[svn] / branches / dev-api-3 / xvidcore / src / image / interpolate8x8.c Repository:
ViewVC logotype

Annotation of /branches/dev-api-3/xvidcore/src/image/interpolate8x8.c

Parent Directory Parent Directory | Revision Log Revision Log


Revision 588 - (view) (download)

1 : Isibaar 333 /**************************************************************************
2 :     *
3 :     * XVID MPEG-4 VIDEO CODEC
4 :     * 8x8 block-based halfpel interpolation
5 :     *
6 :     * This program is free software; you can redistribute it and/or modify
7 :     * it under the terms of the GNU General Public License as published by
8 :     * the Free Software Foundation; either version 2 of the License, or
9 :     * (at your option) any later version.
10 :     *
11 :     * This program is distributed in the hope that it will be useful,
12 :     * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 :     * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 :     * GNU General Public License for more details.
15 :     *
16 :     * You should have received a copy of the GNU General Public License
17 :     * along with this program; if not, write to the Free Software
18 :     * Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
19 :     *
20 :     *************************************************************************/
21 :    
22 :     /**************************************************************************
23 :     *
24 :     * History:
25 : Isibaar 579 *
26 :     * 05.10.2002 new bilinear and qpel interpolation code - Isibaar
27 : Isibaar 333 * 27.12.2001 modified "compensate_halfpel"
28 :     * 05.11.2001 initial version; (c)2001 peter ross <pross@cs.rmit.edu.au>
29 :     *
30 :     *************************************************************************/
31 :    
32 :    
33 :     #include "../portab.h"
34 :     #include "interpolate8x8.h"
35 :    
36 :     // function pointers
37 :     INTERPOLATE8X8_PTR interpolate8x8_halfpel_h;
38 :     INTERPOLATE8X8_PTR interpolate8x8_halfpel_v;
39 :     INTERPOLATE8X8_PTR interpolate8x8_halfpel_hv;
40 :    
41 : Isibaar 579 INTERPOLATE8X8_AVG2_PTR interpolate8x8_avg2;
42 :     INTERPOLATE8X8_AVG4_PTR interpolate8x8_avg4;
43 : Isibaar 333
44 : Isibaar 588 INTERPOLATE_LOWPASS_PTR interpolate8x8_lowpass_h;
45 :     INTERPOLATE_LOWPASS_PTR interpolate8x8_lowpass_v;
46 : Isibaar 579
47 : Isibaar 588 INTERPOLATE_LOWPASS_PTR interpolate16x16_lowpass_h;
48 :     INTERPOLATE_LOWPASS_PTR interpolate16x16_lowpass_v;
49 :    
50 :     INTERPOLATE_LOWPASS_HV_PTR interpolate8x8_lowpass_hv;
51 :     INTERPOLATE_LOWPASS_HV_PTR interpolate16x16_lowpass_hv;
52 :    
53 : Isibaar 579 INTERPOLATE8X8_6TAP_LOWPASS_PTR interpolate8x8_6tap_lowpass_h;
54 :     INTERPOLATE8X8_6TAP_LOWPASS_PTR interpolate8x8_6tap_lowpass_v;
55 :    
56 :     void interpolate8x8_avg2_c(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint32_t stride, const uint32_t rounding)
57 :     {
58 :     int32_t i;
59 :     const int32_t round = 1 - rounding;
60 :    
61 :     for(i = 0; i < 8; i++)
62 :     {
63 :     dst[0] = (src1[0] + src2[0] + round) >> 1;
64 :     dst[1] = (src1[1] + src2[1] + round) >> 1;
65 :     dst[2] = (src1[2] + src2[2] + round) >> 1;
66 :     dst[3] = (src1[3] + src2[3] + round) >> 1;
67 :     dst[4] = (src1[4] + src2[4] + round) >> 1;
68 :     dst[5] = (src1[5] + src2[5] + round) >> 1;
69 :     dst[6] = (src1[6] + src2[6] + round) >> 1;
70 :     dst[7] = (src1[7] + src2[7] + round) >> 1;
71 :    
72 :     dst += stride;
73 :     src1 += stride;
74 :     src2 += stride;
75 :     }
76 :     }
77 :    
78 :     void interpolate8x8_avg4_c(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4, const uint32_t stride, const uint32_t rounding)
79 :     {
80 :     int32_t i;
81 :     const int32_t round = 2 - rounding;
82 :    
83 :     for(i = 0; i < 8; i++)
84 :     {
85 :     dst[0] = (src1[0] + src2[0] + src3[0] + src4[0] + round) >> 2;
86 :     dst[1] = (src1[1] + src2[1] + src3[1] + src4[1] + round) >> 2;
87 :     dst[2] = (src1[2] + src2[2] + src3[2] + src4[2] + round) >> 2;
88 :     dst[3] = (src1[3] + src2[3] + src3[3] + src4[3] + round) >> 2;
89 :     dst[4] = (src1[4] + src2[4] + src3[4] + src4[4] + round) >> 2;
90 :     dst[5] = (src1[5] + src2[5] + src3[5] + src4[5] + round) >> 2;
91 :     dst[6] = (src1[6] + src2[6] + src3[6] + src4[6] + round) >> 2;
92 :     dst[7] = (src1[7] + src2[7] + src3[7] + src4[7] + round) >> 2;
93 :    
94 :     dst += stride;
95 :     src1 += stride;
96 :     src2 += stride;
97 :     src3 += stride;
98 :     src4 += stride;
99 :     }
100 :     }
101 :    
102 : Isibaar 333 // dst = interpolate(src)
103 :    
104 :     void
105 :     interpolate8x8_halfpel_h_c(uint8_t * const dst,
106 :     const uint8_t * const src,
107 :     const uint32_t stride,
108 :     const uint32_t rounding)
109 :     {
110 :     uint32_t i, j;
111 :    
112 :     for (j = 0; j < 8; j++) {
113 :     for (i = 0; i < 8; i++) {
114 :    
115 :     int16_t tot =
116 :     (int32_t) src[j * stride + i] + (int32_t) src[j * stride + i +
117 :     1];
118 :    
119 :     tot = (int32_t) ((tot + 1 - rounding) >> 1);
120 :     dst[j * stride + i] = (uint8_t) tot;
121 :     }
122 :     }
123 :     }
124 :    
125 :    
126 :    
127 :     void
128 :     interpolate8x8_halfpel_v_c(uint8_t * const dst,
129 :     const uint8_t * const src,
130 :     const uint32_t stride,
131 :     const uint32_t rounding)
132 :     {
133 :     uint32_t i, j;
134 :    
135 :     for (j = 0; j < 8; j++) {
136 :     for (i = 0; i < 8; i++) {
137 :     int16_t tot = src[j * stride + i] + src[j * stride + i + stride];
138 :    
139 :     tot = ((tot + 1 - rounding) >> 1);
140 :     dst[j * stride + i] = (uint8_t) tot;
141 :     }
142 :     }
143 :     }
144 :    
145 :    
146 :     void
147 :     interpolate8x8_halfpel_hv_c(uint8_t * const dst,
148 :     const uint8_t * const src,
149 :     const uint32_t stride,
150 :     const uint32_t rounding)
151 :     {
152 :     uint32_t i, j;
153 :    
154 :     for (j = 0; j < 8; j++) {
155 :     for (i = 0; i < 8; i++) {
156 :     int16_t tot =
157 :     src[j * stride + i] + src[j * stride + i + 1] +
158 :     src[j * stride + i + stride] + src[j * stride + i + stride +
159 :     1];
160 :     tot = ((tot + 2 - rounding) >> 2);
161 :     dst[j * stride + i] = (uint8_t) tot;
162 :     }
163 :     }
164 :     }
165 :    
166 :     /*************************************************************
167 :     * QPEL STUFF STARTS HERE *
168 :     *************************************************************/
169 :    
170 :     #define CLIP(X,A,B) (X < A) ? (A) : ((X > B) ? (B) : (X))
171 :    
172 : Isibaar 579 void interpolate8x8_6tap_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
173 : Isibaar 333 {
174 :     int32_t i;
175 : Isibaar 579 uint8_t round_add = 16 - rounding;
176 : Isibaar 333
177 :     for(i = 0; i < 8; i++)
178 :     {
179 :    
180 : Isibaar 579 dst[0] = CLIP((((src[-2] + src[3]) + 5 * (((src[0] + src[1])<<2) - (src[-1] + src[2])) + round_add) >> 5), 0, 255);
181 :     dst[1] = CLIP((((src[-1] + src[4]) + 5 * (((src[1] + src[2])<<2) - (src[0] + src[3])) + round_add) >> 5), 0, 255);
182 :     dst[2] = CLIP((((src[0] + src[5]) + 5 * (((src[2] + src[3])<<2) - (src[1] + src[4])) + round_add) >> 5), 0, 255);
183 :     dst[3] = CLIP((((src[1] + src[6]) + 5 * (((src[3] + src[4])<<2) - (src[2] + src[5])) + round_add) >> 5), 0, 255);
184 :     dst[4] = CLIP((((src[2] + src[7]) + 5 * (((src[4] + src[5])<<2) - (src[3] + src[6])) + round_add) >> 5), 0, 255);
185 :     dst[5] = CLIP((((src[3] + src[8]) + 5 * (((src[5] + src[6])<<2) - (src[4] + src[7])) + round_add) >> 5), 0, 255);
186 :     dst[6] = CLIP((((src[4] + src[9]) + 5 * (((src[6] + src[7])<<2) - (src[5] + src[8])) + round_add) >> 5), 0, 255);
187 :     dst[7] = CLIP((((src[5] + src[10]) + 5 * (((src[7] + src[8])<<2) - (src[6] + src[9])) + round_add) >> 5), 0, 255);
188 :    
189 :     dst += stride;
190 :     src += stride;
191 : Isibaar 333 }
192 :     }
193 :    
194 : Isibaar 588 void interpolate16x16_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
195 :     {
196 :     int32_t i;
197 :     uint8_t round_add = 16 - rounding;
198 :    
199 :     for(i = 0; i < 16; i++)
200 :     {
201 :    
202 :     dst[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);
203 :     dst[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);
204 :     dst[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);
205 :    
206 :     dst[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);
207 :     dst[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);
208 :     dst[5] = CLIP(((20 * (src[5] + src[6]) - 3 * (((src[4] + src[7])<<1) - (src[3] + src[8])) - (src[2] + src[9]) + round_add) >> 5), 0, 255);
209 :     dst[6] = CLIP(((20 * (src[6] + src[7]) - 3 * (((src[5] + src[8])<<1) - (src[4] + src[9])) - (src[3] + src[10]) + round_add) >> 5), 0, 255);
210 :     dst[7] = CLIP(((20 * (src[7] + src[8]) - 3 * (((src[6] + src[9])<<1) - (src[5] + src[10])) - (src[4] + src[11]) + round_add) >> 5), 0, 255);
211 :     dst[8] = CLIP(((20 * (src[8] + src[9]) - 3 * (((src[7] + src[10])<<1) - (src[6] + src[11])) - (src[5] + src[12]) + round_add) >> 5), 0, 255);
212 :     dst[9] = CLIP(((20 * (src[9] + src[10]) - 3 * (((src[8] + src[11])<<1) - (src[7] + src[12])) - (src[6] + src[13]) + round_add) >> 5), 0, 255);
213 :     dst[10] = CLIP(((20 * (src[10] + src[11]) - 3 * (((src[9] + src[12])<<1) - (src[8] + src[13])) - (src[7] + src[14]) + round_add) >> 5), 0, 255);
214 :     dst[11] = CLIP(((20 * (src[11] + src[12]) - 3 * (((src[10] + src[13])<<1) - (src[9] + src[14])) - (src[8] + src[15]) + round_add) >> 5), 0, 255);
215 :     dst[12] = CLIP(((20 * (src[12] + src[13]) - 3 * (((src[11] + src[14])<<1) - (src[10] + src[15])) - (src[9] + src[16]) + round_add) >> 5), 0, 255);
216 :    
217 :     dst[13] = CLIP(((20 * (src[13] + src[14]) + (src[16]<<1) + 3 * (src[11] - ((src[12] + src[15]) << 1)) - src[10] + round_add) >> 5), 0, 255);
218 :     dst[14] = CLIP(((19 * src[15] + 20 * src[14] + 3 * (src[12] - src[16] - (src[13] << 1)) - src[11] + round_add) >> 5), 0, 255);
219 :     dst[15] = CLIP(((23 * src[15] + 7 * ((src[16]<<1) - src[14]) + 3 * src[13] - src[12] + round_add) >> 5), 0, 255);
220 :    
221 :     dst += stride;
222 :     src += stride;
223 :     }
224 :     }
225 :    
226 : Isibaar 579 void interpolate8x8_lowpass_h_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
227 : Isibaar 333 {
228 :     int32_t i;
229 : Isibaar 579 uint8_t round_add = 16 - rounding;
230 : Isibaar 333
231 :     for(i = 0; i < 8; i++)
232 :     {
233 : Isibaar 579
234 :     dst[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);
235 :     dst[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);
236 :     dst[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);
237 :     dst[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);
238 :     dst[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);
239 :     dst[5] = CLIP(((20 * (src[5] + src[6]) + (src[8]<<1) + 3 * (src[3] - ((src[4] + src[7]) << 1)) - src[2] + round_add) >> 5), 0, 255);
240 :     dst[6] = CLIP(((19 * src[7] + 20 * src[6] + 3 * (src[4] - src[8] - (src[5] << 1)) - src[3] + round_add) >> 5), 0, 255);
241 :     dst[7] = CLIP(((23 * src[7] + 7 * ((src[8]<<1) - src[6]) + 3 * src[5] - src[4] + round_add) >> 5), 0, 255);
242 :    
243 :     dst += stride;
244 :     src += stride;
245 : Isibaar 333 }
246 :     }
247 :    
248 : Isibaar 579 void interpolate8x8_6tap_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
249 : Isibaar 333 {
250 : Isibaar 579 int32_t i;
251 :     uint8_t round_add = 16 - rounding;
252 : Isibaar 333
253 : Isibaar 579 for(i = 0; i < 8; i++)
254 : Isibaar 333 {
255 : Isibaar 579 int32_t src_2 = src[-2*stride];
256 :     int32_t src_1 = src[-stride];
257 :     int32_t src0 = src[0];
258 :     int32_t src1 = src[stride];
259 :     int32_t src2 = src[2 * stride];
260 :     int32_t src3 = src[3 * stride];
261 :     int32_t src4 = src[4 * stride];
262 :     int32_t src5 = src[5 * stride];
263 :     int32_t src6 = src[6 * stride];
264 :     int32_t src7 = src[7 * stride];
265 :     int32_t src8 = src[8 * stride];
266 :     int32_t src9 = src[9 * stride];
267 :     int32_t src10 = src[10 * stride];
268 : Isibaar 333
269 : Isibaar 579 dst[0] = CLIP((((src_2 + src3) + 5 * (((src0 + src1)<<2) - (src_1 + src2)) + round_add) >> 5), 0, 255);
270 :     dst[stride] = CLIP((((src_1 + src4) + 5 * (((src1 + src2)<<2) - (src0 + src3)) + round_add) >> 5), 0, 255);
271 :     dst[2 * stride] = CLIP((((src0 + src5) + 5 * (((src2 + src3)<<2) - (src1 + src4)) + round_add) >> 5), 0, 255);
272 :     dst[3 * stride] = CLIP((((src1 + src6) + 5 * (((src3 + src4)<<2) - (src2 + src5)) + round_add) >> 5), 0, 255);
273 :     dst[4 * stride] = CLIP((((src2 + src7) + 5 * (((src4 + src5)<<2) - (src3 + src6)) + round_add) >> 5), 0, 255);
274 :     dst[5 * stride] = CLIP((((src3 + src8) + 5 * (((src5 + src6)<<2) - (src4 + src7)) + round_add) >> 5), 0, 255);
275 :     dst[6 * stride] = CLIP((((src4 + src9) + 5 * (((src6 + src7)<<2) - (src5 + src8)) + round_add) >> 5), 0, 255);
276 :     dst[7 * stride] = CLIP((((src5 + src10) + 5 * (((src7 + src8)<<2) - (src6 + src9)) + round_add) >> 5), 0, 255);
277 : Isibaar 333
278 : Isibaar 579 dst++;
279 :     src++;
280 : Isibaar 333 }
281 :     }
282 :    
283 : Isibaar 588 void interpolate16x16_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
284 :     {
285 :     int32_t i;
286 :     uint8_t round_add = 16 - rounding;
287 :    
288 :     for(i = 0; i < 16; i++)
289 :     {
290 :     int32_t src0 = src[0];
291 :     int32_t src1 = src[stride];
292 :     int32_t src2 = src[2 * stride];
293 :     int32_t src3 = src[3 * stride];
294 :     int32_t src4 = src[4 * stride];
295 :     int32_t src5 = src[5 * stride];
296 :     int32_t src6 = src[6 * stride];
297 :     int32_t src7 = src[7 * stride];
298 :     int32_t src8 = src[8 * stride];
299 :     int32_t src9 = src[9 * stride];
300 :     int32_t src10 = src[10 * stride];
301 :     int32_t src11 = src[11 * stride];
302 :     int32_t src12 = src[12 * stride];
303 :     int32_t src13 = src[13 * stride];
304 :     int32_t src14 = src[14 * stride];
305 :     int32_t src15 = src[15 * stride];
306 :     int32_t src16 = src[16 * stride];
307 :    
308 :    
309 :     dst[0] = CLIP(((7 * ((src0<<1) - src2) + 23 * src1 + 3 * src3 - src4 + round_add) >> 5), 0, 255);
310 :     dst[stride] = CLIP(((19 * src1 + 20 * src2 - src5 + 3 * (src4 - src0 - (src3<<1)) + round_add) >> 5), 0, 255);
311 :     dst[2*stride] = CLIP(((20 * (src2 + src3) + (src0<<1) + 3 * (src5 - ((src1 + src4)<<1)) - src6 + round_add) >> 5), 0, 255);
312 :    
313 :     dst[3*stride] = CLIP(((20 * (src3 + src4) + 3 * ((src6 + src1) - ((src2 + src5)<<1)) - (src0 + src7) + round_add) >> 5), 0, 255);
314 :     dst[4*stride] = CLIP(((20 * (src4 + src5) - 3 * (((src3 + src6)<<1) - (src2 + src7)) - (src1 + src8) + round_add) >> 5), 0, 255);
315 :     dst[5*stride] = CLIP(((20 * (src5 + src6) - 3 * (((src4 + src7)<<1) - (src3 + src8)) - (src2 + src9) + round_add) >> 5), 0, 255);
316 :     dst[6*stride] = CLIP(((20 * (src6 + src7) - 3 * (((src5 + src8)<<1) - (src4 + src9)) - (src3 + src10) + round_add) >> 5), 0, 255);
317 :     dst[7*stride] = CLIP(((20 * (src7 + src8) - 3 * (((src6 + src9)<<1) - (src5 + src10)) - (src4 + src11) + round_add) >> 5), 0, 255);
318 :     dst[8*stride] = CLIP(((20 * (src8 + src9) - 3 * (((src7 + src10)<<1) - (src6 + src11)) - (src5 + src12) + round_add) >> 5), 0, 255);
319 :     dst[9*stride] = CLIP(((20 * (src9 + src10) - 3 * (((src8 + src11)<<1) - (src7 + src12)) - (src6 + src13) + round_add) >> 5), 0, 255);
320 :     dst[10*stride] = CLIP(((20 * (src10 + src11) - 3 * (((src9 + src12)<<1) - (src8 + src13)) - (src7 + src14) + round_add) >> 5), 0, 255);
321 :     dst[11*stride] = CLIP(((20 * (src11 + src12) - 3 * (((src10 + src13)<<1) - (src9 + src14)) - (src8 + src15) + round_add) >> 5), 0, 255);
322 :     dst[12*stride] = CLIP(((20 * (src12 + src13) - 3 * (((src11 + src14)<<1) - (src10 + src15)) - (src9 + src16) + round_add) >> 5), 0, 255);
323 :    
324 :     dst[13*stride] = CLIP(((20 * (src13 + src14) + (src16<<1) + 3 * (src11 - ((src12 + src15) << 1)) - src10 + round_add) >> 5), 0, 255);
325 :     dst[14*stride] = CLIP(((19 * src15 + 20 * src14 + 3 * (src12 - src16 - (src13 << 1)) - src11 + round_add) >> 5), 0, 255);
326 :     dst[15*stride] = CLIP(((23 * src15 + 7 * ((src16<<1) - src14) + 3 * src13 - src12 + round_add) >> 5), 0, 255);
327 :    
328 :     dst++;
329 :     src++;
330 :     }
331 :     }
332 :    
333 : Isibaar 579 void interpolate8x8_lowpass_v_c(uint8_t *dst, uint8_t *src, int32_t stride, int32_t rounding)
334 : Isibaar 333 {
335 :     int32_t i;
336 : Isibaar 579 uint8_t round_add = 16 - rounding;
337 : Isibaar 333
338 :     for(i = 0; i < 8; i++)
339 :     {
340 : Isibaar 579 int32_t src0 = src[0];
341 :     int32_t src1 = src[stride];
342 :     int32_t src2 = src[2 * stride];
343 :     int32_t src3 = src[3 * stride];
344 :     int32_t src4 = src[4 * stride];
345 :     int32_t src5 = src[5 * stride];
346 :     int32_t src6 = src[6 * stride];
347 :     int32_t src7 = src[7 * stride];
348 :     int32_t src8 = src[8 * stride];
349 :    
350 :     dst[0] = CLIP(((7 * ((src0<<1) - src2) + 23 * src1 + 3 * src3 - src4 + round_add) >> 5), 0, 255);
351 :     dst[stride] = CLIP(((19 * src1 + 20 * src2 - src5 + 3 * (src4 - src0 - (src3 << 1)) + round_add) >> 5), 0, 255);
352 :     dst[2 * stride] = CLIP(((20 * (src2 + src3) + (src0<<1) + 3 * (src5 - ((src1 + src4) <<1 )) - src6 + round_add) >> 5), 0, 255);
353 :     dst[3 * stride] = CLIP(((20 * (src3 + src4) + 3 * ((src6 + src1) - ((src2 + src5)<<1)) - (src0 + src7) + round_add) >> 5), 0, 255);
354 :     dst[4 * stride] = CLIP(((20 * (src4 + src5) + 3 * ((src2 + src7) - ((src3 + src6)<<1)) - (src1 + src8) + round_add) >> 5), 0, 255);
355 :     dst[5 * stride] = CLIP(((20 * (src5 + src6) + (src8<<1) + 3 * (src3 - ((src4 + src7) << 1)) - src2 + round_add) >> 5), 0, 255);
356 :     dst[6 * stride] = CLIP(((19 * src7 + 20 * src6 - src3 + 3 * (src4 - src8 - (src5 << 1)) + round_add) >> 5), 0, 255);
357 :     dst[7 * stride] = CLIP(((7 * ((src8<<1) - src6) + 23 * src7 + 3 * src5 - src4 + round_add) >> 5), 0, 255);
358 : Isibaar 333
359 : Isibaar 579 dst++;
360 :     src++;
361 : Isibaar 333 }
362 :     }
363 :    
364 : Isibaar 588 void interpolate16x16_lowpass_hv_c(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t stride, int32_t rounding)
365 :     {
366 :     int32_t i;
367 :     uint8_t round_add = 16 - rounding;
368 :     uint8_t *h_ptr = dst2;
369 :    
370 :     for(i = 0; i < 17; i++)
371 :     {
372 :    
373 :     h_ptr[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);
374 :     h_ptr[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);
375 :     h_ptr[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);
376 :    
377 :     h_ptr[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);
378 :     h_ptr[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);
379 :     h_ptr[5] = CLIP(((20 * (src[5] + src[6]) - 3 * (((src[4] + src[7])<<1) - (src[3] + src[8])) - (src[2] + src[9]) + round_add) >> 5), 0, 255);
380 :     h_ptr[6] = CLIP(((20 * (src[6] + src[7]) - 3 * (((src[5] + src[8])<<1) - (src[4] + src[9])) - (src[3] + src[10]) + round_add) >> 5), 0, 255);
381 :     h_ptr[7] = CLIP(((20 * (src[7] + src[8]) - 3 * (((src[6] + src[9])<<1) - (src[5] + src[10])) - (src[4] + src[11]) + round_add) >> 5), 0, 255);
382 :     h_ptr[8] = CLIP(((20 * (src[8] + src[9]) - 3 * (((src[7] + src[10])<<1) - (src[6] + src[11])) - (src[5] + src[12]) + round_add) >> 5), 0, 255);
383 :     h_ptr[9] = CLIP(((20 * (src[9] + src[10]) - 3 * (((src[8] + src[11])<<1) - (src[7] + src[12])) - (src[6] + src[13]) + round_add) >> 5), 0, 255);
384 :     h_ptr[10] = CLIP(((20 * (src[10] + src[11]) - 3 * (((src[9] + src[12])<<1) - (src[8] + src[13])) - (src[7] + src[14]) + round_add) >> 5), 0, 255);
385 :     h_ptr[11] = CLIP(((20 * (src[11] + src[12]) - 3 * (((src[10] + src[13])<<1) - (src[9] + src[14])) - (src[8] + src[15]) + round_add) >> 5), 0, 255);
386 :     h_ptr[12] = CLIP(((20 * (src[12] + src[13]) - 3 * (((src[11] + src[14])<<1) - (src[10] + src[15])) - (src[9] + src[16]) + round_add) >> 5), 0, 255);
387 :    
388 :     h_ptr[13] = CLIP(((20 * (src[13] + src[14]) + (src[16]<<1) + 3 * (src[11] - ((src[12] + src[15]) << 1)) - src[10] + round_add) >> 5), 0, 255);
389 :     h_ptr[14] = CLIP(((19 * src[15] + 20 * src[14] + 3 * (src[12] - src[16] - (src[13] << 1)) - src[11] + round_add) >> 5), 0, 255);
390 :     h_ptr[15] = CLIP(((23 * src[15] + 7 * ((src[16]<<1) - src[14]) + 3 * src[13] - src[12] + round_add) >> 5), 0, 255);
391 :    
392 :     h_ptr += stride;
393 :     src += stride;
394 :     }
395 :    
396 :     interpolate16x16_lowpass_v_c(dst1, dst2, stride, rounding);
397 :    
398 :     }
399 :    
400 : Isibaar 579 void interpolate8x8_lowpass_hv_c(uint8_t *dst1, uint8_t *dst2, uint8_t *src, int32_t stride, int32_t rounding)
401 : Isibaar 333 {
402 : Isibaar 579 int32_t i;
403 :     uint8_t round_add = 16 - rounding;
404 :     uint8_t *h_ptr = dst2;
405 : Isibaar 333
406 : Isibaar 579 for(i = 0; i < 9; i++)
407 : Isibaar 333 {
408 : Isibaar 579
409 :     h_ptr[0] = CLIP(((7 * ((src[0]<<1) - src[2]) + 23 * src[1] + 3 * src[3] - src[4] + round_add) >> 5), 0, 255);
410 :     h_ptr[1] = CLIP(((19 * src[1] + 20 * src[2] - src[5] + 3 * (src[4] - src[0] - (src[3]<<1)) + round_add) >> 5), 0, 255);
411 :     h_ptr[2] = CLIP(((20 * (src[2] + src[3]) + (src[0]<<1) + 3 * (src[5] - ((src[1] + src[4])<<1)) - src[6] + round_add) >> 5), 0, 255);
412 :     h_ptr[3] = CLIP(((20 * (src[3] + src[4]) + 3 * ((src[6] + src[1]) - ((src[2] + src[5])<<1)) - (src[0] + src[7]) + round_add) >> 5), 0, 255);
413 :     h_ptr[4] = CLIP(((20 * (src[4] + src[5]) - 3 * (((src[3] + src[6])<<1) - (src[2] + src[7])) - (src[1] + src[8]) + round_add) >> 5), 0, 255);
414 :     h_ptr[5] = CLIP(((20 * (src[5] + src[6]) + (src[8]<<1) + 3 * (src[3] - ((src[4] + src[7]) << 1)) - src[2] + round_add) >> 5), 0, 255);
415 :     h_ptr[6] = CLIP(((19 * src[7] + 20 * src[6] + 3 * (src[4] - src[8] - (src[5] << 1)) - src[3] + round_add) >> 5), 0, 255);
416 :     h_ptr[7] = CLIP(((23 * src[7] + 7 * ((src[8]<<1) - src[6]) + 3 * src[5] - src[4] + round_add) >> 5), 0, 255);
417 :    
418 :     h_ptr += stride;
419 :     src += stride;
420 : Isibaar 333 }
421 : Isibaar 579
422 :     interpolate8x8_lowpass_v_c(dst1, dst2, stride, rounding);
423 :    
424 :     }

No admin address has been configured
ViewVC Help
Powered by ViewVC 1.0.4