Commit | Line | Data |
---|---|---|
2ba45a60 DM |
1 | /* |
2 | * Template for the Discrete Cosine Transform for 32 samples | |
3 | * Copyright (c) 2001, 2002 Fabrice Bellard | |
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 "dct32.h" | |
23 | #include "mathops.h" | |
24 | ||
25 | #if DCT32_FLOAT | |
26 | # define dct32 ff_dct32_float | |
27 | # define FIXHR(x) ((float)(x)) | |
28 | # define MULH3(x, y, s) ((s)*(y)*(x)) | |
29 | # define INTFLOAT float | |
30 | #else | |
31 | # define dct32 ff_dct32_fixed | |
32 | # define FIXHR(a) ((int)((a) * (1LL<<32) + 0.5)) | |
33 | # define MULH3(x, y, s) MULH((s)*(x), y) | |
34 | # define INTFLOAT int | |
35 | #endif | |
36 | ||
37 | ||
38 | /* tab[i][j] = 1.0 / (2.0 * cos(pi*(2*k+1) / 2^(6 - j))) */ | |
39 | ||
40 | /* cos(i*pi/64) */ | |
41 | ||
42 | #define COS0_0 FIXHR(0.50060299823519630134/2) | |
43 | #define COS0_1 FIXHR(0.50547095989754365998/2) | |
44 | #define COS0_2 FIXHR(0.51544730992262454697/2) | |
45 | #define COS0_3 FIXHR(0.53104259108978417447/2) | |
46 | #define COS0_4 FIXHR(0.55310389603444452782/2) | |
47 | #define COS0_5 FIXHR(0.58293496820613387367/2) | |
48 | #define COS0_6 FIXHR(0.62250412303566481615/2) | |
49 | #define COS0_7 FIXHR(0.67480834145500574602/2) | |
50 | #define COS0_8 FIXHR(0.74453627100229844977/2) | |
51 | #define COS0_9 FIXHR(0.83934964541552703873/2) | |
52 | #define COS0_10 FIXHR(0.97256823786196069369/2) | |
53 | #define COS0_11 FIXHR(1.16943993343288495515/4) | |
54 | #define COS0_12 FIXHR(1.48416461631416627724/4) | |
55 | #define COS0_13 FIXHR(2.05778100995341155085/8) | |
56 | #define COS0_14 FIXHR(3.40760841846871878570/8) | |
57 | #define COS0_15 FIXHR(10.19000812354805681150/32) | |
58 | ||
59 | #define COS1_0 FIXHR(0.50241928618815570551/2) | |
60 | #define COS1_1 FIXHR(0.52249861493968888062/2) | |
61 | #define COS1_2 FIXHR(0.56694403481635770368/2) | |
62 | #define COS1_3 FIXHR(0.64682178335999012954/2) | |
63 | #define COS1_4 FIXHR(0.78815462345125022473/2) | |
64 | #define COS1_5 FIXHR(1.06067768599034747134/4) | |
65 | #define COS1_6 FIXHR(1.72244709823833392782/4) | |
66 | #define COS1_7 FIXHR(5.10114861868916385802/16) | |
67 | ||
68 | #define COS2_0 FIXHR(0.50979557910415916894/2) | |
69 | #define COS2_1 FIXHR(0.60134488693504528054/2) | |
70 | #define COS2_2 FIXHR(0.89997622313641570463/2) | |
71 | #define COS2_3 FIXHR(2.56291544774150617881/8) | |
72 | ||
73 | #define COS3_0 FIXHR(0.54119610014619698439/2) | |
74 | #define COS3_1 FIXHR(1.30656296487637652785/4) | |
75 | ||
76 | #define COS4_0 FIXHR(0.70710678118654752439/2) | |
77 | ||
78 | /* butterfly operator */ | |
79 | #define BF(a, b, c, s)\ | |
80 | {\ | |
81 | tmp0 = val##a + val##b;\ | |
82 | tmp1 = val##a - val##b;\ | |
83 | val##a = tmp0;\ | |
84 | val##b = MULH3(tmp1, c, 1<<(s));\ | |
85 | } | |
86 | ||
87 | #define BF0(a, b, c, s)\ | |
88 | {\ | |
89 | tmp0 = tab[a] + tab[b];\ | |
90 | tmp1 = tab[a] - tab[b];\ | |
91 | val##a = tmp0;\ | |
92 | val##b = MULH3(tmp1, c, 1<<(s));\ | |
93 | } | |
94 | ||
95 | #define BF1(a, b, c, d)\ | |
96 | {\ | |
97 | BF(a, b, COS4_0, 1);\ | |
98 | BF(c, d,-COS4_0, 1);\ | |
99 | val##c += val##d;\ | |
100 | } | |
101 | ||
102 | #define BF2(a, b, c, d)\ | |
103 | {\ | |
104 | BF(a, b, COS4_0, 1);\ | |
105 | BF(c, d,-COS4_0, 1);\ | |
106 | val##c += val##d;\ | |
107 | val##a += val##c;\ | |
108 | val##c += val##b;\ | |
109 | val##b += val##d;\ | |
110 | } | |
111 | ||
112 | #define ADD(a, b) val##a += val##b | |
113 | ||
114 | /* DCT32 without 1/sqrt(2) coef zero scaling. */ | |
115 | void dct32(INTFLOAT *out, const INTFLOAT *tab) | |
116 | { | |
117 | INTFLOAT tmp0, tmp1; | |
118 | ||
119 | INTFLOAT val0 , val1 , val2 , val3 , val4 , val5 , val6 , val7 , | |
120 | val8 , val9 , val10, val11, val12, val13, val14, val15, | |
121 | val16, val17, val18, val19, val20, val21, val22, val23, | |
122 | val24, val25, val26, val27, val28, val29, val30, val31; | |
123 | ||
124 | /* pass 1 */ | |
125 | BF0( 0, 31, COS0_0 , 1); | |
126 | BF0(15, 16, COS0_15, 5); | |
127 | /* pass 2 */ | |
128 | BF( 0, 15, COS1_0 , 1); | |
129 | BF(16, 31,-COS1_0 , 1); | |
130 | /* pass 1 */ | |
131 | BF0( 7, 24, COS0_7 , 1); | |
132 | BF0( 8, 23, COS0_8 , 1); | |
133 | /* pass 2 */ | |
134 | BF( 7, 8, COS1_7 , 4); | |
135 | BF(23, 24,-COS1_7 , 4); | |
136 | /* pass 3 */ | |
137 | BF( 0, 7, COS2_0 , 1); | |
138 | BF( 8, 15,-COS2_0 , 1); | |
139 | BF(16, 23, COS2_0 , 1); | |
140 | BF(24, 31,-COS2_0 , 1); | |
141 | /* pass 1 */ | |
142 | BF0( 3, 28, COS0_3 , 1); | |
143 | BF0(12, 19, COS0_12, 2); | |
144 | /* pass 2 */ | |
145 | BF( 3, 12, COS1_3 , 1); | |
146 | BF(19, 28,-COS1_3 , 1); | |
147 | /* pass 1 */ | |
148 | BF0( 4, 27, COS0_4 , 1); | |
149 | BF0(11, 20, COS0_11, 2); | |
150 | /* pass 2 */ | |
151 | BF( 4, 11, COS1_4 , 1); | |
152 | BF(20, 27,-COS1_4 , 1); | |
153 | /* pass 3 */ | |
154 | BF( 3, 4, COS2_3 , 3); | |
155 | BF(11, 12,-COS2_3 , 3); | |
156 | BF(19, 20, COS2_3 , 3); | |
157 | BF(27, 28,-COS2_3 , 3); | |
158 | /* pass 4 */ | |
159 | BF( 0, 3, COS3_0 , 1); | |
160 | BF( 4, 7,-COS3_0 , 1); | |
161 | BF( 8, 11, COS3_0 , 1); | |
162 | BF(12, 15,-COS3_0 , 1); | |
163 | BF(16, 19, COS3_0 , 1); | |
164 | BF(20, 23,-COS3_0 , 1); | |
165 | BF(24, 27, COS3_0 , 1); | |
166 | BF(28, 31,-COS3_0 , 1); | |
167 | ||
168 | ||
169 | ||
170 | /* pass 1 */ | |
171 | BF0( 1, 30, COS0_1 , 1); | |
172 | BF0(14, 17, COS0_14, 3); | |
173 | /* pass 2 */ | |
174 | BF( 1, 14, COS1_1 , 1); | |
175 | BF(17, 30,-COS1_1 , 1); | |
176 | /* pass 1 */ | |
177 | BF0( 6, 25, COS0_6 , 1); | |
178 | BF0( 9, 22, COS0_9 , 1); | |
179 | /* pass 2 */ | |
180 | BF( 6, 9, COS1_6 , 2); | |
181 | BF(22, 25,-COS1_6 , 2); | |
182 | /* pass 3 */ | |
183 | BF( 1, 6, COS2_1 , 1); | |
184 | BF( 9, 14,-COS2_1 , 1); | |
185 | BF(17, 22, COS2_1 , 1); | |
186 | BF(25, 30,-COS2_1 , 1); | |
187 | ||
188 | /* pass 1 */ | |
189 | BF0( 2, 29, COS0_2 , 1); | |
190 | BF0(13, 18, COS0_13, 3); | |
191 | /* pass 2 */ | |
192 | BF( 2, 13, COS1_2 , 1); | |
193 | BF(18, 29,-COS1_2 , 1); | |
194 | /* pass 1 */ | |
195 | BF0( 5, 26, COS0_5 , 1); | |
196 | BF0(10, 21, COS0_10, 1); | |
197 | /* pass 2 */ | |
198 | BF( 5, 10, COS1_5 , 2); | |
199 | BF(21, 26,-COS1_5 , 2); | |
200 | /* pass 3 */ | |
201 | BF( 2, 5, COS2_2 , 1); | |
202 | BF(10, 13,-COS2_2 , 1); | |
203 | BF(18, 21, COS2_2 , 1); | |
204 | BF(26, 29,-COS2_2 , 1); | |
205 | /* pass 4 */ | |
206 | BF( 1, 2, COS3_1 , 2); | |
207 | BF( 5, 6,-COS3_1 , 2); | |
208 | BF( 9, 10, COS3_1 , 2); | |
209 | BF(13, 14,-COS3_1 , 2); | |
210 | BF(17, 18, COS3_1 , 2); | |
211 | BF(21, 22,-COS3_1 , 2); | |
212 | BF(25, 26, COS3_1 , 2); | |
213 | BF(29, 30,-COS3_1 , 2); | |
214 | ||
215 | /* pass 5 */ | |
216 | BF1( 0, 1, 2, 3); | |
217 | BF2( 4, 5, 6, 7); | |
218 | BF1( 8, 9, 10, 11); | |
219 | BF2(12, 13, 14, 15); | |
220 | BF1(16, 17, 18, 19); | |
221 | BF2(20, 21, 22, 23); | |
222 | BF1(24, 25, 26, 27); | |
223 | BF2(28, 29, 30, 31); | |
224 | ||
225 | /* pass 6 */ | |
226 | ||
227 | ADD( 8, 12); | |
228 | ADD(12, 10); | |
229 | ADD(10, 14); | |
230 | ADD(14, 9); | |
231 | ADD( 9, 13); | |
232 | ADD(13, 11); | |
233 | ADD(11, 15); | |
234 | ||
235 | out[ 0] = val0; | |
236 | out[16] = val1; | |
237 | out[ 8] = val2; | |
238 | out[24] = val3; | |
239 | out[ 4] = val4; | |
240 | out[20] = val5; | |
241 | out[12] = val6; | |
242 | out[28] = val7; | |
243 | out[ 2] = val8; | |
244 | out[18] = val9; | |
245 | out[10] = val10; | |
246 | out[26] = val11; | |
247 | out[ 6] = val12; | |
248 | out[22] = val13; | |
249 | out[14] = val14; | |
250 | out[30] = val15; | |
251 | ||
252 | ADD(24, 28); | |
253 | ADD(28, 26); | |
254 | ADD(26, 30); | |
255 | ADD(30, 25); | |
256 | ADD(25, 29); | |
257 | ADD(29, 27); | |
258 | ADD(27, 31); | |
259 | ||
260 | out[ 1] = val16 + val24; | |
261 | out[17] = val17 + val25; | |
262 | out[ 9] = val18 + val26; | |
263 | out[25] = val19 + val27; | |
264 | out[ 5] = val20 + val28; | |
265 | out[21] = val21 + val29; | |
266 | out[13] = val22 + val30; | |
267 | out[29] = val23 + val31; | |
268 | out[ 3] = val24 + val20; | |
269 | out[19] = val25 + val21; | |
270 | out[11] = val26 + val22; | |
271 | out[27] = val27 + val23; | |
272 | out[ 7] = val28 + val18; | |
273 | out[23] = val29 + val19; | |
274 | out[15] = val30 + val17; | |
275 | out[31] = val31; | |
276 | } |