Parent Directory
| 
Revision Log
Revision 434 - (view) (download)
| 1 : | chl | 434 | ;/***************************************************************************** |
| 2 : | Isibaar | 3 | ; * |
| 3 : | chl | 434 | ; * XVID MPEG-4 VIDEO CODEC |
| 4 : | ; * mmx yuyv/uyvy to yuv planar conversion | ||
| 5 : | Isibaar | 3 | ; * |
| 6 : | chl | 434 | ; * Copyright(C) 2002 Peter Ross <pross@xvid.org> |
| 7 : | Isibaar | 3 | ; * |
| 8 : | chl | 434 | ; * 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 : | Isibaar | 3 | ; * |
| 17 : | chl | 434 | ; * 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 : | Isibaar | 3 | ; * |
| 22 : | chl | 434 | ; * 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 : | Isibaar | 3 | ; * |
| 27 : | chl | 434 | ; * 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 : | Isibaar | 3 | ; * |
| 31 : | chl | 434 | ; ****************************************************************************/ |
| 32 : | Isibaar | 3 | |
| 33 : | bits 32 | ||
| 34 : | |||
| 35 : | |||
| 36 : | section .data | ||
| 37 : | |||
| 38 : | %macro cglobal 1 | ||
| 39 : | %ifdef PREFIX | ||
| 40 : | global _%1 | ||
| 41 : | %define %1 _%1 | ||
| 42 : | %else | ||
| 43 : | global %1 | ||
| 44 : | %endif | ||
| 45 : | %endmacro | ||
| 46 : | |||
| 47 : | align 16 | ||
| 48 : | |||
| 49 : | |||
| 50 : | section .data | ||
| 51 : | |||
| 52 : | |||
| 53 : | ;=========================================================================== | ||
| 54 : | ; masks for extracting yuv components | ||
| 55 : | ;=========================================================================== | ||
| 56 : | ; y u y v y u y v | ||
| 57 : | |||
| 58 : | mask1 db 0xff, 0, 0xff, 0, 0xff, 0, 0xff, 0 | ||
| 59 : | mask2 db 0, 0xff, 0, 0xff, 0, 0xff, 0, 0xff | ||
| 60 : | |||
| 61 : | |||
| 62 : | section .text | ||
| 63 : | |||
| 64 : | ;=========================================================================== | ||
| 65 : | ; | ||
| 66 : | ; void yuyv_to_yv12_mmx(uint8_t * const y_out, | ||
| 67 : | ; uint8_t * const u_out, | ||
| 68 : | ; uint8_t * const v_out, | ||
| 69 : | ; const uint8_t * const src, | ||
| 70 : | ; const uint32_t width, | ||
| 71 : | ; const uint32_t height, | ||
| 72 : | ; const uint32_t stride); | ||
| 73 : | ; | ||
| 74 : | ; width must be multiple of 8 | ||
| 75 : | ; does not flip | ||
| 76 : | ; ~30% faster than plain c | ||
| 77 : | ; | ||
| 78 : | ;=========================================================================== | ||
| 79 : | |||
| 80 : | align 16 | ||
| 81 : | cglobal yuyv_to_yv12_mmx | ||
| 82 : | yuyv_to_yv12_mmx | ||
| 83 : | |||
| 84 : | push ebx | ||
| 85 : | push ecx | ||
| 86 : | push esi | ||
| 87 : | push edi | ||
| 88 : | push ebp ; STACK BASE = 20 | ||
| 89 : | |||
| 90 : | ; some global consants | ||
| 91 : | |||
| 92 : | mov ecx, [esp + 20 + 20] ; width | ||
| 93 : | mov eax, [esp + 20 + 28] ; stride | ||
| 94 : | sub eax, ecx ; eax = stride - width | ||
| 95 : | mov edx, eax | ||
| 96 : | add edx, [esp + 20 + 28] ; edx = y_dif + stride | ||
| 97 : | push edx ; [esp + 12] = y_dif | ||
| 98 : | |||
| 99 : | shr eax, 1 | ||
| 100 : | push eax ; [esp + 8] = uv_dif | ||
| 101 : | |||
| 102 : | shr ecx, 3 | ||
| 103 : | push ecx ; [esp + 4] = width/8 | ||
| 104 : | |||
| 105 : | sub esp, 4 ; [esp + 0] = tmp_height_counter | ||
| 106 : | ; STACK_BASE = 36 | ||
| 107 : | |||
| 108 : | movq mm6, [mask1] | ||
| 109 : | movq mm7, [mask2] | ||
| 110 : | |||
| 111 : | mov edi, [esp + 36 + 4] ; y_out | ||
| 112 : | mov ebx, [esp + 36 + 8] ; u_out | ||
| 113 : | mov edx, [esp + 36 + 12] ; v_out | ||
| 114 : | mov esi, [esp + 36 + 16] ; src | ||
| 115 : | |||
| 116 : | mov eax, [esp + 36 + 20] | ||
| 117 : | mov ebp, [esp + 36 + 24] | ||
| 118 : | mov ecx, [esp + 36 + 28] ; ecx = stride | ||
| 119 : | shr ebp, 1 ; ebp = height /= 2 | ||
| 120 : | add eax, eax ; eax = 2 * width | ||
| 121 : | |||
| 122 : | .yloop | ||
| 123 : | mov [esp], ebp | ||
| 124 : | mov ebp, [esp + 4] ; width/8 | ||
| 125 : | .xloop | ||
| 126 : | movq mm2, [esi] ; y 1st row | ||
| 127 : | movq mm3, [esi + 8] | ||
| 128 : | movq mm0, mm2 | ||
| 129 : | movq mm1, mm3 | ||
| 130 : | pand mm2, mm6 ; mask1 | ||
| 131 : | pand mm3, mm6 ; mask1 | ||
| 132 : | pand mm0, mm7 ; mask2 | ||
| 133 : | pand mm1, mm7 ; mask2 | ||
| 134 : | packuswb mm2, mm3 | ||
| 135 : | psrlq mm0, 8 | ||
| 136 : | psrlq mm1, 8 | ||
| 137 : | movq [edi], mm2 | ||
| 138 : | |||
| 139 : | movq mm4, [esi + eax] ; y 2nd row | ||
| 140 : | movq mm5, [esi + eax + 8] | ||
| 141 : | movq mm2, mm4 | ||
| 142 : | movq mm3, mm5 | ||
| 143 : | pand mm4, mm6 ; mask1 | ||
| 144 : | pand mm5, mm6 ; mask1 | ||
| 145 : | pand mm2, mm7 ; mask2 | ||
| 146 : | pand mm3, mm7 ; mask2 | ||
| 147 : | packuswb mm4, mm5 | ||
| 148 : | psrlq mm2, 8 | ||
| 149 : | psrlq mm3, 8 | ||
| 150 : | movq [edi + ecx], mm4 | ||
| 151 : | |||
| 152 : | paddw mm0, mm2 ; uv avg 1st & 2nd | ||
| 153 : | paddw mm1, mm3 | ||
| 154 : | psrlw mm0, 1 | ||
| 155 : | psrlw mm1, 1 | ||
| 156 : | packuswb mm0, mm1 | ||
| 157 : | movq mm2, mm0 | ||
| 158 : | pand mm0, mm6 ; mask1 | ||
| 159 : | pand mm2, mm7 ; mask2 | ||
| 160 : | packuswb mm0, mm0 | ||
| 161 : | psrlq mm2, 8 | ||
| 162 : | movd [ebx], mm0 | ||
| 163 : | packuswb mm2, mm2 | ||
| 164 : | movd [edx], mm2 | ||
| 165 : | |||
| 166 : | add esi, 16 | ||
| 167 : | add edi, 8 | ||
| 168 : | add ebx, 4 | ||
| 169 : | add edx, 4 | ||
| 170 : | dec ebp | ||
| 171 : | jnz near .xloop | ||
| 172 : | |||
| 173 : | mov ebp, [esp] | ||
| 174 : | |||
| 175 : | add esi, eax ; += width2 | ||
| 176 : | add edi, [esp + 12] ; += y_dif + stride | ||
| 177 : | add ebx, [esp + 8] ; += uv_dif | ||
| 178 : | add edx, [esp + 8] ; += uv_dif | ||
| 179 : | |||
| 180 : | dec ebp | ||
| 181 : | jnz near .yloop | ||
| 182 : | |||
| 183 : | emms | ||
| 184 : | |||
| 185 : | add esp, 16 | ||
| 186 : | pop ebp | ||
| 187 : | pop edi | ||
| 188 : | pop esi | ||
| 189 : | pop ecx | ||
| 190 : | pop ebx | ||
| 191 : | |||
| 192 : | ret | ||
| 193 : | |||
| 194 : | |||
| 195 : | ;=========================================================================== | ||
| 196 : | ; | ||
| 197 : | ; void uyvy_to_yv12_mmx(uint8_t * const y_out, | ||
| 198 : | ; uint8_t * const u_out, | ||
| 199 : | ; uint8_t * const v_out, | ||
| 200 : | ; const uint8_t * const src, | ||
| 201 : | ; const uint32_t width, | ||
| 202 : | ; const uint32_t height, | ||
| 203 : | ; const uint32_t stride); | ||
| 204 : | ; | ||
| 205 : | ; width must be multiple of 8 | ||
| 206 : | ; does not flip | ||
| 207 : | ; ~30% faster than plain c | ||
| 208 : | ; | ||
| 209 : | ;=========================================================================== | ||
| 210 : | |||
| 211 : | align 16 | ||
| 212 : | cglobal uyvy_to_yv12_mmx | ||
| 213 : | uyvy_to_yv12_mmx | ||
| 214 : | |||
| 215 : | push ebx | ||
| 216 : | push ecx | ||
| 217 : | push esi | ||
| 218 : | push edi | ||
| 219 : | push ebp ; STACK BASE = 20 | ||
| 220 : | |||
| 221 : | ; some global consants | ||
| 222 : | |||
| 223 : | mov ecx, [esp + 20 + 20] ; width | ||
| 224 : | mov eax, [esp + 20 + 28] ; stride | ||
| 225 : | sub eax, ecx ; eax = stride - width | ||
| 226 : | mov edx, eax | ||
| 227 : | add edx, [esp + 20 + 28] ; edx = y_dif + stride | ||
| 228 : | push edx ; [esp + 12] = y_dif | ||
| 229 : | |||
| 230 : | shr eax, 1 | ||
| 231 : | push eax ; [esp + 8] = uv_dif | ||
| 232 : | |||
| 233 : | shr ecx, 3 | ||
| 234 : | push ecx ; [esp + 4] = width/8 | ||
| 235 : | |||
| 236 : | sub esp, 4 ; [esp + 0] = tmp_height_counter | ||
| 237 : | ; STACK_BASE = 36 | ||
| 238 : | |||
| 239 : | movq mm6, [mask1] | ||
| 240 : | movq mm7, [mask2] | ||
| 241 : | |||
| 242 : | mov edi, [esp + 36 + 4] ; y_out | ||
| 243 : | mov ebx, [esp + 36 + 8] ; u_out | ||
| 244 : | mov edx, [esp + 36 + 12] ; v_out | ||
| 245 : | mov esi, [esp + 36 + 16] ; src | ||
| 246 : | |||
| 247 : | mov eax, [esp + 36 + 20] | ||
| 248 : | mov ebp, [esp + 36 + 24] | ||
| 249 : | mov ecx, [esp + 36 + 28] ; ecx = stride | ||
| 250 : | shr ebp, 1 ; ebp = height /= 2 | ||
| 251 : | add eax, eax ; eax = 2 * width | ||
| 252 : | |||
| 253 : | .yloop | ||
| 254 : | mov [esp], ebp | ||
| 255 : | mov ebp, [esp + 4] ; width/8 | ||
| 256 : | .xloop | ||
| 257 : | movq mm2, [esi] ; y 1st row | ||
| 258 : | movq mm3, [esi + 8] | ||
| 259 : | movq mm0, mm2 | ||
| 260 : | movq mm1, mm3 | ||
| 261 : | pand mm2, mm7 ; mask2 | ||
| 262 : | pand mm3, mm7 ; mask2 | ||
| 263 : | psrlq mm2, 8 | ||
| 264 : | psrlq mm3, 8 | ||
| 265 : | pand mm0, mm6 ; mask1 | ||
| 266 : | pand mm1, mm6 ; mask1 | ||
| 267 : | packuswb mm2, mm3 | ||
| 268 : | movq [edi], mm2 | ||
| 269 : | |||
| 270 : | movq mm4, [esi + eax] ; y 2nd row | ||
| 271 : | movq mm5, [esi + eax + 8] | ||
| 272 : | movq mm2, mm4 | ||
| 273 : | movq mm3, mm5 | ||
| 274 : | pand mm4, mm7 ; mask2 | ||
| 275 : | pand mm5, mm7 ; mask2 | ||
| 276 : | psrlq mm4, 8 | ||
| 277 : | psrlq mm5, 8 | ||
| 278 : | pand mm2, mm6 ; mask1 | ||
| 279 : | pand mm3, mm6 ; mask1 | ||
| 280 : | packuswb mm4, mm5 | ||
| 281 : | movq [edi + ecx], mm4 | ||
| 282 : | |||
| 283 : | paddw mm0, mm2 ; uv avg 1st & 2nd | ||
| 284 : | paddw mm1, mm3 | ||
| 285 : | psrlw mm0, 1 | ||
| 286 : | psrlw mm1, 1 | ||
| 287 : | packuswb mm0, mm1 | ||
| 288 : | movq mm2, mm0 | ||
| 289 : | pand mm0, mm6 ; mask1 | ||
| 290 : | pand mm2, mm7 ; mask2 | ||
| 291 : | packuswb mm0, mm0 | ||
| 292 : | psrlq mm2, 8 | ||
| 293 : | movd [ebx], mm0 | ||
| 294 : | packuswb mm2, mm2 | ||
| 295 : | movd [edx], mm2 | ||
| 296 : | |||
| 297 : | add esi, 16 | ||
| 298 : | add edi, 8 | ||
| 299 : | add ebx, 4 | ||
| 300 : | add edx, 4 | ||
| 301 : | dec ebp | ||
| 302 : | jnz near .xloop | ||
| 303 : | |||
| 304 : | mov ebp, [esp] | ||
| 305 : | |||
| 306 : | add esi, eax ; += width2 | ||
| 307 : | add edi, [esp + 12] ; += y_dif + stride | ||
| 308 : | add ebx, [esp + 8] ; += uv_dif | ||
| 309 : | add edx, [esp + 8] ; += uv_dif | ||
| 310 : | |||
| 311 : | dec ebp | ||
| 312 : | jnz near .yloop | ||
| 313 : | |||
| 314 : | emms | ||
| 315 : | |||
| 316 : | add esp, 16 | ||
| 317 : | pop ebp | ||
| 318 : | pop edi | ||
| 319 : | pop esi | ||
| 320 : | pop ecx | ||
| 321 : | pop ebx | ||
| 322 : | |||
| 323 : | ret |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |