Parent Directory
| 
Revision Log
Revision 1.4 - (view) (download)
| 1 : | chl | 1.3 | /***************************************************************************** |
| 2 : | * | ||
| 3 : | * XVID MPEG-4 VIDEO CODEC | ||
| 4 : | * - inverse fast disrete cosine transformation - integer C version | ||
| 5 : | * | ||
| 6 : | * These routines are from Independent JPEG Group's free JPEG software | ||
| 7 : | * Copyright (C) 1991-1998, Thomas G. Lane (see the file README.IJG) | ||
| 8 : | * | ||
| 9 : | edgomez | 1.4 | * This file is part of XviD, a free MPEG-4 video encoder/decoder |
| 10 : | chl | 1.3 | * |
| 11 : | edgomez | 1.4 | * XviD is free software; you can redistribute it and/or modify it |
| 12 : | * under the terms of the GNU General Public License as published by | ||
| 13 : | chl | 1.3 | * the Free Software Foundation; either version 2 of the License, or |
| 14 : | * (at your option) any later version. | ||
| 15 : | * | ||
| 16 : | * This program is distributed in the hope that it will be useful, | ||
| 17 : | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 18 : | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | ||
| 19 : | * GNU General Public License for more details. | ||
| 20 : | * | ||
| 21 : | * You should have received a copy of the GNU General Public License | ||
| 22 : | * along with this program; if not, write to the Free Software | ||
| 23 : | * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | ||
| 24 : | edgomez | 1.4 | * |
| 25 : | * Under section 8 of the GNU General Public License, the copyright | ||
| 26 : | * holders of XVID explicitly forbid distribution in the following | ||
| 27 : | * countries: | ||
| 28 : | * | ||
| 29 : | * - Japan | ||
| 30 : | * - United States of America | ||
| 31 : | * | ||
| 32 : | * Linking XviD statically or dynamically with other modules is making a | ||
| 33 : | * combined work based on XviD. Thus, the terms and conditions of the | ||
| 34 : | * GNU General Public License cover the whole combination. | ||
| 35 : | * | ||
| 36 : | * As a special exception, the copyright holders of XviD give you | ||
| 37 : | * permission to link XviD with independent modules that communicate with | ||
| 38 : | * XviD solely through the VFW1.1 and DShow interfaces, regardless of the | ||
| 39 : | * license terms of these independent modules, and to copy and distribute | ||
| 40 : | * the resulting combined work under terms of your choice, provided that | ||
| 41 : | * every copy of the combined work is accompanied by a complete copy of | ||
| 42 : | * the source code of XviD (the version of XviD used to produce the | ||
| 43 : | * combined work), being distributed under the terms of the GNU General | ||
| 44 : | * Public License plus this exception. An independent module is a module | ||
| 45 : | * which is not derived from or based on XviD. | ||
| 46 : | * | ||
| 47 : | * Note that people who make modified versions of XviD are not obligated | ||
| 48 : | * to grant this special exception for their modified versions; it is | ||
| 49 : | * their choice whether to do so. The GNU General Public License gives | ||
| 50 : | * permission to release a modified version without this exception; this | ||
| 51 : | * exception also makes it possible to release a modified version which | ||
| 52 : | * carries forward this exception. | ||
| 53 : | * | ||
| 54 : | * $Id$ | ||
| 55 : | Isibaar | 1.1 | * |
| 56 : | chl | 1.3 | *************************************************************************/ |
| 57 : | Isibaar | 1.1 | |
| 58 : | /**********************************************************/ | ||
| 59 : | /* inverse two dimensional DCT, Chen-Wang algorithm */ | ||
| 60 : | /* (cf. IEEE ASSP-32, pp. 803-816, Aug. 1984) */ | ||
| 61 : | /* 32-bit integer arithmetic (8 bit coefficients) */ | ||
| 62 : | /* 11 mults, 29 adds per DCT */ | ||
| 63 : | /* sE, 18.8.91 */ | ||
| 64 : | /**********************************************************/ | ||
| 65 : | /* coefficients extended to 12 bit for IEEE1180-1990 */ | ||
| 66 : | /* compliance sE, 2.1.94 */ | ||
| 67 : | /**********************************************************/ | ||
| 68 : | |||
| 69 : | /* this code assumes >> to be a two's-complement arithmetic */ | ||
| 70 : | /* right shift: (-2)>>1 == -1 , (-3)>>1 == -2 */ | ||
| 71 : | |||
| 72 : | #include "idct.h" | ||
| 73 : | |||
| 74 : | edgomez | 1.2 | #define W1 2841 /* 2048*sqrt(2)*cos(1*pi/16) */ |
| 75 : | #define W2 2676 /* 2048*sqrt(2)*cos(2*pi/16) */ | ||
| 76 : | #define W3 2408 /* 2048*sqrt(2)*cos(3*pi/16) */ | ||
| 77 : | #define W5 1609 /* 2048*sqrt(2)*cos(5*pi/16) */ | ||
| 78 : | #define W6 1108 /* 2048*sqrt(2)*cos(6*pi/16) */ | ||
| 79 : | #define W7 565 /* 2048*sqrt(2)*cos(7*pi/16) */ | ||
| 80 : | Isibaar | 1.1 | |
| 81 : | |||
| 82 : | /* global declarations */ | ||
| 83 : | //void init_idct_int32 (void); | ||
| 84 : | //void idct_int32 (short *block); | ||
| 85 : | |||
| 86 : | /* private data */ | ||
| 87 : | edgomez | 1.2 | static short iclip[1024]; /* clipping table */ |
| 88 : | Isibaar | 1.1 | static short *iclp; |
| 89 : | |||
| 90 : | /* private prototypes */ | ||
| 91 : | //static void idctrow _ANSI_ARGS_((short *blk)); | ||
| 92 : | //static void idctcol _ANSI_ARGS_((short *blk)); | ||
| 93 : | |||
| 94 : | /* row (horizontal) IDCT | ||
| 95 : | * | ||
| 96 : | * 7 pi 1 | ||
| 97 : | * dst[k] = sum c[l] * src[l] * cos( -- * ( k + - ) * l ) | ||
| 98 : | * l=0 8 2 | ||
| 99 : | * | ||
| 100 : | * where: c[0] = 128 | ||
| 101 : | * c[1..7] = 128*sqrt(2) | ||
| 102 : | */ | ||
| 103 : | |||
| 104 : | /* | ||
| 105 : | static void idctrow(blk) | ||
| 106 : | short *blk; | ||
| 107 : | { | ||
| 108 : | int X0, X1, X2, X3, X4, X5, X6, X7, X8; | ||
| 109 : | |||
| 110 : | // shortcut | ||
| 111 : | if (!((X1 = blk[4]<<11) | (X2 = blk[6]) | (X3 = blk[2]) | | ||
| 112 : | (X4 = blk[1]) | (X5 = blk[7]) | (X6 = blk[5]) | (X7 = blk[3]))) | ||
| 113 : | { | ||
| 114 : | blk[0]=blk[1]=blk[2]=blk[3]=blk[4]=blk[5]=blk[6]=blk[7]=blk[0]<<3; | ||
| 115 : | return; | ||
| 116 : | } | ||
| 117 : | |||
| 118 : | X0 = (blk[0]<<11) + 128; // for proper rounding in the fourth stage | ||
| 119 : | |||
| 120 : | // first stage | ||
| 121 : | X8 = W7*(X4+X5); | ||
| 122 : | X4 = X8 + (W1-W7)*X4; | ||
| 123 : | X5 = X8 - (W1+W7)*X5; | ||
| 124 : | X8 = W3*(X6+X7); | ||
| 125 : | X6 = X8 - (W3-W5)*X6; | ||
| 126 : | X7 = X8 - (W3+W5)*X7; | ||
| 127 : | |||
| 128 : | // second stage | ||
| 129 : | X8 = X0 + X1; | ||
| 130 : | X0 -= X1; | ||
| 131 : | X1 = W6*(X3+X2); | ||
| 132 : | X2 = X1 - (W2+W6)*X2; | ||
| 133 : | X3 = X1 + (W2-W6)*X3; | ||
| 134 : | X1 = X4 + X6; | ||
| 135 : | X4 -= X6; | ||
| 136 : | X6 = X5 + X7; | ||
| 137 : | X5 -= X7; | ||
| 138 : | |||
| 139 : | // third stage | ||
| 140 : | X7 = X8 + X3; | ||
| 141 : | X8 -= X3; | ||
| 142 : | X3 = X0 + X2; | ||
| 143 : | X0 -= X2; | ||
| 144 : | X2 = (181*(X4+X5)+128)>>8; | ||
| 145 : | X4 = (181*(X4-X5)+128)>>8; | ||
| 146 : | |||
| 147 : | // fourth stage | ||
| 148 : | blk[0] = (X7+X1)>>8; | ||
| 149 : | blk[1] = (X3+X2)>>8; | ||
| 150 : | blk[2] = (X0+X4)>>8; | ||
| 151 : | blk[3] = (X8+X6)>>8; | ||
| 152 : | blk[4] = (X8-X6)>>8; | ||
| 153 : | blk[5] = (X0-X4)>>8; | ||
| 154 : | blk[6] = (X3-X2)>>8; | ||
| 155 : | blk[7] = (X7-X1)>>8; | ||
| 156 : | }*/ | ||
| 157 : | |||
| 158 : | /* column (vertical) IDCT | ||
| 159 : | * | ||
| 160 : | * 7 pi 1 | ||
| 161 : | * dst[8*k] = sum c[l] * src[8*l] * cos( -- * ( k + - ) * l ) | ||
| 162 : | * l=0 8 2 | ||
| 163 : | * | ||
| 164 : | * where: c[0] = 1/1024 | ||
| 165 : | * c[1..7] = (1/1024)*sqrt(2) | ||
| 166 : | */ | ||
| 167 : | /* | ||
| 168 : | static void idctcol(blk) | ||
| 169 : | short *blk; | ||
| 170 : | { | ||
| 171 : | int X0, X1, X2, X3, X4, X5, X6, X7, X8; | ||
| 172 : | |||
| 173 : | // shortcut | ||
| 174 : | if (!((X1 = (blk[8*4]<<8)) | (X2 = blk[8*6]) | (X3 = blk[8*2]) | | ||
| 175 : | (X4 = blk[8*1]) | (X5 = blk[8*7]) | (X6 = blk[8*5]) | (X7 = blk[8*3]))) | ||
| 176 : | { | ||
| 177 : | blk[8*0]=blk[8*1]=blk[8*2]=blk[8*3]=blk[8*4]=blk[8*5]=blk[8*6]=blk[8*7]= | ||
| 178 : | iclp[(blk[8*0]+32)>>6]; | ||
| 179 : | return; | ||
| 180 : | } | ||
| 181 : | |||
| 182 : | X0 = (blk[8*0]<<8) + 8192; | ||
| 183 : | |||
| 184 : | // first stage | ||
| 185 : | X8 = W7*(X4+X5) + 4; | ||
| 186 : | X4 = (X8+(W1-W7)*X4)>>3; | ||
| 187 : | X5 = (X8-(W1+W7)*X5)>>3; | ||
| 188 : | X8 = W3*(X6+X7) + 4; | ||
| 189 : | X6 = (X8-(W3-W5)*X6)>>3; | ||
| 190 : | X7 = (X8-(W3+W5)*X7)>>3; | ||
| 191 : | |||
| 192 : | // second stage | ||
| 193 : | X8 = X0 + X1; | ||
| 194 : | X0 -= X1; | ||
| 195 : | X1 = W6*(X3+X2) + 4; | ||
| 196 : | X2 = (X1-(W2+W6)*X2)>>3; | ||
| 197 : | X3 = (X1+(W2-W6)*X3)>>3; | ||
| 198 : | X1 = X4 + X6; | ||
| 199 : | X4 -= X6; | ||
| 200 : | X6 = X5 + X7; | ||
| 201 : | X5 -= X7; | ||
| 202 : | |||
| 203 : | // third stage | ||
| 204 : | X7 = X8 + X3; | ||
| 205 : | X8 -= X3; | ||
| 206 : | X3 = X0 + X2; | ||
| 207 : | X0 -= X2; | ||
| 208 : | X2 = (181*(X4+X5)+128)>>8; | ||
| 209 : | X4 = (181*(X4-X5)+128)>>8; | ||
| 210 : | |||
| 211 : | // fourth stage | ||
| 212 : | blk[8*0] = iclp[(X7+X1)>>14]; | ||
| 213 : | blk[8*1] = iclp[(X3+X2)>>14]; | ||
| 214 : | blk[8*2] = iclp[(X0+X4)>>14]; | ||
| 215 : | blk[8*3] = iclp[(X8+X6)>>14]; | ||
| 216 : | blk[8*4] = iclp[(X8-X6)>>14]; | ||
| 217 : | blk[8*5] = iclp[(X0-X4)>>14]; | ||
| 218 : | blk[8*6] = iclp[(X3-X2)>>14]; | ||
| 219 : | blk[8*7] = iclp[(X7-X1)>>14]; | ||
| 220 : | }*/ | ||
| 221 : | |||
| 222 : | // function pointer | ||
| 223 : | idctFuncPtr idct; | ||
| 224 : | |||
| 225 : | /* two dimensional inverse discrete cosine transform */ | ||
| 226 : | //void j_rev_dct(block) | ||
| 227 : | //short *block; | ||
| 228 : | edgomez | 1.2 | void |
| 229 : | idct_int32(short *const block) | ||
| 230 : | Isibaar | 1.1 | { |
| 231 : | |||
| 232 : | edgomez | 1.2 | // idct_int32_init() must be called before the first call to this function! |
| 233 : | Isibaar | 1.1 | |
| 234 : | |||
| 235 : | edgomez | 1.2 | /*int i; |
| 236 : | long i; | ||
| 237 : | Isibaar | 1.1 | |
| 238 : | edgomez | 1.2 | for (i=0; i<8; i++) |
| 239 : | idctrow(block+8*i); | ||
| 240 : | Isibaar | 1.1 | |
| 241 : | edgomez | 1.2 | for (i=0; i<8; i++) |
| 242 : | idctcol(block+i); */ | ||
| 243 : | static short *blk; | ||
| 244 : | static long i; | ||
| 245 : | static long X0, X1, X2, X3, X4, X5, X6, X7, X8; | ||
| 246 : | Isibaar | 1.1 | |
| 247 : | |||
| 248 : | edgomez | 1.2 | for (i = 0; i < 8; i++) // idct rows |
| 249 : | Isibaar | 1.1 | { |
| 250 : | edgomez | 1.2 | blk = block + (i << 3); |
| 251 : | if (! | ||
| 252 : | ((X1 = blk[4] << 11) | (X2 = blk[6]) | (X3 = blk[2]) | (X4 = | ||
| 253 : | blk[1]) | | ||
| 254 : | (X5 = blk[7]) | (X6 = blk[5]) | (X7 = blk[3]))) { | ||
| 255 : | blk[0] = blk[1] = blk[2] = blk[3] = blk[4] = blk[5] = blk[6] = | ||
| 256 : | blk[7] = blk[0] << 3; | ||
| 257 : | continue; | ||
| 258 : | } | ||
| 259 : | |||
| 260 : | X0 = (blk[0] << 11) + 128; // for proper rounding in the fourth stage | ||
| 261 : | |||
| 262 : | // first stage | ||
| 263 : | X8 = W7 * (X4 + X5); | ||
| 264 : | X4 = X8 + (W1 - W7) * X4; | ||
| 265 : | X5 = X8 - (W1 + W7) * X5; | ||
| 266 : | X8 = W3 * (X6 + X7); | ||
| 267 : | X6 = X8 - (W3 - W5) * X6; | ||
| 268 : | X7 = X8 - (W3 + W5) * X7; | ||
| 269 : | |||
| 270 : | // second stage | ||
| 271 : | X8 = X0 + X1; | ||
| 272 : | X0 -= X1; | ||
| 273 : | X1 = W6 * (X3 + X2); | ||
| 274 : | X2 = X1 - (W2 + W6) * X2; | ||
| 275 : | X3 = X1 + (W2 - W6) * X3; | ||
| 276 : | X1 = X4 + X6; | ||
| 277 : | X4 -= X6; | ||
| 278 : | X6 = X5 + X7; | ||
| 279 : | X5 -= X7; | ||
| 280 : | |||
| 281 : | // third stage | ||
| 282 : | X7 = X8 + X3; | ||
| 283 : | X8 -= X3; | ||
| 284 : | X3 = X0 + X2; | ||
| 285 : | X0 -= X2; | ||
| 286 : | X2 = (181 * (X4 + X5) + 128) >> 8; | ||
| 287 : | X4 = (181 * (X4 - X5) + 128) >> 8; | ||
| 288 : | |||
| 289 : | // fourth stage | ||
| 290 : | |||
| 291 : | blk[0] = (short) ((X7 + X1) >> 8); | ||
| 292 : | blk[1] = (short) ((X3 + X2) >> 8); | ||
| 293 : | blk[2] = (short) ((X0 + X4) >> 8); | ||
| 294 : | blk[3] = (short) ((X8 + X6) >> 8); | ||
| 295 : | blk[4] = (short) ((X8 - X6) >> 8); | ||
| 296 : | blk[5] = (short) ((X0 - X4) >> 8); | ||
| 297 : | blk[6] = (short) ((X3 - X2) >> 8); | ||
| 298 : | blk[7] = (short) ((X7 - X1) >> 8); | ||
| 299 : | Isibaar | 1.1 | |
| 300 : | edgomez | 1.2 | } // end for ( i = 0; i < 8; ++i ) IDCT-rows |
| 301 : | Isibaar | 1.1 | |
| 302 : | |||
| 303 : | |||
| 304 : | edgomez | 1.2 | for (i = 0; i < 8; i++) // idct columns |
| 305 : | Isibaar | 1.1 | { |
| 306 : | edgomez | 1.2 | blk = block + i; |
| 307 : | // shortcut | ||
| 308 : | if (! | ||
| 309 : | ((X1 = (blk[8 * 4] << 8)) | (X2 = blk[8 * 6]) | (X3 = | ||
| 310 : | blk[8 * | ||
| 311 : | 2]) | (X4 = | ||
| 312 : | blk[8 * | ||
| 313 : | 1]) | ||
| 314 : | | (X5 = blk[8 * 7]) | (X6 = blk[8 * 5]) | (X7 = blk[8 * 3]))) { | ||
| 315 : | blk[8 * 0] = blk[8 * 1] = blk[8 * 2] = blk[8 * 3] = blk[8 * 4] = | ||
| 316 : | blk[8 * 5] = blk[8 * 6] = blk[8 * 7] = | ||
| 317 : | iclp[(blk[8 * 0] + 32) >> 6]; | ||
| 318 : | continue; | ||
| 319 : | } | ||
| 320 : | |||
| 321 : | X0 = (blk[8 * 0] << 8) + 8192; | ||
| 322 : | |||
| 323 : | // first stage | ||
| 324 : | X8 = W7 * (X4 + X5) + 4; | ||
| 325 : | X4 = (X8 + (W1 - W7) * X4) >> 3; | ||
| 326 : | X5 = (X8 - (W1 + W7) * X5) >> 3; | ||
| 327 : | X8 = W3 * (X6 + X7) + 4; | ||
| 328 : | X6 = (X8 - (W3 - W5) * X6) >> 3; | ||
| 329 : | X7 = (X8 - (W3 + W5) * X7) >> 3; | ||
| 330 : | |||
| 331 : | // second stage | ||
| 332 : | X8 = X0 + X1; | ||
| 333 : | X0 -= X1; | ||
| 334 : | X1 = W6 * (X3 + X2) + 4; | ||
| 335 : | X2 = (X1 - (W2 + W6) * X2) >> 3; | ||
| 336 : | X3 = (X1 + (W2 - W6) * X3) >> 3; | ||
| 337 : | X1 = X4 + X6; | ||
| 338 : | X4 -= X6; | ||
| 339 : | X6 = X5 + X7; | ||
| 340 : | X5 -= X7; | ||
| 341 : | |||
| 342 : | // third stage | ||
| 343 : | X7 = X8 + X3; | ||
| 344 : | X8 -= X3; | ||
| 345 : | X3 = X0 + X2; | ||
| 346 : | X0 -= X2; | ||
| 347 : | X2 = (181 * (X4 + X5) + 128) >> 8; | ||
| 348 : | X4 = (181 * (X4 - X5) + 128) >> 8; | ||
| 349 : | |||
| 350 : | // fourth stage | ||
| 351 : | blk[8 * 0] = iclp[(X7 + X1) >> 14]; | ||
| 352 : | blk[8 * 1] = iclp[(X3 + X2) >> 14]; | ||
| 353 : | blk[8 * 2] = iclp[(X0 + X4) >> 14]; | ||
| 354 : | blk[8 * 3] = iclp[(X8 + X6) >> 14]; | ||
| 355 : | blk[8 * 4] = iclp[(X8 - X6) >> 14]; | ||
| 356 : | blk[8 * 5] = iclp[(X0 - X4) >> 14]; | ||
| 357 : | blk[8 * 6] = iclp[(X3 - X2) >> 14]; | ||
| 358 : | blk[8 * 7] = iclp[(X7 - X1) >> 14]; | ||
| 359 : | Isibaar | 1.1 | } |
| 360 : | |||
| 361 : | edgomez | 1.2 | } // end function idct_int32(block) |
| 362 : | Isibaar | 1.1 | |
| 363 : | |||
| 364 : | //void | ||
| 365 : | //idct_int32_init() | ||
| 366 : | edgomez | 1.2 | void |
| 367 : | idct_int32_init() | ||
| 368 : | Isibaar | 1.1 | { |
| 369 : | edgomez | 1.2 | int i; |
| 370 : | Isibaar | 1.1 | |
| 371 : | edgomez | 1.2 | iclp = iclip + 512; |
| 372 : | for (i = -512; i < 512; i++) | ||
| 373 : | iclp[i] = (i < -256) ? -256 : ((i > 255) ? 255 : i); | ||
| 374 : | Isibaar | 1.1 | } |
| No admin address has been configured | ViewVC Help |
| Powered by ViewVC 1.0.4 |