Line | Branch | Exec | Source |
---|---|---|---|
1 | /* | ||
2 | * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder | ||
3 | * Copyright (c) 2003-2010 Michael Niedermayer <michaelni@gmx.at> | ||
4 | * | ||
5 | * This file is part of FFmpeg. | ||
6 | * | ||
7 | * FFmpeg is free software; you can redistribute it and/or | ||
8 | * modify it under the terms of the GNU Lesser General Public | ||
9 | * License as published by the Free Software Foundation; either | ||
10 | * version 2.1 of the License, or (at your option) any later version. | ||
11 | * | ||
12 | * FFmpeg is distributed in the hope that it will be useful, | ||
13 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
14 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
15 | * Lesser General Public License for more details. | ||
16 | * | ||
17 | * You should have received a copy of the GNU Lesser General Public | ||
18 | * License along with FFmpeg; if not, write to the Free Software | ||
19 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
20 | */ | ||
21 | |||
22 | #include "libavutil/common.h" | ||
23 | #include "libavutil/intreadwrite.h" | ||
24 | |||
25 | #include "bit_depth_template.c" | ||
26 | #include "hpel_template.c" | ||
27 | #include "pel_template.c" | ||
28 | |||
29 | 130268 | static inline void FUNC(copy_block2)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride, int h) | |
30 | { | ||
31 | int i; | ||
32 |
2/2✓ Branch 0 taken 455938 times.
✓ Branch 1 taken 65134 times.
|
1042144 | for(i=0; i<h; i++) |
33 | { | ||
34 | 911876 | AV_WN2P(dst , AV_RN2P(src )); | |
35 | 911876 | dst+=dstStride; | |
36 | 911876 | src+=srcStride; | |
37 | } | ||
38 | 130268 | } | |
39 | |||
40 | 25368772 | static inline void FUNC(copy_block4)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride, int h) | |
41 | { | ||
42 | int i; | ||
43 |
2/2✓ Branch 0 taken 114159474 times.
✓ Branch 1 taken 12684386 times.
|
253687720 | for(i=0; i<h; i++) |
44 | { | ||
45 | 228318948 | AV_WN4P(dst , AV_RN4P(src )); | |
46 | 228318948 | dst+=dstStride; | |
47 | 228318948 | src+=srcStride; | |
48 | } | ||
49 | 25368772 | } | |
50 | |||
51 | 13852342 | static inline void FUNC(copy_block8)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride, int h) | |
52 | { | ||
53 | int i; | ||
54 |
2/2✓ Branch 0 taken 90040223 times.
✓ Branch 1 taken 6926171 times.
|
193932788 | for(i=0; i<h; i++) |
55 | { | ||
56 | 180080446 | AV_WN4P(dst , AV_RN4P(src )); | |
57 | 180080446 | AV_WN4P(dst+4*sizeof(pixel), AV_RN4P(src+4*sizeof(pixel))); | |
58 | 180080446 | dst+=dstStride; | |
59 | 180080446 | src+=srcStride; | |
60 | } | ||
61 | 13852342 | } | |
62 | |||
63 | 6676878 | static inline void FUNC(copy_block16)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride, int h) | |
64 | { | ||
65 | int i; | ||
66 |
2/2✓ Branch 0 taken 70107219 times.
✓ Branch 1 taken 3338439 times.
|
146891316 | for(i=0; i<h; i++) |
67 | { | ||
68 | 140214438 | AV_WN4P(dst , AV_RN4P(src )); | |
69 | 140214438 | AV_WN4P(dst+ 4*sizeof(pixel), AV_RN4P(src+ 4*sizeof(pixel))); | |
70 | 140214438 | AV_WN4P(dst+ 8*sizeof(pixel), AV_RN4P(src+ 8*sizeof(pixel))); | |
71 | 140214438 | AV_WN4P(dst+12*sizeof(pixel), AV_RN4P(src+12*sizeof(pixel))); | |
72 | 140214438 | dst+=dstStride; | |
73 | 140214438 | src+=srcStride; | |
74 | } | ||
75 | 6676878 | } | |
76 | |||
77 | #define H264_LOWPASS(OPNAME, OP, OP2) \ | ||
78 | static av_unused void FUNC(OPNAME ## h264_qpel2_h_lowpass)(uint8_t *p_dst, const uint8_t *restrict p_src, int dstStride, int srcStride)\ | ||
79 | {\ | ||
80 | const int h=2;\ | ||
81 | INIT_CLIP\ | ||
82 | int i;\ | ||
83 | pixel *dst = (pixel*)p_dst;\ | ||
84 | const pixel *restrict src = (const pixel*)p_src;\ | ||
85 | dstStride >>= sizeof(pixel)-1;\ | ||
86 | srcStride >>= sizeof(pixel)-1;\ | ||
87 | for(i=0; i<h; i++)\ | ||
88 | {\ | ||
89 | OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\ | ||
90 | OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\ | ||
91 | dst+=dstStride;\ | ||
92 | src+=srcStride;\ | ||
93 | }\ | ||
94 | }\ | ||
95 | \ | ||
96 | static av_unused void FUNC(OPNAME ## h264_qpel2_v_lowpass)(uint8_t *_dst, const uint8_t *restrict _src, int dstStride, int srcStride)\ | ||
97 | {\ | ||
98 | const int w=2;\ | ||
99 | INIT_CLIP\ | ||
100 | int i;\ | ||
101 | pixel *dst = (pixel*)_dst;\ | ||
102 | const pixel *restrict src = (const pixel*)_src;\ | ||
103 | dstStride >>= sizeof(pixel)-1;\ | ||
104 | srcStride >>= sizeof(pixel)-1;\ | ||
105 | for(i=0; i<w; i++)\ | ||
106 | {\ | ||
107 | const int srcB= src[-2*srcStride];\ | ||
108 | const int srcA= src[-1*srcStride];\ | ||
109 | const int src0= src[0 *srcStride];\ | ||
110 | const int src1= src[1 *srcStride];\ | ||
111 | const int src2= src[2 *srcStride];\ | ||
112 | const int src3= src[3 *srcStride];\ | ||
113 | const int src4= src[4 *srcStride];\ | ||
114 | OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\ | ||
115 | OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\ | ||
116 | dst++;\ | ||
117 | src++;\ | ||
118 | }\ | ||
119 | }\ | ||
120 | \ | ||
121 | static av_unused void FUNC(OPNAME ## h264_qpel2_hv_lowpass)(uint8_t *_dst, pixeltmp *tmp, const uint8_t *restrict _src, int dstStride, int tmpStride, int srcStride)\ | ||
122 | {\ | ||
123 | const int h=2;\ | ||
124 | const int w=2;\ | ||
125 | const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\ | ||
126 | INIT_CLIP\ | ||
127 | int i;\ | ||
128 | pixel *dst = (pixel*)_dst;\ | ||
129 | const pixel *restrict src = (const pixel*)_src;\ | ||
130 | dstStride >>= sizeof(pixel)-1;\ | ||
131 | srcStride >>= sizeof(pixel)-1;\ | ||
132 | src -= 2*srcStride;\ | ||
133 | for(i=0; i<h+5; i++)\ | ||
134 | {\ | ||
135 | tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\ | ||
136 | tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\ | ||
137 | tmp+=tmpStride;\ | ||
138 | src+=srcStride;\ | ||
139 | }\ | ||
140 | tmp -= tmpStride*(h+5-2);\ | ||
141 | for(i=0; i<w; i++)\ | ||
142 | {\ | ||
143 | const int tmpB= tmp[-2*tmpStride] - pad;\ | ||
144 | const int tmpA= tmp[-1*tmpStride] - pad;\ | ||
145 | const int tmp0= tmp[0 *tmpStride] - pad;\ | ||
146 | const int tmp1= tmp[1 *tmpStride] - pad;\ | ||
147 | const int tmp2= tmp[2 *tmpStride] - pad;\ | ||
148 | const int tmp3= tmp[3 *tmpStride] - pad;\ | ||
149 | const int tmp4= tmp[4 *tmpStride] - pad;\ | ||
150 | OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\ | ||
151 | OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\ | ||
152 | dst++;\ | ||
153 | tmp++;\ | ||
154 | }\ | ||
155 | }\ | ||
156 | static void FUNC(OPNAME ## h264_qpel4_h_lowpass)(uint8_t *_dst, const uint8_t *restrict _src, int dstStride, int srcStride)\ | ||
157 | {\ | ||
158 | const int h=4;\ | ||
159 | INIT_CLIP\ | ||
160 | int i;\ | ||
161 | pixel *dst = (pixel*)_dst;\ | ||
162 | const pixel *restrict src = (const pixel*)_src;\ | ||
163 | dstStride >>= sizeof(pixel)-1;\ | ||
164 | srcStride >>= sizeof(pixel)-1;\ | ||
165 | for(i=0; i<h; i++)\ | ||
166 | {\ | ||
167 | OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\ | ||
168 | OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\ | ||
169 | OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]));\ | ||
170 | OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]));\ | ||
171 | dst+=dstStride;\ | ||
172 | src+=srcStride;\ | ||
173 | }\ | ||
174 | }\ | ||
175 | \ | ||
176 | static void FUNC(OPNAME ## h264_qpel4_v_lowpass)(uint8_t *_dst, const uint8_t *restrict _src, int dstStride, int srcStride)\ | ||
177 | {\ | ||
178 | const int w=4;\ | ||
179 | INIT_CLIP\ | ||
180 | int i;\ | ||
181 | pixel *dst = (pixel*)_dst;\ | ||
182 | const pixel *restrict src = (const pixel*)_src;\ | ||
183 | dstStride >>= sizeof(pixel)-1;\ | ||
184 | srcStride >>= sizeof(pixel)-1;\ | ||
185 | for(i=0; i<w; i++)\ | ||
186 | {\ | ||
187 | const int srcB= src[-2*srcStride];\ | ||
188 | const int srcA= src[-1*srcStride];\ | ||
189 | const int src0= src[0 *srcStride];\ | ||
190 | const int src1= src[1 *srcStride];\ | ||
191 | const int src2= src[2 *srcStride];\ | ||
192 | const int src3= src[3 *srcStride];\ | ||
193 | const int src4= src[4 *srcStride];\ | ||
194 | const int src5= src[5 *srcStride];\ | ||
195 | const int src6= src[6 *srcStride];\ | ||
196 | OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\ | ||
197 | OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\ | ||
198 | OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\ | ||
199 | OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\ | ||
200 | dst++;\ | ||
201 | src++;\ | ||
202 | }\ | ||
203 | }\ | ||
204 | \ | ||
205 | static void FUNC(OPNAME ## h264_qpel4_hv_lowpass)(uint8_t *_dst, pixeltmp *tmp, const uint8_t *restrict _src, int dstStride, int tmpStride, int srcStride)\ | ||
206 | {\ | ||
207 | const int h=4;\ | ||
208 | const int w=4;\ | ||
209 | const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\ | ||
210 | INIT_CLIP\ | ||
211 | int i;\ | ||
212 | pixel *dst = (pixel*)_dst;\ | ||
213 | const pixel *restrict src = (const pixel*)_src;\ | ||
214 | dstStride >>= sizeof(pixel)-1;\ | ||
215 | srcStride >>= sizeof(pixel)-1;\ | ||
216 | src -= 2*srcStride;\ | ||
217 | for(i=0; i<h+5; i++)\ | ||
218 | {\ | ||
219 | tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\ | ||
220 | tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\ | ||
221 | tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]) + pad;\ | ||
222 | tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]) + pad;\ | ||
223 | tmp+=tmpStride;\ | ||
224 | src+=srcStride;\ | ||
225 | }\ | ||
226 | tmp -= tmpStride*(h+5-2);\ | ||
227 | for(i=0; i<w; i++)\ | ||
228 | {\ | ||
229 | const int tmpB= tmp[-2*tmpStride] - pad;\ | ||
230 | const int tmpA= tmp[-1*tmpStride] - pad;\ | ||
231 | const int tmp0= tmp[0 *tmpStride] - pad;\ | ||
232 | const int tmp1= tmp[1 *tmpStride] - pad;\ | ||
233 | const int tmp2= tmp[2 *tmpStride] - pad;\ | ||
234 | const int tmp3= tmp[3 *tmpStride] - pad;\ | ||
235 | const int tmp4= tmp[4 *tmpStride] - pad;\ | ||
236 | const int tmp5= tmp[5 *tmpStride] - pad;\ | ||
237 | const int tmp6= tmp[6 *tmpStride] - pad;\ | ||
238 | OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\ | ||
239 | OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\ | ||
240 | OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\ | ||
241 | OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\ | ||
242 | dst++;\ | ||
243 | tmp++;\ | ||
244 | }\ | ||
245 | }\ | ||
246 | \ | ||
247 | static void FUNC(OPNAME ## h264_qpel8_h_lowpass)(uint8_t *_dst, const uint8_t *restrict _src, int dstStride, int srcStride)\ | ||
248 | {\ | ||
249 | const int h=8;\ | ||
250 | INIT_CLIP\ | ||
251 | int i;\ | ||
252 | pixel *dst = (pixel*)_dst;\ | ||
253 | const pixel *restrict src = (const pixel*)_src;\ | ||
254 | dstStride >>= sizeof(pixel)-1;\ | ||
255 | srcStride >>= sizeof(pixel)-1;\ | ||
256 | for(i=0; i<h; i++)\ | ||
257 | {\ | ||
258 | OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]));\ | ||
259 | OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]));\ | ||
260 | OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]));\ | ||
261 | OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]));\ | ||
262 | OP(dst[4], (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]));\ | ||
263 | OP(dst[5], (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]));\ | ||
264 | OP(dst[6], (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]));\ | ||
265 | OP(dst[7], (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]));\ | ||
266 | dst+=dstStride;\ | ||
267 | src+=srcStride;\ | ||
268 | }\ | ||
269 | }\ | ||
270 | \ | ||
271 | static void FUNC(OPNAME ## h264_qpel8_v_lowpass)(uint8_t *_dst, const uint8_t *restrict _src, int dstStride, int srcStride)\ | ||
272 | {\ | ||
273 | const int w=8;\ | ||
274 | INIT_CLIP\ | ||
275 | int i;\ | ||
276 | pixel *dst = (pixel*)_dst;\ | ||
277 | const pixel *restrict src = (const pixel*)_src;\ | ||
278 | dstStride >>= sizeof(pixel)-1;\ | ||
279 | srcStride >>= sizeof(pixel)-1;\ | ||
280 | for(i=0; i<w; i++)\ | ||
281 | {\ | ||
282 | const int srcB= src[-2*srcStride];\ | ||
283 | const int srcA= src[-1*srcStride];\ | ||
284 | const int src0= src[0 *srcStride];\ | ||
285 | const int src1= src[1 *srcStride];\ | ||
286 | const int src2= src[2 *srcStride];\ | ||
287 | const int src3= src[3 *srcStride];\ | ||
288 | const int src4= src[4 *srcStride];\ | ||
289 | const int src5= src[5 *srcStride];\ | ||
290 | const int src6= src[6 *srcStride];\ | ||
291 | const int src7= src[7 *srcStride];\ | ||
292 | const int src8= src[8 *srcStride];\ | ||
293 | const int src9= src[9 *srcStride];\ | ||
294 | const int src10=src[10*srcStride];\ | ||
295 | OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\ | ||
296 | OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\ | ||
297 | OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\ | ||
298 | OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\ | ||
299 | OP(dst[4*dstStride], (src4+src5)*20 - (src3+src6)*5 + (src2+src7));\ | ||
300 | OP(dst[5*dstStride], (src5+src6)*20 - (src4+src7)*5 + (src3+src8));\ | ||
301 | OP(dst[6*dstStride], (src6+src7)*20 - (src5+src8)*5 + (src4+src9));\ | ||
302 | OP(dst[7*dstStride], (src7+src8)*20 - (src6+src9)*5 + (src5+src10));\ | ||
303 | dst++;\ | ||
304 | src++;\ | ||
305 | }\ | ||
306 | }\ | ||
307 | \ | ||
308 | static void FUNC(OPNAME ## h264_qpel8_hv_lowpass)(uint8_t *_dst, pixeltmp *tmp, const uint8_t *restrict _src, int dstStride, int tmpStride, int srcStride)\ | ||
309 | {\ | ||
310 | const int h=8;\ | ||
311 | const int w=8;\ | ||
312 | const int pad = (BIT_DEPTH == 10) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\ | ||
313 | INIT_CLIP\ | ||
314 | int i;\ | ||
315 | pixel *dst = (pixel*)_dst;\ | ||
316 | const pixel *restrict src = (const pixel*)_src;\ | ||
317 | dstStride >>= sizeof(pixel)-1;\ | ||
318 | srcStride >>= sizeof(pixel)-1;\ | ||
319 | src -= 2*srcStride;\ | ||
320 | for(i=0; i<h+5; i++)\ | ||
321 | {\ | ||
322 | tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]) + pad;\ | ||
323 | tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]) + pad;\ | ||
324 | tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]) + pad;\ | ||
325 | tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]) + pad;\ | ||
326 | tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]) + pad;\ | ||
327 | tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]) + pad;\ | ||
328 | tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]) + pad;\ | ||
329 | tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]) + pad;\ | ||
330 | tmp+=tmpStride;\ | ||
331 | src+=srcStride;\ | ||
332 | }\ | ||
333 | tmp -= tmpStride*(h+5-2);\ | ||
334 | for(i=0; i<w; i++)\ | ||
335 | {\ | ||
336 | const int tmpB= tmp[-2*tmpStride] - pad;\ | ||
337 | const int tmpA= tmp[-1*tmpStride] - pad;\ | ||
338 | const int tmp0= tmp[0 *tmpStride] - pad;\ | ||
339 | const int tmp1= tmp[1 *tmpStride] - pad;\ | ||
340 | const int tmp2= tmp[2 *tmpStride] - pad;\ | ||
341 | const int tmp3= tmp[3 *tmpStride] - pad;\ | ||
342 | const int tmp4= tmp[4 *tmpStride] - pad;\ | ||
343 | const int tmp5= tmp[5 *tmpStride] - pad;\ | ||
344 | const int tmp6= tmp[6 *tmpStride] - pad;\ | ||
345 | const int tmp7= tmp[7 *tmpStride] - pad;\ | ||
346 | const int tmp8= tmp[8 *tmpStride] - pad;\ | ||
347 | const int tmp9= tmp[9 *tmpStride] - pad;\ | ||
348 | const int tmp10=tmp[10*tmpStride] - pad;\ | ||
349 | OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\ | ||
350 | OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\ | ||
351 | OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\ | ||
352 | OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\ | ||
353 | OP2(dst[4*dstStride], (tmp4+tmp5)*20 - (tmp3+tmp6)*5 + (tmp2+tmp7));\ | ||
354 | OP2(dst[5*dstStride], (tmp5+tmp6)*20 - (tmp4+tmp7)*5 + (tmp3+tmp8));\ | ||
355 | OP2(dst[6*dstStride], (tmp6+tmp7)*20 - (tmp5+tmp8)*5 + (tmp4+tmp9));\ | ||
356 | OP2(dst[7*dstStride], (tmp7+tmp8)*20 - (tmp6+tmp9)*5 + (tmp5+tmp10));\ | ||
357 | dst++;\ | ||
358 | tmp++;\ | ||
359 | }\ | ||
360 | }\ | ||
361 | \ | ||
362 | static void FUNC(OPNAME ## h264_qpel16_v_lowpass)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride)\ | ||
363 | {\ | ||
364 | FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst , src , dstStride, srcStride);\ | ||
365 | FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\ | ||
366 | src += 8*srcStride;\ | ||
367 | dst += 8*dstStride;\ | ||
368 | FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst , src , dstStride, srcStride);\ | ||
369 | FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\ | ||
370 | }\ | ||
371 | \ | ||
372 | static void FUNC(OPNAME ## h264_qpel16_h_lowpass)(uint8_t *dst, const uint8_t *restrict src, int dstStride, int srcStride)\ | ||
373 | {\ | ||
374 | FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst , src , dstStride, srcStride);\ | ||
375 | FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\ | ||
376 | src += 8*srcStride;\ | ||
377 | dst += 8*dstStride;\ | ||
378 | FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst , src , dstStride, srcStride);\ | ||
379 | FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\ | ||
380 | }\ | ||
381 | \ | ||
382 | static void FUNC(OPNAME ## h264_qpel16_hv_lowpass)(uint8_t *dst, pixeltmp *tmp, const uint8_t *restrict src, int dstStride, int tmpStride, int srcStride){\ | ||
383 | FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst , tmp , src , dstStride, tmpStride, srcStride);\ | ||
384 | FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\ | ||
385 | src += 8*srcStride;\ | ||
386 | dst += 8*dstStride;\ | ||
387 | FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst , tmp , src , dstStride, tmpStride, srcStride);\ | ||
388 | FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\ | ||
389 | }\ | ||
390 | |||
391 | #define H264_MC(OPNAME, SIZE) \ | ||
392 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc00)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
393 | {\ | ||
394 | FUNCC(OPNAME ## pixels ## SIZE)(dst, src, stride, SIZE);\ | ||
395 | }\ | ||
396 | \ | ||
397 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc10)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
398 | {\ | ||
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, half, stride, stride, SIZE*sizeof(pixel), SIZE);\ | ||
402 | }\ | ||
403 | \ | ||
404 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc20)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
405 | {\ | ||
406 | FUNC(OPNAME ## h264_qpel ## SIZE ## _h_lowpass)(dst, src, stride, stride);\ | ||
407 | }\ | ||
408 | \ | ||
409 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc30)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
410 | {\ | ||
411 | uint8_t half[SIZE*SIZE*sizeof(pixel)];\ | ||
412 | FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\ | ||
413 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src+sizeof(pixel), half, stride, stride, SIZE*sizeof(pixel), SIZE);\ | ||
414 | }\ | ||
415 | \ | ||
416 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
417 | {\ | ||
418 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
419 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
420 | uint8_t half[SIZE*SIZE*sizeof(pixel)];\ | ||
421 | FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
422 | FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\ | ||
423 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid, half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
424 | }\ | ||
425 | \ | ||
426 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc02)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
427 | {\ | ||
428 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
429 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
430 | FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
431 | FUNC(OPNAME ## h264_qpel ## SIZE ## _v_lowpass)(dst, full_mid, stride, SIZE*sizeof(pixel));\ | ||
432 | }\ | ||
433 | \ | ||
434 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
435 | {\ | ||
436 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
437 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
438 | uint8_t half[SIZE*SIZE*sizeof(pixel)];\ | ||
439 | FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
440 | FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\ | ||
441 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid+SIZE*sizeof(pixel), half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
442 | }\ | ||
443 | \ | ||
444 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
445 | {\ | ||
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, 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);\ | ||
454 | }\ | ||
455 | \ | ||
456 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
457 | {\ | ||
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, SIZE*sizeof(pixel), stride);\ | ||
463 | FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), 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);\ | ||
466 | }\ | ||
467 | \ | ||
468 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
469 | {\ | ||
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, 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);\ | ||
478 | }\ | ||
479 | \ | ||
480 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
481 | {\ | ||
482 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
483 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
484 | uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\ | ||
485 | uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\ | ||
486 | FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\ | ||
487 | FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
488 | FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\ | ||
489 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
490 | }\ | ||
491 | \ | ||
492 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc22)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
493 | {\ | ||
494 | pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
495 | FUNC(OPNAME ## h264_qpel ## SIZE ## _hv_lowpass)(dst, tmp, src, stride, SIZE*sizeof(pixel), stride);\ | ||
496 | }\ | ||
497 | \ | ||
498 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
499 | {\ | ||
500 | pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
501 | uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\ | ||
502 | uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\ | ||
503 | FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\ | ||
504 | FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\ | ||
505 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
506 | }\ | ||
507 | \ | ||
508 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
509 | {\ | ||
510 | pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
511 | uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\ | ||
512 | uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\ | ||
513 | FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\ | ||
514 | FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\ | ||
515 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
516 | }\ | ||
517 | \ | ||
518 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
519 | {\ | ||
520 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
521 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
522 | pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
523 | uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\ | ||
524 | uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\ | ||
525 | FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
526 | FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\ | ||
527 | FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\ | ||
528 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
529 | }\ | ||
530 | \ | ||
531 | static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc32)(uint8_t *dst, const uint8_t *restrict src, ptrdiff_t stride)\ | ||
532 | {\ | ||
533 | uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
534 | uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\ | ||
535 | pixeltmp tmp[SIZE*(SIZE+5)*sizeof(pixel)];\ | ||
536 | uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\ | ||
537 | uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\ | ||
538 | FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel), stride, SIZE + 5);\ | ||
539 | FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\ | ||
540 | FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\ | ||
541 | FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\ | ||
542 | }\ | ||
543 | |||
544 | #define op_avg(a, b) a = (((a)+CLIP(((b) + 16)>>5)+1)>>1) | ||
545 | //#define op_avg2(a, b) a = (((a)*w1+cm[((b) + 16)>>5]*w2 + o + 64)>>7) | ||
546 | #define op_put(a, b) a = CLIP(((b) + 16)>>5) | ||
547 | #define op2_avg(a, b) a = (((a)+CLIP(((b) + 512)>>10)+1)>>1) | ||
548 | #define op2_put(a, b) a = CLIP(((b) + 512)>>10) | ||
549 | |||
550 |
4/4✓ Branch 0 taken 621173246 times.
✓ Branch 1 taken 84723656 times.
✓ Branch 2 taken 86328684 times.
✓ Branch 3 taken 11784426 times.
|
1603231886 | H264_LOWPASS(put_ , op_put, op2_put) |
551 |
4/4✓ Branch 0 taken 9245568 times.
✓ Branch 1 taken 1133378 times.
✓ Branch 2 taken 1428060 times.
✓ Branch 3 taken 191260 times.
|
23980690 | H264_LOWPASS(avg_ , op_avg, op2_avg) |
552 | 389408 | H264_MC(put_, 2) | |
553 | 29188166 | H264_MC(put_, 4) | |
554 | 27349850 | H264_MC(put_, 8) | |
555 | 22403202 | H264_MC(put_, 16) | |
556 | 7888064 | H264_MC(avg_, 4) | |
557 | 2439084 | H264_MC(avg_, 8) | |
558 | 4778460 | H264_MC(avg_, 16) | |
559 | |||
560 | #undef op_avg | ||
561 | #undef op_put | ||
562 | #undef op2_avg | ||
563 | #undef op2_put | ||
564 |