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/attributes.h"
26 #include "libavutil/avassert.h"
27 #include "libavutil/intreadwrite.h"
28 #include "libavutil/x86/asm.h"
29 #include "libavutil/x86/cpu.h"
30 #include "libavutil/cpu.h"
31 #include "libavutil/pixdesc.h"
37 DECLARE_ASM_CONST(8, uint64_t, bF8)= 0xF8F8F8F8F8F8F8F8LL;
38 DECLARE_ASM_CONST(8, uint64_t, bFC)= 0xFCFCFCFCFCFCFCFCLL;
39 DECLARE_ASM_CONST(8, uint64_t, w10)= 0x0010001000100010LL;
40 DECLARE_ASM_CONST(8, uint64_t, w02)= 0x0002000200020002LL;
42 const DECLARE_ALIGNED(8, uint64_t, ff_dither4)[2] = {
44 0x0200020002000200LL,};
46 const DECLARE_ALIGNED(8, uint64_t, ff_dither8)[2] = {
48 0x0004000400040004LL,};
50 DECLARE_ASM_CONST(8, uint64_t, b16Mask)= 0x001F001F001F001FLL;
51 DECLARE_ASM_CONST(8, uint64_t, g16Mask)= 0x07E007E007E007E0LL;
52 DECLARE_ASM_CONST(8, uint64_t, r16Mask)= 0xF800F800F800F800LL;
53 DECLARE_ASM_CONST(8, uint64_t, b15Mask)= 0x001F001F001F001FLL;
54 DECLARE_ASM_CONST(8, uint64_t, g15Mask)= 0x03E003E003E003E0LL;
55 DECLARE_ASM_CONST(8, uint64_t, r15Mask)= 0x7C007C007C007C00LL;
57 DECLARE_ALIGNED(8, const uint64_t, ff_M24A) = 0x00FF0000FF0000FFLL;
58 DECLARE_ALIGNED(8, const uint64_t, ff_M24B) = 0xFF0000FF0000FF00LL;
59 DECLARE_ALIGNED(8, const uint64_t, ff_M24C) = 0x0000FF0000FF0000LL;
61 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2YOffset) = 0x1010101010101010ULL;
62 DECLARE_ALIGNED(8, const uint64_t, ff_bgr2UVOffset) = 0x8080808080808080ULL;
63 DECLARE_ALIGNED(8, const uint64_t, ff_w1111) = 0x0001000100010001ULL;
69 #define COMPILE_TEMPLATE_MMXEXT 0
70 #define RENAME(a) a ## _mmx
71 #include "swscale_template.c"
75 #if HAVE_MMXEXT_INLINE
77 #undef COMPILE_TEMPLATE_MMXEXT
78 #define COMPILE_TEMPLATE_MMXEXT 1
79 #define RENAME(a) a ## _mmxext
80 #include "swscale_template.c"
83 void ff_updateMMXDitherTables(SwsContext *c, int dstY, int lumBufIndex, int chrBufIndex,
84 int lastInLumBuf, int lastInChrBuf)
86 const int dstH= c->dstH;
87 const int flags= c->flags;
90 SwsPlane *lumPlane = &c->slice[c->numSlice-1].plane[0];
91 SwsPlane *chrUPlane = &c->slice[c->numSlice-1].plane[1];
92 SwsPlane *alpPlane = &c->slice[c->numSlice-1].plane[3];
94 int16_t **lumPixBuf= c->lumPixBuf;
95 int16_t **chrUPixBuf= c->chrUPixBuf;
96 int16_t **alpPixBuf= c->alpPixBuf;
98 int hasAlpha = c->alpPixBuf != NULL;
99 const int vLumBufSize= c->vLumBufSize;
100 const int vChrBufSize= c->vChrBufSize;
101 int32_t *vLumFilterPos= c->vLumFilterPos;
102 int32_t *vChrFilterPos= c->vChrFilterPos;
103 int16_t *vLumFilter= c->vLumFilter;
104 int16_t *vChrFilter= c->vChrFilter;
105 int32_t *lumMmxFilter= c->lumMmxFilter;
106 int32_t *chrMmxFilter= c->chrMmxFilter;
107 int32_t av_unused *alpMmxFilter= c->alpMmxFilter;
108 const int vLumFilterSize= c->vLumFilterSize;
109 const int vChrFilterSize= c->vChrFilterSize;
110 const int chrDstY= dstY>>c->chrDstVSubSample;
111 const int firstLumSrcY= vLumFilterPos[dstY]; //First line needed as input
112 const int firstChrSrcY= vChrFilterPos[chrDstY]; //First line needed as input
114 c->blueDither= ff_dither8[dstY&1];
115 if (c->dstFormat == AV_PIX_FMT_RGB555 || c->dstFormat == AV_PIX_FMT_BGR555)
116 c->greenDither= ff_dither8[dstY&1];
118 c->greenDither= ff_dither4[dstY&1];
119 c->redDither= ff_dither8[(dstY+1)&1];
120 if (dstY < dstH - 2) {
122 const int16_t **lumSrcPtr = (const int16_t **)(void*) lumPlane->line + firstLumSrcY - lumPlane->sliceY;
123 const int16_t **chrUSrcPtr = (const int16_t **)(void*) chrUPlane->line + firstChrSrcY - chrUPlane->sliceY;
124 const int16_t **alpSrcPtr = (CONFIG_SWSCALE_ALPHA && c->alpPixBuf) ? (const int16_t **)(void*) alpPlane->line + firstLumSrcY - alpPlane->sliceY : NULL;
126 const int16_t **lumSrcPtr= (const int16_t **)(void*) lumPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize;
127 const int16_t **chrUSrcPtr= (const int16_t **)(void*) chrUPixBuf + chrBufIndex + firstChrSrcY - lastInChrBuf + vChrBufSize;
128 const int16_t **alpSrcPtr= (CONFIG_SWSCALE_ALPHA && alpPixBuf) ? (const int16_t **)(void*) alpPixBuf + lumBufIndex + firstLumSrcY - lastInLumBuf + vLumBufSize : NULL;
131 if (firstLumSrcY < 0 || firstLumSrcY + vLumFilterSize > c->srcH) {
133 const int16_t **tmpY = (const int16_t **) lumPlane->tmp;
135 const int16_t **tmpY = (const int16_t **) lumPixBuf + 2 * vLumBufSize;
137 int neg = -firstLumSrcY, i, end = FFMIN(c->srcH - firstLumSrcY, vLumFilterSize);
138 for (i = 0; i < neg; i++)
139 tmpY[i] = lumSrcPtr[neg];
140 for ( ; i < end; i++)
141 tmpY[i] = lumSrcPtr[i];
142 for ( ; i < vLumFilterSize; i++)
148 const int16_t **tmpA = (const int16_t **) alpPlane->tmp;
150 const int16_t **tmpA = (const int16_t **) alpPixBuf + 2 * vLumBufSize;
152 for (i = 0; i < neg; i++)
153 tmpA[i] = alpSrcPtr[neg];
154 for ( ; i < end; i++)
155 tmpA[i] = alpSrcPtr[i];
156 for ( ; i < vLumFilterSize; i++)
157 tmpA[i] = tmpA[i - 1];
161 if (firstChrSrcY < 0 || firstChrSrcY + vChrFilterSize > c->chrSrcH) {
163 const int16_t **tmpU = (const int16_t **) chrUPlane->tmp;
165 const int16_t **tmpU = (const int16_t **) chrUPixBuf + 2 * vChrBufSize;
167 int neg = -firstChrSrcY, i, end = FFMIN(c->chrSrcH - firstChrSrcY, vChrFilterSize);
168 for (i = 0; i < neg; i++) {
169 tmpU[i] = chrUSrcPtr[neg];
171 for ( ; i < end; i++) {
172 tmpU[i] = chrUSrcPtr[i];
174 for ( ; i < vChrFilterSize; i++) {
175 tmpU[i] = tmpU[i - 1];
180 if (flags & SWS_ACCURATE_RND) {
181 int s= APCK_SIZE / 8;
182 for (i=0; i<vLumFilterSize; i+=2) {
183 *(const void**)&lumMmxFilter[s*i ]= lumSrcPtr[i ];
184 *(const void**)&lumMmxFilter[s*i+APCK_PTR2/4 ]= lumSrcPtr[i+(vLumFilterSize>1)];
185 lumMmxFilter[s*i+APCK_COEF/4 ]=
186 lumMmxFilter[s*i+APCK_COEF/4+1]= vLumFilter[dstY*vLumFilterSize + i ]
187 + (vLumFilterSize>1 ? vLumFilter[dstY*vLumFilterSize + i + 1]<<16 : 0);
188 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {
189 *(const void**)&alpMmxFilter[s*i ]= alpSrcPtr[i ];
190 *(const void**)&alpMmxFilter[s*i+APCK_PTR2/4 ]= alpSrcPtr[i+(vLumFilterSize>1)];
191 alpMmxFilter[s*i+APCK_COEF/4 ]=
192 alpMmxFilter[s*i+APCK_COEF/4+1]= lumMmxFilter[s*i+APCK_COEF/4 ];
195 for (i=0; i<vChrFilterSize; i+=2) {
196 *(const void**)&chrMmxFilter[s*i ]= chrUSrcPtr[i ];
197 *(const void**)&chrMmxFilter[s*i+APCK_PTR2/4 ]= chrUSrcPtr[i+(vChrFilterSize>1)];
198 chrMmxFilter[s*i+APCK_COEF/4 ]=
199 chrMmxFilter[s*i+APCK_COEF/4+1]= vChrFilter[chrDstY*vChrFilterSize + i ]
200 + (vChrFilterSize>1 ? vChrFilter[chrDstY*vChrFilterSize + i + 1]<<16 : 0);
203 for (i=0; i<vLumFilterSize; i++) {
204 *(const void**)&lumMmxFilter[4*i+0]= lumSrcPtr[i];
207 ((uint16_t)vLumFilter[dstY*vLumFilterSize + i])*0x10001U;
208 if (CONFIG_SWSCALE_ALPHA && hasAlpha) {
209 *(const void**)&alpMmxFilter[4*i+0]= alpSrcPtr[i];
211 alpMmxFilter[4*i+3]= lumMmxFilter[4*i+2];
214 for (i=0; i<vChrFilterSize; i++) {
215 *(const void**)&chrMmxFilter[4*i+0]= chrUSrcPtr[i];
218 ((uint16_t)vChrFilter[chrDstY*vChrFilterSize + i])*0x10001U;
225 static void yuv2yuvX_sse3(const int16_t *filter, int filterSize,
226 const int16_t **src, uint8_t *dest, int dstW,
227 const uint8_t *dither, int offset)
229 if(((uintptr_t)dest) & 15){
230 yuv2yuvX_mmxext(filter, filterSize, src, dest, dstW, dither, offset);
234 #define MAIN_FUNCTION \
235 "pxor %%xmm0, %%xmm0 \n\t" \
236 "punpcklbw %%xmm0, %%xmm3 \n\t" \
237 "movd %4, %%xmm1 \n\t" \
238 "punpcklwd %%xmm1, %%xmm1 \n\t" \
239 "punpckldq %%xmm1, %%xmm1 \n\t" \
240 "punpcklqdq %%xmm1, %%xmm1 \n\t" \
241 "psllw $3, %%xmm1 \n\t" \
242 "paddw %%xmm1, %%xmm3 \n\t" \
243 "psraw $4, %%xmm3 \n\t" \
244 "movdqa %%xmm3, %%xmm4 \n\t" \
245 "movdqa %%xmm3, %%xmm7 \n\t" \
246 "movl %3, %%ecx \n\t" \
247 "mov %0, %%"REG_d" \n\t"\
248 "mov (%%"REG_d"), %%"REG_S" \n\t"\
249 ".p2align 4 \n\t" /* FIXME Unroll? */\
251 "movddup 8(%%"REG_d"), %%xmm0 \n\t" /* filterCoeff */\
252 "movdqa (%%"REG_S", %%"REG_c", 2), %%xmm2 \n\t" /* srcData */\
253 "movdqa 16(%%"REG_S", %%"REG_c", 2), %%xmm5 \n\t" /* srcData */\
254 "add $16, %%"REG_d" \n\t"\
255 "mov (%%"REG_d"), %%"REG_S" \n\t"\
256 "test %%"REG_S", %%"REG_S" \n\t"\
257 "pmulhw %%xmm0, %%xmm2 \n\t"\
258 "pmulhw %%xmm0, %%xmm5 \n\t"\
259 "paddw %%xmm2, %%xmm3 \n\t"\
260 "paddw %%xmm5, %%xmm4 \n\t"\
262 "psraw $3, %%xmm3 \n\t"\
263 "psraw $3, %%xmm4 \n\t"\
264 "packuswb %%xmm4, %%xmm3 \n\t"\
265 "movntdq %%xmm3, (%1, %%"REG_c")\n\t"\
266 "add $16, %%"REG_c" \n\t"\
267 "cmp %2, %%"REG_c" \n\t"\
268 "movdqa %%xmm7, %%xmm3 \n\t" \
269 "movdqa %%xmm7, %%xmm4 \n\t" \
270 "mov %0, %%"REG_d" \n\t"\
271 "mov (%%"REG_d"), %%"REG_S" \n\t"\
276 "movq %5, %%xmm3 \n\t"
277 "movdqa %%xmm3, %%xmm4 \n\t"
278 "psrlq $24, %%xmm3 \n\t"
279 "psllq $40, %%xmm4 \n\t"
280 "por %%xmm4, %%xmm3 \n\t"
283 "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),
284 "m"(filterSize), "m"(((uint64_t *) dither)[0])
285 : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)
286 "%"REG_d, "%"REG_S, "%"REG_c
290 "movq %5, %%xmm3 \n\t"
293 "r" (dest-offset), "g" ((x86_reg)(dstW+offset)), "m" (offset),
294 "m"(filterSize), "m"(((uint64_t *) dither)[0])
295 : XMM_CLOBBERS("%xmm0" , "%xmm1" , "%xmm2" , "%xmm3" , "%xmm4" , "%xmm5" , "%xmm7" ,)
296 "%"REG_d, "%"REG_S, "%"REG_c
302 #endif /* HAVE_INLINE_ASM */
304 #define SCALE_FUNC(filter_n, from_bpc, to_bpc, opt) \
305 void ff_hscale ## from_bpc ## to ## to_bpc ## _ ## filter_n ## _ ## opt( \
306 SwsContext *c, int16_t *data, \
307 int dstW, const uint8_t *src, \
308 const int16_t *filter, \
309 const int32_t *filterPos, int filterSize)
311 #define SCALE_FUNCS(filter_n, opt) \
312 SCALE_FUNC(filter_n, 8, 15, opt); \
313 SCALE_FUNC(filter_n, 9, 15, opt); \
314 SCALE_FUNC(filter_n, 10, 15, opt); \
315 SCALE_FUNC(filter_n, 12, 15, opt); \
316 SCALE_FUNC(filter_n, 14, 15, opt); \
317 SCALE_FUNC(filter_n, 16, 15, opt); \
318 SCALE_FUNC(filter_n, 8, 19, opt); \
319 SCALE_FUNC(filter_n, 9, 19, opt); \
320 SCALE_FUNC(filter_n, 10, 19, opt); \
321 SCALE_FUNC(filter_n, 12, 19, opt); \
322 SCALE_FUNC(filter_n, 14, 19, opt); \
323 SCALE_FUNC(filter_n, 16, 19, opt)
325 #define SCALE_FUNCS_MMX(opt) \
326 SCALE_FUNCS(4, opt); \
327 SCALE_FUNCS(8, opt); \
330 #define SCALE_FUNCS_SSE(opt) \
331 SCALE_FUNCS(4, opt); \
332 SCALE_FUNCS(8, opt); \
333 SCALE_FUNCS(X4, opt); \
337 SCALE_FUNCS_MMX(mmx);
339 SCALE_FUNCS_SSE(sse2);
340 SCALE_FUNCS_SSE(ssse3);
341 SCALE_FUNCS_SSE(sse4);
343 #define VSCALEX_FUNC(size, opt) \
344 void ff_yuv2planeX_ ## size ## _ ## opt(const int16_t *filter, int filterSize, \
345 const int16_t **src, uint8_t *dest, int dstW, \
346 const uint8_t *dither, int offset)
347 #define VSCALEX_FUNCS(opt) \
348 VSCALEX_FUNC(8, opt); \
349 VSCALEX_FUNC(9, opt); \
350 VSCALEX_FUNC(10, opt)
353 VSCALEX_FUNCS(mmxext);
357 VSCALEX_FUNC(16, sse4);
360 #define VSCALE_FUNC(size, opt) \
361 void ff_yuv2plane1_ ## size ## _ ## opt(const int16_t *src, uint8_t *dst, int dstW, \
362 const uint8_t *dither, int offset)
363 #define VSCALE_FUNCS(opt1, opt2) \
364 VSCALE_FUNC(8, opt1); \
365 VSCALE_FUNC(9, opt2); \
366 VSCALE_FUNC(10, opt2); \
367 VSCALE_FUNC(16, opt1)
370 VSCALE_FUNCS(mmx, mmxext);
372 VSCALE_FUNCS(sse2, sse2);
373 VSCALE_FUNC(16, sse4);
374 VSCALE_FUNCS(avx, avx);
376 #define INPUT_Y_FUNC(fmt, opt) \
377 void ff_ ## fmt ## ToY_ ## opt(uint8_t *dst, const uint8_t *src, \
378 const uint8_t *unused1, const uint8_t *unused2, \
379 int w, uint32_t *unused)
380 #define INPUT_UV_FUNC(fmt, opt) \
381 void ff_ ## fmt ## ToUV_ ## opt(uint8_t *dstU, uint8_t *dstV, \
382 const uint8_t *unused0, \
383 const uint8_t *src1, \
384 const uint8_t *src2, \
385 int w, uint32_t *unused)
386 #define INPUT_FUNC(fmt, opt) \
387 INPUT_Y_FUNC(fmt, opt); \
388 INPUT_UV_FUNC(fmt, opt)
389 #define INPUT_FUNCS(opt) \
390 INPUT_FUNC(uyvy, opt); \
391 INPUT_FUNC(yuyv, opt); \
392 INPUT_UV_FUNC(nv12, opt); \
393 INPUT_UV_FUNC(nv21, opt); \
394 INPUT_FUNC(rgba, opt); \
395 INPUT_FUNC(bgra, opt); \
396 INPUT_FUNC(argb, opt); \
397 INPUT_FUNC(abgr, opt); \
398 INPUT_FUNC(rgb24, opt); \
399 INPUT_FUNC(bgr24, opt)
408 av_cold void ff_sws_init_swscale_x86(SwsContext *c)
410 int cpu_flags = av_get_cpu_flags();
413 if (INLINE_MMX(cpu_flags))
414 sws_init_swscale_mmx(c);
416 #if HAVE_MMXEXT_INLINE
417 if (INLINE_MMXEXT(cpu_flags))
418 sws_init_swscale_mmxext(c);
419 if (cpu_flags & AV_CPU_FLAG_SSE3){
420 if(c->use_mmx_vfilter && !(c->flags & SWS_ACCURATE_RND))
421 c->yuv2planeX = yuv2yuvX_sse3;
425 #define ASSIGN_SCALE_FUNC2(hscalefn, filtersize, opt1, opt2) do { \
426 if (c->srcBpc == 8) { \
427 hscalefn = c->dstBpc <= 14 ? ff_hscale8to15_ ## filtersize ## _ ## opt2 : \
428 ff_hscale8to19_ ## filtersize ## _ ## opt1; \
429 } else if (c->srcBpc == 9) { \
430 hscalefn = c->dstBpc <= 14 ? ff_hscale9to15_ ## filtersize ## _ ## opt2 : \
431 ff_hscale9to19_ ## filtersize ## _ ## opt1; \
432 } else if (c->srcBpc == 10) { \
433 hscalefn = c->dstBpc <= 14 ? ff_hscale10to15_ ## filtersize ## _ ## opt2 : \
434 ff_hscale10to19_ ## filtersize ## _ ## opt1; \
435 } else if (c->srcBpc == 12) { \
436 hscalefn = c->dstBpc <= 14 ? ff_hscale12to15_ ## filtersize ## _ ## opt2 : \
437 ff_hscale12to19_ ## filtersize ## _ ## opt1; \
438 } else if (c->srcBpc == 14 || ((c->srcFormat==AV_PIX_FMT_PAL8||isAnyRGB(c->srcFormat)) && av_pix_fmt_desc_get(c->srcFormat)->comp[0].depth_minus1<15)) { \
439 hscalefn = c->dstBpc <= 14 ? ff_hscale14to15_ ## filtersize ## _ ## opt2 : \
440 ff_hscale14to19_ ## filtersize ## _ ## opt1; \
441 } else { /* c->srcBpc == 16 */ \
442 av_assert0(c->srcBpc == 16);\
443 hscalefn = c->dstBpc <= 14 ? ff_hscale16to15_ ## filtersize ## _ ## opt2 : \
444 ff_hscale16to19_ ## filtersize ## _ ## opt1; \
447 #define ASSIGN_MMX_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
448 switch (filtersize) { \
449 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
450 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
451 default: ASSIGN_SCALE_FUNC2(hscalefn, X, opt1, opt2); break; \
453 #define ASSIGN_VSCALEX_FUNC(vscalefn, opt, do_16_case, condition_8bit) \
455 case 16: do_16_case; break; \
456 case 10: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_10_ ## opt; break; \
457 case 9: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2planeX_9_ ## opt; break; \
458 case 8: if ((condition_8bit) && !c->use_mmx_vfilter) vscalefn = ff_yuv2planeX_8_ ## opt; break; \
460 #define ASSIGN_VSCALE_FUNC(vscalefn, opt1, opt2, opt2chk) \
462 case 16: if (!isBE(c->dstFormat)) vscalefn = ff_yuv2plane1_16_ ## opt1; break; \
463 case 10: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_10_ ## opt2; break; \
464 case 9: if (!isBE(c->dstFormat) && opt2chk) vscalefn = ff_yuv2plane1_9_ ## opt2; break; \
465 case 8: vscalefn = ff_yuv2plane1_8_ ## opt1; break; \
466 default: av_assert0(c->dstBpc>8); \
468 #define case_rgb(x, X, opt) \
469 case AV_PIX_FMT_ ## X: \
470 c->lumToYV12 = ff_ ## x ## ToY_ ## opt; \
471 if (!c->chrSrcHSubSample) \
472 c->chrToYV12 = ff_ ## x ## ToUV_ ## opt; \
475 if (EXTERNAL_MMX(cpu_flags)) {
476 ASSIGN_MMX_SCALE_FUNC(c->hyScale, c->hLumFilterSize, mmx, mmx);
477 ASSIGN_MMX_SCALE_FUNC(c->hcScale, c->hChrFilterSize, mmx, mmx);
478 ASSIGN_VSCALE_FUNC(c->yuv2plane1, mmx, mmxext, cpu_flags & AV_CPU_FLAG_MMXEXT);
480 switch (c->srcFormat) {
482 c->lumToYV12 = ff_yuyvToY_mmx;
484 c->alpToYV12 = ff_uyvyToY_mmx;
486 case AV_PIX_FMT_YUYV422:
487 c->lumToYV12 = ff_yuyvToY_mmx;
488 c->chrToYV12 = ff_yuyvToUV_mmx;
490 case AV_PIX_FMT_UYVY422:
491 c->lumToYV12 = ff_uyvyToY_mmx;
492 c->chrToYV12 = ff_uyvyToUV_mmx;
494 case AV_PIX_FMT_NV12:
495 c->chrToYV12 = ff_nv12ToUV_mmx;
497 case AV_PIX_FMT_NV21:
498 c->chrToYV12 = ff_nv21ToUV_mmx;
500 case_rgb(rgb24, RGB24, mmx);
501 case_rgb(bgr24, BGR24, mmx);
502 case_rgb(bgra, BGRA, mmx);
503 case_rgb(rgba, RGBA, mmx);
504 case_rgb(abgr, ABGR, mmx);
505 case_rgb(argb, ARGB, mmx);
510 if (EXTERNAL_MMXEXT(cpu_flags)) {
511 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, mmxext, , 1);
513 #endif /* ARCH_X86_32 */
514 #define ASSIGN_SSE_SCALE_FUNC(hscalefn, filtersize, opt1, opt2) \
515 switch (filtersize) { \
516 case 4: ASSIGN_SCALE_FUNC2(hscalefn, 4, opt1, opt2); break; \
517 case 8: ASSIGN_SCALE_FUNC2(hscalefn, 8, opt1, opt2); break; \
518 default: if (filtersize & 4) ASSIGN_SCALE_FUNC2(hscalefn, X4, opt1, opt2); \
519 else ASSIGN_SCALE_FUNC2(hscalefn, X8, opt1, opt2); \
522 if (EXTERNAL_SSE2(cpu_flags)) {
523 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse2, sse2);
524 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse2, sse2);
525 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse2, ,
526 HAVE_ALIGNED_STACK || ARCH_X86_64);
527 ASSIGN_VSCALE_FUNC(c->yuv2plane1, sse2, sse2, 1);
529 switch (c->srcFormat) {
531 c->lumToYV12 = ff_yuyvToY_sse2;
533 c->alpToYV12 = ff_uyvyToY_sse2;
535 case AV_PIX_FMT_YUYV422:
536 c->lumToYV12 = ff_yuyvToY_sse2;
537 c->chrToYV12 = ff_yuyvToUV_sse2;
539 case AV_PIX_FMT_UYVY422:
540 c->lumToYV12 = ff_uyvyToY_sse2;
541 c->chrToYV12 = ff_uyvyToUV_sse2;
543 case AV_PIX_FMT_NV12:
544 c->chrToYV12 = ff_nv12ToUV_sse2;
546 case AV_PIX_FMT_NV21:
547 c->chrToYV12 = ff_nv21ToUV_sse2;
549 case_rgb(rgb24, RGB24, sse2);
550 case_rgb(bgr24, BGR24, sse2);
551 case_rgb(bgra, BGRA, sse2);
552 case_rgb(rgba, RGBA, sse2);
553 case_rgb(abgr, ABGR, sse2);
554 case_rgb(argb, ARGB, sse2);
559 if (EXTERNAL_SSSE3(cpu_flags)) {
560 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, ssse3, ssse3);
561 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, ssse3, ssse3);
562 switch (c->srcFormat) {
563 case_rgb(rgb24, RGB24, ssse3);
564 case_rgb(bgr24, BGR24, ssse3);
569 if (EXTERNAL_SSE4(cpu_flags)) {
570 /* Xto15 don't need special sse4 functions */
571 ASSIGN_SSE_SCALE_FUNC(c->hyScale, c->hLumFilterSize, sse4, ssse3);
572 ASSIGN_SSE_SCALE_FUNC(c->hcScale, c->hChrFilterSize, sse4, ssse3);
573 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, sse4,
574 if (!isBE(c->dstFormat)) c->yuv2planeX = ff_yuv2planeX_16_sse4,
575 HAVE_ALIGNED_STACK || ARCH_X86_64);
576 if (c->dstBpc == 16 && !isBE(c->dstFormat))
577 c->yuv2plane1 = ff_yuv2plane1_16_sse4;
580 if (EXTERNAL_AVX(cpu_flags)) {
581 ASSIGN_VSCALEX_FUNC(c->yuv2planeX, avx, ,
582 HAVE_ALIGNED_STACK || ARCH_X86_64);
583 ASSIGN_VSCALE_FUNC(c->yuv2plane1, avx, avx, 1);
585 switch (c->srcFormat) {
586 case AV_PIX_FMT_YUYV422:
587 c->chrToYV12 = ff_yuyvToUV_avx;
589 case AV_PIX_FMT_UYVY422:
590 c->chrToYV12 = ff_uyvyToUV_avx;
592 case AV_PIX_FMT_NV12:
593 c->chrToYV12 = ff_nv12ToUV_avx;
595 case AV_PIX_FMT_NV21:
596 c->chrToYV12 = ff_nv21ToUV_avx;
598 case_rgb(rgb24, RGB24, avx);
599 case_rgb(bgr24, BGR24, avx);
600 case_rgb(bgra, BGRA, avx);
601 case_rgb(rgba, RGBA, avx);
602 case_rgb(abgr, ABGR, avx);
603 case_rgb(argb, ARGB, avx);