Parent Directory | Revision Log
Revision 434 - (view) (download)
1 : | chl | 434 | ;/***************************************************************************** |
2 : | ; * | ||
3 : | ; * XVID MPEG-4 VIDEO CODEC | ||
4 : | ; * mmx yuv planar to rgb conversion | ||
5 : | ; * | ||
6 : | ; * Copyright (C) 2001 - Michael Militzer <isibaar@xvid.org> | ||
7 : | ; * | ||
8 : | ; * This program is an implementation of a part of one or more MPEG-4 | ||
9 : | ; * Video tools as specified in ISO/IEC 14496-2 standard. Those intending | ||
10 : | ; * to use this software module in hardware or software products are | ||
11 : | ; * advised that its use may infringe existing patents or copyrights, and | ||
12 : | ; * any such use would be at such party's own risk. The original | ||
13 : | ; * developer of this software module and his/her company, and subsequent | ||
14 : | ; * editors and their companies, will have no liability for use of this | ||
15 : | ; * software or modifications or derivatives thereof. | ||
16 : | ; * | ||
17 : | ; * This program is free software; you can redistribute it and/or modify | ||
18 : | ; * it under the terms of the GNU General Public License as published by | ||
19 : | ; * the Free Software Foundation; either version 2 of the License, or | ||
20 : | ; * (at your option) any later version. | ||
21 : | ; * | ||
22 : | ; * This program is distributed in the hope that it will be useful, | ||
23 : | ; * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
24 : | ; * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
25 : | ; * GNU General Public License for more details. | ||
26 : | ; * | ||
27 : | ; * You should have received a copy of the GNU General Public License | ||
28 : | ; * along with this program; if not, write to the Free Software | ||
29 : | ; * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
30 : | ; * | ||
31 : | ; * $Id: yv12_to_rgb32_mmx.asm,v 1.3 2002-09-06 17:48:58 chl Exp $ | ||
32 : | ; * | ||
33 : | ; ****************************************************************************/ | ||
34 : | edgomez | 332 | ; |
35 : | ;------------------------------------------------------------------------------ | ||
36 : | ; NB: n contrary to the c implementation this code does the conversion | ||
37 : | ; using direct calculations. Input data width must be a multiple of 8 | ||
38 : | ; and height must be even. | ||
39 : | ; This implementation is less precise than the c version but is | ||
40 : | ; more than twice as fast :-) | ||
41 : | ;------------------------------------------------------------------------------ | ||
42 : | |||
43 : | BITS 32 | ||
44 : | |||
45 : | |||
46 : | %macro cglobal 1 | ||
47 : | %ifdef PREFIX | ||
48 : | global _%1 | ||
49 : | %define %1 _%1 | ||
50 : | %else | ||
51 : | global %1 | ||
52 : | %endif | ||
53 : | %endmacro | ||
54 : | |||
55 : | |||
56 : | %define SCALEBITS 6 | ||
57 : | |||
58 : | |||
59 : | ALIGN 16 | ||
60 : | |||
61 : | SECTION .data | ||
62 : | |||
63 : | Y_SUB dw 16, 16, 16, 16 | ||
64 : | U_SUB dw 128, 128, 128, 128 | ||
65 : | V_SUB dw 128, 128, 128, 128 | ||
66 : | |||
67 : | Y_MUL dw 74, 74, 74, 74 | ||
68 : | |||
69 : | UG_MUL dw 25, 25, 25, 25 | ||
70 : | VG_MUL dw 52, 52, 52, 52 | ||
71 : | |||
72 : | UB_MUL dw 129, 129, 129, 129 | ||
73 : | VR_MUL dw 102, 102, 102, 102 | ||
74 : | |||
75 : | |||
76 : | ALIGN 16 | ||
77 : | |||
78 : | SECTION .text | ||
79 : | |||
80 : | ;------------------------------------------------------------------------------ | ||
81 : | ; | ||
82 : | ; void yv12_to_rgb32_mmx(uint8_t *dst, | ||
83 : | ; int dst_stride, | ||
84 : | ; uint8_t *y_src, | ||
85 : | ; uint8_t *u_src, | ||
86 : | ; uint8_t *v_src, | ||
87 : | ; int y_stride, int uv_stride, | ||
88 : | ; int width, int height); | ||
89 : | ; | ||
90 : | ;------------------------------------------------------------------------------ | ||
91 : | |||
92 : | cglobal yv12_to_rgb32_mmx | ||
93 : | yv12_to_rgb32_mmx: | ||
94 : | |||
95 : | push ebx | ||
96 : | push esi | ||
97 : | push edi | ||
98 : | push ebp | ||
99 : | |||
100 : | ; local vars alloc | ||
101 : | %define localsize 72 | ||
102 : | %define TEMP_Y1 esp | ||
103 : | %define TEMP_Y2 esp + 8 | ||
104 : | %define TEMP_G1 esp + 16 | ||
105 : | %define TEMP_G2 esp + 24 | ||
106 : | %define TEMP_B1 esp + 32 | ||
107 : | %define TEMP_B2 esp + 40 | ||
108 : | %define y_dif esp + 48 | ||
109 : | %define dst_dif esp + 52 | ||
110 : | %define uv_dif esp + 56 | ||
111 : | %define height esp + 60 | ||
112 : | %define width_8 esp + 64 | ||
113 : | %define height_2 esp + 68 | ||
114 : | sub esp, localsize | ||
115 : | |||
116 : | ; function code | ||
117 : | mov eax, [esp + 52 + localsize] ; height -> eax | ||
118 : | cmp eax, 0x00 | ||
119 : | jge near dont_flip ; flip? | ||
120 : | |||
121 : | neg eax ; neg height | ||
122 : | mov [height], eax | ||
123 : | |||
124 : | mov esi, [esp + 48 + localsize] ; width -> esi | ||
125 : | |||
126 : | mov ebp, [esp + 40 + localsize] ; y_stride -> ebp | ||
127 : | mov ebx, ebp | ||
128 : | shl ebx, 1 ; 2 * y_stride -> ebx | ||
129 : | neg ebx | ||
130 : | sub ebx, esi ; y_dif -> eax | ||
131 : | |||
132 : | mov [y_dif], ebx | ||
133 : | |||
134 : | sub eax, 1 ; height - 1 -> eax | ||
135 : | mul ebp ; (height - 1) * y_stride -> ebp | ||
136 : | mov ecx, eax | ||
137 : | mov eax, [esp + 28 + localsize] ; y_src -> eax | ||
138 : | add eax, ecx ; y_src -> eax | ||
139 : | mov ebx, eax | ||
140 : | sub ebx, ebp ; y_src2 -> ebx | ||
141 : | |||
142 : | mov ecx, [esp + 24 + localsize] ; dst_stride -> ecx | ||
143 : | mov edx, ecx | ||
144 : | shl edx, 3 | ||
145 : | mov ecx, edx ; 8 * dst_stride -> ecx | ||
146 : | shl esi, 2 | ||
147 : | sub ecx, esi ; 8 * dst_stride - 4 * width -> ecx | ||
148 : | |||
149 : | mov [dst_dif], ecx | ||
150 : | |||
151 : | mov esi, [esp + 20 + localsize] ; dst -> esi | ||
152 : | mov edi, esi | ||
153 : | shr edx, 1 | ||
154 : | add edi, edx ; dst2 -> edi | ||
155 : | |||
156 : | mov ebp, [esp + 48 + localsize] ; width -> ebp | ||
157 : | mov ecx, ebp ; width -> ecx | ||
158 : | shr ecx, 1 | ||
159 : | shr ebp, 3 ; width / 8 -> ebp | ||
160 : | mov [width_8], ebp | ||
161 : | |||
162 : | mov ebp, [esp + 44 + localsize] ; uv_stride -> ebp | ||
163 : | mov edx, ebp | ||
164 : | neg edx | ||
165 : | sub edx, ecx | ||
166 : | mov [uv_dif], edx | ||
167 : | |||
168 : | mov edx, ebp | ||
169 : | mov ebp, eax | ||
170 : | mov eax, [height] ; height -> eax | ||
171 : | shr eax, 1 ; height / 2 -> eax | ||
172 : | |||
173 : | mov ecx, [esp + 32 + localsize] ; u_src -> ecx | ||
174 : | sub eax, 1 | ||
175 : | mul edx | ||
176 : | add ecx, eax | ||
177 : | |||
178 : | mov edx, [esp + 36 + localsize] ; v_src -> edx | ||
179 : | add edx, eax | ||
180 : | |||
181 : | mov eax, ebp | ||
182 : | |||
183 : | mov ebp, [height] ; height -> ebp | ||
184 : | shr ebp, 1 ; height / 2 -> ebp | ||
185 : | |||
186 : | pxor mm7, mm7 | ||
187 : | jmp y_loop | ||
188 : | |||
189 : | |||
190 : | dont_flip: | ||
191 : | mov esi, [esp + 48 + localsize] ; width -> esi | ||
192 : | |||
193 : | mov ebp, [esp + 40 + localsize] ; y_stride -> ebp | ||
194 : | mov ebx, ebp | ||
195 : | shl ebx, 1 ; 2 * y_stride -> ebx | ||
196 : | sub ebx, esi ; y_dif -> ebx | ||
197 : | |||
198 : | mov [y_dif], ebx | ||
199 : | |||
200 : | mov eax, [esp + 28 + localsize] ; y_src -> eax | ||
201 : | mov ebx, eax | ||
202 : | add ebx, ebp ; y_src2 -> ebp | ||
203 : | |||
204 : | mov ecx, [esp + 24 + localsize] ; dst_stride -> ecx | ||
205 : | shl ecx, 3 | ||
206 : | mov edx, ecx ; 8 * dst_stride -> edx | ||
207 : | shl esi, 2 | ||
208 : | sub ecx, esi ; 8 * dst_stride - 4 * width -> ecx | ||
209 : | |||
210 : | mov [dst_dif], ecx | ||
211 : | |||
212 : | mov esi, [esp + 20 + localsize] ; dst -> esi | ||
213 : | mov edi, esi | ||
214 : | shr edx, 1 | ||
215 : | add edi, edx ; dst2 -> edi | ||
216 : | |||
217 : | mov ebp, [esp + 48 + localsize] ; width -> ebp | ||
218 : | mov ecx, ebp ; width -> ecx | ||
219 : | shr ecx, 1 | ||
220 : | shr ebp, 3 ; width / 8 -> ebp | ||
221 : | mov [width_8], ebp | ||
222 : | |||
223 : | mov ebp, [esp + 44 + localsize] ; uv_stride -> ebp | ||
224 : | sub ebp, ecx | ||
225 : | mov [uv_dif], ebp | ||
226 : | |||
227 : | mov ecx, [esp + 32 + localsize] ; u_src -> ecx | ||
228 : | mov edx, [esp + 36 + localsize] ; v_src -> edx | ||
229 : | |||
230 : | mov ebp, [esp + 52 + localsize] ; height -> ebp | ||
231 : | shr ebp, 1 ; height / 2 -> ebp | ||
232 : | |||
233 : | pxor mm7, mm7 | ||
234 : | |||
235 : | y_loop: | ||
236 : | mov [height_2], ebp | ||
237 : | mov ebp, [width_8] | ||
238 : | |||
239 : | x_loop: | ||
240 : | movd mm2, [ecx] | ||
241 : | movd mm3, [edx] | ||
242 : | |||
243 : | punpcklbw mm2, mm7 ; u3u2u1u0 -> mm2 | ||
244 : | punpcklbw mm3, mm7 ; v3v2v1v0 -> mm3 | ||
245 : | |||
246 : | psubsw mm2, [U_SUB] ; U - 128 | ||
247 : | psubsw mm3, [V_SUB] ; V - 128 | ||
248 : | |||
249 : | movq mm4, mm2 | ||
250 : | movq mm5, mm3 | ||
251 : | |||
252 : | pmullw mm2, [UG_MUL] | ||
253 : | pmullw mm3, [VG_MUL] | ||
254 : | |||
255 : | movq mm6, mm2 ; u3u2u1u0 -> mm6 | ||
256 : | punpckhwd mm2, mm2 ; u3u3u2u2 -> mm2 | ||
257 : | punpcklwd mm6, mm6 ; u1u1u0u0 -> mm6 | ||
258 : | |||
259 : | pmullw mm4, [UB_MUL] ; B_ADD -> mm4 | ||
260 : | |||
261 : | movq mm0, mm3 | ||
262 : | punpckhwd mm3, mm3 ; v3v3v2v2 -> mm2 | ||
263 : | punpcklwd mm0, mm0 ; v1v1v0v0 -> mm6 | ||
264 : | |||
265 : | paddsw mm2, mm3 | ||
266 : | paddsw mm6, mm0 | ||
267 : | |||
268 : | pmullw mm5, [VR_MUL] ; R_ADD -> mm5 | ||
269 : | |||
270 : | movq mm0, [eax] ; y7y6y5y4y3y2y1y0 -> mm0 | ||
271 : | |||
272 : | movq mm1, mm0 | ||
273 : | punpckhbw mm1, mm7 ; y7y6y5y4 -> mm1 | ||
274 : | punpcklbw mm0, mm7 ; y3y2y1y0 -> mm0 | ||
275 : | |||
276 : | psubsw mm0, [Y_SUB] ; Y - Y_SUB | ||
277 : | psubsw mm1, [Y_SUB] ; Y - Y_SUB | ||
278 : | |||
279 : | pmullw mm1, [Y_MUL] | ||
280 : | pmullw mm0, [Y_MUL] | ||
281 : | |||
282 : | movq [TEMP_Y2], mm1 ; y7y6y5y4 -> mm3 | ||
283 : | movq [TEMP_Y1], mm0 ; y3y2y1y0 -> mm7 | ||
284 : | |||
285 : | psubsw mm1, mm2 ; g7g6g5g4 -> mm1 | ||
286 : | psubsw mm0, mm6 ; g3g2g1g0 -> mm0 | ||
287 : | |||
288 : | psraw mm1, SCALEBITS | ||
289 : | psraw mm0, SCALEBITS | ||
290 : | |||
291 : | packuswb mm0, mm1 ;g7g6g5g4g3g2g1g0 -> mm0 | ||
292 : | |||
293 : | movq [TEMP_G1], mm0 | ||
294 : | |||
295 : | movq mm0, [ebx] ; y7y6y5y4y3y2y1y0 -> mm0 | ||
296 : | |||
297 : | movq mm1, mm0 | ||
298 : | |||
299 : | punpckhbw mm1, mm7 ; y7y6y5y4 -> mm1 | ||
300 : | punpcklbw mm0, mm7 ; y3y2y1y0 -> mm0 | ||
301 : | |||
302 : | psubsw mm0, [Y_SUB] ; Y - Y_SUB | ||
303 : | psubsw mm1, [Y_SUB] ; Y - Y_SUB | ||
304 : | |||
305 : | pmullw mm1, [Y_MUL] | ||
306 : | pmullw mm0, [Y_MUL] | ||
307 : | |||
308 : | movq mm3, mm1 | ||
309 : | psubsw mm1, mm2 ; g7g6g5g4 -> mm1 | ||
310 : | |||
311 : | movq mm2, mm0 | ||
312 : | psubsw mm0, mm6 ; g3g2g1g0 -> mm0 | ||
313 : | |||
314 : | psraw mm1, SCALEBITS | ||
315 : | psraw mm0, SCALEBITS | ||
316 : | |||
317 : | packuswb mm0, mm1 ; g7g6g5g4g3g2g1g0 -> mm0 | ||
318 : | |||
319 : | movq [TEMP_G2], mm0 | ||
320 : | |||
321 : | movq mm0, mm4 | ||
322 : | punpckhwd mm4, mm4 ; u3u3u2u2 -> mm2 | ||
323 : | punpcklwd mm0, mm0 ; u1u1u0u0 -> mm6 | ||
324 : | |||
325 : | movq mm1, mm3 ; y7y6y5y4 -> mm1 | ||
326 : | paddsw mm3, mm4 ; b7b6b5b4 -> mm3 | ||
327 : | |||
328 : | movq mm7, mm2 ; y3y2y1y0 -> mm7 | ||
329 : | |||
330 : | paddsw mm2, mm0 ; b3b2b1b0 -> mm2 | ||
331 : | |||
332 : | psraw mm3, SCALEBITS | ||
333 : | psraw mm2, SCALEBITS | ||
334 : | |||
335 : | packuswb mm2, mm3 ; b7b6b5b4b3b2b1b0 -> mm2 | ||
336 : | |||
337 : | movq [TEMP_B2], mm2 | ||
338 : | |||
339 : | movq mm3, [TEMP_Y2] | ||
340 : | movq mm2, [TEMP_Y1] | ||
341 : | |||
342 : | movq mm6, mm3 ; TEMP_Y2 -> mm6 | ||
343 : | paddsw mm3, mm4 ; b7b6b5b4 -> mm3 | ||
344 : | |||
345 : | movq mm4, mm2 ; TEMP_Y1 -> mm4 | ||
346 : | paddsw mm2, mm0 ; b3b2b1b0 -> mm2 | ||
347 : | |||
348 : | psraw mm3, SCALEBITS | ||
349 : | psraw mm2, SCALEBITS | ||
350 : | |||
351 : | packuswb mm2, mm3 ; b7b6b5b4b3b2b1b0 -> mm2 | ||
352 : | |||
353 : | movq [TEMP_B1], mm2 | ||
354 : | |||
355 : | movq mm0, mm5 | ||
356 : | punpckhwd mm5, mm5 ; v3v3v2v2 -> mm5 | ||
357 : | punpcklwd mm0, mm0 ; v1v1v0v0 -> mm0 | ||
358 : | |||
359 : | paddsw mm1, mm5 ; r7r6r5r4 -> mm1 | ||
360 : | paddsw mm7, mm0 ; r3r2r1r0 -> mm7 | ||
361 : | |||
362 : | psraw mm1, SCALEBITS | ||
363 : | psraw mm7, SCALEBITS | ||
364 : | |||
365 : | packuswb mm7, mm1 ; r7r6r5r4r3r2r1r0 -> mm7 (TEMP_R2) | ||
366 : | |||
367 : | paddsw mm6, mm5 ; r7r6r5r4 -> mm6 | ||
368 : | paddsw mm4, mm0 ; r3r2r1r0 -> mm4 | ||
369 : | |||
370 : | psraw mm6, SCALEBITS | ||
371 : | psraw mm4, SCALEBITS | ||
372 : | |||
373 : | packuswb mm4, mm6 ; r7r6r5r4r3r2r1r0 -> mm4 (TEMP_R1) | ||
374 : | |||
375 : | movq mm0, [TEMP_B1] | ||
376 : | movq mm1, [TEMP_G1] | ||
377 : | |||
378 : | movq mm6, mm7 | ||
379 : | |||
380 : | movq mm2, mm0 | ||
381 : | punpcklbw mm2, mm4 ; r3b3r2b2r1b1r0b0 -> mm2 | ||
382 : | punpckhbw mm0, mm4 ; r7b7r6b6r5b5r4b4 -> mm0 | ||
383 : | |||
384 : | pxor mm7, mm7 | ||
385 : | |||
386 : | movq mm3, mm1 | ||
387 : | punpcklbw mm1, mm7 ; 0g30g20g10g0 -> mm1 | ||
388 : | punpckhbw mm3, mm7 ; 0g70g60g50g4 -> mm3 | ||
389 : | |||
390 : | movq mm4, mm2 | ||
391 : | punpcklbw mm2, mm1 ; 0r1g1b10r0g0b0 -> mm2 | ||
392 : | punpckhbw mm4, mm1 ; 0r3g3b30r2g2b2 -> mm4 | ||
393 : | |||
394 : | movq mm5, mm0 | ||
395 : | punpcklbw mm0, mm3 ; 0r5g5b50r4g4b4 -> mm0 | ||
396 : | punpckhbw mm5, mm3 ; 0r7g7b70r6g6b6 -> mm5 | ||
397 : | |||
398 : | movq [esi], mm2 | ||
399 : | movq [esi + 8], mm4 | ||
400 : | movq [esi + 16], mm0 | ||
401 : | movq [esi + 24], mm5 | ||
402 : | |||
403 : | movq mm0, [TEMP_B2] | ||
404 : | movq mm1, [TEMP_G2] | ||
405 : | |||
406 : | movq mm2, mm0 | ||
407 : | punpcklbw mm2, mm6 ; r3b3r2b2r1b1r0b0 -> mm2 | ||
408 : | punpckhbw mm0, mm6 ; r7b7r6b6r5b5r4b4 -> mm0 | ||
409 : | |||
410 : | movq mm3, mm1 | ||
411 : | punpcklbw mm1, mm7 ; 0g30g20g10g0 -> mm1 | ||
412 : | punpckhbw mm3, mm7 ; 0g70g60g50g4 -> mm3 | ||
413 : | |||
414 : | movq mm4, mm2 | ||
415 : | punpcklbw mm2, mm1 ; 0r1g1b10r0g0b0 -> mm2 | ||
416 : | punpckhbw mm4, mm1 ; 0r3g3b30r2g2b2 -> mm4 | ||
417 : | |||
418 : | movq mm5, mm0 | ||
419 : | punpcklbw mm0, mm3 ; 0r5g5b50r4g4b4 -> mm0 | ||
420 : | punpckhbw mm5, mm3 ; 0r7g7b70r6g6b6 -> mm5 | ||
421 : | |||
422 : | movq [edi], mm2 | ||
423 : | movq [edi + 8], mm4 | ||
424 : | movq [edi + 16], mm0 | ||
425 : | movq [edi + 24], mm5 | ||
426 : | |||
427 : | add esi, 32 | ||
428 : | add edi, 32 | ||
429 : | |||
430 : | add eax, 8 | ||
431 : | add ebx, 8 | ||
432 : | add ecx, 4 | ||
433 : | add edx, 4 | ||
434 : | |||
435 : | dec ebp | ||
436 : | |||
437 : | jnz near x_loop | ||
438 : | |||
439 : | add esi, [dst_dif] | ||
440 : | add edi, [dst_dif] | ||
441 : | |||
442 : | add eax, [y_dif] | ||
443 : | add ebx, [y_dif] | ||
444 : | |||
445 : | add ecx, [uv_dif] | ||
446 : | add edx, [uv_dif] | ||
447 : | |||
448 : | mov ebp, [height_2] | ||
449 : | dec ebp | ||
450 : | jnz near y_loop | ||
451 : | |||
452 : | emms | ||
453 : | |||
454 : | ;; Local vars deallocation | ||
455 : | add esp, localsize | ||
456 : | %undef TEMP_Y1 | ||
457 : | %undef TEMP_Y2 | ||
458 : | %undef TEMP_G1 | ||
459 : | %undef TEMP_G2 | ||
460 : | %undef TEMP_B1 | ||
461 : | %undef TEMP_B2 | ||
462 : | %undef y_dif | ||
463 : | %undef dst_dif | ||
464 : | %undef uv_dif | ||
465 : | %undef height | ||
466 : | %undef width_8 | ||
467 : | %undef height_2 | ||
468 : | %undef localsize | ||
469 : | |||
470 : | pop ebp | ||
471 : | pop edi | ||
472 : | pop esi | ||
473 : | pop ebx | ||
474 : | |||
475 : | ret |
No admin address has been configured | ViewVC Help |
Powered by ViewVC 1.0.4 |