->subframes = get_bits(&s->gb, 4) + 1; s->prim_channels = get_bits(&s->gb, 3) + 1; for (i = 0; i < s->prim_channels; i++) { s->subband_activity[i] = get_bits(&s->gb, 5) + 2; if (s->subband_activity[i] > DCA_SUBBANDS) s->subband_activity[i] = DCA_SUBBANDS; } for (i = 0; i < s->prim_channels; i++) { s->vq_start_subband[i] = get_bits(&s->gb, 5) + 1; if (s->vq_start_subband[i] > DCA_SUBBANDS) s->vq_start_subband[i] = DCA_SUBBANDS; } get_array(&s->gb, s->joint_intensity, s->prim_channels, 3); get_array(&s->gb, s->transient_huffman, s->prim_channels, 2); get_array(&s->gb, s->scalefactor_huffman, s->prim_channels, 3); get_array(&s->gb, s->bitalloc_huffman, s->prim_channels, 3); /* Get codebooks quantization indexes */ memset(s->quant_index_huffman, 0, sizeof(s->quant_index_huffman)); for (j = 1; j < 11; j++) for (i = 0; i < s->prim_channels; i++) s->quant_index_huffman[i][j] = get_bits(&s->gb, bitlen[j]); /* Get scale factor adjustment */ for (j = 0; j < 11; j++) for (i = 0; i < s->prim_channels; i++) s->scalefactor_adj[i][j] = 1; for (j = 1; j < 11; j++) for (i = 0; i < s->prim_channels; i++) if (s->quant_index_huffman[i][j] < thr[j]) s->scalefactor_adj[i][j] = adj_table[get_bits(&s->gb, 2)]; if (s->crc_present) { /* Audio header CRC check */ get_bits(&s->gb, 16); } s->current_subframe = 0; s->current_subsubframe = 0;#ifdef TRACE av_log(s->avctx, AV_LOG_DEBUG, "subframes: %i\n", s->subframes); av_log(s->avctx, AV_LOG_DEBUG, "prim channels: %i\n", s->prim_channels); for(i = 0; i < s->prim_channels; i++){ av_log(s->avctx, AV_LOG_DEBUG, "subband activity: %i\n", s->subband_activity[i]); av_log(s->avctx, AV_LOG_DEBUG, "vq start subband: %i\n", s->vq_start_subband[i]); av_log(s->avctx, AV_LOG_DEBUG, "joint intensity: %i\n", s->joint_intensity[i]); av_log(s->avctx, AV_LOG_DEBUG, "transient mode codebook: %i\n", s->transient_huffman[i]); av_log(s->avctx, AV_LOG_DEBUG, "scale factor codebook: %i\n", s->scalefactor_huffman[i]); av_log(s->avctx, AV_LOG_DEBUG, "bit allocation quantizer: %i\n", s->bitalloc_huffman[i]); av_log(s->avctx, AV_LOG_DEBUG, "quant index huff:"); for (j = 0; j < 11; j++) av_log(s->avctx, AV_LOG_DEBUG, " %i", s->quant_index_huffman[i][j]); av_log(s->avctx, AV_LOG_DEBUG, "\n"); av_log(s->avctx, AV_LOG_DEBUG, "scalefac adj:"); for (j = 0; j < 11; j++) av_log(s->avctx, AV_LOG_DEBUG, " %1.3f", s->scalefactor_adj[i][j]); av_log(s->avctx, AV_LOG_DEBUG, "\n"); }#endif return 0;}static inline int get_scale(GetBitContext *gb, int level, int value){ if (level < 5) { /* huffman encoded */ value += get_bitalloc(gb, &dca_scalefactor, level); } else if(level < 8) value = get_bits(gb, level + 1); return value;}static int dca_subframe_header(DCAContext * s){ /* Primary audio coding side information */ int j, k; s->subsubframes = get_bits(&s->gb, 2) + 1; s->partial_samples = get_bits(&s->gb, 3); for (j = 0; j < s->prim_channels; j++) { for (k = 0; k < s->subband_activity[j]; k++) s->prediction_mode[j][k] = get_bits(&s->gb, 1); } /* Get prediction codebook */ for (j = 0; j < s->prim_channels; j++) { for (k = 0; k < s->subband_activity[j]; k++) { if (s->prediction_mode[j][k] > 0) { /* (Prediction coefficient VQ address) */ s->prediction_vq[j][k] = get_bits(&s->gb, 12); }
⌨️ 快捷键说明
复制代码
Ctrl + C
搜索代码
Ctrl + F
全屏模式
F11
切换主题
Ctrl + Shift + D
显示快捷键
?
增大字号
Ctrl + =
减小字号
Ctrl + -