2 * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
3 * Copyright (c) 2003-2010 Michael Niedermayer <michaelni@gmx.at>
5 * This file is part of FFmpeg.
7 * FFmpeg is free software; you can redistribute it and/or
8 * modify it under the terms of the GNU Lesser General Public
9 * License as published by the Free Software Foundation; either
10 * version 2.1 of the License, or (at your option) any later version.
12 * FFmpeg is distributed in the hope that it will be useful,
13 * but WITHOUT ANY WARRANTY; without even the implied warranty of
14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15 * Lesser General Public License for more details.
17 * You should have received a copy of the GNU Lesser General Public
18 * License along with FFmpeg; if not, write to the Free Software
19 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
22 #include "libavutil/common.h"
23 #include "libavutil/intreadwrite.h"
25 #include "bit_depth_template.c"
26 #include "hpel_template.c"
28 static inline void FUNC(copy_block2)(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride, int h)
33 AV_WN2P(dst , AV_RN2P(src ));
39 static inline void FUNC(copy_block4)(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride, int h)
44 AV_WN4P(dst , AV_RN4P(src ));
50 static inline void FUNC(copy_block8)(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride, int h)
55 AV_WN4P(dst , AV_RN4P(src ));
56 AV_WN4P(dst+4*sizeof(pixel), AV_RN4P(src+4*sizeof(pixel)));
62 static inline void FUNC(copy_block16)(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride, int h)
67 AV_WN4P(dst , AV_RN4P(src ));
68 AV_WN4P(dst+ 4*sizeof(pixel), AV_RN4P(src+ 4*sizeof(pixel)));
69 AV_WN4P(dst+ 8*sizeof(pixel), AV_RN4P(src+ 8*sizeof(pixel)));
70 AV_WN4P(dst+12*sizeof(pixel), AV_RN4P(src+12*sizeof(pixel)));
76 #define H264_LOWPASS(OPNAME, OP, OP2) \
77 static av_unused void FUNC(OPNAME ## h264_qpel2_h_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
81 pixel *dst = (pixel*)p_dst;\
82 pixel *src = (pixel*)p_src;\
83 dstStride >>= sizeof(pixel)-1;\
84 srcStride >>= sizeof(pixel)-1;\
87 OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
88 OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\
94 static av_unused void FUNC(OPNAME ## h264_qpel2_v_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
98 pixel *dst = (pixel*)p_dst;\
99 pixel *src = (pixel*)p_src;\
100 dstStride >>= sizeof(pixel)-1;\
101 srcStride >>= sizeof(pixel)-1;\
104 const int srcB= src[-2*srcStride];\
105 const int srcA= src[-1*srcStride];\
106 const int src0= src[0 *srcStride];\
107 const int src1= src[1 *srcStride];\
108 const int src2= src[2 *srcStride];\
109 const int src3= src[3 *srcStride];\
110 const int src4= src[4 *srcStride];\
111 OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\
112 OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\
118 static av_unused void FUNC(OPNAME ## h264_qpel2_hv_lowpass)(uint8_t *p_dst, pixeltmp *tmp, uint8_t *p_src, int dstStride, int tmpStride, int srcStride){\
121 const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
124 pixel *dst = (pixel*)p_dst;\
125 pixel *src = (pixel*)p_src;\
126 dstStride >>= sizeof(pixel)-1;\
127 srcStride >>= sizeof(pixel)-1;\
129 for(i=0; i<h+5; i++)\
131 tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\
132 tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\
136 tmp -= tmpStride*(h+5-2);\
139 const int tmpB= tmp[-2*tmpStride] - pad;\
140 const int tmpA= tmp[-1*tmpStride] - pad;\
141 const int tmp0= tmp[0 *tmpStride] - pad;\
142 const int tmp1= tmp[1 *tmpStride] - pad;\
143 const int tmp2= tmp[2 *tmpStride] - pad;\
144 const int tmp3= tmp[3 *tmpStride] - pad;\
145 const int tmp4= tmp[4 *tmpStride] - pad;\
146 OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
147 OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
152 static void FUNC(OPNAME ## h264_qpel4_h_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
156 pixel *dst = (pixel*)p_dst;\
157 pixel *src = (pixel*)p_src;\
158 dstStride >>= sizeof(pixel)-1;\
159 srcStride >>= sizeof(pixel)-1;\
162 OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
163 OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\
164 OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]));\
165 OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]));\
171 static void FUNC(OPNAME ## h264_qpel4_v_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
175 pixel *dst = (pixel*)p_dst;\
176 pixel *src = (pixel*)p_src;\
177 dstStride >>= sizeof(pixel)-1;\
178 srcStride >>= sizeof(pixel)-1;\
181 const int srcB= src[-2*srcStride];\
182 const int srcA= src[-1*srcStride];\
183 const int src0= src[0 *srcStride];\
184 const int src1= src[1 *srcStride];\
185 const int src2= src[2 *srcStride];\
186 const int src3= src[3 *srcStride];\
187 const int src4= src[4 *srcStride];\
188 const int src5= src[5 *srcStride];\
189 const int src6= src[6 *srcStride];\
190 OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\
191 OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\
192 OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\
193 OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\
199 static void FUNC(OPNAME ## h264_qpel4_hv_lowpass)(uint8_t *p_dst, pixeltmp *tmp, uint8_t *p_src, int dstStride, int tmpStride, int srcStride){\
202 const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
205 pixel *dst = (pixel*)p_dst;\
206 pixel *src = (pixel*)p_src;\
207 dstStride >>= sizeof(pixel)-1;\
208 srcStride >>= sizeof(pixel)-1;\
210 for(i=0; i<h+5; i++)\
212 tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\
213 tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\
214 tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]) + pad;\
215 tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]) + pad;\
219 tmp -= tmpStride*(h+5-2);\
222 const int tmpB= tmp[-2*tmpStride] - pad;\
223 const int tmpA= tmp[-1*tmpStride] - pad;\
224 const int tmp0= tmp[0 *tmpStride] - pad;\
225 const int tmp1= tmp[1 *tmpStride] - pad;\
226 const int tmp2= tmp[2 *tmpStride] - pad;\
227 const int tmp3= tmp[3 *tmpStride] - pad;\
228 const int tmp4= tmp[4 *tmpStride] - pad;\
229 const int tmp5= tmp[5 *tmpStride] - pad;\
230 const int tmp6= tmp[6 *tmpStride] - pad;\
231 OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
232 OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
233 OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
234 OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\
240 static void FUNC(OPNAME ## h264_qpel8_h_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
244 pixel *dst = (pixel*)p_dst;\
245 pixel *src = (pixel*)p_src;\
246 dstStride >>= sizeof(pixel)-1;\
247 srcStride >>= sizeof(pixel)-1;\
250 OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]));\
251 OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]));\
252 OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]));\
253 OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]));\
254 OP(dst[4], (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]));\
255 OP(dst[5], (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]));\
256 OP(dst[6], (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]));\
257 OP(dst[7], (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]));\
263 static void FUNC(OPNAME ## h264_qpel8_v_lowpass)(uint8_t *p_dst, uint8_t *p_src, int dstStride, int srcStride){\
267 pixel *dst = (pixel*)p_dst;\
268 pixel *src = (pixel*)p_src;\
269 dstStride >>= sizeof(pixel)-1;\
270 srcStride >>= sizeof(pixel)-1;\
273 const int srcB= src[-2*srcStride];\
274 const int srcA= src[-1*srcStride];\
275 const int src0= src[0 *srcStride];\
276 const int src1= src[1 *srcStride];\
277 const int src2= src[2 *srcStride];\
278 const int src3= src[3 *srcStride];\
279 const int src4= src[4 *srcStride];\
280 const int src5= src[5 *srcStride];\
281 const int src6= src[6 *srcStride];\
282 const int src7= src[7 *srcStride];\
283 const int src8= src[8 *srcStride];\
284 const int src9= src[9 *srcStride];\
285 const int src10=src[10*srcStride];\
286 OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\
287 OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\
288 OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\
289 OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\
290 OP(dst[4*dstStride], (src4+src5)*20 - (src3+src6)*5 + (src2+src7));\
291 OP(dst[5*dstStride], (src5+src6)*20 - (src4+src7)*5 + (src3+src8));\
292 OP(dst[6*dstStride], (src6+src7)*20 - (src5+src8)*5 + (src4+src9));\
293 OP(dst[7*dstStride], (src7+src8)*20 - (src6+src9)*5 + (src5+src10));\
299 static void FUNC(OPNAME ## h264_qpel8_hv_lowpass)(uint8_t *p_dst, pixeltmp *tmp, uint8_t *p_src, int dstStride, int tmpStride, int srcStride){\
302 const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
305 pixel *dst = (pixel*)p_dst;\
306 pixel *src = (pixel*)p_src;\
307 dstStride >>= sizeof(pixel)-1;\
308 srcStride >>= sizeof(pixel)-1;\
310 for(i=0; i<h+5; i++)\
312 tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]) + pad;\
313 tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]) + pad;\
314 tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]) + pad;\
315 tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]) + pad;\
316 tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]) + pad;\
317 tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]) + pad;\
318 tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]) + pad;\
319 tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]) + pad;\
323 tmp -= tmpStride*(h+5-2);\
326 const int tmpB= tmp[-2*tmpStride] - pad;\
327 const int tmpA= tmp[-1*tmpStride] - pad;\
328 const int tmp0= tmp[0 *tmpStride] - pad;\
329 const int tmp1= tmp[1 *tmpStride] - pad;\
330 const int tmp2= tmp[2 *tmpStride] - pad;\
331 const int tmp3= tmp[3 *tmpStride] - pad;\
332 const int tmp4= tmp[4 *tmpStride] - pad;\
333 const int tmp5= tmp[5 *tmpStride] - pad;\
334 const int tmp6= tmp[6 *tmpStride] - pad;\
335 const int tmp7= tmp[7 *tmpStride] - pad;\
336 const int tmp8= tmp[8 *tmpStride] - pad;\
337 const int tmp9= tmp[9 *tmpStride] - pad;\
338 const int tmp10=tmp[10*tmpStride] - pad;\
339 OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
340 OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
341 OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
342 OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\
343 OP2(dst[4*dstStride], (tmp4+tmp5)*20 - (tmp3+tmp6)*5 + (tmp2+tmp7));\
344 OP2(dst[5*dstStride], (tmp5+tmp6)*20 - (tmp4+tmp7)*5 + (tmp3+tmp8));\
345 OP2(dst[6*dstStride], (tmp6+tmp7)*20 - (tmp5+tmp8)*5 + (tmp4+tmp9));\
346 OP2(dst[7*dstStride], (tmp7+tmp8)*20 - (tmp6+tmp9)*5 + (tmp5+tmp10));\
352 static void FUNC(OPNAME ## h264_qpel16_v_lowpass)(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
353 FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst , src , dstStride, srcStride);\
354 FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
357 FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst , src , dstStride, srcStride);\
358 FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
361 static void FUNC(OPNAME ## h264_qpel16_h_lowpass)(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
362 FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst , src , dstStride, srcStride);\
363 FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
366 FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst , src , dstStride, srcStride);\
367 FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
370 static void FUNC(OPNAME ## h264_qpel16_hv_lowpass)(uint8_t *dst, pixeltmp *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
371 FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst , tmp , src , dstStride, tmpStride, srcStride);\
372 FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\
375 FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst , tmp , src , dstStride, tmpStride, srcStride);\
376 FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\
379 #define H264_MC(OPNAME, SIZE) \
380 static av_unused void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc00)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
382 FUNCC(OPNAME ## pixels ## SIZE)(dst, src, stride, SIZE);\
385 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc10)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
387 uint8_t half[SIZE*SIZE*sizeof(pixel)];\
388 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
389 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src, half, stride, stride, SIZE*sizeof(pixel), SIZE);\
392 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc20)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
394 FUNC(OPNAME ## h264_qpel ## SIZE ## _h_lowpass)(dst, src, stride, stride);\
397 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc30)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
399 uint8_t half[SIZE*SIZE*sizeof(pixel)];\
400 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
401 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src+sizeof(pixel), half, stride, stride, SIZE*sizeof(pixel), SIZE);\
404 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
406 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
407 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
408 uint8_t half[SIZE*SIZE*sizeof(pixel)];\
409 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
410 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
411 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid, half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
414 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc02)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
416 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
417 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
418 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
419 FUNC(OPNAME ## h264_qpel ## SIZE ## _v_lowpass)(dst, full_mid, stride, SIZE*sizeof(pixel));\
422 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
424 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
425 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
426 uint8_t half[SIZE*SIZE*sizeof(pixel)];\
427 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
428 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
429 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid+SIZE*sizeof(pixel), half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
432 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
434 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
435 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
436 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
437 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
438 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
439 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
440 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
441 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
444 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
446 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
447 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
448 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
449 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
450 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
451 FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel), stride, SIZE + 5);\
452 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
453 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
456 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
458 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
459 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
460 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
461 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
462 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
463 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
464 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
465 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
468 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
470 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
471 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
472 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
473 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
474 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
475 FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel), stride, SIZE + 5);\
476 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
477 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
480 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc22)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
482 pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
483 FUNC(OPNAME ## h264_qpel ## SIZE ## _hv_lowpass)(dst, tmp, src, stride, SIZE*sizeof(pixel), stride);\
486 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
488 pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
489 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
490 uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
491 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
492 FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
493 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
496 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
498 pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
499 uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
500 uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
501 FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
502 FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
503 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
506 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
508 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
509 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
510 pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
511 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
512 uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
513 FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\
514 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
515 FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
516 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
519 static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc32)(uint8_t *dst, uint8_t *src, ptrdiff_t stride)\
521 uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
522 uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
523 pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
524 uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
525 uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
526 FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel), stride, SIZE + 5);\
527 FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
528 FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
529 FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
532 #define op_avg(a, b) a = (((a)+CLIP(((b) + 16)>>5)+1)>>1)
533 //#define op_avg2(a, b) a = (((a)*w1+cm[((b) + 16)>>5]*w2 + o + 64)>>7)
534 #define op_put(a, b) a = CLIP(((b) + 16)>>5)
535 #define op2_avg(a, b) a = (((a)+CLIP(((b) + 512)>>10)+1)>>1)
536 #define op2_put(a, b) a = CLIP(((b) + 512)>>10)
538 H264_LOWPASS(put_ , op_put, op2_put)
539 H264_LOWPASS(avg_ , op_avg, op2_avg)