simple_idct.c 15.9 KB
Newer Older
1
/*
2 3 4 5
 * Simple IDCT
 *
 * Copyright (c) 2001 Michael Niedermayer <michaelni@gmx.at>
 *
6 7 8
 * This file is part of FFmpeg.
 *
 * FFmpeg is free software; you can redistribute it and/or
9 10
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
11
 * version 2.1 of the License, or (at your option) any later version.
12
 *
13
 * FFmpeg is distributed in the hope that it will be useful,
14 15 16 17 18
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
19
 * License along with FFmpeg; if not, write to the Free Software
20
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21
 */
22

Michael Niedermayer's avatar
Michael Niedermayer committed
23 24 25 26
/**
 * @file simple_idct.c
 * simpleidct in C.
 */
27

28
/*
29
  based upon some outcommented c code from mpeg2dec (idct_mmx.c
30
  written by Aaron Holtzman <aholtzma@ess.engr.uvic.ca>)
31
 */
32
#include "avcodec.h"
33
#include "dsputil.h"
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49
#include "simple_idct.h"

#if 0
#define W1 2841 /* 2048*sqrt (2)*cos (1*pi/16) */
#define W2 2676 /* 2048*sqrt (2)*cos (2*pi/16) */
#define W3 2408 /* 2048*sqrt (2)*cos (3*pi/16) */
#define W4 2048 /* 2048*sqrt (2)*cos (4*pi/16) */
#define W5 1609 /* 2048*sqrt (2)*cos (5*pi/16) */
#define W6 1108 /* 2048*sqrt (2)*cos (6*pi/16) */
#define W7 565  /* 2048*sqrt (2)*cos (7*pi/16) */
#define ROW_SHIFT 8
#define COL_SHIFT 17
#else
#define W1  22725  //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W2  21407  //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W3  19266  //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
50
#define W4  16383  //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
51 52 53 54 55 56
#define W5  12873  //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W6  8867   //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define W7  4520   //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
#define ROW_SHIFT 11
#define COL_SHIFT 20 // 6
#endif
57

58
#if defined(ARCH_POWERPC_405)
59

60 61 62
/* signed 16x16 -> 32 multiply add accumulate */
#define MAC16(rt, ra, rb) \
    asm ("maclhw %0, %2, %3" : "=r" (rt) : "0" (rt), "r" (ra), "r" (rb));
63

64 65 66 67 68 69 70 71 72 73 74 75 76
/* signed 16x16 -> 32 multiply */
#define MUL16(rt, ra, rb) \
    asm ("mullhw %0, %1, %2" : "=r" (rt) : "r" (ra), "r" (rb));

#else

/* signed 16x16 -> 32 multiply add accumulate */
#define MAC16(rt, ra, rb) rt += (ra) * (rb)

/* signed 16x16 -> 32 multiply */
#define MUL16(rt, ra, rb) rt = (ra) * (rb)

#endif
77

78
static inline void idctRowCondDC (DCTELEM * row)
79
{
80
        int a0, a1, a2, a3, b0, b1, b2, b3;
81
#ifdef HAVE_FAST_64BIT
82 83 84 85
        uint64_t temp;
#else
        uint32_t temp;
#endif
86

87
#ifdef HAVE_FAST_64BIT
88 89 90 91 92
#ifdef WORDS_BIGENDIAN
#define ROW0_MASK 0xffff000000000000LL
#else
#define ROW0_MASK 0xffffLL
#endif
93
        if(sizeof(DCTELEM)==2){
94
            if ( ((((uint64_t *)row)[0] & ~ROW0_MASK) |
95 96 97 98 99 100 101
                  ((uint64_t *)row)[1]) == 0) {
                temp = (row[0] << 3) & 0xffff;
                temp += temp << 16;
                temp += temp << 32;
                ((uint64_t *)row)[0] = temp;
                ((uint64_t *)row)[1] = temp;
                return;
102
            }
103 104 105 106 107 108
        }else{
            if (!(row[1]|row[2]|row[3]|row[4]|row[5]|row[6]|row[7])) {
                row[0]=row[1]=row[2]=row[3]=row[4]=row[5]=row[6]=row[7]= row[0] << 3;
                return;
            }
        }
109
#else
110 111 112
        if(sizeof(DCTELEM)==2){
            if (!(((uint32_t*)row)[1] |
                  ((uint32_t*)row)[2] |
113
                  ((uint32_t*)row)[3] |
114 115 116 117 118 119 120 121 122 123 124 125 126
                  row[1])) {
                temp = (row[0] << 3) & 0xffff;
                temp += temp << 16;
                ((uint32_t*)row)[0]=((uint32_t*)row)[1] =
                ((uint32_t*)row)[2]=((uint32_t*)row)[3] = temp;
                return;
            }
        }else{
            if (!(row[1]|row[2]|row[3]|row[4]|row[5]|row[6]|row[7])) {
                row[0]=row[1]=row[2]=row[3]=row[4]=row[5]=row[6]=row[7]= row[0] << 3;
                return;
            }
        }
127
#endif
128

129
        a0 = (W4 * row[0]) + (1 << (ROW_SHIFT - 1));
130 131 132
        a1 = a0;
        a2 = a0;
        a3 = a0;
133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148

        /* no need to optimize : gcc does it */
        a0 += W2 * row[2];
        a1 += W6 * row[2];
        a2 -= W6 * row[2];
        a3 -= W2 * row[2];

        MUL16(b0, W1, row[1]);
        MAC16(b0, W3, row[3]);
        MUL16(b1, W3, row[1]);
        MAC16(b1, -W7, row[3]);
        MUL16(b2, W5, row[1]);
        MAC16(b2, -W1, row[3]);
        MUL16(b3, W7, row[1]);
        MAC16(b3, -W5, row[3]);

149
#ifdef HAVE_FAST_64BIT
150 151 152 153
        temp = ((uint64_t*)row)[1];
#else
        temp = ((uint32_t*)row)[2] | ((uint32_t*)row)[3];
#endif
154
        if (temp != 0) {
155 156 157 158 159 160 161
            a0 += W4*row[4] + W6*row[6];
            a1 += - W4*row[4] - W2*row[6];
            a2 += - W4*row[4] + W2*row[6];
            a3 += W4*row[4] - W6*row[6];

            MAC16(b0, W5, row[5]);
            MAC16(b0, W7, row[7]);
162

163 164
            MAC16(b1, -W1, row[5]);
            MAC16(b1, -W5, row[7]);
165

166 167
            MAC16(b2, W7, row[5]);
            MAC16(b2, W3, row[7]);
168

169 170
            MAC16(b3, W3, row[5]);
            MAC16(b3, -W1, row[7]);
171 172 173 174 175 176 177 178 179 180
        }

        row[0] = (a0 + b0) >> ROW_SHIFT;
        row[7] = (a0 - b0) >> ROW_SHIFT;
        row[1] = (a1 + b1) >> ROW_SHIFT;
        row[6] = (a1 - b1) >> ROW_SHIFT;
        row[2] = (a2 + b2) >> ROW_SHIFT;
        row[5] = (a2 - b2) >> ROW_SHIFT;
        row[3] = (a3 + b3) >> ROW_SHIFT;
        row[4] = (a3 - b3) >> ROW_SHIFT;
181 182
}

183
static inline void idctSparseColPut (uint8_t *dest, int line_size,
184
                                     DCTELEM * col)
185
{
186
        int a0, a1, a2, a3, b0, b1, b2, b3;
187
        uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
188 189

        /* XXX: I did that only to give same values as previous code */
190 191 192 193
        a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4));
        a1 = a0;
        a2 = a0;
        a3 = a0;
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209

        a0 +=  + W2*col[8*2];
        a1 +=  + W6*col[8*2];
        a2 +=  - W6*col[8*2];
        a3 +=  - W2*col[8*2];

        MUL16(b0, W1, col[8*1]);
        MUL16(b1, W3, col[8*1]);
        MUL16(b2, W5, col[8*1]);
        MUL16(b3, W7, col[8*1]);

        MAC16(b0, + W3, col[8*3]);
        MAC16(b1, - W7, col[8*3]);
        MAC16(b2, - W1, col[8*3]);
        MAC16(b3, - W5, col[8*3]);

210
        if(col[8*4]){
211 212 213 214
            a0 += + W4*col[8*4];
            a1 += - W4*col[8*4];
            a2 += - W4*col[8*4];
            a3 += + W4*col[8*4];
215
        }
216

217
        if (col[8*5]) {
218 219 220 221
            MAC16(b0, + W5, col[8*5]);
            MAC16(b1, - W1, col[8*5]);
            MAC16(b2, + W7, col[8*5]);
            MAC16(b3, + W3, col[8*5]);
222
        }
223

224
        if(col[8*6]){
225 226 227 228
            a0 += + W6*col[8*6];
            a1 += - W2*col[8*6];
            a2 += + W2*col[8*6];
            a3 += - W6*col[8*6];
229
        }
230

231
        if (col[8*7]) {
232 233 234 235
            MAC16(b0, + W7, col[8*7]);
            MAC16(b1, - W5, col[8*7]);
            MAC16(b2, + W3, col[8*7]);
            MAC16(b3, - W1, col[8*7]);
236
        }
237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254

        dest[0] = cm[(a0 + b0) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a1 + b1) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a2 + b2) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a3 + b3) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a3 - b3) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a2 - b2) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a1 - b1) >> COL_SHIFT];
        dest += line_size;
        dest[0] = cm[(a0 - b0) >> COL_SHIFT];
}

255
static inline void idctSparseColAdd (uint8_t *dest, int line_size,
256
                                     DCTELEM * col)
257
{
258
        int a0, a1, a2, a3, b0, b1, b2, b3;
259
        uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
260

261
        /* XXX: I did that only to give same values as previous code */
262 263 264 265
        a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4));
        a1 = a0;
        a2 = a0;
        a3 = a0;
266

267 268 269 270
        a0 +=  + W2*col[8*2];
        a1 +=  + W6*col[8*2];
        a2 +=  - W6*col[8*2];
        a3 +=  - W2*col[8*2];
271

272 273 274 275
        MUL16(b0, W1, col[8*1]);
        MUL16(b1, W3, col[8*1]);
        MUL16(b2, W5, col[8*1]);
        MUL16(b3, W7, col[8*1]);
276

277 278 279 280
        MAC16(b0, + W3, col[8*3]);
        MAC16(b1, - W7, col[8*3]);
        MAC16(b2, - W1, col[8*3]);
        MAC16(b3, - W5, col[8*3]);
281

282
        if(col[8*4]){
283 284 285 286
            a0 += + W4*col[8*4];
            a1 += - W4*col[8*4];
            a2 += - W4*col[8*4];
            a3 += + W4*col[8*4];
287
        }
288

289
        if (col[8*5]) {
290 291 292 293
            MAC16(b0, + W5, col[8*5]);
            MAC16(b1, - W1, col[8*5]);
            MAC16(b2, + W7, col[8*5]);
            MAC16(b3, + W3, col[8*5]);
294
        }
295

296
        if(col[8*6]){
297 298 299 300
            a0 += + W6*col[8*6];
            a1 += - W2*col[8*6];
            a2 += + W2*col[8*6];
            a3 += - W6*col[8*6];
301
        }
302

303
        if (col[8*7]) {
304 305 306 307
            MAC16(b0, + W7, col[8*7]);
            MAC16(b1, - W5, col[8*7]);
            MAC16(b2, + W3, col[8*7]);
            MAC16(b3, - W1, col[8*7]);
308
        }
309

310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
        dest[0] = cm[dest[0] + ((a0 + b0) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a1 + b1) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a2 + b2) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a3 + b3) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a3 - b3) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a2 - b2) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a1 - b1) >> COL_SHIFT)];
        dest += line_size;
        dest[0] = cm[dest[0] + ((a0 - b0) >> COL_SHIFT)];
325 326
}

327
static inline void idctSparseCol (DCTELEM * col)
Michael Niedermayer's avatar
Michael Niedermayer committed
328
{
329
        int a0, a1, a2, a3, b0, b1, b2, b3;
Michael Niedermayer's avatar
Michael Niedermayer committed
330 331

        /* XXX: I did that only to give same values as previous code */
332 333 334 335
        a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4));
        a1 = a0;
        a2 = a0;
        a3 = a0;
Michael Niedermayer's avatar
Michael Niedermayer committed
336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351

        a0 +=  + W2*col[8*2];
        a1 +=  + W6*col[8*2];
        a2 +=  - W6*col[8*2];
        a3 +=  - W2*col[8*2];

        MUL16(b0, W1, col[8*1]);
        MUL16(b1, W3, col[8*1]);
        MUL16(b2, W5, col[8*1]);
        MUL16(b3, W7, col[8*1]);

        MAC16(b0, + W3, col[8*3]);
        MAC16(b1, - W7, col[8*3]);
        MAC16(b2, - W1, col[8*3]);
        MAC16(b3, - W5, col[8*3]);

352
        if(col[8*4]){
Michael Niedermayer's avatar
Michael Niedermayer committed
353 354 355 356
            a0 += + W4*col[8*4];
            a1 += - W4*col[8*4];
            a2 += - W4*col[8*4];
            a3 += + W4*col[8*4];
357
        }
Michael Niedermayer's avatar
Michael Niedermayer committed
358

359
        if (col[8*5]) {
Michael Niedermayer's avatar
Michael Niedermayer committed
360 361 362 363
            MAC16(b0, + W5, col[8*5]);
            MAC16(b1, - W1, col[8*5]);
            MAC16(b2, + W7, col[8*5]);
            MAC16(b3, + W3, col[8*5]);
364
        }
Michael Niedermayer's avatar
Michael Niedermayer committed
365

366
        if(col[8*6]){
Michael Niedermayer's avatar
Michael Niedermayer committed
367 368 369 370
            a0 += + W6*col[8*6];
            a1 += - W2*col[8*6];
            a2 += + W2*col[8*6];
            a3 += - W6*col[8*6];
371
        }
Michael Niedermayer's avatar
Michael Niedermayer committed
372

373
        if (col[8*7]) {
Michael Niedermayer's avatar
Michael Niedermayer committed
374 375 376 377
            MAC16(b0, + W7, col[8*7]);
            MAC16(b1, - W5, col[8*7]);
            MAC16(b2, + W3, col[8*7]);
            MAC16(b3, - W1, col[8*7]);
378
        }
Michael Niedermayer's avatar
Michael Niedermayer committed
379 380 381 382 383 384 385 386 387 388 389

        col[0 ] = ((a0 + b0) >> COL_SHIFT);
        col[8 ] = ((a1 + b1) >> COL_SHIFT);
        col[16] = ((a2 + b2) >> COL_SHIFT);
        col[24] = ((a3 + b3) >> COL_SHIFT);
        col[32] = ((a3 - b3) >> COL_SHIFT);
        col[40] = ((a2 - b2) >> COL_SHIFT);
        col[48] = ((a1 - b1) >> COL_SHIFT);
        col[56] = ((a0 - b0) >> COL_SHIFT);
}

390
void simple_idct_put(uint8_t *dest, int line_size, DCTELEM *block)
391 392 393 394
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);
395

396 397 398 399
    for(i=0; i<8; i++)
        idctSparseColPut(dest + i, line_size, block + i);
}

400
void simple_idct_add(uint8_t *dest, int line_size, DCTELEM *block)
401 402 403 404
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);
405

406
    for(i=0; i<8; i++)
407
        idctSparseColAdd(dest + i, line_size, block + i);
408
}
409

410
void simple_idct(DCTELEM *block)
Michael Niedermayer's avatar
Michael Niedermayer committed
411 412 413 414
{
    int i;
    for(i=0; i<8; i++)
        idctRowCondDC(block + i*8);
415

Michael Niedermayer's avatar
Michael Niedermayer committed
416 417 418 419
    for(i=0; i<8; i++)
        idctSparseCol(block + i);
}

420 421 422 423
/* 2x4x8 idct */

#define CN_SHIFT 12
#define C_FIX(x) ((int)((x) * (1 << CN_SHIFT) + 0.5))
424 425
#define C1 C_FIX(0.6532814824)
#define C2 C_FIX(0.2705980501)
426

427 428 429
/* row idct is multiple by 16 * sqrt(2.0), col idct4 is normalized,
   and the butterfly must be multiplied by 0.5 * sqrt(2.0) */
#define C_SHIFT (4+1+12)
430

431
static inline void idct4col(uint8_t *dest, int line_size, const DCTELEM *col)
432 433
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
434
    const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
435 436 437 438 439

    a0 = col[8*0];
    a1 = col[8*2];
    a2 = col[8*4];
    a3 = col[8*6];
440 441
    c0 = ((a0 + a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
    c2 = ((a0 - a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467
    c1 = a1 * C1 + a3 * C2;
    c3 = a1 * C2 - a3 * C1;
    dest[0] = cm[(c0 + c1) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c2 + c3) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c2 - c3) >> C_SHIFT];
    dest += line_size;
    dest[0] = cm[(c0 - c1) >> C_SHIFT];
}

#define BF(k) \
{\
    int a0, a1;\
    a0 = ptr[k];\
    a1 = ptr[8 + k];\
    ptr[k] = a0 + a1;\
    ptr[8 + k] = a0 - a1;\
}

/* only used by DV codec. The input must be interlaced. 128 is added
   to the pixels before clamping to avoid systematic error
   (1024*sqrt(2)) offset would be needed otherwise. */
/* XXX: I think a 1.0/sqrt(2) normalization should be needed to
   compensate the extra butterfly stage - I don't have the full DV
   specification */
468
void simple_idct248_put(uint8_t *dest, int line_size, DCTELEM *block)
469 470
{
    int i;
471
    DCTELEM *ptr;
472

473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497
    /* butterfly */
    ptr = block;
    for(i=0;i<4;i++) {
        BF(0);
        BF(1);
        BF(2);
        BF(3);
        BF(4);
        BF(5);
        BF(6);
        BF(7);
        ptr += 2 * 8;
    }

    /* IDCT8 on each line */
    for(i=0; i<8; i++) {
        idctRowCondDC(block + i*8);
    }

    /* IDCT4 and store */
    for(i=0;i<8;i++) {
        idct4col(dest + i, 2 * line_size, block + i);
        idct4col(dest + line_size + i, 2 * line_size, block + 8 + i);
    }
}
Michael Niedermayer's avatar
Michael Niedermayer committed
498 499 500 501 502 503 504 505 506 507 508 509 510

/* 8x4 & 4x8 WMV2 IDCT */
#undef CN_SHIFT
#undef C_SHIFT
#undef C_FIX
#undef C1
#undef C2
#define CN_SHIFT 12
#define C_FIX(x) ((int)((x) * 1.414213562 * (1 << CN_SHIFT) + 0.5))
#define C1 C_FIX(0.6532814824)
#define C2 C_FIX(0.2705980501)
#define C3 C_FIX(0.5)
#define C_SHIFT (4+1+12)
511
static inline void idct4col_add(uint8_t *dest, int line_size, const DCTELEM *col)
Michael Niedermayer's avatar
Michael Niedermayer committed
512 513
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
514
    const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
Michael Niedermayer's avatar
Michael Niedermayer committed
515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538

    a0 = col[8*0];
    a1 = col[8*1];
    a2 = col[8*2];
    a3 = col[8*3];
    c0 = (a0 + a2)*C3 + (1 << (C_SHIFT - 1));
    c2 = (a0 - a2)*C3 + (1 << (C_SHIFT - 1));
    c1 = a1 * C1 + a3 * C2;
    c3 = a1 * C2 - a3 * C1;
    dest[0] = cm[dest[0] + ((c0 + c1) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c2 + c3) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c2 - c3) >> C_SHIFT)];
    dest += line_size;
    dest[0] = cm[dest[0] + ((c0 - c1) >> C_SHIFT)];
}

#define RN_SHIFT 15
#define R_FIX(x) ((int)((x) * 1.414213562 * (1 << RN_SHIFT) + 0.5))
#define R1 R_FIX(0.6532814824)
#define R2 R_FIX(0.2705980501)
#define R3 R_FIX(0.5)
#define R_SHIFT 11
539
static inline void idct4row(DCTELEM *row)
Michael Niedermayer's avatar
Michael Niedermayer committed
540 541
{
    int c0, c1, c2, c3, a0, a1, a2, a3;
542
    //const uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
Michael Niedermayer's avatar
Michael Niedermayer committed
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557

    a0 = row[0];
    a1 = row[1];
    a2 = row[2];
    a3 = row[3];
    c0 = (a0 + a2)*R3 + (1 << (R_SHIFT - 1));
    c2 = (a0 - a2)*R3 + (1 << (R_SHIFT - 1));
    c1 = a1 * R1 + a3 * R2;
    c3 = a1 * R2 - a3 * R1;
    row[0]= (c0 + c1) >> R_SHIFT;
    row[1]= (c2 + c3) >> R_SHIFT;
    row[2]= (c2 - c3) >> R_SHIFT;
    row[3]= (c0 - c1) >> R_SHIFT;
}

558
void simple_idct84_add(uint8_t *dest, int line_size, DCTELEM *block)
Michael Niedermayer's avatar
Michael Niedermayer committed
559 560 561 562 563 564 565 566 567 568 569 570 571 572
{
    int i;

    /* IDCT8 on each line */
    for(i=0; i<4; i++) {
        idctRowCondDC(block + i*8);
    }

    /* IDCT4 and store */
    for(i=0;i<8;i++) {
        idct4col_add(dest + i, line_size, block + i);
    }
}

573
void simple_idct48_add(uint8_t *dest, int line_size, DCTELEM *block)
Michael Niedermayer's avatar
Michael Niedermayer committed
574 575 576 577 578 579 580 581 582 583 584 585 586 587
{
    int i;

    /* IDCT4 on each line */
    for(i=0; i<8; i++) {
        idct4row(block + i*8);
    }

    /* IDCT8 and store */
    for(i=0; i<4; i++){
        idctSparseColAdd(dest + i, line_size, block + i);
    }
}