summaryrefslogtreecommitdiffstats
path: root/media/libvpx/vp8/common/reconinter.c
blob: e3025955871319b9347b7c45407180206f2f5444 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
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
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
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
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
/*
 *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */


#include <limits.h>
#include <string.h>

#include "vpx_config.h"
#include "vp8_rtcd.h"
#include "vpx/vpx_integer.h"
#include "blockd.h"
#include "reconinter.h"
#if CONFIG_RUNTIME_CPU_DETECT
#include "onyxc_int.h"
#endif

void vp8_copy_mem16x16_c(
    unsigned char *src,
    int src_stride,
    unsigned char *dst,
    int dst_stride)
{

    int r;

    for (r = 0; r < 16; r++)
    {
        memcpy(dst, src, 16);

        src += src_stride;
        dst += dst_stride;

    }

}

void vp8_copy_mem8x8_c(
    unsigned char *src,
    int src_stride,
    unsigned char *dst,
    int dst_stride)
{
    int r;

    for (r = 0; r < 8; r++)
    {
        memcpy(dst, src, 8);

        src += src_stride;
        dst += dst_stride;

    }

}

void vp8_copy_mem8x4_c(
    unsigned char *src,
    int src_stride,
    unsigned char *dst,
    int dst_stride)
{
    int r;

    for (r = 0; r < 4; r++)
    {
        memcpy(dst, src, 8);

        src += src_stride;
        dst += dst_stride;

    }

}


void vp8_build_inter_predictors_b(BLOCKD *d, int pitch, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
{
    int r;
    unsigned char *pred_ptr = d->predictor;
    unsigned char *ptr;
    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);

    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
    {
        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, pred_ptr, pitch);
    }
    else
    {
        for (r = 0; r < 4; r++)
        {
            pred_ptr[0]  = ptr[0];
            pred_ptr[1]  = ptr[1];
            pred_ptr[2]  = ptr[2];
            pred_ptr[3]  = ptr[3];
            pred_ptr     += pitch;
            ptr         += pre_stride;
        }
    }
}

static void build_inter_predictors4b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
{
    unsigned char *ptr;
    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);

    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
    {
        x->subpixel_predict8x8(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
    }
    else
    {
        vp8_copy_mem8x8(ptr, pre_stride, dst, dst_stride);
    }
}

static void build_inter_predictors2b(MACROBLOCKD *x, BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride)
{
    unsigned char *ptr;
    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);

    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
    {
        x->subpixel_predict8x4(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
    }
    else
    {
        vp8_copy_mem8x4(ptr, pre_stride, dst, dst_stride);
    }
}

static void build_inter_predictors_b(BLOCKD *d, unsigned char *dst, int dst_stride, unsigned char *base_pre, int pre_stride, vp8_subpix_fn_t sppf)
{
    int r;
    unsigned char *ptr;
    ptr = base_pre + d->offset + (d->bmi.mv.as_mv.row >> 3) * pre_stride + (d->bmi.mv.as_mv.col >> 3);

    if (d->bmi.mv.as_mv.row & 7 || d->bmi.mv.as_mv.col & 7)
    {
        sppf(ptr, pre_stride, d->bmi.mv.as_mv.col & 7, d->bmi.mv.as_mv.row & 7, dst, dst_stride);
    }
    else
    {
        for (r = 0; r < 4; r++)
        {
          dst[0]  = ptr[0];
          dst[1]  = ptr[1];
          dst[2]  = ptr[2];
          dst[3]  = ptr[3];
          dst     += dst_stride;
          ptr     += pre_stride;
        }
    }
}


/*encoder only*/
void vp8_build_inter16x16_predictors_mbuv(MACROBLOCKD *x)
{
    unsigned char *uptr, *vptr;
    unsigned char *upred_ptr = &x->predictor[256];
    unsigned char *vpred_ptr = &x->predictor[320];

    int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
    int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
    int offset;
    int pre_stride = x->pre.uv_stride;

    /* calc uv motion vectors */
    mv_row += 1 | (mv_row >> (sizeof(int) * CHAR_BIT - 1));
    mv_col += 1 | (mv_col >> (sizeof(int) * CHAR_BIT - 1));
    mv_row /= 2;
    mv_col /= 2;
    mv_row &= x->fullpixel_mask;
    mv_col &= x->fullpixel_mask;

    offset = (mv_row >> 3) * pre_stride + (mv_col >> 3);
    uptr = x->pre.u_buffer + offset;
    vptr = x->pre.v_buffer + offset;

    if ((mv_row | mv_col) & 7)
    {
        x->subpixel_predict8x8(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr, 8);
        x->subpixel_predict8x8(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr, 8);
    }
    else
    {
        vp8_copy_mem8x8(uptr, pre_stride, upred_ptr, 8);
        vp8_copy_mem8x8(vptr, pre_stride, vpred_ptr, 8);
    }
}

/*encoder only*/
void vp8_build_inter4x4_predictors_mbuv(MACROBLOCKD *x)
{
    int i, j;
    int pre_stride = x->pre.uv_stride;
    unsigned char *base_pre;

    /* build uv mvs */
    for (i = 0; i < 2; i++)
    {
        for (j = 0; j < 2; j++)
        {
            int yoffset = i * 8 + j * 2;
            int uoffset = 16 + i * 2 + j;
            int voffset = 20 + i * 2 + j;

            int temp;

            temp = x->block[yoffset  ].bmi.mv.as_mv.row
                   + x->block[yoffset+1].bmi.mv.as_mv.row
                   + x->block[yoffset+4].bmi.mv.as_mv.row
                   + x->block[yoffset+5].bmi.mv.as_mv.row;

            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

            x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;

            temp = x->block[yoffset  ].bmi.mv.as_mv.col
                   + x->block[yoffset+1].bmi.mv.as_mv.col
                   + x->block[yoffset+4].bmi.mv.as_mv.col
                   + x->block[yoffset+5].bmi.mv.as_mv.col;

            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

            x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;

            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
        }
    }

    base_pre = x->pre.u_buffer;
    for (i = 16; i < 20; i += 2)
    {
        BLOCKD *d0 = &x->block[i];
        BLOCKD *d1 = &x->block[i+1];

        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
        else
        {
            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
        }
    }

    base_pre = x->pre.v_buffer;
    for (i = 20; i < 24; i += 2)
    {
        BLOCKD *d0 = &x->block[i];
        BLOCKD *d1 = &x->block[i+1];

        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
            build_inter_predictors2b(x, d0, d0->predictor, 8, base_pre, pre_stride);
        else
        {
            vp8_build_inter_predictors_b(d0, 8, base_pre, pre_stride, x->subpixel_predict);
            vp8_build_inter_predictors_b(d1, 8, base_pre, pre_stride, x->subpixel_predict);
        }
    }
}


/*encoder only*/
void vp8_build_inter16x16_predictors_mby(MACROBLOCKD *x,
                                         unsigned char *dst_y,
                                         int dst_ystride)
{
    unsigned char *ptr_base;
    unsigned char *ptr;
    int mv_row = x->mode_info_context->mbmi.mv.as_mv.row;
    int mv_col = x->mode_info_context->mbmi.mv.as_mv.col;
    int pre_stride = x->pre.y_stride;

    ptr_base = x->pre.y_buffer;
    ptr = ptr_base + (mv_row >> 3) * pre_stride + (mv_col >> 3);

    if ((mv_row | mv_col) & 7)
    {
        x->subpixel_predict16x16(ptr, pre_stride, mv_col & 7, mv_row & 7,
                                 dst_y, dst_ystride);
    }
    else
    {
        vp8_copy_mem16x16(ptr, pre_stride, dst_y,
            dst_ystride);
    }
}

static void clamp_mv_to_umv_border(MV *mv, const MACROBLOCKD *xd)
{
    /* If the MV points so far into the UMV border that no visible pixels
     * are used for reconstruction, the subpel part of the MV can be
     * discarded and the MV limited to 16 pixels with equivalent results.
     *
     * This limit kicks in at 19 pixels for the top and left edges, for
     * the 16 pixels plus 3 taps right of the central pixel when subpel
     * filtering. The bottom and right edges use 16 pixels plus 2 pixels
     * left of the central pixel when filtering.
     */
    if (mv->col < (xd->mb_to_left_edge - (19 << 3)))
        mv->col = xd->mb_to_left_edge - (16 << 3);
    else if (mv->col > xd->mb_to_right_edge + (18 << 3))
        mv->col = xd->mb_to_right_edge + (16 << 3);

    if (mv->row < (xd->mb_to_top_edge - (19 << 3)))
        mv->row = xd->mb_to_top_edge - (16 << 3);
    else if (mv->row > xd->mb_to_bottom_edge + (18 << 3))
        mv->row = xd->mb_to_bottom_edge + (16 << 3);
}

/* A version of the above function for chroma block MVs.*/
static void clamp_uvmv_to_umv_border(MV *mv, const MACROBLOCKD *xd)
{
    mv->col = (2*mv->col < (xd->mb_to_left_edge - (19 << 3))) ?
        (xd->mb_to_left_edge - (16 << 3)) >> 1 : mv->col;
    mv->col = (2*mv->col > xd->mb_to_right_edge + (18 << 3)) ?
        (xd->mb_to_right_edge + (16 << 3)) >> 1 : mv->col;

    mv->row = (2*mv->row < (xd->mb_to_top_edge - (19 << 3))) ?
        (xd->mb_to_top_edge - (16 << 3)) >> 1 : mv->row;
    mv->row = (2*mv->row > xd->mb_to_bottom_edge + (18 << 3)) ?
        (xd->mb_to_bottom_edge + (16 << 3)) >> 1 : mv->row;
}

void vp8_build_inter16x16_predictors_mb(MACROBLOCKD *x,
                                        unsigned char *dst_y,
                                        unsigned char *dst_u,
                                        unsigned char *dst_v,
                                        int dst_ystride,
                                        int dst_uvstride)
{
    int offset;
    unsigned char *ptr;
    unsigned char *uptr, *vptr;

    int_mv _16x16mv;

    unsigned char *ptr_base = x->pre.y_buffer;
    int pre_stride = x->pre.y_stride;

    _16x16mv.as_int = x->mode_info_context->mbmi.mv.as_int;

    if (x->mode_info_context->mbmi.need_to_clamp_mvs)
    {
        clamp_mv_to_umv_border(&_16x16mv.as_mv, x);
    }

    ptr = ptr_base + ( _16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);

    if ( _16x16mv.as_int & 0x00070007)
    {
        x->subpixel_predict16x16(ptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_y, dst_ystride);
    }
    else
    {
        vp8_copy_mem16x16(ptr, pre_stride, dst_y, dst_ystride);
    }

    /* calc uv motion vectors */
    _16x16mv.as_mv.row += 1 | (_16x16mv.as_mv.row >> (sizeof(int) * CHAR_BIT - 1));
    _16x16mv.as_mv.col += 1 | (_16x16mv.as_mv.col >> (sizeof(int) * CHAR_BIT - 1));
    _16x16mv.as_mv.row /= 2;
    _16x16mv.as_mv.col /= 2;
    _16x16mv.as_mv.row &= x->fullpixel_mask;
    _16x16mv.as_mv.col &= x->fullpixel_mask;

    pre_stride >>= 1;
    offset = ( _16x16mv.as_mv.row >> 3) * pre_stride + (_16x16mv.as_mv.col >> 3);
    uptr = x->pre.u_buffer + offset;
    vptr = x->pre.v_buffer + offset;

    if ( _16x16mv.as_int & 0x00070007)
    {
        x->subpixel_predict8x8(uptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_u, dst_uvstride);
        x->subpixel_predict8x8(vptr, pre_stride, _16x16mv.as_mv.col & 7,  _16x16mv.as_mv.row & 7, dst_v, dst_uvstride);
    }
    else
    {
        vp8_copy_mem8x8(uptr, pre_stride, dst_u, dst_uvstride);
        vp8_copy_mem8x8(vptr, pre_stride, dst_v, dst_uvstride);
    }
}

static void build_inter4x4_predictors_mb(MACROBLOCKD *x)
{
    int i;
    unsigned char *base_dst = x->dst.y_buffer;
    unsigned char *base_pre = x->pre.y_buffer;

    if (x->mode_info_context->mbmi.partitioning < 3)
    {
        BLOCKD *b;
        int dst_stride = x->dst.y_stride;

        x->block[ 0].bmi = x->mode_info_context->bmi[ 0];
        x->block[ 2].bmi = x->mode_info_context->bmi[ 2];
        x->block[ 8].bmi = x->mode_info_context->bmi[ 8];
        x->block[10].bmi = x->mode_info_context->bmi[10];
        if (x->mode_info_context->mbmi.need_to_clamp_mvs)
        {
            clamp_mv_to_umv_border(&x->block[ 0].bmi.mv.as_mv, x);
            clamp_mv_to_umv_border(&x->block[ 2].bmi.mv.as_mv, x);
            clamp_mv_to_umv_border(&x->block[ 8].bmi.mv.as_mv, x);
            clamp_mv_to_umv_border(&x->block[10].bmi.mv.as_mv, x);
        }

        b = &x->block[ 0];
        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
        b = &x->block[ 2];
        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
        b = &x->block[ 8];
        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
        b = &x->block[10];
        build_inter_predictors4b(x, b, base_dst + b->offset, dst_stride, base_pre, dst_stride);
    }
    else
    {
        for (i = 0; i < 16; i += 2)
        {
            BLOCKD *d0 = &x->block[i];
            BLOCKD *d1 = &x->block[i+1];
            int dst_stride = x->dst.y_stride;

            x->block[i+0].bmi = x->mode_info_context->bmi[i+0];
            x->block[i+1].bmi = x->mode_info_context->bmi[i+1];
            if (x->mode_info_context->mbmi.need_to_clamp_mvs)
            {
                clamp_mv_to_umv_border(&x->block[i+0].bmi.mv.as_mv, x);
                clamp_mv_to_umv_border(&x->block[i+1].bmi.mv.as_mv, x);
            }

            if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
                build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
            else
            {
                build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
                build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
            }

        }

    }
    base_dst = x->dst.u_buffer;
    base_pre = x->pre.u_buffer;
    for (i = 16; i < 20; i += 2)
    {
        BLOCKD *d0 = &x->block[i];
        BLOCKD *d1 = &x->block[i+1];
        int dst_stride = x->dst.uv_stride;

        /* Note: uv mvs already clamped in build_4x4uvmvs() */

        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
        else
        {
            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
        }
    }

    base_dst = x->dst.v_buffer;
    base_pre = x->pre.v_buffer;
    for (i = 20; i < 24; i += 2)
    {
        BLOCKD *d0 = &x->block[i];
        BLOCKD *d1 = &x->block[i+1];
        int dst_stride = x->dst.uv_stride;

        /* Note: uv mvs already clamped in build_4x4uvmvs() */

        if (d0->bmi.mv.as_int == d1->bmi.mv.as_int)
            build_inter_predictors2b(x, d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride);
        else
        {
            build_inter_predictors_b(d0, base_dst + d0->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
            build_inter_predictors_b(d1, base_dst + d1->offset, dst_stride, base_pre, dst_stride, x->subpixel_predict);
        }
    }
}

static
void build_4x4uvmvs(MACROBLOCKD *x)
{
    int i, j;

    for (i = 0; i < 2; i++)
    {
        for (j = 0; j < 2; j++)
        {
            int yoffset = i * 8 + j * 2;
            int uoffset = 16 + i * 2 + j;
            int voffset = 20 + i * 2 + j;

            int temp;

            temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.row
                 + x->mode_info_context->bmi[yoffset + 1].mv.as_mv.row
                 + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.row
                 + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.row;

            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

            x->block[uoffset].bmi.mv.as_mv.row = (temp / 8) & x->fullpixel_mask;

            temp = x->mode_info_context->bmi[yoffset + 0].mv.as_mv.col
                 + x->mode_info_context->bmi[yoffset + 1].mv.as_mv.col
                 + x->mode_info_context->bmi[yoffset + 4].mv.as_mv.col
                 + x->mode_info_context->bmi[yoffset + 5].mv.as_mv.col;

            temp += 4 + ((temp >> (sizeof(temp) * CHAR_BIT - 1)) * 8);

            x->block[uoffset].bmi.mv.as_mv.col = (temp / 8) & x->fullpixel_mask;

            if (x->mode_info_context->mbmi.need_to_clamp_mvs)
                clamp_uvmv_to_umv_border(&x->block[uoffset].bmi.mv.as_mv, x);

            x->block[voffset].bmi.mv.as_int = x->block[uoffset].bmi.mv.as_int;
        }
    }
}

void vp8_build_inter_predictors_mb(MACROBLOCKD *xd)
{
    if (xd->mode_info_context->mbmi.mode != SPLITMV)
    {
        vp8_build_inter16x16_predictors_mb(xd, xd->dst.y_buffer,
                                           xd->dst.u_buffer, xd->dst.v_buffer,
                                           xd->dst.y_stride, xd->dst.uv_stride);
    }
    else
    {
        build_4x4uvmvs(xd);
        build_inter4x4_predictors_mb(xd);
    }
}