2 * audio encoder psychoacoustic model
3 * Copyright (C) 2008 Konstantin Shishkov
5 * This file is part of FFmpeg.
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.
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.
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
22 #ifndef AVCODEC_PSYMODEL_H
23 #define AVCODEC_PSYMODEL_H
27 /** maximum possible number of bands */
28 #define PSY_MAX_BANDS 128
29 /** maximum number of channels */
30 #define PSY_MAX_CHANS 20
32 #define AAC_CUTOFF(s) ((s)->bit_rate ? FFMIN3(4000 + (s)->bit_rate/8, 12000 + (s)->bit_rate/32, (s)->sample_rate / 2) : ((s)->sample_rate / 2))
35 * single band psychoacoustic information
37 typedef struct FFPsyBand {
41 float spread; /* Energy spread over the band */
45 * single channel psychoacoustic information
47 typedef struct FFPsyChannel {
48 FFPsyBand psy_bands[PSY_MAX_BANDS]; ///< channel bands information
49 float entropy; ///< total PE for this channel
53 * psychoacoustic information for an arbitrary group of channels
55 typedef struct FFPsyChannelGroup {
56 FFPsyChannel *ch[PSY_MAX_CHANS]; ///< pointers to the individual channels in the group
57 uint8_t num_ch; ///< number of channels in this group
58 uint8_t coupling[PSY_MAX_BANDS]; ///< allow coupling for this band in the group
62 * windowing related information
64 typedef struct FFPsyWindowInfo {
65 int window_type[3]; ///< window type (short/long/transitional, etc.) - current, previous and next
66 int window_shape; ///< window shape (sine/KBD/whatever)
67 int num_windows; ///< number of windows in a frame
68 int grouping[8]; ///< window grouping (for e.g. AAC)
69 int *window_sizes; ///< sequence of window sizes inside one frame (for eg. WMA)
73 * context used by psychoacoustic model
75 typedef struct FFPsyContext {
76 AVCodecContext *avctx; ///< encoder context
77 const struct FFPsyModel *model; ///< encoder-specific model functions
79 FFPsyChannel *ch; ///< single channel information
80 FFPsyChannelGroup *group; ///< channel group information
81 int num_groups; ///< number of channel groups
83 uint8_t **bands; ///< scalefactor band sizes for possible frame sizes
84 int *num_bands; ///< number of scalefactor bands for possible frame sizes
85 int num_lens; ///< number of scalefactor band sets
88 int size; ///< size of the bitresevoir in bits
89 int bits; ///< number of bits used in the bitresevoir
92 void* model_priv_data; ///< psychoacoustic model implementation private data
96 * codec-specific psychoacoustic model implementation
98 typedef struct FFPsyModel {
100 int (*init) (FFPsyContext *apc);
103 * Suggest window sequence for channel.
105 * @param ctx model context
106 * @param audio samples for the current frame
107 * @param la lookahead samples (NULL when unavailable)
108 * @param channel number of channel element to analyze
109 * @param prev_type previous window type
111 * @return suggested window information in a structure
113 FFPsyWindowInfo (*window)(FFPsyContext *ctx, const float *audio, const float *la, int channel, int prev_type);
116 * Perform psychoacoustic analysis and set band info (threshold, energy) for a group of channels.
118 * @param ctx model context
119 * @param channel channel number of the first channel in the group to perform analysis on
120 * @param coeffs array of pointers to the transformed coefficients
121 * @param wi window information for the channels in the group
123 void (*analyze)(FFPsyContext *ctx, int channel, const float **coeffs, const FFPsyWindowInfo *wi);
125 void (*end) (FFPsyContext *apc);
129 * Initialize psychoacoustic model.
131 * @param ctx model context
132 * @param avctx codec context
133 * @param num_lens number of possible frame lengths
134 * @param bands scalefactor band lengths for all frame lengths
135 * @param num_bands number of scalefactor bands for all frame lengths
136 * @param num_groups number of channel groups
137 * @param group_map array with # of channels in group - 1, for each group
139 * @return zero if successful, a negative value if not
141 int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, int num_lens,
142 const uint8_t **bands, const int *num_bands,
143 int num_groups, const uint8_t *group_map);
146 * Determine what group a channel belongs to.
148 * @param ctx psymodel context
149 * @param channel channel to locate the group for
151 * @return pointer to the FFPsyChannelGroup this channel belongs to
153 FFPsyChannelGroup *ff_psy_find_group(FFPsyContext *ctx, int channel);
156 * Cleanup model context at the end.
158 * @param ctx model context
160 void ff_psy_end(FFPsyContext *ctx);
163 /**************************************************************************
164 * Audio preprocessing stuff. *
165 * This should be moved into some audio filter eventually. *
166 **************************************************************************/
167 struct FFPsyPreprocessContext;
170 * psychoacoustic model audio preprocessing initialization
172 struct FFPsyPreprocessContext *ff_psy_preprocess_init(AVCodecContext *avctx);
175 * Preprocess several channel in audio frame in order to compress it better.
177 * @param ctx preprocessing context
178 * @param audio samples to be filtered (in place)
179 * @param channels number of channel to preprocess
181 void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels);
184 * Cleanup audio preprocessing module.
186 void ff_psy_preprocess_end(struct FFPsyPreprocessContext *ctx);
188 #endif /* AVCODEC_PSYMODEL_H */