h264pred.c 21.2 KB
Newer Older
1 2 3 4
/*
 * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
 * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
 *
5
 * This file is part of Libav.
6
 *
7
 * Libav is free software; you can redistribute it and/or
8 9 10 11
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
12
 * Libav is distributed in the hope that it will be useful,
13 14 15 16 17
 * 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
18
 * License along with Libav; if not, write to the Free Software
19 20 21 22
 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
 */

/**
23
 * @file
24 25 26 27 28
 * H.264 / AVC / MPEG4 part10 prediction functions.
 * @author Michael Niedermayer <michaelni@gmx.at>
 */

#include "h264pred.h"
29 30 31 32 33 34

#define BIT_DEPTH 8
#include "h264pred_template.c"
#undef BIT_DEPTH

#define BIT_DEPTH 9
35
#include "h264pred_template.c"
36 37 38 39 40
#undef BIT_DEPTH

#define BIT_DEPTH 10
#include "h264pred_template.c"
#undef BIT_DEPTH
41

42 43 44
static void pred4x4_vertical_vp8_c(uint8_t *src, const uint8_t *topright,
                                   ptrdiff_t stride)
{
45
    const unsigned lt = src[-1-1*stride];
46 47
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE
48
    uint32_t v = PACK_4U8((lt + 2*t0 + t1 + 2) >> 2,
49 50 51
                          (t0 + 2*t1 + t2 + 2) >> 2,
                          (t1 + 2*t2 + t3 + 2) >> 2,
                          (t2 + 2*t3 + t4 + 2) >> 2);
52 53 54 55 56 57 58

    AV_WN32A(src+0*stride, v);
    AV_WN32A(src+1*stride, v);
    AV_WN32A(src+2*stride, v);
    AV_WN32A(src+3*stride, v);
}

59 60 61
static void pred4x4_horizontal_vp8_c(uint8_t *src, const uint8_t *topright,
                                     ptrdiff_t stride)
{
62
    const unsigned lt = src[-1-1*stride];
63 64 65 66 67 68 69 70
    LOAD_LEFT_EDGE

    AV_WN32A(src+0*stride, ((lt + 2*l0 + l1 + 2) >> 2)*0x01010101);
    AV_WN32A(src+1*stride, ((l0 + 2*l1 + l2 + 2) >> 2)*0x01010101);
    AV_WN32A(src+2*stride, ((l1 + 2*l2 + l3 + 2) >> 2)*0x01010101);
    AV_WN32A(src+3*stride, ((l2 + 2*l3 + l3 + 2) >> 2)*0x01010101);
}

71 72 73
static void pred4x4_down_left_svq3_c(uint8_t *src, const uint8_t *topright,
                                     ptrdiff_t stride)
{
74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
    LOAD_TOP_EDGE
    LOAD_LEFT_EDGE

    src[0+0*stride]=(l1 + t1)>>1;
    src[1+0*stride]=
    src[0+1*stride]=(l2 + t2)>>1;
    src[2+0*stride]=
    src[1+1*stride]=
    src[0+2*stride]=
    src[3+0*stride]=
    src[2+1*stride]=
    src[1+2*stride]=
    src[0+3*stride]=
    src[3+1*stride]=
    src[2+2*stride]=
    src[1+3*stride]=
    src[3+2*stride]=
    src[2+3*stride]=
    src[3+3*stride]=(l3 + t3)>>1;
}

95 96 97
static void pred4x4_down_left_rv40_c(uint8_t *src, const uint8_t *topright,
                                     ptrdiff_t stride)
{
98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE
    LOAD_LEFT_EDGE
    LOAD_DOWN_LEFT_EDGE

    src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
    src[1+0*stride]=
    src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
    src[2+0*stride]=
    src[1+1*stride]=
    src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + l4 + 2*l3 + 2)>>3;
    src[3+0*stride]=
    src[2+1*stride]=
    src[1+2*stride]=
    src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3 + l5 + 2*l4 + 2)>>3;
    src[3+1*stride]=
    src[2+2*stride]=
    src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l4 + l6 + 2*l5 + 2)>>3;
    src[3+2*stride]=
    src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l5 + l7 + 2*l6 + 2)>>3;
    src[3+3*stride]=(t6 + t7 + 1 + l6 + l7 + 1)>>2;
}

121 122 123 124
static void pred4x4_down_left_rv40_nodown_c(uint8_t *src,
                                            const uint8_t *topright,
                                            ptrdiff_t stride)
{
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE
    LOAD_LEFT_EDGE

    src[0+0*stride]=(t0 + t2 + 2*t1 + 2 + l0 + l2 + 2*l1 + 2)>>3;
    src[1+0*stride]=
    src[0+1*stride]=(t1 + t3 + 2*t2 + 2 + l1 + l3 + 2*l2 + 2)>>3;
    src[2+0*stride]=
    src[1+1*stride]=
    src[0+2*stride]=(t2 + t4 + 2*t3 + 2 + l2 + 3*l3 + 2)>>3;
    src[3+0*stride]=
    src[2+1*stride]=
    src[1+2*stride]=
    src[0+3*stride]=(t3 + t5 + 2*t4 + 2 + l3*4 + 2)>>3;
    src[3+1*stride]=
    src[2+2*stride]=
    src[1+3*stride]=(t4 + t6 + 2*t5 + 2 + l3*4 + 2)>>3;
    src[3+2*stride]=
    src[2+3*stride]=(t5 + t7 + 2*t6 + 2 + l3*4 + 2)>>3;
    src[3+3*stride]=(t6 + t7 + 1 + 2*l3 + 1)>>2;
}

147 148 149 150 151
static void pred4x4_vertical_left_rv40(uint8_t *src, const uint8_t *topright,
                                       ptrdiff_t stride,
                                       const int l0, const int l1, const int l2,
                                       const int l3, const int l4)
{
152 153 154 155 156 157 158 159 160 161 162
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE

    src[0+0*stride]=(2*t0 + 2*t1 + l1 + 2*l2 + l3 + 4)>>3;
    src[1+0*stride]=
    src[0+2*stride]=(t1 + t2 + 1)>>1;
    src[2+0*stride]=
    src[1+2*stride]=(t2 + t3 + 1)>>1;
    src[3+0*stride]=
    src[2+2*stride]=(t3 + t4+ 1)>>1;
    src[3+2*stride]=(t4 + t5+ 1)>>1;
163
    src[0+1*stride]=(t0 + 2*t1 + t2 + l2 + 2*l3 + l4 + 4)>>3;
164 165 166 167 168 169 170 171 172
    src[1+1*stride]=
    src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
    src[2+1*stride]=
    src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
    src[3+1*stride]=
    src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
    src[3+3*stride]=(t4 + 2*t5 + t6 + 2)>>2;
}

173 174 175
static void pred4x4_vertical_left_rv40_c(uint8_t *src, const uint8_t *topright,
                                         ptrdiff_t stride)
{
176 177 178 179 180 181
    LOAD_LEFT_EDGE
    LOAD_DOWN_LEFT_EDGE

    pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l4);
}

182 183 184 185
static void pred4x4_vertical_left_rv40_nodown_c(uint8_t *src,
                                                const uint8_t *topright,
                                                ptrdiff_t stride)
{
186 187 188 189 190
    LOAD_LEFT_EDGE

    pred4x4_vertical_left_rv40(src, topright, stride, l0, l1, l2, l3, l3);
}

191 192 193
static void pred4x4_vertical_left_vp8_c(uint8_t *src, const uint8_t *topright,
                                        ptrdiff_t stride)
{
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE

    src[0+0*stride]=(t0 + t1 + 1)>>1;
    src[1+0*stride]=
    src[0+2*stride]=(t1 + t2 + 1)>>1;
    src[2+0*stride]=
    src[1+2*stride]=(t2 + t3 + 1)>>1;
    src[3+0*stride]=
    src[2+2*stride]=(t3 + t4 + 1)>>1;
    src[0+1*stride]=(t0 + 2*t1 + t2 + 2)>>2;
    src[1+1*stride]=
    src[0+3*stride]=(t1 + 2*t2 + t3 + 2)>>2;
    src[2+1*stride]=
    src[1+3*stride]=(t2 + 2*t3 + t4 + 2)>>2;
    src[3+1*stride]=
    src[2+3*stride]=(t3 + 2*t4 + t5 + 2)>>2;
    src[3+2*stride]=(t4 + 2*t5 + t6 + 2)>>2;
    src[3+3*stride]=(t5 + 2*t6 + t7 + 2)>>2;
}

215 216 217
static void pred4x4_horizontal_up_rv40_c(uint8_t *src, const uint8_t *topright,
                                         ptrdiff_t stride)
{
218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240
    LOAD_LEFT_EDGE
    LOAD_DOWN_LEFT_EDGE
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE

    src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
    src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
    src[2+0*stride]=
    src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
    src[3+0*stride]=
    src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
    src[2+1*stride]=
    src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
    src[3+1*stride]=
    src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
    src[3+2*stride]=
    src[1+3*stride]=(l3 + 2*l4 + l5 + 2)>>2;
    src[0+3*stride]=
    src[2+2*stride]=(t6 + t7 + l3 + l4 + 2)>>2;
    src[2+3*stride]=(l4 + l5 + 1)>>1;
    src[3+3*stride]=(l4 + 2*l5 + l6 + 2)>>2;
}

241 242 243 244
static void pred4x4_horizontal_up_rv40_nodown_c(uint8_t *src,
                                                const uint8_t *topright,
                                                ptrdiff_t stride)
{
245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266
    LOAD_LEFT_EDGE
    LOAD_TOP_EDGE
    LOAD_TOP_RIGHT_EDGE

    src[0+0*stride]=(t1 + 2*t2 + t3 + 2*l0 + 2*l1 + 4)>>3;
    src[1+0*stride]=(t2 + 2*t3 + t4 + l0 + 2*l1 + l2 + 4)>>3;
    src[2+0*stride]=
    src[0+1*stride]=(t3 + 2*t4 + t5 + 2*l1 + 2*l2 + 4)>>3;
    src[3+0*stride]=
    src[1+1*stride]=(t4 + 2*t5 + t6 + l1 + 2*l2 + l3 + 4)>>3;
    src[2+1*stride]=
    src[0+2*stride]=(t5 + 2*t6 + t7 + 2*l2 + 2*l3 + 4)>>3;
    src[3+1*stride]=
    src[1+2*stride]=(t6 + 3*t7 + l2 + 3*l3 + 4)>>3;
    src[3+2*stride]=
    src[1+3*stride]=l3;
    src[0+3*stride]=
    src[2+2*stride]=(t6 + t7 + 2*l3 + 2)>>2;
    src[2+3*stride]=
    src[3+3*stride]=l3;
}

267 268 269
static void pred4x4_tm_vp8_c(uint8_t *src, const uint8_t *topright,
                             ptrdiff_t stride)
{
270 271 272 273 274 275 276 277 278 279 280 281 282 283
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP - src[-1-stride];
    uint8_t *top = src-stride;
    int y;

    for (y = 0; y < 4; y++) {
        uint8_t *cm_in = cm + src[-1];
        src[0] = cm_in[top[0]];
        src[1] = cm_in[top[1]];
        src[2] = cm_in[top[2]];
        src[3] = cm_in[top[3]];
        src += stride;
    }
}

284 285
static void pred16x16_plane_svq3_c(uint8_t *src, ptrdiff_t stride)
{
286
    pred16x16_plane_compat_8_c(src, stride, 1, 0);
287 288
}

289 290
static void pred16x16_plane_rv40_c(uint8_t *src, ptrdiff_t stride)
{
291
    pred16x16_plane_compat_8_c(src, stride, 0, 1);
292 293
}

294 295
static void pred16x16_tm_vp8_c(uint8_t *src, ptrdiff_t stride)
{
296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP - src[-1-stride];
    uint8_t *top = src-stride;
    int y;

    for (y = 0; y < 16; y++) {
        uint8_t *cm_in = cm + src[-1];
        src[0]  = cm_in[top[0]];
        src[1]  = cm_in[top[1]];
        src[2]  = cm_in[top[2]];
        src[3]  = cm_in[top[3]];
        src[4]  = cm_in[top[4]];
        src[5]  = cm_in[top[5]];
        src[6]  = cm_in[top[6]];
        src[7]  = cm_in[top[7]];
        src[8]  = cm_in[top[8]];
        src[9]  = cm_in[top[9]];
        src[10] = cm_in[top[10]];
        src[11] = cm_in[top[11]];
        src[12] = cm_in[top[12]];
        src[13] = cm_in[top[13]];
        src[14] = cm_in[top[14]];
        src[15] = cm_in[top[15]];
        src += stride;
    }
}

322 323
static void pred8x8_left_dc_rv40_c(uint8_t *src, ptrdiff_t stride)
{
324
    int i;
325
    unsigned dc0;
326 327 328 329 330 331 332 333 334 335 336 337

    dc0=0;
    for(i=0;i<8; i++)
        dc0+= src[-1+i*stride];
    dc0= 0x01010101*((dc0 + 4)>>3);

    for(i=0; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]= dc0;
    }
}

338 339
static void pred8x8_top_dc_rv40_c(uint8_t *src, ptrdiff_t stride)
{
340
    int i;
341
    unsigned dc0;
342 343 344 345 346 347 348 349 350 351 352 353

    dc0=0;
    for(i=0;i<8; i++)
        dc0+= src[i-stride];
    dc0= 0x01010101*((dc0 + 4)>>3);

    for(i=0; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]=
        ((uint32_t*)(src+i*stride))[1]= dc0;
    }
}

354 355
static void pred8x8_dc_rv40_c(uint8_t *src, ptrdiff_t stride)
{
356
    int i;
357
    unsigned dc0 = 0;
358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375

    for(i=0;i<4; i++){
        dc0+= src[-1+i*stride] + src[i-stride];
        dc0+= src[4+i-stride];
        dc0+= src[-1+(i+4)*stride];
    }
    dc0= 0x01010101*((dc0 + 8)>>4);

    for(i=0; i<4; i++){
        ((uint32_t*)(src+i*stride))[0]= dc0;
        ((uint32_t*)(src+i*stride))[1]= dc0;
    }
    for(i=4; i<8; i++){
        ((uint32_t*)(src+i*stride))[0]= dc0;
        ((uint32_t*)(src+i*stride))[1]= dc0;
    }
}

376 377
static void pred8x8_tm_vp8_c(uint8_t *src, ptrdiff_t stride)
{
378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395
    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP - src[-1-stride];
    uint8_t *top = src-stride;
    int y;

    for (y = 0; y < 8; y++) {
        uint8_t *cm_in = cm + src[-1];
        src[0] = cm_in[top[0]];
        src[1] = cm_in[top[1]];
        src[2] = cm_in[top[2]];
        src[3] = cm_in[top[3]];
        src[4] = cm_in[top[4]];
        src[5] = cm_in[top[5]];
        src[6] = cm_in[top[6]];
        src[7] = cm_in[top[7]];
        src += stride;
    }
}

396
/**
397
 * Set the intra prediction function pointers.
398
 */
399 400 401
void ff_h264_pred_init(H264PredContext *h, int codec_id, const int bit_depth,
                       const int chroma_format_idc)
{
402 403
//    MpegEncContext * const s = &h->s;

404 405 406 407 408 409 410
#undef FUNC
#undef FUNCC
#define FUNC(a, depth) a ## _ ## depth
#define FUNCC(a, depth) a ## _ ## depth ## _c
#define FUNCD(a) a ## _c

#define H264_PRED(depth) \
411 412
    if(codec_id != AV_CODEC_ID_RV40){\
        if(codec_id == AV_CODEC_ID_VP8) {\
413 414 415 416 417 418 419
            h->pred4x4[VERT_PRED       ]= FUNCD(pred4x4_vertical_vp8);\
            h->pred4x4[HOR_PRED        ]= FUNCD(pred4x4_horizontal_vp8);\
        } else {\
            h->pred4x4[VERT_PRED       ]= FUNCC(pred4x4_vertical          , depth);\
            h->pred4x4[HOR_PRED        ]= FUNCC(pred4x4_horizontal        , depth);\
        }\
        h->pred4x4[DC_PRED             ]= FUNCC(pred4x4_dc                , depth);\
420
        if(codec_id == AV_CODEC_ID_SVQ3)\
421 422 423 424 425 426
            h->pred4x4[DIAG_DOWN_LEFT_PRED ]= FUNCD(pred4x4_down_left_svq3);\
        else\
            h->pred4x4[DIAG_DOWN_LEFT_PRED ]= FUNCC(pred4x4_down_left     , depth);\
        h->pred4x4[DIAG_DOWN_RIGHT_PRED]= FUNCC(pred4x4_down_right        , depth);\
        h->pred4x4[VERT_RIGHT_PRED     ]= FUNCC(pred4x4_vertical_right    , depth);\
        h->pred4x4[HOR_DOWN_PRED       ]= FUNCC(pred4x4_horizontal_down   , depth);\
427
        if (codec_id == AV_CODEC_ID_VP8) {\
428 429 430 431
            h->pred4x4[VERT_LEFT_PRED  ]= FUNCD(pred4x4_vertical_left_vp8);\
        } else\
            h->pred4x4[VERT_LEFT_PRED  ]= FUNCC(pred4x4_vertical_left     , depth);\
        h->pred4x4[HOR_UP_PRED         ]= FUNCC(pred4x4_horizontal_up     , depth);\
432
        if(codec_id != AV_CODEC_ID_VP8) {\
433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473
            h->pred4x4[LEFT_DC_PRED    ]= FUNCC(pred4x4_left_dc           , depth);\
            h->pred4x4[TOP_DC_PRED     ]= FUNCC(pred4x4_top_dc            , depth);\
            h->pred4x4[DC_128_PRED     ]= FUNCC(pred4x4_128_dc            , depth);\
        } else {\
            h->pred4x4[TM_VP8_PRED     ]= FUNCD(pred4x4_tm_vp8);\
            h->pred4x4[DC_127_PRED     ]= FUNCC(pred4x4_127_dc            , depth);\
            h->pred4x4[DC_129_PRED     ]= FUNCC(pred4x4_129_dc            , depth);\
            h->pred4x4[VERT_VP8_PRED   ]= FUNCC(pred4x4_vertical          , depth);\
            h->pred4x4[HOR_VP8_PRED    ]= FUNCC(pred4x4_horizontal        , depth);\
        }\
    }else{\
        h->pred4x4[VERT_PRED           ]= FUNCC(pred4x4_vertical          , depth);\
        h->pred4x4[HOR_PRED            ]= FUNCC(pred4x4_horizontal        , depth);\
        h->pred4x4[DC_PRED             ]= FUNCC(pred4x4_dc                , depth);\
        h->pred4x4[DIAG_DOWN_LEFT_PRED ]= FUNCD(pred4x4_down_left_rv40);\
        h->pred4x4[DIAG_DOWN_RIGHT_PRED]= FUNCC(pred4x4_down_right        , depth);\
        h->pred4x4[VERT_RIGHT_PRED     ]= FUNCC(pred4x4_vertical_right    , depth);\
        h->pred4x4[HOR_DOWN_PRED       ]= FUNCC(pred4x4_horizontal_down   , depth);\
        h->pred4x4[VERT_LEFT_PRED      ]= FUNCD(pred4x4_vertical_left_rv40);\
        h->pred4x4[HOR_UP_PRED         ]= FUNCD(pred4x4_horizontal_up_rv40);\
        h->pred4x4[LEFT_DC_PRED        ]= FUNCC(pred4x4_left_dc           , depth);\
        h->pred4x4[TOP_DC_PRED         ]= FUNCC(pred4x4_top_dc            , depth);\
        h->pred4x4[DC_128_PRED         ]= FUNCC(pred4x4_128_dc            , depth);\
        h->pred4x4[DIAG_DOWN_LEFT_PRED_RV40_NODOWN]= FUNCD(pred4x4_down_left_rv40_nodown);\
        h->pred4x4[HOR_UP_PRED_RV40_NODOWN]= FUNCD(pred4x4_horizontal_up_rv40_nodown);\
        h->pred4x4[VERT_LEFT_PRED_RV40_NODOWN]= FUNCD(pred4x4_vertical_left_rv40_nodown);\
    }\
\
    h->pred8x8l[VERT_PRED           ]= FUNCC(pred8x8l_vertical            , depth);\
    h->pred8x8l[HOR_PRED            ]= FUNCC(pred8x8l_horizontal          , depth);\
    h->pred8x8l[DC_PRED             ]= FUNCC(pred8x8l_dc                  , depth);\
    h->pred8x8l[DIAG_DOWN_LEFT_PRED ]= FUNCC(pred8x8l_down_left           , depth);\
    h->pred8x8l[DIAG_DOWN_RIGHT_PRED]= FUNCC(pred8x8l_down_right          , depth);\
    h->pred8x8l[VERT_RIGHT_PRED     ]= FUNCC(pred8x8l_vertical_right      , depth);\
    h->pred8x8l[HOR_DOWN_PRED       ]= FUNCC(pred8x8l_horizontal_down     , depth);\
    h->pred8x8l[VERT_LEFT_PRED      ]= FUNCC(pred8x8l_vertical_left       , depth);\
    h->pred8x8l[HOR_UP_PRED         ]= FUNCC(pred8x8l_horizontal_up       , depth);\
    h->pred8x8l[LEFT_DC_PRED        ]= FUNCC(pred8x8l_left_dc             , depth);\
    h->pred8x8l[TOP_DC_PRED         ]= FUNCC(pred8x8l_top_dc              , depth);\
    h->pred8x8l[DC_128_PRED         ]= FUNCC(pred8x8l_128_dc              , depth);\
\
474 475 476 477 478 479 480
    if (chroma_format_idc == 1) {\
        h->pred8x8[VERT_PRED8x8   ]= FUNCC(pred8x8_vertical               , depth);\
        h->pred8x8[HOR_PRED8x8    ]= FUNCC(pred8x8_horizontal             , depth);\
    } else {\
        h->pred8x8[VERT_PRED8x8   ]= FUNCC(pred8x16_vertical              , depth);\
        h->pred8x8[HOR_PRED8x8    ]= FUNCC(pred8x16_horizontal            , depth);\
    }\
481
    if (codec_id != AV_CODEC_ID_VP8) {\
482 483 484 485 486
        if (chroma_format_idc == 1) {\
            h->pred8x8[PLANE_PRED8x8]= FUNCC(pred8x8_plane                , depth);\
        } else {\
            h->pred8x8[PLANE_PRED8x8]= FUNCC(pred8x16_plane               , depth);\
        }\
487 488
    } else\
        h->pred8x8[PLANE_PRED8x8]= FUNCD(pred8x8_tm_vp8);\
489
    if(codec_id != AV_CODEC_ID_RV40 && codec_id != AV_CODEC_ID_VP8){\
490 491 492 493 494 495 496 497 498 499 500 501
        if (chroma_format_idc == 1) {\
            h->pred8x8[DC_PRED8x8     ]= FUNCC(pred8x8_dc                     , depth);\
            h->pred8x8[LEFT_DC_PRED8x8]= FUNCC(pred8x8_left_dc                , depth);\
            h->pred8x8[TOP_DC_PRED8x8 ]= FUNCC(pred8x8_top_dc                 , depth);\
            h->pred8x8[ALZHEIMER_DC_L0T_PRED8x8 ]= FUNC(pred8x8_mad_cow_dc_l0t, depth);\
            h->pred8x8[ALZHEIMER_DC_0LT_PRED8x8 ]= FUNC(pred8x8_mad_cow_dc_0lt, depth);\
            h->pred8x8[ALZHEIMER_DC_L00_PRED8x8 ]= FUNC(pred8x8_mad_cow_dc_l00, depth);\
            h->pred8x8[ALZHEIMER_DC_0L0_PRED8x8 ]= FUNC(pred8x8_mad_cow_dc_0l0, depth);\
        } else {\
            h->pred8x8[DC_PRED8x8     ]= FUNCC(pred8x16_dc                    , depth);\
            h->pred8x8[LEFT_DC_PRED8x8]= FUNCC(pred8x16_left_dc               , depth);\
            h->pred8x8[TOP_DC_PRED8x8 ]= FUNCC(pred8x16_top_dc                , depth);\
502 503 504 505
            h->pred8x8[ALZHEIMER_DC_L0T_PRED8x8 ]= FUNC(pred8x16_mad_cow_dc_l0t, depth);\
            h->pred8x8[ALZHEIMER_DC_0LT_PRED8x8 ]= FUNC(pred8x16_mad_cow_dc_0lt, depth);\
            h->pred8x8[ALZHEIMER_DC_L00_PRED8x8 ]= FUNC(pred8x16_mad_cow_dc_l00, depth);\
            h->pred8x8[ALZHEIMER_DC_0L0_PRED8x8 ]= FUNC(pred8x16_mad_cow_dc_0l0, depth);\
506
        }\
507 508 509 510
    }else{\
        h->pred8x8[DC_PRED8x8     ]= FUNCD(pred8x8_dc_rv40);\
        h->pred8x8[LEFT_DC_PRED8x8]= FUNCD(pred8x8_left_dc_rv40);\
        h->pred8x8[TOP_DC_PRED8x8 ]= FUNCD(pred8x8_top_dc_rv40);\
511
        if (codec_id == AV_CODEC_ID_VP8) {\
512 513 514 515
            h->pred8x8[DC_127_PRED8x8]= FUNCC(pred8x8_127_dc              , depth);\
            h->pred8x8[DC_129_PRED8x8]= FUNCC(pred8x8_129_dc              , depth);\
        }\
    }\
516 517 518 519 520
    if (chroma_format_idc == 1) {\
        h->pred8x8[DC_128_PRED8x8 ]= FUNCC(pred8x8_128_dc                 , depth);\
    } else {\
        h->pred8x8[DC_128_PRED8x8 ]= FUNCC(pred8x16_128_dc                , depth);\
    }\
521 522 523 524 525
\
    h->pred16x16[DC_PRED8x8     ]= FUNCC(pred16x16_dc                     , depth);\
    h->pred16x16[VERT_PRED8x8   ]= FUNCC(pred16x16_vertical               , depth);\
    h->pred16x16[HOR_PRED8x8    ]= FUNCC(pred16x16_horizontal             , depth);\
    switch(codec_id){\
526
    case AV_CODEC_ID_SVQ3:\
527 528
       h->pred16x16[PLANE_PRED8x8  ]= FUNCD(pred16x16_plane_svq3);\
       break;\
529
    case AV_CODEC_ID_RV40:\
530 531
       h->pred16x16[PLANE_PRED8x8  ]= FUNCD(pred16x16_plane_rv40);\
       break;\
532
    case AV_CODEC_ID_VP8:\
533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549
       h->pred16x16[PLANE_PRED8x8  ]= FUNCD(pred16x16_tm_vp8);\
       h->pred16x16[DC_127_PRED8x8]= FUNCC(pred16x16_127_dc               , depth);\
       h->pred16x16[DC_129_PRED8x8]= FUNCC(pred16x16_129_dc               , depth);\
       break;\
    default:\
       h->pred16x16[PLANE_PRED8x8  ]= FUNCC(pred16x16_plane               , depth);\
       break;\
    }\
    h->pred16x16[LEFT_DC_PRED8x8]= FUNCC(pred16x16_left_dc                , depth);\
    h->pred16x16[TOP_DC_PRED8x8 ]= FUNCC(pred16x16_top_dc                 , depth);\
    h->pred16x16[DC_128_PRED8x8 ]= FUNCC(pred16x16_128_dc                 , depth);\
\
    /* special lossless h/v prediction for h264 */ \
    h->pred4x4_add  [VERT_PRED   ]= FUNCC(pred4x4_vertical_add            , depth);\
    h->pred4x4_add  [ HOR_PRED   ]= FUNCC(pred4x4_horizontal_add          , depth);\
    h->pred8x8l_add [VERT_PRED   ]= FUNCC(pred8x8l_vertical_add           , depth);\
    h->pred8x8l_add [ HOR_PRED   ]= FUNCC(pred8x8l_horizontal_add         , depth);\
550
    if (chroma_format_idc == 1) {\
551 552
    h->pred8x8_add  [VERT_PRED8x8]= FUNCC(pred8x8_vertical_add            , depth);\
    h->pred8x8_add  [ HOR_PRED8x8]= FUNCC(pred8x8_horizontal_add          , depth);\
553 554 555 556
    } else {\
        h->pred8x8_add  [VERT_PRED8x8]= FUNCC(pred8x16_vertical_add            , depth);\
        h->pred8x8_add  [ HOR_PRED8x8]= FUNCC(pred8x16_horizontal_add          , depth);\
    }\
557 558 559 560 561 562 563 564 565 566 567 568 569
    h->pred16x16_add[VERT_PRED8x8]= FUNCC(pred16x16_vertical_add          , depth);\
    h->pred16x16_add[ HOR_PRED8x8]= FUNCC(pred16x16_horizontal_add        , depth);\

    switch (bit_depth) {
        case 9:
            H264_PRED(9)
            break;
        case 10:
            H264_PRED(10)
            break;
        default:
            H264_PRED(8)
            break;
570 571
    }

572
    if (ARCH_ARM) ff_h264_pred_init_arm(h, codec_id, bit_depth, chroma_format_idc);
573
    if (ARCH_X86) ff_h264_pred_init_x86(h, codec_id, bit_depth, chroma_format_idc);
574
}