12027
|
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 #ifdef DCT32_FLOAT
|
|
23 # define FIXHR(x) ((float)(x))
|
|
24 # define MULH3(x, y, s) ((s)*(y)*(x))
|
|
25 # define INTFLOAT float
|
|
26 #endif
|
|
27
|
|
28
|
|
29 /* tab[i][j] = 1.0 / (2.0 * cos(pi*(2*k+1) / 2^(6 - j))) */
|
|
30
|
|
31 /* cos(i*pi/64) */
|
|
32
|
|
33 #define COS0_0 FIXHR(0.50060299823519630134/2)
|
|
34 #define COS0_1 FIXHR(0.50547095989754365998/2)
|
|
35 #define COS0_2 FIXHR(0.51544730992262454697/2)
|
|
36 #define COS0_3 FIXHR(0.53104259108978417447/2)
|
|
37 #define COS0_4 FIXHR(0.55310389603444452782/2)
|
|
38 #define COS0_5 FIXHR(0.58293496820613387367/2)
|
|
39 #define COS0_6 FIXHR(0.62250412303566481615/2)
|
|
40 #define COS0_7 FIXHR(0.67480834145500574602/2)
|
|
41 #define COS0_8 FIXHR(0.74453627100229844977/2)
|
|
42 #define COS0_9 FIXHR(0.83934964541552703873/2)
|
|
43 #define COS0_10 FIXHR(0.97256823786196069369/2)
|
|
44 #define COS0_11 FIXHR(1.16943993343288495515/4)
|
|
45 #define COS0_12 FIXHR(1.48416461631416627724/4)
|
|
46 #define COS0_13 FIXHR(2.05778100995341155085/8)
|
|
47 #define COS0_14 FIXHR(3.40760841846871878570/8)
|
|
48 #define COS0_15 FIXHR(10.19000812354805681150/32)
|
|
49
|
|
50 #define COS1_0 FIXHR(0.50241928618815570551/2)
|
|
51 #define COS1_1 FIXHR(0.52249861493968888062/2)
|
|
52 #define COS1_2 FIXHR(0.56694403481635770368/2)
|
|
53 #define COS1_3 FIXHR(0.64682178335999012954/2)
|
|
54 #define COS1_4 FIXHR(0.78815462345125022473/2)
|
|
55 #define COS1_5 FIXHR(1.06067768599034747134/4)
|
|
56 #define COS1_6 FIXHR(1.72244709823833392782/4)
|
|
57 #define COS1_7 FIXHR(5.10114861868916385802/16)
|
|
58
|
|
59 #define COS2_0 FIXHR(0.50979557910415916894/2)
|
|
60 #define COS2_1 FIXHR(0.60134488693504528054/2)
|
|
61 #define COS2_2 FIXHR(0.89997622313641570463/2)
|
|
62 #define COS2_3 FIXHR(2.56291544774150617881/8)
|
|
63
|
|
64 #define COS3_0 FIXHR(0.54119610014619698439/2)
|
|
65 #define COS3_1 FIXHR(1.30656296487637652785/4)
|
|
66
|
|
67 #define COS4_0 FIXHR(0.70710678118654752439/2)
|
|
68
|
|
69 /* butterfly operator */
|
|
70 #define BF(a, b, c, s)\
|
|
71 {\
|
|
72 tmp0 = val##a + val##b;\
|
|
73 tmp1 = val##a - val##b;\
|
|
74 val##a = tmp0;\
|
|
75 val##b = MULH3(tmp1, c, 1<<(s));\
|
|
76 }
|
|
77
|
|
78 #define BF0(a, b, c, s)\
|
|
79 {\
|
|
80 tmp0 = tab[a] + tab[b];\
|
|
81 tmp1 = tab[a] - tab[b];\
|
|
82 val##a = tmp0;\
|
|
83 val##b = MULH3(tmp1, c, 1<<(s));\
|
|
84 }
|
|
85
|
|
86 #define BF1(a, b, c, d)\
|
|
87 {\
|
|
88 BF(a, b, COS4_0, 1);\
|
|
89 BF(c, d,-COS4_0, 1);\
|
|
90 val##c += val##d;\
|
|
91 }
|
|
92
|
|
93 #define BF2(a, b, c, d)\
|
|
94 {\
|
|
95 BF(a, b, COS4_0, 1);\
|
|
96 BF(c, d,-COS4_0, 1);\
|
|
97 val##c += val##d;\
|
|
98 val##a += val##c;\
|
|
99 val##c += val##b;\
|
|
100 val##b += val##d;\
|
|
101 }
|
|
102
|
|
103 #define ADD(a, b) val##a += val##b
|
|
104
|
|
105 /* DCT32 without 1/sqrt(2) coef zero scaling. */
|
|
106 static void dct32(INTFLOAT *out, const INTFLOAT *tab)
|
|
107 {
|
|
108 INTFLOAT tmp0, tmp1;
|
|
109
|
|
110 INTFLOAT val0 , val1 , val2 , val3 , val4 , val5 , val6 , val7 ,
|
|
111 val8 , val9 , val10, val11, val12, val13, val14, val15,
|
|
112 val16, val17, val18, val19, val20, val21, val22, val23,
|
|
113 val24, val25, val26, val27, val28, val29, val30, val31;
|
|
114
|
|
115 /* pass 1 */
|
|
116 BF0( 0, 31, COS0_0 , 1);
|
|
117 BF0(15, 16, COS0_15, 5);
|
|
118 /* pass 2 */
|
|
119 BF( 0, 15, COS1_0 , 1);
|
|
120 BF(16, 31,-COS1_0 , 1);
|
|
121 /* pass 1 */
|
|
122 BF0( 7, 24, COS0_7 , 1);
|
|
123 BF0( 8, 23, COS0_8 , 1);
|
|
124 /* pass 2 */
|
|
125 BF( 7, 8, COS1_7 , 4);
|
|
126 BF(23, 24,-COS1_7 , 4);
|
|
127 /* pass 3 */
|
|
128 BF( 0, 7, COS2_0 , 1);
|
|
129 BF( 8, 15,-COS2_0 , 1);
|
|
130 BF(16, 23, COS2_0 , 1);
|
|
131 BF(24, 31,-COS2_0 , 1);
|
|
132 /* pass 1 */
|
|
133 BF0( 3, 28, COS0_3 , 1);
|
|
134 BF0(12, 19, COS0_12, 2);
|
|
135 /* pass 2 */
|
|
136 BF( 3, 12, COS1_3 , 1);
|
|
137 BF(19, 28,-COS1_3 , 1);
|
|
138 /* pass 1 */
|
|
139 BF0( 4, 27, COS0_4 , 1);
|
|
140 BF0(11, 20, COS0_11, 2);
|
|
141 /* pass 2 */
|
|
142 BF( 4, 11, COS1_4 , 1);
|
|
143 BF(20, 27,-COS1_4 , 1);
|
|
144 /* pass 3 */
|
|
145 BF( 3, 4, COS2_3 , 3);
|
|
146 BF(11, 12,-COS2_3 , 3);
|
|
147 BF(19, 20, COS2_3 , 3);
|
|
148 BF(27, 28,-COS2_3 , 3);
|
|
149 /* pass 4 */
|
|
150 BF( 0, 3, COS3_0 , 1);
|
|
151 BF( 4, 7,-COS3_0 , 1);
|
|
152 BF( 8, 11, COS3_0 , 1);
|
|
153 BF(12, 15,-COS3_0 , 1);
|
|
154 BF(16, 19, COS3_0 , 1);
|
|
155 BF(20, 23,-COS3_0 , 1);
|
|
156 BF(24, 27, COS3_0 , 1);
|
|
157 BF(28, 31,-COS3_0 , 1);
|
|
158
|
|
159
|
|
160
|
|
161 /* pass 1 */
|
|
162 BF0( 1, 30, COS0_1 , 1);
|
|
163 BF0(14, 17, COS0_14, 3);
|
|
164 /* pass 2 */
|
|
165 BF( 1, 14, COS1_1 , 1);
|
|
166 BF(17, 30,-COS1_1 , 1);
|
|
167 /* pass 1 */
|
|
168 BF0( 6, 25, COS0_6 , 1);
|
|
169 BF0( 9, 22, COS0_9 , 1);
|
|
170 /* pass 2 */
|
|
171 BF( 6, 9, COS1_6 , 2);
|
|
172 BF(22, 25,-COS1_6 , 2);
|
|
173 /* pass 3 */
|
|
174 BF( 1, 6, COS2_1 , 1);
|
|
175 BF( 9, 14,-COS2_1 , 1);
|
|
176 BF(17, 22, COS2_1 , 1);
|
|
177 BF(25, 30,-COS2_1 , 1);
|
|
178
|
|
179 /* pass 1 */
|
|
180 BF0( 2, 29, COS0_2 , 1);
|
|
181 BF0(13, 18, COS0_13, 3);
|
|
182 /* pass 2 */
|
|
183 BF( 2, 13, COS1_2 , 1);
|
|
184 BF(18, 29,-COS1_2 , 1);
|
|
185 /* pass 1 */
|
|
186 BF0( 5, 26, COS0_5 , 1);
|
|
187 BF0(10, 21, COS0_10, 1);
|
|
188 /* pass 2 */
|
|
189 BF( 5, 10, COS1_5 , 2);
|
|
190 BF(21, 26,-COS1_5 , 2);
|
|
191 /* pass 3 */
|
|
192 BF( 2, 5, COS2_2 , 1);
|
|
193 BF(10, 13,-COS2_2 , 1);
|
|
194 BF(18, 21, COS2_2 , 1);
|
|
195 BF(26, 29,-COS2_2 , 1);
|
|
196 /* pass 4 */
|
|
197 BF( 1, 2, COS3_1 , 2);
|
|
198 BF( 5, 6,-COS3_1 , 2);
|
|
199 BF( 9, 10, COS3_1 , 2);
|
|
200 BF(13, 14,-COS3_1 , 2);
|
|
201 BF(17, 18, COS3_1 , 2);
|
|
202 BF(21, 22,-COS3_1 , 2);
|
|
203 BF(25, 26, COS3_1 , 2);
|
|
204 BF(29, 30,-COS3_1 , 2);
|
|
205
|
|
206 /* pass 5 */
|
|
207 BF1( 0, 1, 2, 3);
|
|
208 BF2( 4, 5, 6, 7);
|
|
209 BF1( 8, 9, 10, 11);
|
|
210 BF2(12, 13, 14, 15);
|
|
211 BF1(16, 17, 18, 19);
|
|
212 BF2(20, 21, 22, 23);
|
|
213 BF1(24, 25, 26, 27);
|
|
214 BF2(28, 29, 30, 31);
|
|
215
|
|
216 /* pass 6 */
|
|
217
|
|
218 ADD( 8, 12);
|
|
219 ADD(12, 10);
|
|
220 ADD(10, 14);
|
|
221 ADD(14, 9);
|
|
222 ADD( 9, 13);
|
|
223 ADD(13, 11);
|
|
224 ADD(11, 15);
|
|
225
|
|
226 out[ 0] = val0;
|
|
227 out[16] = val1;
|
|
228 out[ 8] = val2;
|
|
229 out[24] = val3;
|
|
230 out[ 4] = val4;
|
|
231 out[20] = val5;
|
|
232 out[12] = val6;
|
|
233 out[28] = val7;
|
|
234 out[ 2] = val8;
|
|
235 out[18] = val9;
|
|
236 out[10] = val10;
|
|
237 out[26] = val11;
|
|
238 out[ 6] = val12;
|
|
239 out[22] = val13;
|
|
240 out[14] = val14;
|
|
241 out[30] = val15;
|
|
242
|
|
243 ADD(24, 28);
|
|
244 ADD(28, 26);
|
|
245 ADD(26, 30);
|
|
246 ADD(30, 25);
|
|
247 ADD(25, 29);
|
|
248 ADD(29, 27);
|
|
249 ADD(27, 31);
|
|
250
|
|
251 out[ 1] = val16 + val24;
|
|
252 out[17] = val17 + val25;
|
|
253 out[ 9] = val18 + val26;
|
|
254 out[25] = val19 + val27;
|
|
255 out[ 5] = val20 + val28;
|
|
256 out[21] = val21 + val29;
|
|
257 out[13] = val22 + val30;
|
|
258 out[29] = val23 + val31;
|
|
259 out[ 3] = val24 + val20;
|
|
260 out[19] = val25 + val21;
|
|
261 out[11] = val26 + val22;
|
|
262 out[27] = val27 + val23;
|
|
263 out[ 7] = val28 + val18;
|
|
264 out[23] = val29 + val19;
|
|
265 out[15] = val30 + val17;
|
|
266 out[31] = val31;
|
|
267 }
|