1/* 2 * audio encoder psychoacoustic model 3 * Copyright (C) 2008 Konstantin Shishkov 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#ifndef AVCODEC_PSYMODEL_H 23#define AVCODEC_PSYMODEL_H 24 25#include "avcodec.h" 26 27/** maximum possible number of bands */ 28#define PSY_MAX_BANDS 128 29/** maximum number of channels */ 30#define PSY_MAX_CHANS 20 31 32/* cutoff for VBR is purposely increased, since LP filtering actually 33 * hinders VBR performance rather than the opposite 34 */ 35#define AAC_CUTOFF_FROM_BITRATE(bit_rate,channels,sample_rate) (bit_rate ? FFMIN3(FFMIN3( \ 36 FFMAX(bit_rate/channels/5, bit_rate/channels*15/32 - 5500), \ 37 3000 + bit_rate/channels/4, \ 38 12000 + bit_rate/channels/16), \ 39 22000, \ 40 sample_rate / 2): (sample_rate / 2)) 41#define AAC_CUTOFF(s) ( \ 42 (s->flags & AV_CODEC_FLAG_QSCALE) \ 43 ? s->sample_rate / 2 \ 44 : AAC_CUTOFF_FROM_BITRATE(s->bit_rate, s->ch_layout.nb_channels, s->sample_rate) \ 45) 46 47/** 48 * single band psychoacoustic information 49 */ 50typedef struct FFPsyBand { 51 int bits; 52 float energy; 53 float threshold; 54 float spread; /* Energy spread over the band */ 55} FFPsyBand; 56 57/** 58 * single channel psychoacoustic information 59 */ 60typedef struct FFPsyChannel { 61 FFPsyBand psy_bands[PSY_MAX_BANDS]; ///< channel bands information 62 float entropy; ///< total PE for this channel 63} FFPsyChannel; 64 65/** 66 * psychoacoustic information for an arbitrary group of channels 67 */ 68typedef struct FFPsyChannelGroup { 69 FFPsyChannel *ch[PSY_MAX_CHANS]; ///< pointers to the individual channels in the group 70 uint8_t num_ch; ///< number of channels in this group 71 uint8_t coupling[PSY_MAX_BANDS]; ///< allow coupling for this band in the group 72} FFPsyChannelGroup; 73 74/** 75 * windowing related information 76 */ 77typedef struct FFPsyWindowInfo { 78 int window_type[3]; ///< window type (short/long/transitional, etc.) - current, previous and next 79 int window_shape; ///< window shape (sine/KBD/whatever) 80 int num_windows; ///< number of windows in a frame 81 int grouping[8]; ///< window grouping (for e.g. AAC) 82 float clipping[8]; ///< maximum absolute normalized intensity in the given window for clip avoidance 83 int *window_sizes; ///< sequence of window sizes inside one frame (for eg. WMA) 84} FFPsyWindowInfo; 85 86/** 87 * context used by psychoacoustic model 88 */ 89typedef struct FFPsyContext { 90 AVCodecContext *avctx; ///< encoder context 91 const struct FFPsyModel *model; ///< encoder-specific model functions 92 93 FFPsyChannel *ch; ///< single channel information 94 FFPsyChannelGroup *group; ///< channel group information 95 int num_groups; ///< number of channel groups 96 int cutoff; ///< lowpass frequency cutoff for analysis 97 98 uint8_t **bands; ///< scalefactor band sizes for possible frame sizes 99 int *num_bands; ///< number of scalefactor bands for possible frame sizes 100 int num_lens; ///< number of scalefactor band sets 101 102 struct { 103 int size; ///< size of the bitresevoir in bits 104 int bits; ///< number of bits used in the bitresevoir 105 int alloc; ///< number of bits allocated by the psy, or -1 if no allocation was done 106 } bitres; 107 108 void* model_priv_data; ///< psychoacoustic model implementation private data 109} FFPsyContext; 110 111/** 112 * codec-specific psychoacoustic model implementation 113 */ 114typedef struct FFPsyModel { 115 const char *name; 116 int (*init) (FFPsyContext *apc); 117 118 /** 119 * Suggest window sequence for channel. 120 * 121 * @param ctx model context 122 * @param audio samples for the current frame 123 * @param la lookahead samples (NULL when unavailable) 124 * @param channel number of channel element to analyze 125 * @param prev_type previous window type 126 * 127 * @return suggested window information in a structure 128 */ 129 FFPsyWindowInfo (*window)(FFPsyContext *ctx, const float *audio, const float *la, int channel, int prev_type); 130 131 /** 132 * Perform psychoacoustic analysis and set band info (threshold, energy) for a group of channels. 133 * 134 * @param ctx model context 135 * @param channel channel number of the first channel in the group to perform analysis on 136 * @param coeffs array of pointers to the transformed coefficients 137 * @param wi window information for the channels in the group 138 */ 139 void (*analyze)(FFPsyContext *ctx, int channel, const float **coeffs, const FFPsyWindowInfo *wi); 140 141 void (*end) (FFPsyContext *apc); 142} FFPsyModel; 143 144/** 145 * Initialize psychoacoustic model. 146 * 147 * @param ctx model context 148 * @param avctx codec context 149 * @param num_lens number of possible frame lengths 150 * @param bands scalefactor band lengths for all frame lengths 151 * @param num_bands number of scalefactor bands for all frame lengths 152 * @param num_groups number of channel groups 153 * @param group_map array with # of channels in group - 1, for each group 154 * 155 * @return zero if successful, a negative value if not 156 */ 157int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, int num_lens, 158 const uint8_t **bands, const int *num_bands, 159 int num_groups, const uint8_t *group_map); 160 161/** 162 * Determine what group a channel belongs to. 163 * 164 * @param ctx psymodel context 165 * @param channel channel to locate the group for 166 * 167 * @return pointer to the FFPsyChannelGroup this channel belongs to 168 */ 169FFPsyChannelGroup *ff_psy_find_group(FFPsyContext *ctx, int channel); 170 171/** 172 * Cleanup model context at the end. 173 * 174 * @param ctx model context 175 */ 176void ff_psy_end(FFPsyContext *ctx); 177 178 179/************************************************************************** 180 * Audio preprocessing stuff. * 181 * This should be moved into some audio filter eventually. * 182 **************************************************************************/ 183struct FFPsyPreprocessContext; 184 185/** 186 * psychoacoustic model audio preprocessing initialization 187 */ 188struct FFPsyPreprocessContext *ff_psy_preprocess_init(AVCodecContext *avctx); 189 190/** 191 * Preprocess several channel in audio frame in order to compress it better. 192 * 193 * @param ctx preprocessing context 194 * @param audio samples to be filtered (in place) 195 * @param channels number of channel to preprocess 196 */ 197void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels); 198 199/** 200 * Cleanup audio preprocessing module. 201 */ 202void ff_psy_preprocess_end(struct FFPsyPreprocessContext *ctx); 203 204#endif /* AVCODEC_PSYMODEL_H */ 205