1 |
|
/***************************************************************************** |
2 |
|
* |
3 |
|
* XVID MPEG-4 VIDEO CODEC |
4 |
|
* - Interpolation related header - |
5 |
|
* |
6 |
|
* Copyright(C) 2001-2003 Peter Ross <pross@xvid.org> |
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 |
|
* $Id$ |
23 |
|
* |
24 |
|
****************************************************************************/ |
25 |
|
|
26 |
#ifndef _INTERPOLATE8X8_H_ |
#ifndef _INTERPOLATE8X8_H_ |
27 |
#define _INTERPOLATE8X8_H_ |
#define _INTERPOLATE8X8_H_ |
28 |
|
|
38 |
const uint8_t *src1, |
const uint8_t *src1, |
39 |
const uint8_t *src2, |
const uint8_t *src2, |
40 |
const uint32_t stride, |
const uint32_t stride, |
41 |
const uint32_t rounding); |
const uint32_t rounding, |
42 |
|
const uint32_t height); |
43 |
typedef INTERPOLATE8X8_AVG2 *INTERPOLATE8X8_AVG2_PTR; |
typedef INTERPOLATE8X8_AVG2 *INTERPOLATE8X8_AVG2_PTR; |
44 |
|
|
45 |
typedef void (INTERPOLATE8X8_AVG4) (uint8_t *dst, |
typedef void (INTERPOLATE8X8_AVG4) (uint8_t *dst, |
108 |
INTERPOLATE8X8 interpolate8x8_halfpel_v_3dn; |
INTERPOLATE8X8 interpolate8x8_halfpel_v_3dn; |
109 |
INTERPOLATE8X8 interpolate8x8_halfpel_hv_3dn; |
INTERPOLATE8X8 interpolate8x8_halfpel_hv_3dn; |
110 |
|
|
111 |
|
INTERPOLATE8X8 interpolate8x8_halfpel_h_3dne; |
112 |
|
INTERPOLATE8X8 interpolate8x8_halfpel_v_3dne; |
113 |
|
INTERPOLATE8X8 interpolate8x8_halfpel_hv_3dne; |
114 |
|
|
115 |
INTERPOLATE8X8 interpolate8x8_halfpel_h_ia64; |
INTERPOLATE8X8 interpolate8x8_halfpel_h_ia64; |
116 |
INTERPOLATE8X8 interpolate8x8_halfpel_v_ia64; |
INTERPOLATE8X8 interpolate8x8_halfpel_v_ia64; |
117 |
INTERPOLATE8X8 interpolate8x8_halfpel_hv_ia64; |
INTERPOLATE8X8 interpolate8x8_halfpel_hv_ia64; |
149 |
{ |
{ |
150 |
int32_t ddx, ddy; |
int32_t ddx, ddy; |
151 |
|
|
152 |
switch (((dx & 1) << 1) + (dy & 1)) // ((dx%2)?2:0)+((dy%2)?1:0) |
switch (((dx & 1) << 1) + (dy & 1)) /* ((dx%2)?2:0)+((dy%2)?1:0) */ |
153 |
{ |
{ |
154 |
case 0: |
case 0: |
155 |
ddx = dx / 2; |
ddx = dx / 2; |
184 |
} |
} |
185 |
} |
} |
186 |
|
|
187 |
|
|
188 |
|
static __inline void |
189 |
|
interpolate16x16_switch(uint8_t * const cur, |
190 |
|
const uint8_t * const refn, |
191 |
|
const uint32_t x, |
192 |
|
const uint32_t y, |
193 |
|
const int32_t dx, |
194 |
|
const int dy, |
195 |
|
const uint32_t stride, |
196 |
|
const uint32_t rounding) |
197 |
|
{ |
198 |
|
interpolate8x8_switch(cur, refn, x, y, dx, dy, stride, rounding); |
199 |
|
interpolate8x8_switch(cur, refn, x+8, y, dx, dy, stride, rounding); |
200 |
|
interpolate8x8_switch(cur, refn, x, y+8, dx, dy, stride, rounding); |
201 |
|
interpolate8x8_switch(cur, refn, x+8, y+8, dx, dy, stride, rounding); |
202 |
|
} |
203 |
|
|
204 |
|
|
205 |
|
static __inline void |
206 |
|
interpolate32x32_switch(uint8_t * const cur, |
207 |
|
const uint8_t * const refn, |
208 |
|
const uint32_t x, |
209 |
|
const uint32_t y, |
210 |
|
const int32_t dx, |
211 |
|
const int dy, |
212 |
|
const uint32_t stride, |
213 |
|
const uint32_t rounding) |
214 |
|
{ |
215 |
|
interpolate16x16_switch(cur, refn, x, y, dx, dy, stride, rounding); |
216 |
|
interpolate16x16_switch(cur, refn, x+16, y, dx, dy, stride, rounding); |
217 |
|
interpolate16x16_switch(cur, refn, x, y+16, dx, dy, stride, rounding); |
218 |
|
interpolate16x16_switch(cur, refn, x+16, y+16, dx, dy, stride, rounding); |
219 |
|
} |
220 |
|
|
221 |
|
|
222 |
static __inline uint8_t * |
static __inline uint8_t * |
223 |
interpolate8x8_switch2(uint8_t * const buffer, |
interpolate8x8_switch2(uint8_t * const buffer, |
224 |
const uint8_t * const refn, |
const uint8_t * const refn, |
231 |
{ |
{ |
232 |
int32_t ddx, ddy; |
int32_t ddx, ddy; |
233 |
|
|
234 |
switch (((dx & 1) << 1) + (dy & 1)) // ((dx%2)?2:0)+((dy%2)?1:0) |
switch (((dx & 1) << 1) + (dy & 1)) /* ((dx%2)?2:0)+((dy%2)?1:0) */ |
235 |
{ |
{ |
236 |
case 0: |
case 0: |
237 |
return (uint8_t *)refn + (int)((y + dy/2) * stride + x + dx/2); |
return (uint8_t *)refn + (int)((y + dy/2) * stride + x + dx/2); |
263 |
return buffer; |
return buffer; |
264 |
} |
} |
265 |
|
|
|
|
|
266 |
static __inline void interpolate8x8_quarterpel(uint8_t * const cur, |
static __inline void interpolate8x8_quarterpel(uint8_t * const cur, |
267 |
uint8_t * const refn, |
uint8_t * const refn, |
268 |
uint8_t * const refh, |
uint8_t * const refh, |
307 |
|
|
308 |
case 1: |
case 1: |
309 |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
310 |
interpolate8x8_avg2(dst, src, halfpel_h, stride, rounding); |
interpolate8x8_avg2(dst, src, halfpel_h, stride, rounding, 8); |
311 |
break; |
break; |
312 |
|
|
313 |
case 2: |
case 2: |
316 |
|
|
317 |
case 3: |
case 3: |
318 |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
319 |
interpolate8x8_avg2(dst, src + 1, halfpel_h, stride, rounding); |
interpolate8x8_avg2(dst, src + 1, halfpel_h, stride, rounding, 8); |
320 |
break; |
break; |
321 |
|
|
322 |
case 4: |
case 4: |
323 |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
324 |
interpolate8x8_avg2(dst, src, halfpel_v, stride, rounding); |
interpolate8x8_avg2(dst, src, halfpel_v, stride, rounding, 8); |
325 |
break; |
break; |
326 |
|
|
327 |
case 5: |
case 5: |
328 |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
329 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
330 |
interpolate8x8_avg4(dst, src, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
331 |
|
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding, 8); |
332 |
break; |
break; |
333 |
|
|
334 |
case 6: |
case 6: |
335 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
336 |
interpolate8x8_avg2(dst, halfpel_h, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(dst, halfpel_h, halfpel_hv, stride, rounding, 8); |
337 |
break; |
break; |
338 |
|
|
339 |
case 7: |
case 7: |
340 |
interpolate8x8_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
341 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src + 1, halfpel_h, stride, rounding, 9); |
342 |
interpolate8x8_avg4(dst, src+1, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
343 |
|
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding, 8); |
344 |
break; |
break; |
345 |
|
|
346 |
case 8: |
case 8: |
348 |
break; |
break; |
349 |
|
|
350 |
case 9: |
case 9: |
351 |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
352 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
353 |
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(dst, halfpel_v, stride, rounding); |
354 |
break; |
break; |
355 |
|
|
356 |
case 10: |
case 10: |
358 |
break; |
break; |
359 |
|
|
360 |
case 11: |
case 11: |
361 |
interpolate8x8_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
362 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src + 1, halfpel_h, stride, rounding, 9); |
363 |
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(dst, halfpel_v, stride, rounding); |
364 |
break; |
break; |
365 |
|
|
366 |
case 12: |
case 12: |
367 |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
368 |
interpolate8x8_avg2(dst, src+stride, halfpel_v, stride, rounding); |
interpolate8x8_avg2(dst, src+stride, halfpel_v, stride, rounding, 8); |
369 |
break; |
break; |
370 |
|
|
371 |
case 13: |
case 13: |
372 |
interpolate8x8_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
373 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
374 |
interpolate8x8_avg4(dst, src+stride, halfpel_h+stride, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
375 |
|
interpolate8x8_avg2(dst, halfpel_v+stride, halfpel_hv, stride, rounding, 8); |
376 |
break; |
break; |
377 |
|
|
378 |
case 14: |
case 14: |
379 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
380 |
interpolate8x8_avg2(dst, halfpel_h+stride, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(dst, halfpel_h+stride, halfpel_hv, stride, rounding, 8); |
381 |
break; |
break; |
382 |
|
|
383 |
case 15: |
case 15: |
384 |
interpolate8x8_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate8x8_lowpass_h(halfpel_h, src, stride, rounding); |
385 |
interpolate8x8_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src + 1, halfpel_h, stride, rounding, 9); |
386 |
interpolate8x8_avg4(dst, src+stride+1, halfpel_h+stride, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
387 |
|
interpolate8x8_avg2(dst, halfpel_hv, halfpel_v + stride, stride, rounding, 8); |
388 |
break; |
break; |
389 |
} |
} |
390 |
} |
} |
391 |
|
|
|
|
|
392 |
static __inline void interpolate16x16_quarterpel(uint8_t * const cur, |
static __inline void interpolate16x16_quarterpel(uint8_t * const cur, |
393 |
uint8_t * const refn, |
uint8_t * const refn, |
394 |
uint8_t * const refh, |
uint8_t * const refh, |
428 |
switch((y_frac << 2) | (x_frac)) { |
switch((y_frac << 2) | (x_frac)) { |
429 |
|
|
430 |
case 0: |
case 0: |
431 |
transfer8x8_copy(dst, src, stride); |
transfer16x16_copy(dst, src, stride); |
|
transfer8x8_copy(dst+8, src+8, stride); |
|
|
transfer8x8_copy(dst+8*stride, src+8*stride, stride); |
|
|
transfer8x8_copy(dst+8*stride+8, src+8*stride+8, stride); |
|
432 |
break; |
break; |
433 |
|
|
434 |
case 1: |
case 1: |
435 |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
436 |
interpolate8x8_avg2(dst, src, halfpel_h, stride, rounding); |
interpolate8x8_avg2(dst, src, halfpel_h, stride, rounding, 8); |
437 |
interpolate8x8_avg2(dst+8, src+8, halfpel_h+8, stride, rounding); |
interpolate8x8_avg2(dst+8, src+8, halfpel_h+8, stride, rounding, 8); |
438 |
interpolate8x8_avg2(dst+8*stride, src+8*stride, halfpel_h+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, src+8*stride, halfpel_h+8*stride, stride, rounding, 8); |
439 |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 8); |
440 |
break; |
break; |
441 |
|
|
442 |
case 2: |
case 2: |
445 |
|
|
446 |
case 3: |
case 3: |
447 |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
448 |
interpolate8x8_avg2(dst, src + 1, halfpel_h, stride, rounding); |
interpolate8x8_avg2(dst, src + 1, halfpel_h, stride, rounding, 8); |
449 |
interpolate8x8_avg2(dst+8, src + 8 + 1, halfpel_h+8, stride, rounding); |
interpolate8x8_avg2(dst+8, src + 8 + 1, halfpel_h+8, stride, rounding, 8); |
450 |
interpolate8x8_avg2(dst+8*stride, src + 8*stride + 1, halfpel_h+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, src + 8*stride + 1, halfpel_h+8*stride, stride, rounding, 8); |
451 |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8 + 1, halfpel_h+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8 + 1, halfpel_h+8*stride+8, stride, rounding, 8); |
452 |
break; |
break; |
453 |
|
|
454 |
case 4: |
case 4: |
455 |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
456 |
interpolate8x8_avg2(dst, src, halfpel_v, stride, rounding); |
interpolate8x8_avg2(dst, src, halfpel_v, stride, rounding, 8); |
457 |
interpolate8x8_avg2(dst+8, src+8, halfpel_v+8, stride, rounding); |
interpolate8x8_avg2(dst+8, src+8, halfpel_v+8, stride, rounding, 8); |
458 |
interpolate8x8_avg2(dst+8*stride, src+8*stride, halfpel_v+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, src+8*stride, halfpel_v+8*stride, stride, rounding, 8); |
459 |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8, halfpel_v+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, src+8*stride+8, halfpel_v+8*stride+8, stride, rounding, 8); |
460 |
break; |
break; |
461 |
|
|
462 |
case 5: |
case 5: |
463 |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
464 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
465 |
interpolate8x8_avg4(dst, src, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src + 8, halfpel_h+8, stride, rounding, 9); |
466 |
interpolate8x8_avg4(dst+8, src+8, halfpel_h+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
467 |
interpolate8x8_avg4(dst+8*stride, src+8*stride, halfpel_h+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
468 |
interpolate8x8_avg4(dst+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
|
469 |
|
interpolate16x16_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
470 |
|
interpolate8x8_avg2(dst, halfpel_hv, halfpel_v, stride, rounding, 8); |
471 |
|
interpolate8x8_avg2(dst+8, halfpel_hv+8, halfpel_v+8, stride, rounding, 8); |
472 |
|
interpolate8x8_avg2(dst+8*stride, halfpel_hv+8*stride, halfpel_v+8*stride, stride, rounding, 8); |
473 |
|
interpolate8x8_avg2(dst+8*stride+8, halfpel_hv+8*stride+8, halfpel_v+8*stride+8, stride, rounding, 8); |
474 |
break; |
break; |
475 |
|
|
476 |
case 6: |
case 6: |
477 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
478 |
interpolate8x8_avg2(dst, halfpel_h, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(dst, halfpel_h, halfpel_hv, stride, rounding, 8); |
479 |
interpolate8x8_avg2(dst+8, halfpel_h+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(dst+8, halfpel_h+8, halfpel_hv+8, stride, rounding, 8); |
480 |
interpolate8x8_avg2(dst+8*stride, halfpel_h+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, halfpel_h+8*stride, halfpel_hv+8*stride, stride, rounding, 8); |
481 |
interpolate8x8_avg2(dst+8*stride+8, halfpel_h+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, halfpel_h+8*stride+8, halfpel_hv+8*stride+8, stride, rounding, 8); |
482 |
break; |
break; |
483 |
|
|
484 |
case 7: |
case 7: |
485 |
interpolate16x16_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
486 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src+1, halfpel_h, stride, rounding, 9); |
487 |
interpolate8x8_avg4(dst, src+1, halfpel_h, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src+1 + 8, halfpel_h+8, stride, rounding, 9); |
488 |
interpolate8x8_avg4(dst+8, src+8+1, halfpel_h+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src+1 + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
489 |
interpolate8x8_avg4(dst+8*stride, src+1+8*stride, halfpel_h+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+1+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
490 |
interpolate8x8_avg4(dst+8*stride+8, src+1+8*stride+8, halfpel_h+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
|
491 |
|
interpolate16x16_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
492 |
|
interpolate8x8_avg2(dst, halfpel_hv, halfpel_v, stride, rounding, 8); |
493 |
|
interpolate8x8_avg2(dst+8, halfpel_hv+8, halfpel_v+8, stride, rounding, 8); |
494 |
|
interpolate8x8_avg2(dst+8*stride, halfpel_hv+8*stride, halfpel_v+8*stride, stride, rounding, 8); |
495 |
|
interpolate8x8_avg2(dst+8*stride+8, halfpel_hv+8*stride+8, halfpel_v+8*stride+8, stride, rounding, 8); |
496 |
break; |
break; |
497 |
|
|
498 |
case 8: |
case 8: |
500 |
break; |
break; |
501 |
|
|
502 |
case 9: |
case 9: |
503 |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
504 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
505 |
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src + 8, halfpel_h+8, stride, rounding, 9); |
506 |
interpolate8x8_avg2(dst+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
507 |
interpolate8x8_avg2(dst+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
508 |
interpolate8x8_avg2(dst+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
interpolate16x16_lowpass_v(dst, halfpel_v, stride, rounding); |
509 |
break; |
break; |
510 |
|
|
511 |
case 10: |
case 10: |
513 |
break; |
break; |
514 |
|
|
515 |
case 11: |
case 11: |
516 |
interpolate16x16_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
517 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src+1, halfpel_h, stride, rounding, 9); |
518 |
interpolate8x8_avg2(dst, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src+1 + 8, halfpel_h+8, stride, rounding, 9); |
519 |
interpolate8x8_avg2(dst+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src+1 + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
520 |
interpolate8x8_avg2(dst+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+1+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
521 |
interpolate8x8_avg2(dst+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
interpolate16x16_lowpass_v(dst, halfpel_v, stride, rounding); |
522 |
break; |
break; |
523 |
|
|
524 |
case 12: |
case 12: |
525 |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
526 |
interpolate8x8_avg2(dst, src+stride, halfpel_v, stride, rounding); |
interpolate8x8_avg2(dst, src+stride, halfpel_v, stride, rounding, 8); |
527 |
interpolate8x8_avg2(dst+8, src+stride+8, halfpel_v+8, stride, rounding); |
interpolate8x8_avg2(dst+8, src+stride+8, halfpel_v+8, stride, rounding, 8); |
528 |
interpolate8x8_avg2(dst+8*stride, src+stride+8*stride, halfpel_v+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, src+stride+8*stride, halfpel_v+8*stride, stride, rounding, 8); |
529 |
interpolate8x8_avg2(dst+8*stride+8, src+stride+8*stride+8, halfpel_v+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, src+stride+8*stride+8, halfpel_v+8*stride+8, stride, rounding, 8); |
530 |
break; |
break; |
531 |
|
|
532 |
case 13: |
case 13: |
533 |
interpolate16x16_lowpass_v(halfpel_v, src, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
534 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src, halfpel_h, stride, rounding, 9); |
535 |
interpolate8x8_avg4(dst, src+stride, halfpel_h+stride, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src + 8, halfpel_h+8, stride, rounding, 9); |
536 |
interpolate8x8_avg4(dst+8, src+stride+8, halfpel_h+stride+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
537 |
interpolate8x8_avg4(dst+8*stride, src+stride+8*stride, halfpel_h+stride+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
538 |
interpolate8x8_avg4(dst+8*stride+8, src+stride+8*stride+8, halfpel_h+stride+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
|
539 |
|
interpolate16x16_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
540 |
|
interpolate8x8_avg2(dst, halfpel_hv, halfpel_v+stride, stride, rounding, 8); |
541 |
|
interpolate8x8_avg2(dst+8, halfpel_hv+8, halfpel_v+stride+8, stride, rounding, 8); |
542 |
|
interpolate8x8_avg2(dst+8*stride, halfpel_hv+8*stride, halfpel_v+stride+8*stride, stride, rounding, 8); |
543 |
|
interpolate8x8_avg2(dst+8*stride+8, halfpel_hv+8*stride+8, halfpel_v+stride+8*stride+8, stride, rounding, 8); |
544 |
break; |
break; |
545 |
|
|
546 |
case 14: |
case 14: |
547 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
548 |
interpolate8x8_avg2(dst, halfpel_h+stride, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(dst, halfpel_h+stride, halfpel_hv, stride, rounding, 8); |
549 |
interpolate8x8_avg2(dst+8, halfpel_h+stride+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(dst+8, halfpel_h+stride+8, halfpel_hv+8, stride, rounding, 8); |
550 |
interpolate8x8_avg2(dst+8*stride, halfpel_h+stride+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, halfpel_h+stride+8*stride, halfpel_hv+8*stride, stride, rounding, 8); |
551 |
interpolate8x8_avg2(dst+8*stride+8, halfpel_h+stride+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, halfpel_h+stride+8*stride+8, halfpel_hv+8*stride+8, stride, rounding, 8); |
552 |
break; |
break; |
553 |
|
|
554 |
case 15: |
case 15: |
555 |
interpolate16x16_lowpass_v(halfpel_v, src+1, stride, rounding); |
interpolate16x16_lowpass_h(halfpel_h, src, stride, rounding); |
556 |
interpolate16x16_lowpass_hv(halfpel_hv, halfpel_h, src, stride, rounding); |
interpolate8x8_avg2(halfpel_v, src+1, halfpel_h, stride, rounding, 9); |
557 |
interpolate8x8_avg4(dst, src+stride+1, halfpel_h+stride, halfpel_v, halfpel_hv, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8, src+1 + 8, halfpel_h+8, stride, rounding, 9); |
558 |
interpolate8x8_avg4(dst+8, src+stride+1+8, halfpel_h+stride+8, halfpel_v+8, halfpel_hv+8, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride, src+1 + 8*stride, halfpel_h+8*stride, stride, rounding, 9); |
559 |
interpolate8x8_avg4(dst+8*stride, src+stride+1+8*stride, halfpel_h+stride+8*stride, halfpel_v+8*stride, halfpel_hv+8*stride, stride, rounding); |
interpolate8x8_avg2(halfpel_v+8*stride+8, src+1+8*stride+8, halfpel_h+8*stride+8, stride, rounding, 9); |
560 |
interpolate8x8_avg4(dst+8*stride+8, src+stride+1+8*stride+8, halfpel_h+stride+8*stride+8, halfpel_v+8*stride+8, halfpel_hv+8*stride+8, stride, rounding); |
|
561 |
break; |
interpolate16x16_lowpass_v(halfpel_hv, halfpel_v, stride, rounding); |
562 |
default: |
interpolate8x8_avg2(dst, halfpel_hv, halfpel_v+stride, stride, rounding, 8); |
563 |
interpolate8x8_quarterpel(cur, refn, refh, refv, refhv, x, y, dx, dy, stride, rounding); |
interpolate8x8_avg2(dst+8, halfpel_hv+8, halfpel_v+stride+8, stride, rounding, 8); |
564 |
interpolate8x8_quarterpel(cur, refn, refh, refv, refhv, x+8, y, dx, dy, stride, rounding); |
interpolate8x8_avg2(dst+8*stride, halfpel_hv+8*stride, halfpel_v+stride+8*stride, stride, rounding, 8); |
565 |
interpolate8x8_quarterpel(cur, refn, refh, refv, refhv, x, y+8, dx, dy, stride, rounding); |
interpolate8x8_avg2(dst+8*stride+8, halfpel_hv+8*stride+8, halfpel_v+stride+8*stride+8, stride, rounding, 8); |
|
interpolate8x8_quarterpel(cur, refn, refh, refv, refhv, x+8, y+8, dx, dy, stride, rounding); |
|
|
return; |
|
566 |
break; |
break; |
567 |
} |
} |
568 |
} |
} |