blob: 2b5f111fbe636ce6e7eba0efef0e99c72ee588d4
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 | #include <string.h> |
23 | |
24 | #include "avcodec.h" |
25 | #include "psymodel.h" |
26 | #include "iirfilter.h" |
27 | #include "libavutil/mem.h" |
28 | |
29 | extern const FFPsyModel ff_aac_psy_model; |
30 | |
31 | av_cold int ff_psy_init(FFPsyContext *ctx, AVCodecContext *avctx, int num_lens, |
32 | const uint8_t **bands, const int* num_bands, |
33 | int num_groups, const uint8_t *group_map) |
34 | { |
35 | int i, j, k = 0; |
36 | |
37 | ctx->avctx = avctx; |
38 | ctx->ch = av_mallocz_array(sizeof(ctx->ch[0]), avctx->channels * 2); |
39 | ctx->group = av_mallocz_array(sizeof(ctx->group[0]), num_groups); |
40 | ctx->bands = av_malloc_array (sizeof(ctx->bands[0]), num_lens); |
41 | ctx->num_bands = av_malloc_array (sizeof(ctx->num_bands[0]), num_lens); |
42 | ctx->cutoff = avctx->cutoff; |
43 | |
44 | if (!ctx->ch || !ctx->group || !ctx->bands || !ctx->num_bands) { |
45 | ff_psy_end(ctx); |
46 | return AVERROR(ENOMEM); |
47 | } |
48 | |
49 | memcpy(ctx->bands, bands, sizeof(ctx->bands[0]) * num_lens); |
50 | memcpy(ctx->num_bands, num_bands, sizeof(ctx->num_bands[0]) * num_lens); |
51 | |
52 | /* assign channels to groups (with virtual channels for coupling) */ |
53 | for (i = 0; i < num_groups; i++) { |
54 | /* NOTE: Add 1 to handle the AAC chan_config without modification. |
55 | * This has the side effect of allowing an array of 0s to map |
56 | * to one channel per group. |
57 | */ |
58 | ctx->group[i].num_ch = group_map[i] + 1; |
59 | for (j = 0; j < ctx->group[i].num_ch * 2; j++) |
60 | ctx->group[i].ch[j] = &ctx->ch[k++]; |
61 | } |
62 | |
63 | switch (ctx->avctx->codec_id) { |
64 | case AV_CODEC_ID_AAC: |
65 | ctx->model = &ff_aac_psy_model; |
66 | break; |
67 | } |
68 | if (ctx->model->init) |
69 | return ctx->model->init(ctx); |
70 | return 0; |
71 | } |
72 | |
73 | FFPsyChannelGroup *ff_psy_find_group(FFPsyContext *ctx, int channel) |
74 | { |
75 | int i = 0, ch = 0; |
76 | |
77 | while (ch <= channel) |
78 | ch += ctx->group[i++].num_ch; |
79 | |
80 | return &ctx->group[i-1]; |
81 | } |
82 | |
83 | av_cold void ff_psy_end(FFPsyContext *ctx) |
84 | { |
85 | if (ctx->model && ctx->model->end) |
86 | ctx->model->end(ctx); |
87 | av_freep(&ctx->bands); |
88 | av_freep(&ctx->num_bands); |
89 | av_freep(&ctx->group); |
90 | av_freep(&ctx->ch); |
91 | } |
92 | |
93 | typedef struct FFPsyPreprocessContext{ |
94 | AVCodecContext *avctx; |
95 | float stereo_att; |
96 | struct FFIIRFilterCoeffs *fcoeffs; |
97 | struct FFIIRFilterState **fstate; |
98 | struct FFIIRFilterContext fiir; |
99 | }FFPsyPreprocessContext; |
100 | |
101 | #define FILT_ORDER 4 |
102 | |
103 | av_cold struct FFPsyPreprocessContext* ff_psy_preprocess_init(AVCodecContext *avctx) |
104 | { |
105 | FFPsyPreprocessContext *ctx; |
106 | int i; |
107 | float cutoff_coeff = 0; |
108 | ctx = av_mallocz(sizeof(FFPsyPreprocessContext)); |
109 | if (!ctx) |
110 | return NULL; |
111 | ctx->avctx = avctx; |
112 | |
113 | /* AAC has its own LP method */ |
114 | if (avctx->codec_id != AV_CODEC_ID_AAC) { |
115 | if (avctx->cutoff > 0) |
116 | cutoff_coeff = 2.0 * avctx->cutoff / avctx->sample_rate; |
117 | |
118 | if (cutoff_coeff && cutoff_coeff < 0.98) |
119 | ctx->fcoeffs = ff_iir_filter_init_coeffs(avctx, FF_FILTER_TYPE_BUTTERWORTH, |
120 | FF_FILTER_MODE_LOWPASS, FILT_ORDER, |
121 | cutoff_coeff, 0.0, 0.0); |
122 | if (ctx->fcoeffs) { |
123 | ctx->fstate = av_mallocz_array(sizeof(ctx->fstate[0]), avctx->channels); |
124 | if (!ctx->fstate) { |
125 | av_free(ctx->fcoeffs); |
126 | av_free(ctx); |
127 | return NULL; |
128 | } |
129 | for (i = 0; i < avctx->channels; i++) |
130 | ctx->fstate[i] = ff_iir_filter_init_state(FILT_ORDER); |
131 | } |
132 | } |
133 | |
134 | ff_iir_filter_init(&ctx->fiir); |
135 | |
136 | return ctx; |
137 | } |
138 | |
139 | void ff_psy_preprocess(struct FFPsyPreprocessContext *ctx, float **audio, int channels) |
140 | { |
141 | int ch; |
142 | int frame_size = ctx->avctx->frame_size; |
143 | FFIIRFilterContext *iir = &ctx->fiir; |
144 | |
145 | if (ctx->fstate) { |
146 | for (ch = 0; ch < channels; ch++) |
147 | iir->filter_flt(ctx->fcoeffs, ctx->fstate[ch], frame_size, |
148 | &audio[ch][frame_size], 1, &audio[ch][frame_size], 1); |
149 | } |
150 | } |
151 | |
152 | av_cold void ff_psy_preprocess_end(struct FFPsyPreprocessContext *ctx) |
153 | { |
154 | int i; |
155 | ff_iir_filter_free_coeffsp(&ctx->fcoeffs); |
156 | if (ctx->fstate) |
157 | for (i = 0; i < ctx->avctx->channels; i++) |
158 | ff_iir_filter_free_statep(&ctx->fstate[i]); |
159 | av_freep(&ctx->fstate); |
160 | av_free(ctx); |
161 | } |
162 |