2 * Header file for hardcoded Parametric Stereo tables
4 * Copyright (c) 2010 Alex Converse <alex.converse@gmail.com>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 #ifndef AVCODEC_AACPS_TABLEGEN_H
24 #define AVCODEC_AACPS_TABLEGEN_H
29 #if CONFIG_HARDCODED_TABLES
30 #define ps_tableinit()
31 #define TABLE_CONST const
32 #include "libavcodec/aacps_tables.h"
34 #include "libavutil/common.h"
35 #include "libavutil/libm.h"
36 #include "libavutil/mathematics.h"
37 #include "libavutil/mem.h"
38 #define NR_ALLPASS_BANDS20 30
39 #define NR_ALLPASS_BANDS34 50
42 static float pd_re_smooth[8*8*8];
43 static float pd_im_smooth[8*8*8];
44 static float HA[46][8][4];
45 static float HB[46][8][4];
46 static DECLARE_ALIGNED(16, float, f20_0_8) [ 8][8][2];
47 static DECLARE_ALIGNED(16, float, f34_0_12)[12][8][2];
48 static DECLARE_ALIGNED(16, float, f34_1_8) [ 8][8][2];
49 static DECLARE_ALIGNED(16, float, f34_2_4) [ 4][8][2];
50 static TABLE_CONST DECLARE_ALIGNED(16, float, Q_fract_allpass)[2][50][3][2];
51 static DECLARE_ALIGNED(16, float, phi_fract)[2][50][2];
53 static const float g0_Q8[] = {
54 0.00746082949812f, 0.02270420949825f, 0.04546865930473f, 0.07266113929591f,
55 0.09885108575264f, 0.11793710567217f, 0.125f
58 static const float g0_Q12[] = {
59 0.04081179924692f, 0.03812810994926f, 0.05144908135699f, 0.06399831151592f,
60 0.07428313801106f, 0.08100347892914f, 0.08333333333333f
63 static const float g1_Q8[] = {
64 0.01565675600122f, 0.03752716391991f, 0.05417891378782f, 0.08417044116767f,
65 0.10307344158036f, 0.12222452249753f, 0.125f
68 static const float g2_Q4[] = {
69 -0.05908211155639f, -0.04871498374946f, 0.0f, 0.07778723915851f,
70 0.16486303567403f, 0.23279856662996f, 0.25f
73 static av_cold void make_filters_from_proto(float (*filter)[8][2], const float *proto, int bands)
76 for (q = 0; q < bands; q++) {
77 for (n = 0; n < 7; n++) {
78 double theta = 2 * M_PI * (q + 0.5) * (n - 6) / bands;
79 filter[q][n][0] = proto[n] * cos(theta);
80 filter[q][n][1] = proto[n] * -sin(theta);
85 static av_cold void ps_tableinit(void)
87 static const float ipdopd_sin[] = { 0, M_SQRT1_2, 1, M_SQRT1_2, 0, -M_SQRT1_2, -1, -M_SQRT1_2 };
88 static const float ipdopd_cos[] = { 1, M_SQRT1_2, 0, -M_SQRT1_2, -1, -M_SQRT1_2, 0, M_SQRT1_2 };
91 static const float iid_par_dequant[] = {
92 //iid_par_dequant_default
93 0.05623413251903, 0.12589254117942, 0.19952623149689, 0.31622776601684,
94 0.44668359215096, 0.63095734448019, 0.79432823472428, 1,
95 1.25892541179417, 1.58489319246111, 2.23872113856834, 3.16227766016838,
96 5.01187233627272, 7.94328234724282, 17.7827941003892,
97 //iid_par_dequant_fine
98 0.00316227766017, 0.00562341325190, 0.01, 0.01778279410039,
99 0.03162277660168, 0.05623413251903, 0.07943282347243, 0.11220184543020,
100 0.15848931924611, 0.22387211385683, 0.31622776601684, 0.39810717055350,
101 0.50118723362727, 0.63095734448019, 0.79432823472428, 1,
102 1.25892541179417, 1.58489319246111, 1.99526231496888, 2.51188643150958,
103 3.16227766016838, 4.46683592150963, 6.30957344480193, 8.91250938133745,
104 12.5892541179417, 17.7827941003892, 31.6227766016838, 56.2341325190349,
105 100, 177.827941003892, 316.227766016837,
107 static const float icc_invq[] = {
108 1, 0.937, 0.84118, 0.60092, 0.36764, 0, -0.589, -1
110 static const float acos_icc_invq[] = {
111 0, 0.35685527, 0.57133466, 0.92614472, 1.1943263, M_PI/2, 2.2006171, M_PI
116 static const int8_t f_center_20[] = {
117 -3, -1, 1, 3, 5, 7, 10, 14, 18, 22,
119 static const int8_t f_center_34[] = {
120 2, 6, 10, 14, 18, 22, 26, 30,
121 34,-10, -6, -2, 51, 57, 15, 21,
122 27, 33, 39, 45, 54, 66, 78, 42,
123 102, 66, 78, 90,102,114,126, 90,
125 static const float fractional_delay_links[] = { 0.43f, 0.75f, 0.347f };
126 const float fractional_delay_gain = 0.39f;
128 for (pd0 = 0; pd0 < 8; pd0++) {
129 float pd0_re = ipdopd_cos[pd0];
130 float pd0_im = ipdopd_sin[pd0];
131 for (pd1 = 0; pd1 < 8; pd1++) {
132 float pd1_re = ipdopd_cos[pd1];
133 float pd1_im = ipdopd_sin[pd1];
134 for (pd2 = 0; pd2 < 8; pd2++) {
135 float pd2_re = ipdopd_cos[pd2];
136 float pd2_im = ipdopd_sin[pd2];
137 float re_smooth = 0.25f * pd0_re + 0.5f * pd1_re + pd2_re;
138 float im_smooth = 0.25f * pd0_im + 0.5f * pd1_im + pd2_im;
139 float pd_mag = 1 / hypot(im_smooth, re_smooth);
140 pd_re_smooth[pd0*64+pd1*8+pd2] = re_smooth * pd_mag;
141 pd_im_smooth[pd0*64+pd1*8+pd2] = im_smooth * pd_mag;
146 for (iid = 0; iid < 46; iid++) {
147 float c = iid_par_dequant[iid]; ///< Linear Inter-channel Intensity Difference
148 float c1 = (float)M_SQRT2 / sqrtf(1.0f + c*c);
150 for (icc = 0; icc < 8; icc++) {
151 /*if (PS_BASELINE || ps->icc_mode < 3)*/ {
152 float alpha = 0.5f * acos_icc_invq[icc];
153 float beta = alpha * (c1 - c2) * (float)M_SQRT1_2;
154 HA[iid][icc][0] = c2 * cosf(beta + alpha);
155 HA[iid][icc][1] = c1 * cosf(beta - alpha);
156 HA[iid][icc][2] = c2 * sinf(beta + alpha);
157 HA[iid][icc][3] = c1 * sinf(beta - alpha);
159 float alpha, gamma, mu, rho;
160 float alpha_c, alpha_s, gamma_c, gamma_s;
161 rho = FFMAX(icc_invq[icc], 0.05f);
162 alpha = 0.5f * atan2f(2.0f * c * rho, c*c - 1.0f);
164 mu = sqrtf(1 + (4 * rho * rho - 4)/(mu * mu));
165 gamma = atanf(sqrtf((1.0f - mu)/(1.0f + mu)));
166 if (alpha < 0) alpha += M_PI/2;
167 alpha_c = cosf(alpha);
168 alpha_s = sinf(alpha);
169 gamma_c = cosf(gamma);
170 gamma_s = sinf(gamma);
171 HB[iid][icc][0] = M_SQRT2 * alpha_c * gamma_c;
172 HB[iid][icc][1] = M_SQRT2 * alpha_s * gamma_c;
173 HB[iid][icc][2] = -M_SQRT2 * alpha_s * gamma_s;
174 HB[iid][icc][3] = M_SQRT2 * alpha_c * gamma_s;
179 for (k = 0; k < NR_ALLPASS_BANDS20; k++) {
180 double f_center, theta;
181 if (k < FF_ARRAY_ELEMS(f_center_20))
182 f_center = f_center_20[k] * 0.125;
185 for (m = 0; m < PS_AP_LINKS; m++) {
186 theta = -M_PI * fractional_delay_links[m] * f_center;
187 Q_fract_allpass[0][k][m][0] = cos(theta);
188 Q_fract_allpass[0][k][m][1] = sin(theta);
190 theta = -M_PI*fractional_delay_gain*f_center;
191 phi_fract[0][k][0] = cos(theta);
192 phi_fract[0][k][1] = sin(theta);
194 for (k = 0; k < NR_ALLPASS_BANDS34; k++) {
195 double f_center, theta;
196 if (k < FF_ARRAY_ELEMS(f_center_34))
197 f_center = f_center_34[k] / 24.0;
199 f_center = k - 26.5f;
200 for (m = 0; m < PS_AP_LINKS; m++) {
201 theta = -M_PI * fractional_delay_links[m] * f_center;
202 Q_fract_allpass[1][k][m][0] = cos(theta);
203 Q_fract_allpass[1][k][m][1] = sin(theta);
205 theta = -M_PI*fractional_delay_gain*f_center;
206 phi_fract[1][k][0] = cos(theta);
207 phi_fract[1][k][1] = sin(theta);
210 make_filters_from_proto(f20_0_8, g0_Q8, 8);
211 make_filters_from_proto(f34_0_12, g0_Q12, 12);
212 make_filters_from_proto(f34_1_8, g1_Q8, 8);
213 make_filters_from_proto(f34_2_4, g2_Q4, 4);
215 #endif /* CONFIG_HARDCODED_TABLES */
217 #endif /* AVCODEC_AACPS_TABLEGEN_H */