2 * Copyright (C) 2001-2011 Michael Niedermayer <michaelni@gmx.at>
4 * This file is part of FFmpeg.
6 * FFmpeg is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU Lesser General Public
8 * License as published by the Free Software Foundation; either
9 * version 2.1 of the License, or (at your option) any later version.
11 * FFmpeg 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 GNU
14 * Lesser General Public License for more details.
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with FFmpeg; if not, write to the Free Software
18 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
23 #include "libswscale/swscale.h"
24 #include "libswscale/swscale_internal.h"
25 #include "libavutil/avassert.h"
26 #include "libavutil/intreadwrite.h"
27 #include "libavutil/x86_cpu.h"
28 #include "libavutil/cpu.h"
29 #include "libavutil/pixdesc.h"
33 DECLARE_ASM_CONST(8, uint64_t, bF8)= 0xF8F8F8F8F8F8F8F8LL;
34 DECLARE_ASM_CONST(8, uint64_t, bFC)= 0xFCFCFCFCFCFCFCFCLL;
35 DECLARE_ASM_CONST(8, uint64_t, w10)= 0x0010001000100010LL;
36 DECLARE_ASM_CONST(8, uint64_t, w02)= 0x0002000200020002LL;
38 const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {
40 0x0200020002000200LL,};
42 const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {
44 0x0004000400040004LL,};
46 DECLARE_ASM_CONST(8, uint64_t, b16Mask)= 0x001F001F001F001FLL;
47 DECLARE_ASM_CONST(8, uint64_t, g16Mask)= 0x07E007E007E007E0LL;
48 DECLARE_ASM_CONST(8, uint64_t, r16Mask)= 0xF800F800F800F800LL;
49 DECLARE_ASM_CONST(8, uint64_t, b15Mask)= 0x001F001F001F001FLL;
50 DECLARE_ASM_CONST(8, uint64_t, g15Mask)= 0x03E003E003E003E0LL;
51 DECLARE_ASM_CONST(8, uint64_t, r15Mask)= 0x7C007C007C007C00LL;
53 DECLARE_ALIGNED(8, const uint64_t, ff_M24A) = 0x00FF0000FF0000FFLL;
54 DECLARE_ALIGNED(8, const uint64_t, ff_M24B) = 0xFF0000FF0000FF00LL;
55 DECLARE_ALIGNED(8, const uint64_t, ff_M24C) = 0x0000FF0000FF0000LL;
58 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff) = 0x000000210041000DULL;
59 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff) = 0x0000FFEEFFDC0038ULL;
60 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff) = 0x00000038FFD2FFF8ULL;
62 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YCoeff) = 0x000020E540830C8BULL;
63 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UCoeff) = 0x0000ED0FDAC23831ULL;
64 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2VCoeff) = 0x00003831D0E6F6EAULL;
65 #endif /* FAST_BGR2YV12 */
66 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YOffset) = 0x1010101010101010ULL;
67 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;
68 DECLARE_ALIGNED(8, const uint64_t, ff_w1111) = 0x0001000100010001ULL;
74 #define COMPILE_TEMPLATE_MMX2 0
75 #define RENAME(a) a ## _MMX
76 #include "swscale_template.c"
82 #undef COMPILE_TEMPLATE_MMX2
83 #define COMPILE_TEMPLATE_MMX2 1
84 #define RENAME(a) a ## _MMX2
85 #include "swscale_template.c"
88 void updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
89 int lastInLumBuf, int lastInChrBuf)
91 const int dstH= c->dstH;
92 const int flags= c->flags;
93 int16_t **lumPixBuf= c->lumPixBuf;
94 int16_t **chrUPixBuf= c->chrUPixBuf;
95 int16_t **alpPixBuf= c->alpPixBuf;
96 const int vLumBufSize= c->vLumBufSize;
97 const int vChrBufSize= c->vChrBufSize;
98 int32_t *vLumFilterPos= c->vLumFilterPos;
99 int32_t *vChrFilterPos= c->vChrFilterPos;
100 int16_t *vLumFilter= c->vLumFilter;
101 int16_t *vChrFilter= c->vChrFilter;
102 int32_t *lumMmxFilter= c->lumMmxFilter;
103 int32_t *chrMmxFilter= c->chrMmxFilter;
104 int32_t av_unused *alpMmxFilter= c->alpMmxFilter;
105 const int vLumFilterSize= c->vLumFilterSize;
106 const int vChrFilterSize= c->vChrFilterSize;
107 const int chrDstY= dstY>>c->chrDstVSubSample;
108 const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
109 const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
111 c->blueDither= ff_dither8[dstY&1];
112 if (c->dstFormat == PIX_FMT_RGB555 || c->dstFormat == PIX_FMT_BGR555)
113 c->greenDither= ff_dither8[dstY&1];
115 c->greenDither= ff_dither4[dstY&1];
116 c->redDither= ff_dither8[(dstY+1)&1];
117 if (dstY < dstH - 2) {
118 const int16_t **lumSrcPtr= (const int16_t **)(void*) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
119 const int16_t **chrUSrcPtr= (const int16_t **)(void*) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
120 const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **)(void*) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
123 if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
124 const int16_t **tmpY = (const int16_t **) lumPixBuf + 2 * vLumBufSize;
125 int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
126 for (i = 0; i < neg; i++)
127 tmpY[i] = lumSrcPtr[neg];
128 for ( ; i < end; i++)
129 tmpY[i] = lumSrcPtr[i];
130 for ( ; i < vLumFilterSize; i++)
135 const int16_t **tmpA = (const int16_t **) alpPixBuf + 2 * vLumBufSize;
136 for (i = 0; i < neg; i++)
137 tmpA[i] = alpSrcPtr[neg];
138 for ( ; i < end; i++)
139 tmpA[i] = alpSrcPtr[i];
140 for ( ; i < vLumFilterSize; i++)
141 tmpA[i] = tmpA[i - 1];
145 if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
146 const int16_t **tmpU = (const int16_t **) chrUPixBuf + 2 * vChrBufSize;
147 int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
148 for (i = 0; i < neg; i++) {
149 tmpU[i] = chrUSrcPtr[neg];
151 for ( ; i < end; i++) {
152 tmpU[i] = chrUSrcPtr[i];
154 for ( ; i < vChrFilterSize; i++) {
155 tmpU[i] = tmpU[i - 1];
160 if (flags & SWS_ACCURATE_RND) {
161 int s= APCK_SIZE / 8;
162 for (i=0; i<vLumFilterSize; i+=2) {
163 *(const void**)&lumMmxFilter[s*i ]= lumSrcPtr[i ];
164 *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4 ]= lumSrcPtr[i+(vLumFilterSize>1)];
165 lumMmxFilter[s*i+APCK_COEF/4 ]=
166 lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i ]
167 + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1]<<16 : 0);
168 if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
169 *(const void**)&alpMmxFilter[s*i ]= alpSrcPtr[i ];
170 *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4 ]= alpSrcPtr[i+(vLumFilterSize>1)];
171 alpMmxFilter[s*i+APCK_COEF/4 ]=
172 alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4 ];
175 for (i=0; i<vChrFilterSize; i+=2) {
176 *(const void**)&chrMmxFilter[s*i ]= chrUSrcPtr[i ];
177 *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4 ]= chrUSrcPtr[i+(vChrFilterSize>1)];
178 chrMmxFilter[s*i+APCK_COEF/4 ]=
179 chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i ]
180 + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1]<<16 : 0);
183 for (i=0; i<vLumFilterSize; i++) {
184 *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];
187 ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001;
188 if (CONFIG_SWSCALE_ALPHA && alpPixBuf) {
189 *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];
191 alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];
194 for (i=0; i<vChrFilterSize; i++) {
195 *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];
198 ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001;
205 static void yuv2yuvX_sse3(const int16_t *filter, int filterSize,
206 const int16_t **src, uint8_t *dest, int dstW,
207 const uint8_t *dither, int offset)
209 if(((int)dest) & 15){
210 return yuv2yuvX_MMX2(filter, filterSize, src, dest, dstW, dither, offset);
213 __asm__ volatile("movq (%0), %%xmm3\n\t"
214 "movdqa %%xmm3, %%xmm4\n\t"
215 "psrlq $24, %%xmm3\n\t"
216 "psllq $40, %%xmm4\n\t"
217 "por %%xmm4, %%xmm3\n\t"
221 __asm__ volatile("movq (%0), %%xmm3\n\t"
226 "pxor %%xmm0, %%xmm0\n\t"
227 "punpcklbw %%xmm0, %%xmm3\n\t"
228 "psraw $4, %%xmm3\n\t"
229 "movdqa %%xmm3, %%xmm4\n\t"
230 "movdqa %%xmm3, %%xmm7\n\t"
232 "mov %0, %%"REG_d" \n\t"\
233 "mov (%%"REG_d"), %%"REG_S" \n\t"\
234 ".p2align 4 \n\t" /* FIXME Unroll? */\
236 "movddup 8(%%"REG_d"), %%xmm0 \n\t" /* filterCoeff */\
237 "movdqa (%%"REG_S", %%"REG_c", 2), %%xmm2 \n\t" /* srcData */\
238 "movdqa 16(%%"REG_S", %%"REG_c", 2), %%xmm5 \n\t" /* srcData */\
239 "add $16, %%"REG_d" \n\t"\
240 "mov (%%"REG_d"), %%"REG_S" \n\t"\
241 "test %%"REG_S", %%"REG_S" \n\t"\
242 "pmulhw %%xmm0, %%xmm2 \n\t"\
243 "pmulhw %%xmm0, %%xmm5 \n\t"\
244 "paddw %%xmm2, %%xmm3 \n\t"\
245 "paddw %%xmm5, %%xmm4 \n\t"\
247 "psraw $3, %%xmm3 \n\t"\
248 "psraw $3, %%xmm4 \n\t"\
249 "packuswb %%xmm4, %%xmm3 \n\t"
250 "movntdq %%xmm3, (%1, %%"REG_c")\n\t"
251 "add $16, %%"REG_c" \n\t"\
252 "cmp %2, %%"REG_c" \n\t"\
253 "movdqa %%xmm7, %%xmm3\n\t"
254 "movdqa %%xmm7, %%xmm4\n\t"
255 "mov %0, %%"REG_d" \n\t"\
256 "mov (%%"REG_d"), %%"REG_S" \n\t"\
259 "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset)
260 : "%"REG_d, "%"REG_S, "%"REG_c
265 #define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
266 extern void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
267 SwsContext *c, int16_t *data, \
268 int dstW, const uint8_t *src, \
269 const int16_t *filter, \
270 const int32_t *filterPos, int filterSize)
272 #define SCALE_FUNCS(filter_n, opt) \
273 SCALE_FUNC(filter_n, 8, 15, opt); \
274 SCALE_FUNC(filter_n, 9, 15, opt); \
275 SCALE_FUNC(filter_n, 10, 15, opt); \
276 SCALE_FUNC(filter_n, 12, 15, opt); \
277 SCALE_FUNC(filter_n, 14, 15, opt); \
278 SCALE_FUNC(filter_n, 16, 15, opt); \
279 SCALE_FUNC(filter_n, 8, 19, opt); \
280 SCALE_FUNC(filter_n, 9, 19, opt); \
281 SCALE_FUNC(filter_n, 10, 19, opt); \
282 SCALE_FUNC(filter_n, 12, 19, opt); \
283 SCALE_FUNC(filter_n, 14, 19, opt); \
284 SCALE_FUNC(filter_n, 16, 19, opt)
286 #define SCALE_FUNCS_MMX(opt) \
287 SCALE_FUNCS(4, opt); \
288 SCALE_FUNCS(8, opt); \
291 #define SCALE_FUNCS_SSE(opt) \
292 SCALE_FUNCS(4, opt); \
293 SCALE_FUNCS(8, opt); \
294 SCALE_FUNCS(X4, opt); \
298 SCALE_FUNCS_MMX(mmx);
300 SCALE_FUNCS_SSE(sse2);
301 SCALE_FUNCS_SSE(ssse3);
302 SCALE_FUNCS_SSE(sse4);
304 #define VSCALEX_FUNC(size, opt) \
305 extern void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \
306 const int16_t **src, uint8_t *dest, int dstW, \
307 const uint8_t *dither, int offset)
308 #define VSCALEX_FUNCS(opt) \
309 VSCALEX_FUNC(8, opt); \
310 VSCALEX_FUNC(9, opt); \
311 VSCALEX_FUNC(10, opt)
318 VSCALEX_FUNC(16, sse4);
321 #define VSCALE_FUNC(size, opt) \
322 extern void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \
323 const uint8_t *dither, int offset)
324 #define VSCALE_FUNCS(opt1, opt2) \
325 VSCALE_FUNC(8, opt1); \
326 VSCALE_FUNC(9, opt2); \
327 VSCALE_FUNC(10, opt2); \
328 VSCALE_FUNC(16, opt1)
331 VSCALE_FUNCS(mmx, mmx2);
333 VSCALE_FUNCS(sse2, sse2);
334 VSCALE_FUNC(16, sse4);
335 VSCALE_FUNCS(avx, avx);
337 #define INPUT_Y_FUNC(fmt, opt) \
338 extern void ff_ ## fmt ## ToY_ ## opt(uint8_t *dst, const uint8_t *src, \
339 int w, uint32_t *unused)
340 #define INPUT_UV_FUNC(fmt, opt) \
341 extern void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \
342 const uint8_t *src, const uint8_t *unused1, \
343 int w, uint32_t *unused2)
344 #define INPUT_FUNC(fmt, opt) \
345 INPUT_Y_FUNC(fmt, opt); \
346 INPUT_UV_FUNC(fmt, opt)
347 #define INPUT_FUNCS(opt) \
348 INPUT_FUNC(uyvy, opt); \
349 INPUT_FUNC(yuyv, opt); \
350 INPUT_UV_FUNC(nv12, opt); \
351 INPUT_UV_FUNC(nv21, opt); \
352 INPUT_FUNC(rgba, opt); \
353 INPUT_FUNC(bgra, opt); \
354 INPUT_FUNC(argb, opt); \
355 INPUT_FUNC(abgr, opt); \
356 INPUT_FUNC(rgb24, opt); \
357 INPUT_FUNC(bgr24, opt)
366 void ff_sws_init_swScale_mmx(SwsContext *c)
368 int cpu_flags = av_get_cpu_flags();
370 if (cpu_flags & AV_CPU_FLAG_MMX)
371 sws_init_swScale_MMX(c);
373 if (cpu_flags & AV_CPU_FLAG_MMX2)
374 sws_init_swScale_MMX2(c);
375 if (cpu_flags & AV_CPU_FLAG_SSE3){
376 if(c->use_mmx_vfilter && !(c->flags & SWS_ACCURATE_RND))
377 c->yuv2planeX = yuv2yuvX_sse3;
382 #define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \
383 if (c->srcBpc == 8) { \
384 hscalefn = c->dstBpc <= 14 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \
385 ff_hscale8to19_ ## filtersize ## _ ## opt1; \
386 } else if (c->srcBpc == 9) { \
387 hscalefn = c->dstBpc <= 14 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \
388 ff_hscale9to19_ ## filtersize ## _ ## opt1; \
389 } else if (c->srcBpc == 10) { \
390 hscalefn = c->dstBpc <= 14 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
391 ff_hscale10to19_ ## filtersize ## _ ## opt1; \
392 } else if (c->srcBpc == 12) { \
393 hscalefn = c->dstBpc <= 14 ? ff_hscale12to15_ ## filtersize ## _ ## opt2 : \
394 ff_hscale12to19_ ## filtersize ## _ ## opt1; \
395 } else if (c->srcBpc == 14 || ((c->srcFormat==PIX_FMT_PAL8||isAnyRGB(c->srcFormat)) && av_pix_fmt_descriptors[c->srcFormat].comp[0].depth_minus1<15)) { \
396 hscalefn = c->dstBpc <= 14 ? ff_hscale14to15_ ## filtersize ## _ ## opt2 : \
397 ff_hscale14to19_ ## filtersize ## _ ## opt1; \
398 } else { /* c->srcBpc == 16 */ \
399 av_assert0(c->srcBpc == 16);\
400 hscalefn = c->dstBpc <= 14 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
401 ff_hscale16to19_ ## filtersize ## _ ## opt1; \
404 #define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
405 switch (filtersize) { \
406 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
407 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
408 default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \
410 #define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \
412 case 16: do_16_case; break; \
413 case 10: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_10_ ## opt; break; \
414 case 9: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_ ## opt; break; \
415 default: if (condition_8bit) /*vscalefn = ff_yuv2planeX_8_ ## opt;*/ break; \
417 #define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
419 case 16: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2plane1_16_ ## opt1; break; \
420 case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \
421 case 9: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_ ## opt2; break; \
422 case 8: vscalefn = ff_yuv2plane1_8_ ## opt1; break; \
423 default: av_assert0(c->dstBpc>8); \
425 #define case_rgb(x, X, opt) \
426 case PIX_FMT_ ## X: \
427 c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \
428 if (!c->chrSrcHSubSample) \
429 c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \
432 if (cpu_flags & AV_CPU_FLAG_MMX) {
433 ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);
434 ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);
435 ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmx2, cpu_flags & AV_CPU_FLAG_MMX2);
437 switch (c->srcFormat) {
439 c->lumToYV12 = ff_yuyvToY_mmx;
441 c->alpToYV12 = ff_uyvyToY_mmx;
443 case PIX_FMT_YUYV422:
444 c->lumToYV12 = ff_yuyvToY_mmx;
445 c->chrToYV12 = ff_yuyvToUV_mmx;
447 case PIX_FMT_UYVY422:
448 c->lumToYV12 = ff_uyvyToY_mmx;
449 c->chrToYV12 = ff_uyvyToUV_mmx;
452 c->chrToYV12 = ff_nv12ToUV_mmx;
455 c->chrToYV12 = ff_nv21ToUV_mmx;
457 case_rgb(rgb24, RGB24, mmx);
458 case_rgb(bgr24, BGR24, mmx);
459 case_rgb(bgra, BGRA, mmx);
460 case_rgb(rgba, RGBA, mmx);
461 case_rgb(abgr, ABGR, mmx);
462 case_rgb(argb, ARGB, mmx);
467 if (cpu_flags & AV_CPU_FLAG_MMX2) {
468 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmx2, , 1);
471 #define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
472 switch (filtersize) { \
473 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
474 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
475 default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \
476 else ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \
479 if (cpu_flags & AV_CPU_FLAG_SSE2) {
480 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);
481 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);
482 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,
483 HAVE_ALIGNED_STACK || ARCH_X86_64);
484 ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);
486 switch (c->srcFormat) {
488 c->lumToYV12 = ff_yuyvToY_sse2;
490 c->alpToYV12 = ff_uyvyToY_sse2;
492 case PIX_FMT_YUYV422:
493 c->lumToYV12 = ff_yuyvToY_sse2;
494 c->chrToYV12 = ff_yuyvToUV_sse2;
496 case PIX_FMT_UYVY422:
497 c->lumToYV12 = ff_uyvyToY_sse2;
498 c->chrToYV12 = ff_uyvyToUV_sse2;
501 c->chrToYV12 = ff_nv12ToUV_sse2;
504 c->chrToYV12 = ff_nv21ToUV_sse2;
506 case_rgb(rgb24, RGB24, sse2);
507 case_rgb(bgr24, BGR24, sse2);
508 case_rgb(bgra, BGRA, sse2);
509 case_rgb(rgba, RGBA, sse2);
510 case_rgb(abgr, ABGR, sse2);
511 case_rgb(argb, ARGB, sse2);
516 if (cpu_flags & AV_CPU_FLAG_SSSE3) {
517 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);
518 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);
519 switch (c->srcFormat) {
520 case_rgb(rgb24, RGB24, ssse3);
521 case_rgb(bgr24, BGR24, ssse3);
526 if (cpu_flags & AV_CPU_FLAG_SSE4) {
527 /* Xto15 don't need special sse4 functions */
528 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);
529 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);
530 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,
531 if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,
532 HAVE_ALIGNED_STACK || ARCH_X86_64);
533 if (c->dstBpc == 16 && !isBE(c->dstFormat))
534 c->yuv2plane1 = ff_yuv2plane1_16_sse4;
537 if (HAVE_AVX && cpu_flags & AV_CPU_FLAG_AVX) {
538 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,
539 HAVE_ALIGNED_STACK || ARCH_X86_64);
540 ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);
542 switch (c->srcFormat) {
543 case PIX_FMT_YUYV422:
544 c->chrToYV12 = ff_yuyvToUV_avx;
546 case PIX_FMT_UYVY422:
547 c->chrToYV12 = ff_uyvyToUV_avx;
550 c->chrToYV12 = ff_nv12ToUV_avx;
553 c->chrToYV12 = ff_nv21ToUV_avx;
555 case_rgb(rgb24, RGB24, avx);
556 case_rgb(bgr24, BGR24, avx);
557 case_rgb(bgra, BGRA, avx);
558 case_rgb(rgba, RGBA, avx);
559 case_rgb(abgr, ABGR, avx);
560 case_rgb(argb, ARGB, avx);