Parent Directory | Revision Log
Revision 1626 - (view) (download)
1 : | edgomez | 1382 | /***************************************************************************** |
2 : | * | ||
3 : | * XVID MPEG-4 VIDEO CODEC | ||
4 : | * - QPel interpolation - | ||
5 : | * | ||
6 : | * Copyright(C) 2003 Pascal Massimino <skal@planet-d.net> | ||
7 : | * | ||
8 : | * This program is free software ; you can redistribute it and/or modify | ||
9 : | * it under the terms of the GNU General Public License as published by | ||
10 : | * the Free Software Foundation ; either version 2 of the License, or | ||
11 : | * (at your option) any later version. | ||
12 : | * | ||
13 : | * This program 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 | ||
16 : | * GNU General Public License for more details. | ||
17 : | * | ||
18 : | * You should have received a copy of the GNU General Public License | ||
19 : | * along with this program ; if not, write to the Free Software | ||
20 : | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
21 : | * | ||
22 : | Skal | 1626 | * $Id: qpel.c,v 1.6 2005-07-11 22:10:24 Skal Exp $ |
23 : | edgomez | 1382 | * |
24 : | ****************************************************************************/ | ||
25 : | |||
26 : | #ifndef XVID_AUTO_INCLUDE | ||
27 : | |||
28 : | #include "../portab.h" | ||
29 : | #include "qpel.h" | ||
30 : | |||
31 : | /* Quarterpel FIR definition | ||
32 : | ****************************************************************************/ | ||
33 : | |||
34 : | static const int32_t FIR_Tab_8[9][8] = { | ||
35 : | { 14, -3, 2, -1, 0, 0, 0, 0 }, | ||
36 : | { 23, 19, -6, 3, -1, 0, 0, 0 }, | ||
37 : | { -7, 20, 20, -6, 3, -1, 0, 0 }, | ||
38 : | { 3, -6, 20, 20, -6, 3, -1, 0 }, | ||
39 : | { -1, 3, -6, 20, 20, -6, 3, -1 }, | ||
40 : | { 0, -1, 3, -6, 20, 20, -6, 3 }, | ||
41 : | { 0, 0, -1, 3, -6, 20, 20, -7 }, | ||
42 : | { 0, 0, 0, -1, 3, -6, 19, 23 }, | ||
43 : | { 0, 0, 0, 0, -1, 2, -3, 14 } | ||
44 : | }; | ||
45 : | |||
46 : | static const int32_t FIR_Tab_16[17][16] = { | ||
47 : | { 14, -3, 2, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, | ||
48 : | { 23, 19, -6, 3, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, | ||
49 : | { -7, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, | ||
50 : | { 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0 }, | ||
51 : | { -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0, 0, 0, 0 }, | ||
52 : | { 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0, 0, 0 }, | ||
53 : | { 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0, 0 }, | ||
54 : | { 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0, 0 }, | ||
55 : | { 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0, 0 }, | ||
56 : | { 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0, 0 }, | ||
57 : | { 0, 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0, 0 }, | ||
58 : | { 0, 0, 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1, 0 }, | ||
59 : | { 0, 0, 0, 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3, -1 }, | ||
60 : | { 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -6, 3 }, | ||
61 : | { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 3, -6, 20, 20, -7 }, | ||
62 : | { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 3, -6, 19, 23 }, | ||
63 : | { 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, -1, 2, -3, 14 } | ||
64 : | }; | ||
65 : | |||
66 : | /* Implementation | ||
67 : | ****************************************************************************/ | ||
68 : | |||
69 : | #define XVID_AUTO_INCLUDE | ||
70 : | edgomez | 1530 | /* First auto include this file to generate reference code for SIMD versions |
71 : | * This set of functions are good for educational purpose, because they're | ||
72 : | * straightforward to understand, use loops and so on... But obviously they | ||
73 : | * sux when it comes to speed */ | ||
74 : | #define REFERENCE_CODE | ||
75 : | edgomez | 1382 | |
76 : | /* 16x? filters */ | ||
77 : | |||
78 : | #define SIZE 16 | ||
79 : | #define TABLE FIR_Tab_16 | ||
80 : | |||
81 : | #define STORE(d,s) (d) = (s) | ||
82 : | edgomez | 1530 | #define FUNC_H H_Pass_16_C_ref |
83 : | #define FUNC_V V_Pass_16_C_ref | ||
84 : | #define FUNC_HA H_Pass_Avrg_16_C_ref | ||
85 : | #define FUNC_VA V_Pass_Avrg_16_C_ref | ||
86 : | #define FUNC_HA_UP H_Pass_Avrg_Up_16_C_ref | ||
87 : | #define FUNC_VA_UP V_Pass_Avrg_Up_16_C_ref | ||
88 : | |||
89 : | #include "qpel.c" | ||
90 : | |||
91 : | /* note: B-frame always uses Rnd=0... */ | ||
92 : | #define STORE(d,s) (d) = ( (s)+(d)+1 ) >> 1 | ||
93 : | #define FUNC_H H_Pass_16_Add_C_ref | ||
94 : | #define FUNC_V V_Pass_16_Add_C_ref | ||
95 : | #define FUNC_HA H_Pass_Avrg_16_Add_C_ref | ||
96 : | #define FUNC_VA V_Pass_Avrg_16_Add_C_ref | ||
97 : | #define FUNC_HA_UP H_Pass_Avrg_Up_16_Add_C_ref | ||
98 : | #define FUNC_VA_UP V_Pass_Avrg_Up_16_Add_C_ref | ||
99 : | |||
100 : | #include "qpel.c" | ||
101 : | |||
102 : | #undef SIZE | ||
103 : | #undef TABLE | ||
104 : | |||
105 : | /* 8x? filters */ | ||
106 : | |||
107 : | #define SIZE 8 | ||
108 : | #define TABLE FIR_Tab_8 | ||
109 : | |||
110 : | #define STORE(d,s) (d) = (s) | ||
111 : | #define FUNC_H H_Pass_8_C_ref | ||
112 : | #define FUNC_V V_Pass_8_C_ref | ||
113 : | #define FUNC_HA H_Pass_Avrg_8_C_ref | ||
114 : | #define FUNC_VA V_Pass_Avrg_8_C_ref | ||
115 : | #define FUNC_HA_UP H_Pass_Avrg_Up_8_C_ref | ||
116 : | #define FUNC_VA_UP V_Pass_Avrg_Up_8_C_ref | ||
117 : | |||
118 : | #include "qpel.c" | ||
119 : | |||
120 : | /* note: B-frame always uses Rnd=0... */ | ||
121 : | #define STORE(d,s) (d) = ( (s)+(d)+1 ) >> 1 | ||
122 : | #define FUNC_H H_Pass_8_Add_C_ref | ||
123 : | #define FUNC_V V_Pass_8_Add_C_ref | ||
124 : | #define FUNC_HA H_Pass_Avrg_8_Add_C_ref | ||
125 : | #define FUNC_VA V_Pass_Avrg_8_Add_C_ref | ||
126 : | #define FUNC_HA_UP H_Pass_Avrg_Up_8_Add_C_ref | ||
127 : | #define FUNC_VA_UP V_Pass_Avrg_Up_8_Add_C_ref | ||
128 : | |||
129 : | #include "qpel.c" | ||
130 : | |||
131 : | #undef SIZE | ||
132 : | #undef TABLE | ||
133 : | |||
134 : | /* Then we define more optimized C version where loops are unrolled, where | ||
135 : | * FIR coeffcients are not read from memory but are hardcoded in instructions | ||
136 : | * They should be faster */ | ||
137 : | #undef REFERENCE_CODE | ||
138 : | |||
139 : | /* 16x? filters */ | ||
140 : | |||
141 : | #define SIZE 16 | ||
142 : | |||
143 : | #define STORE(d,s) (d) = (s) | ||
144 : | edgomez | 1382 | #define FUNC_H H_Pass_16_C |
145 : | #define FUNC_V V_Pass_16_C | ||
146 : | #define FUNC_HA H_Pass_Avrg_16_C | ||
147 : | #define FUNC_VA V_Pass_Avrg_16_C | ||
148 : | #define FUNC_HA_UP H_Pass_Avrg_Up_16_C | ||
149 : | #define FUNC_VA_UP V_Pass_Avrg_Up_16_C | ||
150 : | |||
151 : | edgomez | 1530 | #include "qpel.c" |
152 : | edgomez | 1382 | |
153 : | /* note: B-frame always uses Rnd=0... */ | ||
154 : | #define STORE(d,s) (d) = ( (s)+(d)+1 ) >> 1 | ||
155 : | #define FUNC_H H_Pass_16_Add_C | ||
156 : | #define FUNC_V V_Pass_16_Add_C | ||
157 : | #define FUNC_HA H_Pass_Avrg_16_Add_C | ||
158 : | #define FUNC_VA V_Pass_Avrg_16_Add_C | ||
159 : | #define FUNC_HA_UP H_Pass_Avrg_Up_16_Add_C | ||
160 : | #define FUNC_VA_UP V_Pass_Avrg_Up_16_Add_C | ||
161 : | |||
162 : | edgomez | 1530 | #include "qpel.c" |
163 : | edgomez | 1382 | |
164 : | #undef SIZE | ||
165 : | #undef TABLE | ||
166 : | |||
167 : | /* 8x? filters */ | ||
168 : | |||
169 : | #define SIZE 8 | ||
170 : | #define TABLE FIR_Tab_8 | ||
171 : | |||
172 : | #define STORE(d,s) (d) = (s) | ||
173 : | #define FUNC_H H_Pass_8_C | ||
174 : | #define FUNC_V V_Pass_8_C | ||
175 : | #define FUNC_HA H_Pass_Avrg_8_C | ||
176 : | #define FUNC_VA V_Pass_Avrg_8_C | ||
177 : | #define FUNC_HA_UP H_Pass_Avrg_Up_8_C | ||
178 : | #define FUNC_VA_UP V_Pass_Avrg_Up_8_C | ||
179 : | |||
180 : | edgomez | 1530 | #include "qpel.c" |
181 : | edgomez | 1382 | |
182 : | /* note: B-frame always uses Rnd=0... */ | ||
183 : | #define STORE(d,s) (d) = ( (s)+(d)+1 ) >> 1 | ||
184 : | #define FUNC_H H_Pass_8_Add_C | ||
185 : | #define FUNC_V V_Pass_8_Add_C | ||
186 : | #define FUNC_HA H_Pass_Avrg_8_Add_C | ||
187 : | #define FUNC_VA V_Pass_Avrg_8_Add_C | ||
188 : | #define FUNC_HA_UP H_Pass_Avrg_Up_8_Add_C | ||
189 : | #define FUNC_VA_UP V_Pass_Avrg_Up_8_Add_C | ||
190 : | |||
191 : | edgomez | 1530 | #include "qpel.c" |
192 : | edgomez | 1382 | |
193 : | #undef SIZE | ||
194 : | #undef TABLE | ||
195 : | #undef XVID_AUTO_INCLUDE | ||
196 : | |||
197 : | edgomez | 1530 | /* Global scope hooks |
198 : | edgomez | 1382 | ****************************************************************************/ |
199 : | |||
200 : | XVID_QP_FUNCS *xvid_QP_Funcs = 0; | ||
201 : | XVID_QP_FUNCS *xvid_QP_Add_Funcs = 0; | ||
202 : | |||
203 : | edgomez | 1530 | /* Reference plain C impl. declaration |
204 : | edgomez | 1382 | ****************************************************************************/ |
205 : | |||
206 : | edgomez | 1530 | XVID_QP_FUNCS xvid_QP_Funcs_C_ref = { |
207 : | H_Pass_16_C_ref, H_Pass_Avrg_16_C_ref, H_Pass_Avrg_Up_16_C_ref, | ||
208 : | V_Pass_16_C_ref, V_Pass_Avrg_16_C_ref, V_Pass_Avrg_Up_16_C_ref, | ||
209 : | |||
210 : | H_Pass_8_C_ref, H_Pass_Avrg_8_C_ref, H_Pass_Avrg_Up_8_C_ref, | ||
211 : | V_Pass_8_C_ref, V_Pass_Avrg_8_C_ref, V_Pass_Avrg_Up_8_C_ref | ||
212 : | }; | ||
213 : | |||
214 : | XVID_QP_FUNCS xvid_QP_Add_Funcs_C_ref = { | ||
215 : | H_Pass_16_Add_C_ref, H_Pass_Avrg_16_Add_C_ref, H_Pass_Avrg_Up_16_Add_C_ref, | ||
216 : | V_Pass_16_Add_C_ref, V_Pass_Avrg_16_Add_C_ref, V_Pass_Avrg_Up_16_Add_C_ref, | ||
217 : | |||
218 : | H_Pass_8_Add_C_ref, H_Pass_Avrg_8_Add_C_ref, H_Pass_Avrg_Up_8_Add_C_ref, | ||
219 : | V_Pass_8_Add_C_ref, V_Pass_Avrg_8_Add_C_ref, V_Pass_Avrg_Up_8_Add_C_ref | ||
220 : | }; | ||
221 : | |||
222 : | /* Plain C impl. declaration (faster than ref one) | ||
223 : | ****************************************************************************/ | ||
224 : | |||
225 : | edgomez | 1382 | XVID_QP_FUNCS xvid_QP_Funcs_C = { |
226 : | H_Pass_16_C, H_Pass_Avrg_16_C, H_Pass_Avrg_Up_16_C, | ||
227 : | V_Pass_16_C, V_Pass_Avrg_16_C, V_Pass_Avrg_Up_16_C, | ||
228 : | |||
229 : | H_Pass_8_C, H_Pass_Avrg_8_C, H_Pass_Avrg_Up_8_C, | ||
230 : | V_Pass_8_C, V_Pass_Avrg_8_C, V_Pass_Avrg_Up_8_C | ||
231 : | }; | ||
232 : | |||
233 : | XVID_QP_FUNCS xvid_QP_Add_Funcs_C = { | ||
234 : | H_Pass_16_Add_C, H_Pass_Avrg_16_Add_C, H_Pass_Avrg_Up_16_Add_C, | ||
235 : | V_Pass_16_Add_C, V_Pass_Avrg_16_Add_C, V_Pass_Avrg_Up_16_Add_C, | ||
236 : | |||
237 : | H_Pass_8_Add_C, H_Pass_Avrg_8_Add_C, H_Pass_Avrg_Up_8_Add_C, | ||
238 : | V_Pass_8_Add_C, V_Pass_Avrg_8_Add_C, V_Pass_Avrg_Up_8_Add_C | ||
239 : | }; | ||
240 : | |||
241 : | /* mmx impl. declaration (see. qpel_mmx.asm | ||
242 : | ****************************************************************************/ | ||
243 : | |||
244 : | #ifdef ARCH_IS_IA32 | ||
245 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_16_mmx); | ||
246 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_16_mmx); | ||
247 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_16_mmx); | ||
248 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_16_mmx); | ||
249 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_16_mmx); | ||
250 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_16_mmx); | ||
251 : | |||
252 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_8_mmx); | ||
253 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_8_mmx); | ||
254 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_8_mmx); | ||
255 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_8_mmx); | ||
256 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_8_mmx); | ||
257 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_8_mmx); | ||
258 : | |||
259 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Add_16_mmx); | ||
260 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Add_16_mmx); | ||
261 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_Add_16_mmx); | ||
262 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Add_16_mmx); | ||
263 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Add_16_mmx); | ||
264 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_Add_16_mmx); | ||
265 : | |||
266 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_8_Add_mmx); | ||
267 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_8_Add_mmx); | ||
268 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_8_Add_mmx); | ||
269 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_8_Add_mmx); | ||
270 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_8_Add_mmx); | ||
271 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_8_Add_mmx); | ||
272 : | |||
273 : | edgomez | 1530 | XVID_QP_FUNCS xvid_QP_Funcs_mmx = { |
274 : | xvid_H_Pass_16_mmx, xvid_H_Pass_Avrg_16_mmx, xvid_H_Pass_Avrg_Up_16_mmx, | ||
275 : | xvid_V_Pass_16_mmx, xvid_V_Pass_Avrg_16_mmx, xvid_V_Pass_Avrg_Up_16_mmx, | ||
276 : | |||
277 : | xvid_H_Pass_8_mmx, xvid_H_Pass_Avrg_8_mmx, xvid_H_Pass_Avrg_Up_8_mmx, | ||
278 : | xvid_V_Pass_8_mmx, xvid_V_Pass_Avrg_8_mmx, xvid_V_Pass_Avrg_Up_8_mmx | ||
279 : | }; | ||
280 : | |||
281 : | edgomez | 1382 | XVID_QP_FUNCS xvid_QP_Add_Funcs_mmx = { |
282 : | xvid_H_Pass_Add_16_mmx, xvid_H_Pass_Avrg_Add_16_mmx, xvid_H_Pass_Avrg_Up_Add_16_mmx, | ||
283 : | xvid_V_Pass_Add_16_mmx, xvid_V_Pass_Avrg_Add_16_mmx, xvid_V_Pass_Avrg_Up_Add_16_mmx, | ||
284 : | |||
285 : | xvid_H_Pass_8_Add_mmx, xvid_H_Pass_Avrg_8_Add_mmx, xvid_H_Pass_Avrg_Up_8_Add_mmx, | ||
286 : | xvid_V_Pass_8_Add_mmx, xvid_V_Pass_Avrg_8_Add_mmx, xvid_V_Pass_Avrg_Up_8_Add_mmx, | ||
287 : | }; | ||
288 : | #endif /* ARCH_IS_IA32 */ | ||
289 : | |||
290 : | edgomez | 1557 | |
291 : | /* altivec impl. declaration (see qpel_altivec.c) | ||
292 : | ****************************************************************************/ | ||
293 : | |||
294 : | #ifdef ARCH_IS_PPC | ||
295 : | |||
296 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_16_Altivec_C); | ||
297 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_16_Altivec_C); | ||
298 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_Up_16_Altivec_C); | ||
299 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_16_Altivec_C); | ||
300 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_16_Altivec_C); | ||
301 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_Up_16_Altivec_C); | ||
302 : | |||
303 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_8_Altivec_C); | ||
304 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_8_Altivec_C); | ||
305 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_Up_8_Altivec_C); | ||
306 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_8_Altivec_C); | ||
307 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_8_Altivec_C); | ||
308 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_Up_8_Altivec_C); | ||
309 : | |||
310 : | |||
311 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_16_Add_Altivec_C); | ||
312 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_16_Add_Altivec_C); | ||
313 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_Up_16_Add_Altivec_C); | ||
314 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_16_Add_Altivec_C); | ||
315 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_16_Add_Altivec_C); | ||
316 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_Up_16_Add_Altivec_C); | ||
317 : | |||
318 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_8_Add_Altivec_C); | ||
319 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_8_Add_Altivec_C); | ||
320 : | extern XVID_QP_PASS_SIGNATURE(H_Pass_Avrg_Up_8_Add_Altivec_C); | ||
321 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_8_Add_Altivec_C); | ||
322 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_8_Add_Altivec_C); | ||
323 : | extern XVID_QP_PASS_SIGNATURE(V_Pass_Avrg_Up_8_Add_Altivec_C); | ||
324 : | |||
325 : | XVID_QP_FUNCS xvid_QP_Funcs_Altivec_C = { | ||
326 : | H_Pass_16_Altivec_C, H_Pass_Avrg_16_Altivec_C, H_Pass_Avrg_Up_16_Altivec_C, | ||
327 : | V_Pass_16_Altivec_C, V_Pass_Avrg_16_Altivec_C, V_Pass_Avrg_Up_16_Altivec_C, | ||
328 : | |||
329 : | H_Pass_8_Altivec_C, H_Pass_Avrg_8_Altivec_C, H_Pass_Avrg_Up_8_Altivec_C, | ||
330 : | V_Pass_8_Altivec_C, V_Pass_Avrg_8_Altivec_C, V_Pass_Avrg_Up_8_Altivec_C | ||
331 : | }; | ||
332 : | |||
333 : | XVID_QP_FUNCS xvid_QP_Add_Funcs_Altivec_C = { | ||
334 : | H_Pass_16_Add_Altivec_C, H_Pass_Avrg_16_Add_Altivec_C, H_Pass_Avrg_Up_16_Add_Altivec_C, | ||
335 : | V_Pass_16_Add_Altivec_C, V_Pass_Avrg_16_Add_Altivec_C, V_Pass_Avrg_Up_16_Add_Altivec_C, | ||
336 : | |||
337 : | H_Pass_8_Add_Altivec_C, H_Pass_Avrg_8_Add_Altivec_C, H_Pass_Avrg_Up_8_Add_Altivec_C, | ||
338 : | V_Pass_8_Add_Altivec_C, V_Pass_Avrg_8_Add_Altivec_C, V_Pass_Avrg_Up_8_Add_Altivec_C | ||
339 : | }; | ||
340 : | |||
341 : | #endif /* ARCH_IS_PPC */ | ||
342 : | |||
343 : | edgomez | 1586 | /* mmx impl. (for 64bit bus) declaration (see. qpel_mmx.asm |
344 : | ****************************************************************************/ | ||
345 : | |||
346 : | #ifdef ARCH_IS_X86_64 | ||
347 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_16_x86_64); | ||
348 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_16_x86_64); | ||
349 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_16_x86_64); | ||
350 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_16_x86_64); | ||
351 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_16_x86_64); | ||
352 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_16_x86_64); | ||
353 : | |||
354 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_8_x86_64); | ||
355 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_8_x86_64); | ||
356 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_8_x86_64); | ||
357 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_8_x86_64); | ||
358 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_8_x86_64); | ||
359 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_8_x86_64); | ||
360 : | |||
361 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Add_16_x86_64); | ||
362 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Add_16_x86_64); | ||
363 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_Add_16_x86_64); | ||
364 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Add_16_x86_64); | ||
365 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Add_16_x86_64); | ||
366 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_Add_16_x86_64); | ||
367 : | |||
368 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_8_Add_x86_64); | ||
369 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_8_Add_x86_64); | ||
370 : | extern XVID_QP_PASS_SIGNATURE(xvid_H_Pass_Avrg_Up_8_Add_x86_64); | ||
371 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_8_Add_x86_64); | ||
372 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_8_Add_x86_64); | ||
373 : | extern XVID_QP_PASS_SIGNATURE(xvid_V_Pass_Avrg_Up_8_Add_x86_64); | ||
374 : | |||
375 : | XVID_QP_FUNCS xvid_QP_Funcs_x86_64 = { | ||
376 : | xvid_H_Pass_16_x86_64, xvid_H_Pass_Avrg_16_x86_64, xvid_H_Pass_Avrg_Up_16_x86_64, | ||
377 : | xvid_V_Pass_16_x86_64, xvid_V_Pass_Avrg_16_x86_64, xvid_V_Pass_Avrg_Up_16_x86_64, | ||
378 : | |||
379 : | xvid_H_Pass_8_x86_64, xvid_H_Pass_Avrg_8_x86_64, xvid_H_Pass_Avrg_Up_8_x86_64, | ||
380 : | xvid_V_Pass_8_x86_64, xvid_V_Pass_Avrg_8_x86_64, xvid_V_Pass_Avrg_Up_8_x86_64 | ||
381 : | }; | ||
382 : | |||
383 : | XVID_QP_FUNCS xvid_QP_Add_Funcs_x86_64 = { | ||
384 : | xvid_H_Pass_Add_16_x86_64, xvid_H_Pass_Avrg_Add_16_x86_64, xvid_H_Pass_Avrg_Up_Add_16_x86_64, | ||
385 : | xvid_V_Pass_Add_16_x86_64, xvid_V_Pass_Avrg_Add_16_x86_64, xvid_V_Pass_Avrg_Up_Add_16_x86_64, | ||
386 : | |||
387 : | xvid_H_Pass_8_Add_x86_64, xvid_H_Pass_Avrg_8_Add_x86_64, xvid_H_Pass_Avrg_Up_8_Add_x86_64, | ||
388 : | xvid_V_Pass_8_Add_x86_64, xvid_V_Pass_Avrg_8_Add_x86_64, xvid_V_Pass_Avrg_Up_8_Add_x86_64, | ||
389 : | }; | ||
390 : | #endif /* ARCH_IS_X86_64 */ | ||
391 : | |||
392 : | edgomez | 1382 | /* tables for ASM |
393 : | ****************************************************************************/ | ||
394 : | |||
395 : | #ifdef ARCH_IS_IA32 | ||
396 : | uint16_t xvid_Expand_mmx[256][4]; /* 8b -> 64b expansion table */ | ||
397 : | #endif | ||
398 : | |||
399 : | edgomez | 1586 | #ifdef ARCH_IS_X86_64 |
400 : | Skal | 1626 | extern uint16_t xvid_Expand_mmx[256][4]; /* 8b -> 64b expansion table */ |
401 : | edgomez | 1586 | #endif |
402 : | |||
403 : | edgomez | 1382 | /* Alternate way of filtering (cf. USE_TABLES flag in qpel_mmx.asm) |
404 : | * | ||
405 : | * 17 tables, 2K each => 34K | ||
406 : | * Mirroring can be acheived composing 11 basic tables | ||
407 : | * (for instance: (23,19,-6,3)=(20,20,-6,3)+(3,-1,0,0) | ||
408 : | * Using Symmetries (and bswap) could reduce further | ||
409 : | * the memory to 7 tables (->14K). */ | ||
410 : | edgomez | 1586 | #ifdef ARCH_IS_X86_64 |
411 : | Skal | 1626 | extern int16_t xvid_FIR_1_0_0_0[256][4]; |
412 : | extern int16_t xvid_FIR_3_1_0_0[256][4]; | ||
413 : | extern int16_t xvid_FIR_6_3_1_0[256][4]; | ||
414 : | extern int16_t xvid_FIR_14_3_2_1[256][4]; | ||
415 : | extern int16_t xvid_FIR_20_6_3_1[256][4]; | ||
416 : | extern int16_t xvid_FIR_20_20_6_3[256][4]; | ||
417 : | extern int16_t xvid_FIR_23_19_6_3[256][4]; | ||
418 : | extern int16_t xvid_FIR_7_20_20_6[256][4]; | ||
419 : | extern int16_t xvid_FIR_6_20_20_6[256][4]; | ||
420 : | extern int16_t xvid_FIR_6_20_20_7[256][4]; | ||
421 : | extern int16_t xvid_FIR_3_6_20_20[256][4]; | ||
422 : | extern int16_t xvid_FIR_3_6_19_23[256][4]; | ||
423 : | extern int16_t xvid_FIR_1_3_6_20[256][4]; | ||
424 : | extern int16_t xvid_FIR_1_2_3_14[256][4]; | ||
425 : | extern int16_t xvid_FIR_0_1_3_6[256][4]; | ||
426 : | extern int16_t xvid_FIR_0_0_1_3[256][4]; | ||
427 : | extern int16_t xvid_FIR_0_0_0_1[256][4]; | ||
428 : | edgomez | 1586 | #else |
429 : | edgomez | 1382 | int16_t xvid_FIR_1_0_0_0[256][4]; |
430 : | int16_t xvid_FIR_3_1_0_0[256][4]; | ||
431 : | int16_t xvid_FIR_6_3_1_0[256][4]; | ||
432 : | int16_t xvid_FIR_14_3_2_1[256][4]; | ||
433 : | int16_t xvid_FIR_20_6_3_1[256][4]; | ||
434 : | int16_t xvid_FIR_20_20_6_3[256][4]; | ||
435 : | int16_t xvid_FIR_23_19_6_3[256][4]; | ||
436 : | int16_t xvid_FIR_7_20_20_6[256][4]; | ||
437 : | int16_t xvid_FIR_6_20_20_6[256][4]; | ||
438 : | int16_t xvid_FIR_6_20_20_7[256][4]; | ||
439 : | int16_t xvid_FIR_3_6_20_20[256][4]; | ||
440 : | int16_t xvid_FIR_3_6_19_23[256][4]; | ||
441 : | int16_t xvid_FIR_1_3_6_20[256][4]; | ||
442 : | int16_t xvid_FIR_1_2_3_14[256][4]; | ||
443 : | int16_t xvid_FIR_0_1_3_6[256][4]; | ||
444 : | int16_t xvid_FIR_0_0_1_3[256][4]; | ||
445 : | int16_t xvid_FIR_0_0_0_1[256][4]; | ||
446 : | edgomez | 1586 | #endif |
447 : | edgomez | 1382 | |
448 : | static void Init_FIR_Table(int16_t Tab[][4], | ||
449 : | int A, int B, int C, int D) | ||
450 : | { | ||
451 : | int i; | ||
452 : | for(i=0; i<256; ++i) { | ||
453 : | Tab[i][0] = i*A; | ||
454 : | Tab[i][1] = i*B; | ||
455 : | Tab[i][2] = i*C; | ||
456 : | Tab[i][3] = i*D; | ||
457 : | } | ||
458 : | } | ||
459 : | |||
460 : | |||
461 : | void xvid_Init_QP() | ||
462 : | { | ||
463 : | #ifdef ARCH_IS_IA32 | ||
464 : | int i; | ||
465 : | |||
466 : | for(i=0; i<256; ++i) { | ||
467 : | xvid_Expand_mmx[i][0] = i; | ||
468 : | xvid_Expand_mmx[i][1] = i; | ||
469 : | xvid_Expand_mmx[i][2] = i; | ||
470 : | xvid_Expand_mmx[i][3] = i; | ||
471 : | } | ||
472 : | #endif | ||
473 : | |||
474 : | /* Alternate way of filtering (cf. USE_TABLES flag in qpel_mmx.asm) */ | ||
475 : | |||
476 : | Init_FIR_Table(xvid_FIR_1_0_0_0, -1, 0, 0, 0); | ||
477 : | Init_FIR_Table(xvid_FIR_3_1_0_0, 3, -1, 0, 0); | ||
478 : | Init_FIR_Table(xvid_FIR_6_3_1_0, -6, 3, -1, 0); | ||
479 : | Init_FIR_Table(xvid_FIR_14_3_2_1, 14, -3, 2, -1); | ||
480 : | Init_FIR_Table(xvid_FIR_20_6_3_1, 20, -6, 3, -1); | ||
481 : | Init_FIR_Table(xvid_FIR_20_20_6_3, 20, 20, -6, 3); | ||
482 : | Init_FIR_Table(xvid_FIR_23_19_6_3, 23, 19, -6, 3); | ||
483 : | Init_FIR_Table(xvid_FIR_7_20_20_6, -7, 20, 20, -6); | ||
484 : | Init_FIR_Table(xvid_FIR_6_20_20_6, -6, 20, 20, -6); | ||
485 : | Init_FIR_Table(xvid_FIR_6_20_20_7, -6, 20, 20, -7); | ||
486 : | Init_FIR_Table(xvid_FIR_3_6_20_20, 3, -6, 20, 20); | ||
487 : | Init_FIR_Table(xvid_FIR_3_6_19_23, 3, -6, 19, 23); | ||
488 : | Init_FIR_Table(xvid_FIR_1_3_6_20, -1, 3, -6, 20); | ||
489 : | Init_FIR_Table(xvid_FIR_1_2_3_14, -1, 2, -3, 14); | ||
490 : | Init_FIR_Table(xvid_FIR_0_1_3_6, 0, -1, 3, -6); | ||
491 : | Init_FIR_Table(xvid_FIR_0_0_1_3, 0, 0, -1, 3); | ||
492 : | Init_FIR_Table(xvid_FIR_0_0_0_1, 0, 0, 0, -1); | ||
493 : | |||
494 : | } | ||
495 : | |||
496 : | #endif /* !XVID_AUTO_INCLUDE */ | ||
497 : | |||
498 : | edgomez | 1530 | #if defined(XVID_AUTO_INCLUDE) && defined(REFERENCE_CODE) |
499 : | |||
500 : | edgomez | 1382 | /***************************************************************************** |
501 : | * "reference" filters impl. in plain C | ||
502 : | ****************************************************************************/ | ||
503 : | |||
504 : | static | ||
505 : | void FUNC_H(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t Rnd) | ||
506 : | { | ||
507 : | while(H-->0) { | ||
508 : | int32_t i, k; | ||
509 : | int32_t Sums[SIZE] = { 0 }; | ||
510 : | for(i=0; i<=SIZE; ++i) | ||
511 : | for(k=0; k<SIZE; ++k) | ||
512 : | Sums[k] += TABLE[i][k] * Src[i]; | ||
513 : | |||
514 : | for(i=0; i<SIZE; ++i) { | ||
515 : | int32_t C = ( Sums[i] + 16-Rnd ) >> 5; | ||
516 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
517 : | STORE(Dst[i], C); | ||
518 : | } | ||
519 : | Src += BpS; | ||
520 : | Dst += BpS; | ||
521 : | } | ||
522 : | } | ||
523 : | |||
524 : | static | ||
525 : | void FUNC_V(uint8_t *Dst, const uint8_t *Src, int32_t W, int32_t BpS, int32_t Rnd) | ||
526 : | { | ||
527 : | while(W-->0) { | ||
528 : | int32_t i, k; | ||
529 : | int32_t Sums[SIZE] = { 0 }; | ||
530 : | const uint8_t *S = Src++; | ||
531 : | uint8_t *D = Dst++; | ||
532 : | for(i=0; i<=SIZE; ++i) { | ||
533 : | for(k=0; k<SIZE; ++k) | ||
534 : | Sums[k] += TABLE[i][k] * S[0]; | ||
535 : | S += BpS; | ||
536 : | } | ||
537 : | |||
538 : | for(i=0; i<SIZE; ++i) { | ||
539 : | int32_t C = ( Sums[i] + 16-Rnd )>>5; | ||
540 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
541 : | STORE(D[0], C); | ||
542 : | D += BpS; | ||
543 : | } | ||
544 : | } | ||
545 : | } | ||
546 : | |||
547 : | static | ||
548 : | void FUNC_HA(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t Rnd) | ||
549 : | { | ||
550 : | while(H-->0) { | ||
551 : | int32_t i, k; | ||
552 : | int32_t Sums[SIZE] = { 0 }; | ||
553 : | for(i=0; i<=SIZE; ++i) | ||
554 : | for(k=0; k<SIZE; ++k) | ||
555 : | Sums[k] += TABLE[i][k] * Src[i]; | ||
556 : | |||
557 : | for(i=0; i<SIZE; ++i) { | ||
558 : | int32_t C = ( Sums[i] + 16-Rnd ) >> 5; | ||
559 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
560 : | C = (C+Src[i]+1-Rnd) >> 1; | ||
561 : | STORE(Dst[i], C); | ||
562 : | } | ||
563 : | Src += BpS; | ||
564 : | Dst += BpS; | ||
565 : | } | ||
566 : | } | ||
567 : | |||
568 : | static | ||
569 : | void FUNC_HA_UP(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t Rnd) | ||
570 : | { | ||
571 : | while(H-->0) { | ||
572 : | int32_t i, k; | ||
573 : | int32_t Sums[SIZE] = { 0 }; | ||
574 : | for(i=0; i<=SIZE; ++i) | ||
575 : | for(k=0; k<SIZE; ++k) | ||
576 : | Sums[k] += TABLE[i][k] * Src[i]; | ||
577 : | |||
578 : | for(i=0; i<SIZE; ++i) { | ||
579 : | int32_t C = ( Sums[i] + 16-Rnd ) >> 5; | ||
580 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
581 : | C = (C+Src[i+1]+1-Rnd) >> 1; | ||
582 : | STORE(Dst[i], C); | ||
583 : | } | ||
584 : | Src += BpS; | ||
585 : | Dst += BpS; | ||
586 : | } | ||
587 : | } | ||
588 : | |||
589 : | static | ||
590 : | void FUNC_VA(uint8_t *Dst, const uint8_t *Src, int32_t W, int32_t BpS, int32_t Rnd) | ||
591 : | { | ||
592 : | while(W-->0) { | ||
593 : | int32_t i, k; | ||
594 : | int32_t Sums[SIZE] = { 0 }; | ||
595 : | const uint8_t *S = Src; | ||
596 : | uint8_t *D = Dst; | ||
597 : | |||
598 : | for(i=0; i<=SIZE; ++i) { | ||
599 : | for(k=0; k<SIZE; ++k) | ||
600 : | Sums[k] += TABLE[i][k] * S[0]; | ||
601 : | S += BpS; | ||
602 : | } | ||
603 : | |||
604 : | S = Src; | ||
605 : | for(i=0; i<SIZE; ++i) { | ||
606 : | int32_t C = ( Sums[i] + 16-Rnd )>>5; | ||
607 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
608 : | C = ( C+S[0]+1-Rnd ) >> 1; | ||
609 : | STORE(D[0], C); | ||
610 : | D += BpS; | ||
611 : | S += BpS; | ||
612 : | } | ||
613 : | Src++; | ||
614 : | Dst++; | ||
615 : | } | ||
616 : | } | ||
617 : | |||
618 : | static | ||
619 : | void FUNC_VA_UP(uint8_t *Dst, const uint8_t *Src, int32_t W, int32_t BpS, int32_t Rnd) | ||
620 : | { | ||
621 : | while(W-->0) { | ||
622 : | int32_t i, k; | ||
623 : | int32_t Sums[SIZE] = { 0 }; | ||
624 : | const uint8_t *S = Src; | ||
625 : | uint8_t *D = Dst; | ||
626 : | |||
627 : | for(i=0; i<=SIZE; ++i) { | ||
628 : | for(k=0; k<SIZE; ++k) | ||
629 : | Sums[k] += TABLE[i][k] * S[0]; | ||
630 : | S += BpS; | ||
631 : | } | ||
632 : | |||
633 : | S = Src + BpS; | ||
634 : | for(i=0; i<SIZE; ++i) { | ||
635 : | int32_t C = ( Sums[i] + 16-Rnd )>>5; | ||
636 : | if (C<0) C = 0; else if (C>255) C = 255; | ||
637 : | C = ( C+S[0]+1-Rnd ) >> 1; | ||
638 : | STORE(D[0], C); | ||
639 : | D += BpS; | ||
640 : | S += BpS; | ||
641 : | } | ||
642 : | Dst++; | ||
643 : | Src++; | ||
644 : | } | ||
645 : | } | ||
646 : | |||
647 : | #undef STORE | ||
648 : | #undef FUNC_H | ||
649 : | #undef FUNC_V | ||
650 : | #undef FUNC_HA | ||
651 : | #undef FUNC_VA | ||
652 : | #undef FUNC_HA_UP | ||
653 : | #undef FUNC_VA_UP | ||
654 : | |||
655 : | edgomez | 1530 | #elif defined(XVID_AUTO_INCLUDE) && !defined(REFERENCE_CODE) |
656 : | |||
657 : | /***************************************************************************** | ||
658 : | * "fast" filters impl. in plain C | ||
659 : | ****************************************************************************/ | ||
660 : | |||
661 : | #define CLIP_STORE(D,C) \ | ||
662 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
663 : | STORE(D, C) | ||
664 : | |||
665 : | static void | ||
666 : | FUNC_H(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
667 : | { | ||
668 : | #if (SIZE==16) | ||
669 : | while(H-->0) { | ||
670 : | int C; | ||
671 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
672 : | CLIP_STORE(Dst[ 0],C); | ||
673 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
674 : | CLIP_STORE(Dst[ 1],C); | ||
675 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
676 : | CLIP_STORE(Dst[ 2],C); | ||
677 : | C = 16-RND - (Src[0]+Src[7 ]) + 3*(Src[ 1]+Src[ 6])-6*(Src[ 2]+Src[ 5]) + 20*(Src[ 3]+Src[ 4]); | ||
678 : | CLIP_STORE(Dst[ 3],C); | ||
679 : | C = 16-RND - (Src[1]+Src[8 ]) + 3*(Src[ 2]+Src[ 7])-6*(Src[ 3]+Src[ 6]) + 20*(Src[ 4]+Src[ 5]); | ||
680 : | CLIP_STORE(Dst[ 4],C); | ||
681 : | C = 16-RND - (Src[2]+Src[9 ]) + 3*(Src[ 3]+Src[ 8])-6*(Src[ 4]+Src[ 7]) + 20*(Src[ 5]+Src[ 6]); | ||
682 : | CLIP_STORE(Dst[ 5],C); | ||
683 : | C = 16-RND - (Src[3]+Src[10]) + 3*(Src[ 4]+Src[ 9])-6*(Src[ 5]+Src[ 8]) + 20*(Src[ 6]+Src[ 7]); | ||
684 : | CLIP_STORE(Dst[ 6],C); | ||
685 : | C = 16-RND - (Src[4]+Src[11]) + 3*(Src[ 5]+Src[10])-6*(Src[ 6]+Src[ 9]) + 20*(Src[ 7]+Src[ 8]); | ||
686 : | CLIP_STORE(Dst[ 7],C); | ||
687 : | C = 16-RND - (Src[5]+Src[12]) + 3*(Src[ 6]+Src[11])-6*(Src[ 7]+Src[10]) + 20*(Src[ 8]+Src[ 9]); | ||
688 : | CLIP_STORE(Dst[ 8],C); | ||
689 : | C = 16-RND - (Src[6]+Src[13]) + 3*(Src[ 7]+Src[12])-6*(Src[ 8]+Src[11]) + 20*(Src[ 9]+Src[10]); | ||
690 : | CLIP_STORE(Dst[ 9],C); | ||
691 : | C = 16-RND - (Src[7]+Src[14]) + 3*(Src[ 8]+Src[13])-6*(Src[ 9]+Src[12]) + 20*(Src[10]+Src[11]); | ||
692 : | CLIP_STORE(Dst[10],C); | ||
693 : | C = 16-RND - (Src[8]+Src[15]) + 3*(Src[ 9]+Src[14])-6*(Src[10]+Src[13]) + 20*(Src[11]+Src[12]); | ||
694 : | CLIP_STORE(Dst[11],C); | ||
695 : | C = 16-RND - (Src[9]+Src[16]) + 3*(Src[10]+Src[15])-6*(Src[11]+Src[14]) + 20*(Src[12]+Src[13]); | ||
696 : | CLIP_STORE(Dst[12],C); | ||
697 : | C = 16-RND - Src[10] +3*Src[11] -6*(Src[12]+Src[15]) + 20*(Src[13]+Src[14]) +2*Src[16]; | ||
698 : | CLIP_STORE(Dst[13],C); | ||
699 : | C = 16-RND - Src[11] +3*(Src[12]-Src[16]) -6*Src[13] + 20*Src[14] + 19*Src[15]; | ||
700 : | CLIP_STORE(Dst[14],C); | ||
701 : | C = 16-RND - Src[12] +3*Src[13] -7*Src[14] + 23*Src[15] + 14*Src[16]; | ||
702 : | CLIP_STORE(Dst[15],C); | ||
703 : | Src += BpS; | ||
704 : | Dst += BpS; | ||
705 : | } | ||
706 : | #else | ||
707 : | while(H-->0) { | ||
708 : | int C; | ||
709 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
710 : | CLIP_STORE(Dst[0],C); | ||
711 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
712 : | CLIP_STORE(Dst[1],C); | ||
713 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
714 : | CLIP_STORE(Dst[2],C); | ||
715 : | C = 16-RND - (Src[0]+Src[7]) + 3*(Src[1]+Src[6])-6*(Src[2]+Src[5]) + 20*(Src[3]+Src[4]); | ||
716 : | CLIP_STORE(Dst[3],C); | ||
717 : | C = 16-RND - (Src[1]+Src[8]) + 3*(Src[2]+Src[7])-6*(Src[3]+Src[6]) + 20*(Src[4]+Src[5]); | ||
718 : | CLIP_STORE(Dst[4],C); | ||
719 : | C = 16-RND - Src[2] +3*Src[3] -6*(Src[4]+Src[7]) + 20*(Src[5]+Src[6]) +2*Src[8]; | ||
720 : | CLIP_STORE(Dst[5],C); | ||
721 : | C = 16-RND - Src[3] +3*(Src[4]-Src[8]) -6*Src[5] + 20*Src[6] + 19*Src[7]; | ||
722 : | CLIP_STORE(Dst[6],C); | ||
723 : | C = 16-RND - Src[4] +3*Src[5] -7*Src[6] + 23*Src[7] + 14*Src[8]; | ||
724 : | CLIP_STORE(Dst[7],C); | ||
725 : | Src += BpS; | ||
726 : | Dst += BpS; | ||
727 : | } | ||
728 : | #endif | ||
729 : | } | ||
730 : | #undef CLIP_STORE | ||
731 : | |||
732 : | #define CLIP_STORE(i,C) \ | ||
733 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
734 : | C = (C+Src[i]+1-RND) >> 1; \ | ||
735 : | STORE(Dst[i], C) | ||
736 : | |||
737 : | static void | ||
738 : | FUNC_HA(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
739 : | { | ||
740 : | #if (SIZE==16) | ||
741 : | while(H-->0) { | ||
742 : | int C; | ||
743 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
744 : | CLIP_STORE(0,C); | ||
745 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
746 : | CLIP_STORE( 1,C); | ||
747 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
748 : | CLIP_STORE( 2,C); | ||
749 : | C = 16-RND - (Src[0]+Src[7 ]) + 3*(Src[ 1]+Src[ 6])-6*(Src[ 2]+Src[ 5]) + 20*(Src[ 3]+Src[ 4]); | ||
750 : | CLIP_STORE( 3,C); | ||
751 : | C = 16-RND - (Src[1]+Src[8 ]) + 3*(Src[ 2]+Src[ 7])-6*(Src[ 3]+Src[ 6]) + 20*(Src[ 4]+Src[ 5]); | ||
752 : | CLIP_STORE( 4,C); | ||
753 : | C = 16-RND - (Src[2]+Src[9 ]) + 3*(Src[ 3]+Src[ 8])-6*(Src[ 4]+Src[ 7]) + 20*(Src[ 5]+Src[ 6]); | ||
754 : | CLIP_STORE( 5,C); | ||
755 : | C = 16-RND - (Src[3]+Src[10]) + 3*(Src[ 4]+Src[ 9])-6*(Src[ 5]+Src[ 8]) + 20*(Src[ 6]+Src[ 7]); | ||
756 : | CLIP_STORE( 6,C); | ||
757 : | C = 16-RND - (Src[4]+Src[11]) + 3*(Src[ 5]+Src[10])-6*(Src[ 6]+Src[ 9]) + 20*(Src[ 7]+Src[ 8]); | ||
758 : | CLIP_STORE( 7,C); | ||
759 : | C = 16-RND - (Src[5]+Src[12]) + 3*(Src[ 6]+Src[11])-6*(Src[ 7]+Src[10]) + 20*(Src[ 8]+Src[ 9]); | ||
760 : | CLIP_STORE( 8,C); | ||
761 : | C = 16-RND - (Src[6]+Src[13]) + 3*(Src[ 7]+Src[12])-6*(Src[ 8]+Src[11]) + 20*(Src[ 9]+Src[10]); | ||
762 : | CLIP_STORE( 9,C); | ||
763 : | C = 16-RND - (Src[7]+Src[14]) + 3*(Src[ 8]+Src[13])-6*(Src[ 9]+Src[12]) + 20*(Src[10]+Src[11]); | ||
764 : | CLIP_STORE(10,C); | ||
765 : | C = 16-RND - (Src[8]+Src[15]) + 3*(Src[ 9]+Src[14])-6*(Src[10]+Src[13]) + 20*(Src[11]+Src[12]); | ||
766 : | CLIP_STORE(11,C); | ||
767 : | C = 16-RND - (Src[9]+Src[16]) + 3*(Src[10]+Src[15])-6*(Src[11]+Src[14]) + 20*(Src[12]+Src[13]); | ||
768 : | CLIP_STORE(12,C); | ||
769 : | C = 16-RND - Src[10] +3*Src[11] -6*(Src[12]+Src[15]) + 20*(Src[13]+Src[14]) +2*Src[16]; | ||
770 : | CLIP_STORE(13,C); | ||
771 : | C = 16-RND - Src[11] +3*(Src[12]-Src[16]) -6*Src[13] + 20*Src[14] + 19*Src[15]; | ||
772 : | CLIP_STORE(14,C); | ||
773 : | C = 16-RND - Src[12] +3*Src[13] -7*Src[14] + 23*Src[15] + 14*Src[16]; | ||
774 : | CLIP_STORE(15,C); | ||
775 : | Src += BpS; | ||
776 : | Dst += BpS; | ||
777 : | } | ||
778 : | #else | ||
779 : | while(H-->0) { | ||
780 : | int C; | ||
781 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
782 : | CLIP_STORE(0,C); | ||
783 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
784 : | CLIP_STORE(1,C); | ||
785 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
786 : | CLIP_STORE(2,C); | ||
787 : | C = 16-RND - (Src[0]+Src[7]) + 3*(Src[1]+Src[6])-6*(Src[2]+Src[5]) + 20*(Src[3]+Src[4]); | ||
788 : | CLIP_STORE(3,C); | ||
789 : | C = 16-RND - (Src[1]+Src[8]) + 3*(Src[2]+Src[7])-6*(Src[3]+Src[6]) + 20*(Src[4]+Src[5]); | ||
790 : | CLIP_STORE(4,C); | ||
791 : | C = 16-RND - Src[2] +3*Src[3] -6*(Src[4]+Src[7]) + 20*(Src[5]+Src[6]) +2*Src[8]; | ||
792 : | CLIP_STORE(5,C); | ||
793 : | C = 16-RND - Src[3] +3*(Src[4]-Src[8]) -6*Src[5] + 20*Src[6] + 19*Src[7]; | ||
794 : | CLIP_STORE(6,C); | ||
795 : | C = 16-RND - Src[4] +3*Src[5] -7*Src[6] + 23*Src[7] + 14*Src[8]; | ||
796 : | CLIP_STORE(7,C); | ||
797 : | Src += BpS; | ||
798 : | Dst += BpS; | ||
799 : | } | ||
800 : | #endif | ||
801 : | } | ||
802 : | #undef CLIP_STORE | ||
803 : | |||
804 : | #define CLIP_STORE(i,C) \ | ||
805 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
806 : | C = (C+Src[i+1]+1-RND) >> 1; \ | ||
807 : | STORE(Dst[i], C) | ||
808 : | |||
809 : | static void | ||
810 : | FUNC_HA_UP(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
811 : | { | ||
812 : | #if (SIZE==16) | ||
813 : | while(H-->0) { | ||
814 : | int C; | ||
815 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
816 : | CLIP_STORE(0,C); | ||
817 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
818 : | CLIP_STORE( 1,C); | ||
819 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
820 : | CLIP_STORE( 2,C); | ||
821 : | C = 16-RND - (Src[0]+Src[7 ]) + 3*(Src[ 1]+Src[ 6])-6*(Src[ 2]+Src[ 5]) + 20*(Src[ 3]+Src[ 4]); | ||
822 : | CLIP_STORE( 3,C); | ||
823 : | C = 16-RND - (Src[1]+Src[8 ]) + 3*(Src[ 2]+Src[ 7])-6*(Src[ 3]+Src[ 6]) + 20*(Src[ 4]+Src[ 5]); | ||
824 : | CLIP_STORE( 4,C); | ||
825 : | C = 16-RND - (Src[2]+Src[9 ]) + 3*(Src[ 3]+Src[ 8])-6*(Src[ 4]+Src[ 7]) + 20*(Src[ 5]+Src[ 6]); | ||
826 : | CLIP_STORE( 5,C); | ||
827 : | C = 16-RND - (Src[3]+Src[10]) + 3*(Src[ 4]+Src[ 9])-6*(Src[ 5]+Src[ 8]) + 20*(Src[ 6]+Src[ 7]); | ||
828 : | CLIP_STORE( 6,C); | ||
829 : | C = 16-RND - (Src[4]+Src[11]) + 3*(Src[ 5]+Src[10])-6*(Src[ 6]+Src[ 9]) + 20*(Src[ 7]+Src[ 8]); | ||
830 : | CLIP_STORE( 7,C); | ||
831 : | C = 16-RND - (Src[5]+Src[12]) + 3*(Src[ 6]+Src[11])-6*(Src[ 7]+Src[10]) + 20*(Src[ 8]+Src[ 9]); | ||
832 : | CLIP_STORE( 8,C); | ||
833 : | C = 16-RND - (Src[6]+Src[13]) + 3*(Src[ 7]+Src[12])-6*(Src[ 8]+Src[11]) + 20*(Src[ 9]+Src[10]); | ||
834 : | CLIP_STORE( 9,C); | ||
835 : | C = 16-RND - (Src[7]+Src[14]) + 3*(Src[ 8]+Src[13])-6*(Src[ 9]+Src[12]) + 20*(Src[10]+Src[11]); | ||
836 : | CLIP_STORE(10,C); | ||
837 : | C = 16-RND - (Src[8]+Src[15]) + 3*(Src[ 9]+Src[14])-6*(Src[10]+Src[13]) + 20*(Src[11]+Src[12]); | ||
838 : | CLIP_STORE(11,C); | ||
839 : | C = 16-RND - (Src[9]+Src[16]) + 3*(Src[10]+Src[15])-6*(Src[11]+Src[14]) + 20*(Src[12]+Src[13]); | ||
840 : | CLIP_STORE(12,C); | ||
841 : | C = 16-RND - Src[10] +3*Src[11] -6*(Src[12]+Src[15]) + 20*(Src[13]+Src[14]) +2*Src[16]; | ||
842 : | CLIP_STORE(13,C); | ||
843 : | C = 16-RND - Src[11] +3*(Src[12]-Src[16]) -6*Src[13] + 20*Src[14] + 19*Src[15]; | ||
844 : | CLIP_STORE(14,C); | ||
845 : | C = 16-RND - Src[12] +3*Src[13] -7*Src[14] + 23*Src[15] + 14*Src[16]; | ||
846 : | CLIP_STORE(15,C); | ||
847 : | Src += BpS; | ||
848 : | Dst += BpS; | ||
849 : | } | ||
850 : | #else | ||
851 : | while(H-->0) { | ||
852 : | int C; | ||
853 : | C = 16-RND +14*Src[0] +23*Src[1] - 7*Src[2] + 3*Src[3] - Src[4]; | ||
854 : | CLIP_STORE(0,C); | ||
855 : | C = 16-RND - 3*(Src[0]-Src[4]) +19*Src[1] +20*Src[2] - 6*Src[3] - Src[5]; | ||
856 : | CLIP_STORE(1,C); | ||
857 : | C = 16-RND + 2*Src[0] - 6*(Src[1]+Src[4]) +20*(Src[2]+Src[3]) + 3*Src[5] - Src[6]; | ||
858 : | CLIP_STORE(2,C); | ||
859 : | C = 16-RND - (Src[0]+Src[7]) + 3*(Src[1]+Src[6])-6*(Src[2]+Src[5]) + 20*(Src[3]+Src[4]); | ||
860 : | CLIP_STORE(3,C); | ||
861 : | C = 16-RND - (Src[1]+Src[8]) + 3*(Src[2]+Src[7])-6*(Src[3]+Src[6]) + 20*(Src[4]+Src[5]); | ||
862 : | CLIP_STORE(4,C); | ||
863 : | C = 16-RND - Src[2] +3*Src[3] -6*(Src[4]+Src[7]) + 20*(Src[5]+Src[6]) +2*Src[8]; | ||
864 : | CLIP_STORE(5,C); | ||
865 : | C = 16-RND - Src[3] +3*(Src[4]-Src[8]) -6*Src[5] + 20*Src[6] + 19*Src[7]; | ||
866 : | CLIP_STORE(6,C); | ||
867 : | C = 16-RND - Src[4] +3*Src[5] -7*Src[6] + 23*Src[7] + 14*Src[8]; | ||
868 : | CLIP_STORE(7,C); | ||
869 : | Src += BpS; | ||
870 : | Dst += BpS; | ||
871 : | } | ||
872 : | #endif | ||
873 : | } | ||
874 : | #undef CLIP_STORE | ||
875 : | |||
876 : | ////////////////////////////////////////////////////////// | ||
877 : | // vertical passes | ||
878 : | ////////////////////////////////////////////////////////// | ||
879 : | // Note: for vertical passes, width (W) needs only be 8 or 16. | ||
880 : | |||
881 : | #define CLIP_STORE(D,C) \ | ||
882 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
883 : | STORE(D, C) | ||
884 : | |||
885 : | static void | ||
886 : | FUNC_V(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
887 : | { | ||
888 : | #if (SIZE==16) | ||
889 : | while(H-->0) { | ||
890 : | int C; | ||
891 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
892 : | CLIP_STORE(Dst[BpS* 0],C); | ||
893 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
894 : | CLIP_STORE(Dst[BpS* 1],C); | ||
895 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
896 : | CLIP_STORE(Dst[BpS* 2],C); | ||
897 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7 ]) + 3*(Src[BpS* 1]+Src[BpS* 6])-6*(Src[BpS* 2]+Src[BpS* 5]) + 20*(Src[BpS* 3]+Src[BpS* 4]); | ||
898 : | CLIP_STORE(Dst[BpS* 3],C); | ||
899 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8 ]) + 3*(Src[BpS* 2]+Src[BpS* 7])-6*(Src[BpS* 3]+Src[BpS* 6]) + 20*(Src[BpS* 4]+Src[BpS* 5]); | ||
900 : | CLIP_STORE(Dst[BpS* 4],C); | ||
901 : | C = 16-RND - (Src[BpS*2]+Src[BpS*9 ]) + 3*(Src[BpS* 3]+Src[BpS* 8])-6*(Src[BpS* 4]+Src[BpS* 7]) + 20*(Src[BpS* 5]+Src[BpS* 6]); | ||
902 : | CLIP_STORE(Dst[BpS* 5],C); | ||
903 : | C = 16-RND - (Src[BpS*3]+Src[BpS*10]) + 3*(Src[BpS* 4]+Src[BpS* 9])-6*(Src[BpS* 5]+Src[BpS* 8]) + 20*(Src[BpS* 6]+Src[BpS* 7]); | ||
904 : | CLIP_STORE(Dst[BpS* 6],C); | ||
905 : | C = 16-RND - (Src[BpS*4]+Src[BpS*11]) + 3*(Src[BpS* 5]+Src[BpS*10])-6*(Src[BpS* 6]+Src[BpS* 9]) + 20*(Src[BpS* 7]+Src[BpS* 8]); | ||
906 : | CLIP_STORE(Dst[BpS* 7],C); | ||
907 : | C = 16-RND - (Src[BpS*5]+Src[BpS*12]) + 3*(Src[BpS* 6]+Src[BpS*11])-6*(Src[BpS* 7]+Src[BpS*10]) + 20*(Src[BpS* 8]+Src[BpS* 9]); | ||
908 : | CLIP_STORE(Dst[BpS* 8],C); | ||
909 : | C = 16-RND - (Src[BpS*6]+Src[BpS*13]) + 3*(Src[BpS* 7]+Src[BpS*12])-6*(Src[BpS* 8]+Src[BpS*11]) + 20*(Src[BpS* 9]+Src[BpS*10]); | ||
910 : | CLIP_STORE(Dst[BpS* 9],C); | ||
911 : | C = 16-RND - (Src[BpS*7]+Src[BpS*14]) + 3*(Src[BpS* 8]+Src[BpS*13])-6*(Src[BpS* 9]+Src[BpS*12]) + 20*(Src[BpS*10]+Src[BpS*11]); | ||
912 : | CLIP_STORE(Dst[BpS*10],C); | ||
913 : | C = 16-RND - (Src[BpS*8]+Src[BpS*15]) + 3*(Src[BpS* 9]+Src[BpS*14])-6*(Src[BpS*10]+Src[BpS*13]) + 20*(Src[BpS*11]+Src[BpS*12]); | ||
914 : | CLIP_STORE(Dst[BpS*11],C); | ||
915 : | C = 16-RND - (Src[BpS*9]+Src[BpS*16]) + 3*(Src[BpS*10]+Src[BpS*15])-6*(Src[BpS*11]+Src[BpS*14]) + 20*(Src[BpS*12]+Src[BpS*13]); | ||
916 : | CLIP_STORE(Dst[BpS*12],C); | ||
917 : | C = 16-RND - Src[BpS*10] +3*Src[BpS*11] -6*(Src[BpS*12]+Src[BpS*15]) + 20*(Src[BpS*13]+Src[BpS*14]) +2*Src[BpS*16]; | ||
918 : | CLIP_STORE(Dst[BpS*13],C); | ||
919 : | C = 16-RND - Src[BpS*11] +3*(Src[BpS*12]-Src[BpS*16]) -6*Src[BpS*13] + 20*Src[BpS*14] + 19*Src[BpS*15]; | ||
920 : | CLIP_STORE(Dst[BpS*14],C); | ||
921 : | C = 16-RND - Src[BpS*12] +3*Src[BpS*13] -7*Src[BpS*14] + 23*Src[BpS*15] + 14*Src[BpS*16]; | ||
922 : | CLIP_STORE(Dst[BpS*15],C); | ||
923 : | Src += 1; | ||
924 : | Dst += 1; | ||
925 : | } | ||
926 : | #else | ||
927 : | while(H-->0) { | ||
928 : | int C; | ||
929 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
930 : | CLIP_STORE(Dst[BpS*0],C); | ||
931 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
932 : | CLIP_STORE(Dst[BpS*1],C); | ||
933 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
934 : | CLIP_STORE(Dst[BpS*2],C); | ||
935 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7]) + 3*(Src[BpS*1]+Src[BpS*6])-6*(Src[BpS*2]+Src[BpS*5]) + 20*(Src[BpS*3]+Src[BpS*4]); | ||
936 : | CLIP_STORE(Dst[BpS*3],C); | ||
937 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8]) + 3*(Src[BpS*2]+Src[BpS*7])-6*(Src[BpS*3]+Src[BpS*6]) + 20*(Src[BpS*4]+Src[BpS*5]); | ||
938 : | CLIP_STORE(Dst[BpS*4],C); | ||
939 : | C = 16-RND - Src[BpS*2] +3*Src[BpS*3] -6*(Src[BpS*4]+Src[BpS*7]) + 20*(Src[BpS*5]+Src[BpS*6]) +2*Src[BpS*8]; | ||
940 : | CLIP_STORE(Dst[BpS*5],C); | ||
941 : | C = 16-RND - Src[BpS*3] +3*(Src[BpS*4]-Src[BpS*8]) -6*Src[BpS*5] + 20*Src[BpS*6] + 19*Src[BpS*7]; | ||
942 : | CLIP_STORE(Dst[BpS*6],C); | ||
943 : | C = 16-RND - Src[BpS*4] +3*Src[BpS*5] -7*Src[BpS*6] + 23*Src[BpS*7] + 14*Src[BpS*8]; | ||
944 : | CLIP_STORE(Dst[BpS*7],C); | ||
945 : | Src += 1; | ||
946 : | Dst += 1; | ||
947 : | } | ||
948 : | #endif | ||
949 : | } | ||
950 : | #undef CLIP_STORE | ||
951 : | |||
952 : | #define CLIP_STORE(i,C) \ | ||
953 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
954 : | C = (C+Src[BpS*i]+1-RND) >> 1; \ | ||
955 : | STORE(Dst[BpS*i], C) | ||
956 : | |||
957 : | static void | ||
958 : | FUNC_VA(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
959 : | { | ||
960 : | #if (SIZE==16) | ||
961 : | while(H-->0) { | ||
962 : | int C; | ||
963 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
964 : | CLIP_STORE(0,C); | ||
965 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
966 : | CLIP_STORE( 1,C); | ||
967 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
968 : | CLIP_STORE( 2,C); | ||
969 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7 ]) + 3*(Src[BpS* 1]+Src[BpS* 6])-6*(Src[BpS* 2]+Src[BpS* 5]) + 20*(Src[BpS* 3]+Src[BpS* 4]); | ||
970 : | CLIP_STORE( 3,C); | ||
971 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8 ]) + 3*(Src[BpS* 2]+Src[BpS* 7])-6*(Src[BpS* 3]+Src[BpS* 6]) + 20*(Src[BpS* 4]+Src[BpS* 5]); | ||
972 : | CLIP_STORE( 4,C); | ||
973 : | C = 16-RND - (Src[BpS*2]+Src[BpS*9 ]) + 3*(Src[BpS* 3]+Src[BpS* 8])-6*(Src[BpS* 4]+Src[BpS* 7]) + 20*(Src[BpS* 5]+Src[BpS* 6]); | ||
974 : | CLIP_STORE( 5,C); | ||
975 : | C = 16-RND - (Src[BpS*3]+Src[BpS*10]) + 3*(Src[BpS* 4]+Src[BpS* 9])-6*(Src[BpS* 5]+Src[BpS* 8]) + 20*(Src[BpS* 6]+Src[BpS* 7]); | ||
976 : | CLIP_STORE( 6,C); | ||
977 : | C = 16-RND - (Src[BpS*4]+Src[BpS*11]) + 3*(Src[BpS* 5]+Src[BpS*10])-6*(Src[BpS* 6]+Src[BpS* 9]) + 20*(Src[BpS* 7]+Src[BpS* 8]); | ||
978 : | CLIP_STORE( 7,C); | ||
979 : | C = 16-RND - (Src[BpS*5]+Src[BpS*12]) + 3*(Src[BpS* 6]+Src[BpS*11])-6*(Src[BpS* 7]+Src[BpS*10]) + 20*(Src[BpS* 8]+Src[BpS* 9]); | ||
980 : | CLIP_STORE( 8,C); | ||
981 : | C = 16-RND - (Src[BpS*6]+Src[BpS*13]) + 3*(Src[BpS* 7]+Src[BpS*12])-6*(Src[BpS* 8]+Src[BpS*11]) + 20*(Src[BpS* 9]+Src[BpS*10]); | ||
982 : | CLIP_STORE( 9,C); | ||
983 : | C = 16-RND - (Src[BpS*7]+Src[BpS*14]) + 3*(Src[BpS* 8]+Src[BpS*13])-6*(Src[BpS* 9]+Src[BpS*12]) + 20*(Src[BpS*10]+Src[BpS*11]); | ||
984 : | CLIP_STORE(10,C); | ||
985 : | C = 16-RND - (Src[BpS*8]+Src[BpS*15]) + 3*(Src[BpS* 9]+Src[BpS*14])-6*(Src[BpS*10]+Src[BpS*13]) + 20*(Src[BpS*11]+Src[BpS*12]); | ||
986 : | CLIP_STORE(11,C); | ||
987 : | C = 16-RND - (Src[BpS*9]+Src[BpS*16]) + 3*(Src[BpS*10]+Src[BpS*15])-6*(Src[BpS*11]+Src[BpS*14]) + 20*(Src[BpS*12]+Src[BpS*13]); | ||
988 : | CLIP_STORE(12,C); | ||
989 : | C = 16-RND - Src[BpS*10] +3*Src[BpS*11] -6*(Src[BpS*12]+Src[BpS*15]) + 20*(Src[BpS*13]+Src[BpS*14]) +2*Src[BpS*16]; | ||
990 : | CLIP_STORE(13,C); | ||
991 : | C = 16-RND - Src[BpS*11] +3*(Src[BpS*12]-Src[BpS*16]) -6*Src[BpS*13] + 20*Src[BpS*14] + 19*Src[BpS*15]; | ||
992 : | CLIP_STORE(14,C); | ||
993 : | C = 16-RND - Src[BpS*12] +3*Src[BpS*13] -7*Src[BpS*14] + 23*Src[BpS*15] + 14*Src[BpS*16]; | ||
994 : | CLIP_STORE(15,C); | ||
995 : | Src += 1; | ||
996 : | Dst += 1; | ||
997 : | } | ||
998 : | #else | ||
999 : | while(H-->0) { | ||
1000 : | int C; | ||
1001 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
1002 : | CLIP_STORE(0,C); | ||
1003 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
1004 : | CLIP_STORE(1,C); | ||
1005 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
1006 : | CLIP_STORE(2,C); | ||
1007 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7]) + 3*(Src[BpS*1]+Src[BpS*6])-6*(Src[BpS*2]+Src[BpS*5]) + 20*(Src[BpS*3]+Src[BpS*4]); | ||
1008 : | CLIP_STORE(3,C); | ||
1009 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8]) + 3*(Src[BpS*2]+Src[BpS*7])-6*(Src[BpS*3]+Src[BpS*6]) + 20*(Src[BpS*4]+Src[BpS*5]); | ||
1010 : | CLIP_STORE(4,C); | ||
1011 : | C = 16-RND - Src[BpS*2] +3*Src[BpS*3] -6*(Src[BpS*4]+Src[BpS*7]) + 20*(Src[BpS*5]+Src[BpS*6]) +2*Src[BpS*8]; | ||
1012 : | CLIP_STORE(5,C); | ||
1013 : | C = 16-RND - Src[BpS*3] +3*(Src[BpS*4]-Src[BpS*8]) -6*Src[BpS*5] + 20*Src[BpS*6] + 19*Src[BpS*7]; | ||
1014 : | CLIP_STORE(6,C); | ||
1015 : | C = 16-RND - Src[BpS*4] +3*Src[BpS*5] -7*Src[BpS*6] + 23*Src[BpS*7] + 14*Src[BpS*8]; | ||
1016 : | CLIP_STORE(7,C); | ||
1017 : | Src += 1; | ||
1018 : | Dst += 1; | ||
1019 : | } | ||
1020 : | #endif | ||
1021 : | } | ||
1022 : | #undef CLIP_STORE | ||
1023 : | |||
1024 : | #define CLIP_STORE(i,C) \ | ||
1025 : | if (C<0) C = 0; else if (C>(255<<5)) C = 255; else C = C>>5; \ | ||
1026 : | C = (C+Src[BpS*i+BpS]+1-RND) >> 1; \ | ||
1027 : | STORE(Dst[BpS*i], C) | ||
1028 : | |||
1029 : | static void | ||
1030 : | FUNC_VA_UP(uint8_t *Dst, const uint8_t *Src, int32_t H, int32_t BpS, int32_t RND) | ||
1031 : | { | ||
1032 : | #if (SIZE==16) | ||
1033 : | while(H-->0) { | ||
1034 : | int C; | ||
1035 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
1036 : | CLIP_STORE(0,C); | ||
1037 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
1038 : | CLIP_STORE( 1,C); | ||
1039 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
1040 : | CLIP_STORE( 2,C); | ||
1041 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7 ]) + 3*(Src[BpS* 1]+Src[BpS* 6])-6*(Src[BpS* 2]+Src[BpS* 5]) + 20*(Src[BpS* 3]+Src[BpS* 4]); | ||
1042 : | CLIP_STORE( 3,C); | ||
1043 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8 ]) + 3*(Src[BpS* 2]+Src[BpS* 7])-6*(Src[BpS* 3]+Src[BpS* 6]) + 20*(Src[BpS* 4]+Src[BpS* 5]); | ||
1044 : | CLIP_STORE( 4,C); | ||
1045 : | C = 16-RND - (Src[BpS*2]+Src[BpS*9 ]) + 3*(Src[BpS* 3]+Src[BpS* 8])-6*(Src[BpS* 4]+Src[BpS* 7]) + 20*(Src[BpS* 5]+Src[BpS* 6]); | ||
1046 : | CLIP_STORE( 5,C); | ||
1047 : | C = 16-RND - (Src[BpS*3]+Src[BpS*10]) + 3*(Src[BpS* 4]+Src[BpS* 9])-6*(Src[BpS* 5]+Src[BpS* 8]) + 20*(Src[BpS* 6]+Src[BpS* 7]); | ||
1048 : | CLIP_STORE( 6,C); | ||
1049 : | C = 16-RND - (Src[BpS*4]+Src[BpS*11]) + 3*(Src[BpS* 5]+Src[BpS*10])-6*(Src[BpS* 6]+Src[BpS* 9]) + 20*(Src[BpS* 7]+Src[BpS* 8]); | ||
1050 : | CLIP_STORE( 7,C); | ||
1051 : | C = 16-RND - (Src[BpS*5]+Src[BpS*12]) + 3*(Src[BpS* 6]+Src[BpS*11])-6*(Src[BpS* 7]+Src[BpS*10]) + 20*(Src[BpS* 8]+Src[BpS* 9]); | ||
1052 : | CLIP_STORE( 8,C); | ||
1053 : | C = 16-RND - (Src[BpS*6]+Src[BpS*13]) + 3*(Src[BpS* 7]+Src[BpS*12])-6*(Src[BpS* 8]+Src[BpS*11]) + 20*(Src[BpS* 9]+Src[BpS*10]); | ||
1054 : | CLIP_STORE( 9,C); | ||
1055 : | C = 16-RND - (Src[BpS*7]+Src[BpS*14]) + 3*(Src[BpS* 8]+Src[BpS*13])-6*(Src[BpS* 9]+Src[BpS*12]) + 20*(Src[BpS*10]+Src[BpS*11]); | ||
1056 : | CLIP_STORE(10,C); | ||
1057 : | C = 16-RND - (Src[BpS*8]+Src[BpS*15]) + 3*(Src[BpS* 9]+Src[BpS*14])-6*(Src[BpS*10]+Src[BpS*13]) + 20*(Src[BpS*11]+Src[BpS*12]); | ||
1058 : | CLIP_STORE(11,C); | ||
1059 : | C = 16-RND - (Src[BpS*9]+Src[BpS*16]) + 3*(Src[BpS*10]+Src[BpS*15])-6*(Src[BpS*11]+Src[BpS*14]) + 20*(Src[BpS*12]+Src[BpS*13]); | ||
1060 : | CLIP_STORE(12,C); | ||
1061 : | C = 16-RND - Src[BpS*10] +3*Src[BpS*11] -6*(Src[BpS*12]+Src[BpS*15]) + 20*(Src[BpS*13]+Src[BpS*14]) +2*Src[BpS*16]; | ||
1062 : | CLIP_STORE(13,C); | ||
1063 : | C = 16-RND - Src[BpS*11] +3*(Src[BpS*12]-Src[BpS*16]) -6*Src[BpS*13] + 20*Src[BpS*14] + 19*Src[BpS*15]; | ||
1064 : | CLIP_STORE(14,C); | ||
1065 : | C = 16-RND - Src[BpS*12] +3*Src[BpS*13] -7*Src[BpS*14] + 23*Src[BpS*15] + 14*Src[BpS*16]; | ||
1066 : | CLIP_STORE(15,C); | ||
1067 : | Src += 1; | ||
1068 : | Dst += 1; | ||
1069 : | } | ||
1070 : | #else | ||
1071 : | while(H-->0) { | ||
1072 : | int C; | ||
1073 : | C = 16-RND +14*Src[BpS*0] +23*Src[BpS*1] - 7*Src[BpS*2] + 3*Src[BpS*3] - Src[BpS*4]; | ||
1074 : | CLIP_STORE(0,C); | ||
1075 : | C = 16-RND - 3*(Src[BpS*0]-Src[BpS*4]) +19*Src[BpS*1] +20*Src[BpS*2] - 6*Src[BpS*3] - Src[BpS*5]; | ||
1076 : | CLIP_STORE(1,C); | ||
1077 : | C = 16-RND + 2*Src[BpS*0] - 6*(Src[BpS*1]+Src[BpS*4]) +20*(Src[BpS*2]+Src[BpS*3]) + 3*Src[BpS*5] - Src[BpS*6]; | ||
1078 : | CLIP_STORE(2,C); | ||
1079 : | C = 16-RND - (Src[BpS*0]+Src[BpS*7]) + 3*(Src[BpS*1]+Src[BpS*6])-6*(Src[BpS*2]+Src[BpS*5]) + 20*(Src[BpS*3]+Src[BpS*4]); | ||
1080 : | CLIP_STORE(3,C); | ||
1081 : | C = 16-RND - (Src[BpS*1]+Src[BpS*8]) + 3*(Src[BpS*2]+Src[BpS*7])-6*(Src[BpS*3]+Src[BpS*6]) + 20*(Src[BpS*4]+Src[BpS*5]); | ||
1082 : | CLIP_STORE(4,C); | ||
1083 : | C = 16-RND - Src[BpS*2] +3*Src[BpS*3] -6*(Src[BpS*4]+Src[BpS*7]) + 20*(Src[BpS*5]+Src[BpS*6]) +2*Src[BpS*8]; | ||
1084 : | CLIP_STORE(5,C); | ||
1085 : | C = 16-RND - Src[BpS*3] +3*(Src[BpS*4]-Src[BpS*8]) -6*Src[BpS*5] + 20*Src[BpS*6] + 19*Src[BpS*7]; | ||
1086 : | CLIP_STORE(6,C); | ||
1087 : | C = 16-RND - Src[BpS*4] +3*Src[BpS*5] -7*Src[BpS*6] + 23*Src[BpS*7] + 14*Src[BpS*8]; | ||
1088 : | CLIP_STORE(7,C); | ||
1089 : | Src += 1; | ||
1090 : | Dst += 1; | ||
1091 : | } | ||
1092 : | #endif | ||
1093 : | } | ||
1094 : | #undef CLIP_STORE | ||
1095 : | |||
1096 : | #undef STORE | ||
1097 : | #undef FUNC_H | ||
1098 : | #undef FUNC_V | ||
1099 : | #undef FUNC_HA | ||
1100 : | #undef FUNC_VA | ||
1101 : | #undef FUNC_HA_UP | ||
1102 : | #undef FUNC_VA_UP | ||
1103 : | |||
1104 : | |||
1105 : | #endif /* XVID_AUTO_INCLUDE && !defined(REF) */ |
No admin address has been configured | ViewVC Help |
Powered by ViewVC 1.0.4 |