rdopt.c 82.9 KB
Newer Older
John Koleszar's avatar
John Koleszar committed
1
/*
2
 *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
John Koleszar's avatar
John Koleszar committed
3
 *
4
 *  Use of this source code is governed by a BSD-style license
5
6
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
7
 *  in the file PATENTS.  All contributing project authors may
8
 *  be found in the AUTHORS file in the root of the source tree.
John Koleszar's avatar
John Koleszar committed
9
10
11
12
13
14
15
 */


#include <stdio.h>
#include <math.h>
#include <limits.h>
#include <assert.h>
John Koleszar's avatar
John Koleszar committed
16
#include "vp8/common/pragmas.h"
John Koleszar's avatar
John Koleszar committed
17
18
19
20
21
22

#include "tokenize.h"
#include "treewriter.h"
#include "onyx_int.h"
#include "modecosts.h"
#include "encodeintra.h"
John Koleszar's avatar
John Koleszar committed
23
24
25
26
27
#include "vp8/common/entropymode.h"
#include "vp8/common/reconinter.h"
#include "vp8/common/reconintra.h"
#include "vp8/common/reconintra4x4.h"
#include "vp8/common/findnearmv.h"
John Koleszar's avatar
John Koleszar committed
28
29
#include "encodemb.h"
#include "quantize.h"
John Koleszar's avatar
John Koleszar committed
30
31
#include "vp8/common/idct.h"
#include "vp8/common/g_common.h"
John Koleszar's avatar
John Koleszar committed
32
33
#include "variance.h"
#include "mcomp.h"
Yunqing Wang's avatar
Yunqing Wang committed
34
#include "rdopt.h"
John Koleszar's avatar
John Koleszar committed
35
36
#include "vpx_mem/vpx_mem.h"
#include "dct.h"
John Koleszar's avatar
John Koleszar committed
37
#include "vp8/common/systemdependent.h"
John Koleszar's avatar
John Koleszar committed
38
39
40
41
42
43
44
45

#if CONFIG_RUNTIME_CPU_DETECT
#define IF_RTCD(x)  (x)
#else
#define IF_RTCD(x)  NULL
#endif


Scott LaVarnway's avatar
Scott LaVarnway committed
46
47
48
extern void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x);
extern void vp8_update_zbin_extra(VP8_COMP *cpi, MACROBLOCK *x);

John Koleszar's avatar
John Koleszar committed
49
50
#define MAXF(a,b)            (((a) > (b)) ? (a) : (b))

51
static const int auto_speed_thresh[17] =
John Koleszar's avatar
John Koleszar committed
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
{
    1000,
    200,
    150,
    130,
    150,
    125,
    120,
    115,
    115,
    115,
    115,
    115,
    115,
    115,
    115,
    115,
    105
};

const MB_PREDICTION_MODE vp8_mode_order[MAX_MODES] =
{
    ZEROMV,
    DC_PRED,

    NEARESTMV,
    NEARMV,

    ZEROMV,
    NEARESTMV,

    ZEROMV,
    NEARESTMV,

    NEARMV,
    NEARMV,

    V_PRED,
    H_PRED,
    TM_PRED,

    NEWMV,
    NEWMV,
    NEWMV,

    SPLITMV,
    SPLITMV,
    SPLITMV,

    B_PRED,
};

const MV_REFERENCE_FRAME vp8_ref_frame_order[MAX_MODES] =
{
    LAST_FRAME,
    INTRA_FRAME,

    LAST_FRAME,
    LAST_FRAME,

    GOLDEN_FRAME,
    GOLDEN_FRAME,

    ALTREF_FRAME,
    ALTREF_FRAME,

    GOLDEN_FRAME,
    ALTREF_FRAME,

    INTRA_FRAME,
    INTRA_FRAME,
    INTRA_FRAME,

    LAST_FRAME,
    GOLDEN_FRAME,
    ALTREF_FRAME,

    LAST_FRAME,
    GOLDEN_FRAME,
    ALTREF_FRAME,

    INTRA_FRAME,
};

static void fill_token_costs(
    unsigned int c      [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [vp8_coef_tokens],
    const vp8_prob p    [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [vp8_coef_tokens-1]
)
{
    int i, j, k;


    for (i = 0; i < BLOCK_TYPES; i++)
        for (j = 0; j < COEF_BANDS; j++)
            for (k = 0; k < PREV_COEF_CONTEXTS; k++)

                vp8_cost_tokens((int *)(c [i][j][k]), p [i][j][k], vp8_coef_tree);

}

Paul Wilkins's avatar
Paul Wilkins committed
152
static int rd_iifactor [ 32 ] =  {    4,   4,   3,   2,   1,   0,   0,   0,
John Koleszar's avatar
John Koleszar committed
153
154
155
156
157
158
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                 };


159
160
/* values are now correlated to quantizer */
static int sad_per_bit16lut[QINDEX_RANGE] =
John Koleszar's avatar
John Koleszar committed
161
{
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
    5,  5,  5,  5,  5,  5,  6,  6,
    6,  6,  6,  6,  6,  7,  7,  7,
    7,  7,  7,  7,  8,  8,  8,  8,
    8,  8,  8,  8,  8,  8,  9,  9,
    9,  9,  9,  9, 10, 10, 10, 10,
    10, 10, 11, 11, 11, 11, 11, 11,
    12, 12, 12, 12, 12, 12, 12, 13,
    13, 13, 13, 13, 13, 14, 14, 14,
    14, 14, 15, 15, 15, 15, 15, 15,
    16, 16, 16, 16, 16, 16, 17, 17,
    17, 17, 17, 17, 17, 18, 18, 18,
    18, 18, 19, 19, 19, 19, 19, 19,
    20, 20, 20, 21, 21, 21, 21, 22,
    22, 22, 23, 23, 23, 24, 24, 24,
    25, 25, 26, 26, 27, 27, 27, 28,
    28, 28, 29, 29, 30, 30, 31, 31
John Koleszar's avatar
John Koleszar committed
178
};
179
static int sad_per_bit4lut[QINDEX_RANGE] =
John Koleszar's avatar
John Koleszar committed
180
{
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
    5,  5,  5,  5,  5,  5,  7,  7,
    7,  7,  7,  7,  7,  8,  8,  8,
    8,  8,  8,  8,  10, 10, 10, 10,
    10, 10, 10, 10, 10, 10, 11, 11,
    11, 11, 11, 11, 13, 13, 13, 13,
    13, 13, 14, 14, 14, 14, 14, 14,
    16, 16, 16, 16, 16, 16, 16, 17,
    17, 17, 17, 17, 17, 19, 19, 19,
    19, 19, 20, 20, 20, 20, 20, 20,
    22, 22, 22, 22, 22, 22, 23, 23,
    23, 23, 23, 23, 23, 25, 25, 25,
    25, 25, 26, 26, 26, 26, 26, 26,
    28, 28, 28, 29, 29, 29, 29, 31,
    31, 31, 32, 32, 32, 34, 34, 34,
    35, 35, 37, 37, 38, 38, 38, 40,
    40, 40, 41, 41, 43, 43, 44, 44,
John Koleszar's avatar
John Koleszar committed
197
198
199
200
201
202
203
204
205
206
207
208
};

void vp8cx_initialize_me_consts(VP8_COMP *cpi, int QIndex)
{
    cpi->mb.sadperbit16 =  sad_per_bit16lut[QIndex];
    cpi->mb.sadperbit4  =  sad_per_bit4lut[QIndex];
}

void vp8_initialize_rd_consts(VP8_COMP *cpi, int Qvalue)
{
    int q;
    int i;
209
210
    double capped_q = (Qvalue < 160) ? (double)Qvalue : 160.0;
    double rdconst = 3.00;
John Koleszar's avatar
John Koleszar committed
211
212
213

    vp8_clear_system_state();  //__asm emms;

214
215
216
217
218
    // Further tests required to see if optimum is different
    // for key frames, golden frames and arf frames.
    // if (cpi->common.refresh_golden_frame ||
    //     cpi->common.refresh_alt_ref_frame)
    cpi->RDMULT = (int)(rdconst * (capped_q * capped_q));
John Koleszar's avatar
John Koleszar committed
219

220
221
222
223
224
225
226
227
228
229
230
231
    // Extend rate multiplier along side quantizer zbin increases
    if (cpi->zbin_over_quant  > 0)
    {
        double oq_factor;
        double modq;

        // Experimental code using the same basic equation as used for Q above
        // The units of cpi->zbin_over_quant are 1/128 of Q bin size
        oq_factor = 1.0 + ((double)0.0015625 * cpi->zbin_over_quant);
        modq = (int)((double)capped_q * oq_factor);
        cpi->RDMULT = (int)(rdconst * (modq * modq));
    }
John Koleszar's avatar
John Koleszar committed
232

Paul Wilkins's avatar
Paul Wilkins committed
233
    if (cpi->pass == 2 && (cpi->common.frame_type != KEY_FRAME))
John Koleszar's avatar
John Koleszar committed
234
    {
Paul Wilkins's avatar
Paul Wilkins committed
235
236
        if (cpi->next_iiratio > 31)
            cpi->RDMULT += (cpi->RDMULT * rd_iifactor[31]) >> 4;
John Koleszar's avatar
John Koleszar committed
237
        else
Paul Wilkins's avatar
Paul Wilkins committed
238
            cpi->RDMULT += (cpi->RDMULT * rd_iifactor[cpi->next_iiratio]) >> 4;
John Koleszar's avatar
John Koleszar committed
239
240
241
    }

    cpi->mb.errorperbit = (cpi->RDMULT / 100);
242
243
    cpi->mb.errorperbit += (cpi->mb.errorperbit==0);

John Koleszar's avatar
John Koleszar committed
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
    vp8_set_speed_features(cpi);

    if (cpi->common.simpler_lpf)
        cpi->common.filter_type = SIMPLE_LOOPFILTER;

    q = (int)pow(Qvalue, 1.25);

    if (q < 8)
        q = 8;

    if (cpi->RDMULT > 1000)
    {
        cpi->RDDIV = 1;
        cpi->RDMULT /= 100;

        for (i = 0; i < MAX_MODES; i++)
        {
            if (cpi->sf.thresh_mult[i] < INT_MAX)
            {
                cpi->rd_threshes[i] = cpi->sf.thresh_mult[i] * q / 100;
            }
            else
            {
                cpi->rd_threshes[i] = INT_MAX;
            }

            cpi->rd_baseline_thresh[i] = cpi->rd_threshes[i];
        }
    }
    else
    {
        cpi->RDDIV = 100;

        for (i = 0; i < MAX_MODES; i++)
        {
            if (cpi->sf.thresh_mult[i] < (INT_MAX / q))
            {
                cpi->rd_threshes[i] = cpi->sf.thresh_mult[i] * q;
            }
            else
            {
                cpi->rd_threshes[i] = INT_MAX;
            }

            cpi->rd_baseline_thresh[i] = cpi->rd_threshes[i];
        }
    }

    fill_token_costs(
        cpi->mb.token_costs,
        (const vp8_prob( *)[8][3][11]) cpi->common.fc.coef_probs
    );

    vp8_init_mode_costs(cpi);

}

void vp8_auto_select_speed(VP8_COMP *cpi)
{
    int used = cpi->oxcf.cpu_used;

    int milliseconds_for_compress = (int)(1000000 / cpi->oxcf.frame_rate);

    milliseconds_for_compress = milliseconds_for_compress * (16 - cpi->oxcf.cpu_used) / 16;

#if 0

    if (0)
    {
        FILE *f;

        f = fopen("speed.stt", "a");
        fprintf(f, " %8ld %10ld %10ld %10ld\n",
                cpi->common.current_video_frame, cpi->Speed, milliseconds_for_compress, cpi->avg_pick_mode_time);
        fclose(f);
    }

#endif

    /*
    // this is done during parameter valid check
    if( used > 16)
        used = 16;
    if( used < -16)
        used = -16;
    */

    if (cpi->avg_pick_mode_time < milliseconds_for_compress && (cpi->avg_encode_time - cpi->avg_pick_mode_time) < milliseconds_for_compress)
    {
        if (cpi->avg_pick_mode_time == 0)
        {
            cpi->Speed = 4;
        }
        else
        {
            if (milliseconds_for_compress * 100 < cpi->avg_encode_time * 95)
            {
                cpi->Speed          += 2;
                cpi->avg_pick_mode_time = 0;
                cpi->avg_encode_time = 0;

                if (cpi->Speed > 16)
                {
                    cpi->Speed = 16;
                }
            }

351
            if (milliseconds_for_compress * 100 > cpi->avg_encode_time * auto_speed_thresh[cpi->Speed])
John Koleszar's avatar
John Koleszar committed
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
            {
                cpi->Speed          -= 1;
                cpi->avg_pick_mode_time = 0;
                cpi->avg_encode_time = 0;

                // In real-time mode, cpi->speed is in [4, 16].
                if (cpi->Speed < 4)        //if ( cpi->Speed < 0 )
                {
                    cpi->Speed = 4;        //cpi->Speed = 0;
                }
            }
        }
    }
    else
    {
        cpi->Speed += 4;

        if (cpi->Speed > 16)
            cpi->Speed = 16;


        cpi->avg_pick_mode_time = 0;
        cpi->avg_encode_time = 0;
    }
}

int vp8_block_error_c(short *coeff, short *dqcoeff)
{
    int i;
    int error = 0;

    for (i = 0; i < 16; i++)
    {
        int this_diff = coeff[i] - dqcoeff[i];
        error += this_diff * this_diff;
    }

    return error;
}

int vp8_mbblock_error_c(MACROBLOCK *mb, int dc)
{
    BLOCK  *be;
    BLOCKD *bd;
    int i, j;
    int berror, error = 0;

    for (i = 0; i < 16; i++)
    {
        be = &mb->block[i];
        bd = &mb->e_mbd.block[i];

        berror = 0;

        for (j = dc; j < 16; j++)
        {
            int this_diff = be->coeff[j] - bd->dqcoeff[j];
            berror += this_diff * this_diff;
        }

        error += berror;
    }

    return error;
}

int vp8_mbuverror_c(MACROBLOCK *mb)
{

    BLOCK  *be;
    BLOCKD *bd;


    int i;
    int error = 0;

    for (i = 16; i < 24; i++)
    {
        be = &mb->block[i];
        bd = &mb->e_mbd.block[i];

        error += vp8_block_error_c(be->coeff, bd->dqcoeff);
    }

    return error;
}

int VP8_UVSSE(MACROBLOCK *x, const vp8_variance_rtcd_vtable_t *rtcd)
{
    unsigned char *uptr, *vptr;
    unsigned char *upred_ptr = (*(x->block[16].base_src) + x->block[16].src);
    unsigned char *vpred_ptr = (*(x->block[20].base_src) + x->block[20].src);
    int uv_stride = x->block[16].src_stride;

    unsigned int sse1 = 0;
    unsigned int sse2 = 0;
    int mv_row;
    int mv_col;
    int offset;
    int pre_stride = x->e_mbd.block[16].pre_stride;

    vp8_build_uvmvs(&x->e_mbd, 0);
    mv_row = x->e_mbd.block[16].bmi.mv.as_mv.row;
    mv_col = x->e_mbd.block[16].bmi.mv.as_mv.col;

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

    if ((mv_row | mv_col) & 7)
    {
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr, uv_stride, &sse2);
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr, uv_stride, &sse1);
        sse2 += sse1;
    }
    else
    {
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(uptr, pre_stride, mv_col & 7, mv_row & 7, upred_ptr, uv_stride, &sse2);
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(vptr, pre_stride, mv_col & 7, mv_row & 7, vpred_ptr, uv_stride, &sse1);
        sse2 += sse1;
    }

    return sse2;

}

static int cost_coeffs(MACROBLOCK *mb, BLOCKD *b, int type, ENTROPY_CONTEXT *a, ENTROPY_CONTEXT *l)
{
    int c = !type;              /* start at coef 0, unless Y with Y2 */
    int eob = b->eob;
    int pt ;    /* surrounding block/prev coef predictor */
    int cost = 0;
    short *qcoeff_ptr = b->qcoeff;

    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);

# define QC( I)  ( qcoeff_ptr [vp8_default_zig_zag1d[I]] )

    for (; c < eob; c++)
    {
        int v = QC(c);
        int t = vp8_dct_value_tokens_ptr[v].Token;
        cost += mb->token_costs [type] [vp8_coef_bands[c]] [pt] [t];
        cost += vp8_dct_value_cost_ptr[v];
        pt = vp8_prev_token_class[t];
    }

# undef QC

    if (c < 16)
        cost += mb->token_costs [type] [vp8_coef_bands[c]] [pt] [DCT_EOB_TOKEN];

    pt = (c != !type); // is eob first coefficient;
    *a = *l = pt;

    return cost;
}

Scott LaVarnway's avatar
Scott LaVarnway committed
510
static int vp8_rdcost_mby(MACROBLOCK *mb)
John Koleszar's avatar
John Koleszar committed
511
512
513
514
{
    int cost = 0;
    int b;
    MACROBLOCKD *x = &mb->e_mbd;
515
516
517
518
519
520
    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;

    vpx_memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
    vpx_memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
521

522
523
    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
John Koleszar's avatar
John Koleszar committed
524
525

    for (b = 0; b < 16; b++)
Scott LaVarnway's avatar
Scott LaVarnway committed
526
        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_NO_DC,
527
                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
John Koleszar's avatar
John Koleszar committed
528

Scott LaVarnway's avatar
Scott LaVarnway committed
529
    cost += cost_coeffs(mb, x->block + 24, PLANE_TYPE_Y2,
Scott LaVarnway's avatar
Scott LaVarnway committed
530
                ta + vp8_block2above[24], tl + vp8_block2left[24]);
John Koleszar's avatar
John Koleszar committed
531
532
533
534

    return cost;
}

535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
static void macro_block_yrd( MACROBLOCK *mb,
                             int *Rate,
                             int *Distortion,
                             const vp8_encodemb_rtcd_vtable_t *rtcd)
{
    int b;
    MACROBLOCKD *const x = &mb->e_mbd;
    BLOCK   *const mb_y2 = mb->block + 24;
    BLOCKD *const x_y2  = x->block + 24;
    short *Y2DCPtr = mb_y2->src_diff;
    BLOCK *beptr;
    int d;

    ENCODEMB_INVOKE(rtcd, submby)( mb->src_diff, mb->src.y_buffer,
                                   mb->e_mbd.predictor, mb->src.y_stride );

    // Fdct and building the 2nd order block
    for (beptr = mb->block; beptr < mb->block + 16; beptr += 2)
    {
        mb->vp8_short_fdct8x4(beptr->src_diff, beptr->coeff, 32);
        *Y2DCPtr++ = beptr->coeff[0];
        *Y2DCPtr++ = beptr->coeff[16];
    }

    // 2nd order fdct
    mb->short_walsh4x4(mb_y2->src_diff, mb_y2->coeff, 8);

    // Quantization
    for (b = 0; b < 16; b++)
    {
        mb->quantize_b(&mb->block[b], &mb->e_mbd.block[b]);
    }

    // DC predication and Quantization of 2nd Order block
    mb->quantize_b(mb_y2, x_y2);

    // Distortion
    d = ENCODEMB_INVOKE(rtcd, mberr)(mb, 1) << 2;
    d += ENCODEMB_INVOKE(rtcd, berr)(mb_y2->coeff, x_y2->dqcoeff);

    *Distortion = (d >> 4);

    // rate
    *Rate = vp8_rdcost_mby(mb);
}
John Koleszar's avatar
John Koleszar committed
580

581
static void copy_predictor(unsigned char *dst, const unsigned char *predictor)
582
{
583
584
585
586
587
588
    const unsigned int *p = (const unsigned int *)predictor;
    unsigned int *d = (unsigned int *)dst;
    d[0] = p[0];
    d[4] = p[4];
    d[8] = p[8];
    d[12] = p[12];
589
}
590
static int rd_pick_intra4x4block(
John Koleszar's avatar
John Koleszar committed
591
592
593
594
595
    VP8_COMP *cpi,
    MACROBLOCK *x,
    BLOCK *be,
    BLOCKD *b,
    B_PREDICTION_MODE *best_mode,
596
    unsigned int *bmode_costs,
John Koleszar's avatar
John Koleszar committed
597
598
599
600
601
602
603
604
    ENTROPY_CONTEXT *a,
    ENTROPY_CONTEXT *l,

    int *bestrate,
    int *bestratey,
    int *bestdistortion)
{
    B_PREDICTION_MODE mode;
605
    int best_rd = INT_MAX;
John Koleszar's avatar
John Koleszar committed
606
607
608
609
610
    int rate = 0;
    int distortion;

    ENTROPY_CONTEXT ta = *a, tempa = *a;
    ENTROPY_CONTEXT tl = *l, templ = *l;
611
612
613
614
615
616
617
    /*
     * The predictor buffer is a 2d buffer with a stride of 16.  Create
     * a temp buffer that meets the stride requirements, but we are only
     * interested in the left 4x4 block
     * */
    DECLARE_ALIGNED_ARRAY(16, unsigned char,  best_predictor, 16*4);
    DECLARE_ALIGNED_ARRAY(16, short, best_dqcoeff, 16);
John Koleszar's avatar
John Koleszar committed
618
619
620
621
622
623

    for (mode = B_DC_PRED; mode <= B_HU_PRED; mode++)
    {
        int this_rd;
        int ratey;

624
625
626
627
628
629
        rate = bmode_costs[mode];

        vp8_predict_intra4x4(b, mode, b->predictor);
        ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), subb)(be, b, 16);
        x->vp8_short_fdct4x4(be->src_diff, be->coeff, 32);
        x->quantize_b(be, b);
John Koleszar's avatar
John Koleszar committed
630
631
632
633

        tempa = ta;
        templ = tl;

Scott LaVarnway's avatar
Scott LaVarnway committed
634
        ratey = cost_coeffs(x, b, PLANE_TYPE_Y_WITH_DC, &tempa, &templ);
John Koleszar's avatar
John Koleszar committed
635
636
637
638
639
640
641
642
643
644
645
646
647
648
        rate += ratey;
        distortion = ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)(be->coeff, b->dqcoeff) >> 2;

        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);

        if (this_rd < best_rd)
        {
            *bestrate = rate;
            *bestratey = ratey;
            *bestdistortion = distortion;
            best_rd = this_rd;
            *best_mode = mode;
            *a = tempa;
            *l = templ;
649
650
            copy_predictor(best_predictor, b->predictor);
            vpx_memcpy(best_dqcoeff, b->dqcoeff, 32);
John Koleszar's avatar
John Koleszar committed
651
652
653
654
        }
    }

    b->bmi.mode = (B_PREDICTION_MODE)(*best_mode);
655

656
657
    IDCT_INVOKE(IF_RTCD(&cpi->rtcd.common->idct), idct16)(best_dqcoeff, b->diff, 32);
    RECON_INVOKE(IF_RTCD(&cpi->rtcd.common->recon), recon)(best_predictor, b->diff, *(b->base_dst) + b->dst, b->dst_stride);
John Koleszar's avatar
John Koleszar committed
658

659
    return best_rd;
John Koleszar's avatar
John Koleszar committed
660
661
}

662
663
int vp8_rd_pick_intra4x4mby_modes(VP8_COMP *cpi, MACROBLOCK *mb, int *Rate,
                                  int *rate_y, int *Distortion, int best_rd)
John Koleszar's avatar
John Koleszar committed
664
665
666
667
668
669
{
    MACROBLOCKD *const xd = &mb->e_mbd;
    int i;
    int cost = mb->mbmode_cost [xd->frame_type] [B_PRED];
    int distortion = 0;
    int tot_rate_y = 0;
670
    long long total_rd = 0;
671
672
673
    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;
674
    unsigned int *bmode_costs;
675
676
677
678
679
680

    vpx_memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
    vpx_memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));

    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
John Koleszar's avatar
John Koleszar committed
681
682
683

    vp8_intra_prediction_down_copy(xd);

684
685
    bmode_costs = mb->inter_bmode_costs;

John Koleszar's avatar
John Koleszar committed
686
687
688
689
690
691
692
    for (i = 0; i < 16; i++)
    {
        MODE_INFO *const mic = xd->mode_info_context;
        const int mis = xd->mode_info_stride;
        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
        int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(ry), UNINITIALIZED_IS_SAFE(d);

693
694
695
696
697
698
699
700
        if (mb->e_mbd.frame_type == KEY_FRAME)
        {
            const B_PREDICTION_MODE A = vp8_above_bmi(mic, i, mis)->mode;
            const B_PREDICTION_MODE L = vp8_left_bmi(mic, i)->mode;

            bmode_costs  = mb->bmode_costs[A][L];
        }

701
        total_rd += rd_pick_intra4x4block(
702
            cpi, mb, mb->block + i, xd->block + i, &best_mode, bmode_costs,
703
704
            ta + vp8_block2above[i],
            tl + vp8_block2left[i], &r, &ry, &d);
John Koleszar's avatar
John Koleszar committed
705
706
707
708
709

        cost += r;
        distortion += d;
        tot_rate_y += ry;
        mic->bmi[i].mode = xd->block[i].bmi.mode = best_mode;
710

711
712
        if(total_rd >= (long long)best_rd)
            break;
John Koleszar's avatar
John Koleszar committed
713
714
    }

715
716
    if(total_rd >= (long long)best_rd)
        return INT_MAX;
717

John Koleszar's avatar
John Koleszar committed
718
719
720
721
722
723
    *Rate = cost;
    *rate_y += tot_rate_y;
    *Distortion = distortion;

    return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
}
724
725
726
727
728
int vp8_rd_pick_intra16x16mby_mode(VP8_COMP *cpi,
                                   MACROBLOCK *x,
                                   int *Rate,
                                   int *rate_y,
                                   int *Distortion)
John Koleszar's avatar
John Koleszar committed
729
730
731
732
{
    MB_PREDICTION_MODE mode;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
    int rate, ratey;
733
    int distortion;
John Koleszar's avatar
John Koleszar committed
734
    int best_rd = INT_MAX;
735
    int this_rd;
John Koleszar's avatar
John Koleszar committed
736
737
738
739

    //Y Search for 16x16 intra prediction mode
    for (mode = DC_PRED; mode <= TM_PRED; mode++)
    {
740
        x->e_mbd.mode_info_context->mbmi.mode = mode;
John Koleszar's avatar
John Koleszar committed
741

742
743
        RECON_INVOKE(&cpi->common.rtcd.recon, build_intra_predictors_mby)
            (&x->e_mbd);
John Koleszar's avatar
John Koleszar committed
744

745
746
747
        macro_block_yrd(x, &ratey, &distortion, IF_RTCD(&cpi->rtcd.encodemb));
        rate = ratey + x->mbmode_cost[x->e_mbd.frame_type]
                                     [x->e_mbd.mode_info_context->mbmi.mode];
John Koleszar's avatar
John Koleszar committed
748
749
750
751
752
753
754
755
756

        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);

        if (this_rd < best_rd)
        {
            mode_selected = mode;
            best_rd = this_rd;
            *Rate = rate;
            *rate_y = ratey;
757
            *Distortion = distortion;
John Koleszar's avatar
John Koleszar committed
758
759
760
        }
    }

761
    x->e_mbd.mode_info_context->mbmi.mode = mode_selected;
John Koleszar's avatar
John Koleszar committed
762
763
764
765
766
767
768
769
    return best_rd;
}

static int rd_cost_mbuv(MACROBLOCK *mb)
{
    int b;
    int cost = 0;
    MACROBLOCKD *x = &mb->e_mbd;
770
771
772
773
774
775
    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;

    vpx_memcpy(&t_above, mb->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
    vpx_memcpy(&t_left, mb->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
776

777
778
    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
John Koleszar's avatar
John Koleszar committed
779

Scott LaVarnway's avatar
Scott LaVarnway committed
780
781
    for (b = 16; b < 24; b++)
        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_UV,
782
                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
John Koleszar's avatar
John Koleszar committed
783
784
785
786
787
788
789
790
791
792
793
794
795
796

    return cost;
}


static int vp8_rd_inter_uv(VP8_COMP *cpi, MACROBLOCK *x, int *rate, int *distortion, int fullpixel)
{
    vp8_build_uvmvs(&x->e_mbd, fullpixel);
    vp8_encode_inter16x16uvrd(IF_RTCD(&cpi->rtcd), x);


    *rate       = rd_cost_mbuv(x);
    *distortion = ENCODEMB_INVOKE(&cpi->rtcd.encodemb, mbuverr)(x) / 4;

797
    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
John Koleszar's avatar
John Koleszar committed
798
799
}

Paul Wilkins's avatar
Paul Wilkins committed
800
void vp8_rd_pick_intra_mbuv_mode(VP8_COMP *cpi, MACROBLOCK *x, int *rate, int *rate_tokenonly, int *distortion)
John Koleszar's avatar
John Koleszar committed
801
802
803
804
805
806
807
808
809
810
811
812
813
{
    MB_PREDICTION_MODE mode;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
    int best_rd = INT_MAX;
    int UNINITIALIZED_IS_SAFE(d), UNINITIALIZED_IS_SAFE(r);
    int rate_to;

    for (mode = DC_PRED; mode <= TM_PRED; mode++)
    {
        int rate;
        int distortion;
        int this_rd;

814
        x->e_mbd.mode_info_context->mbmi.uv_mode = mode;
815
816
817
818
819
820
        vp8_build_intra_predictors_mbuv(&x->e_mbd);
        ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), submbuv)(x->src_diff,
                      x->src.u_buffer, x->src.v_buffer, x->e_mbd.predictor,
                      x->src.uv_stride);
        vp8_transform_mbuv(x);
        vp8_quantize_mbuv(x);
John Koleszar's avatar
John Koleszar committed
821
822

        rate_to = rd_cost_mbuv(x);
823
        rate = rate_to + x->intra_uv_mode_cost[x->e_mbd.frame_type][x->e_mbd.mode_info_context->mbmi.uv_mode];
John Koleszar's avatar
John Koleszar committed
824

825
        distortion = ENCODEMB_INVOKE(&cpi->rtcd.encodemb, mbuverr)(x) / 4;
John Koleszar's avatar
John Koleszar committed
826

827
        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
John Koleszar's avatar
John Koleszar committed
828
829
830
831
832
833
834
835
836
837
838
839
840
841

        if (this_rd < best_rd)
        {
            best_rd = this_rd;
            d = distortion;
            r = rate;
            *rate_tokenonly = rate_to;
            mode_selected = mode;
        }
    }

    *rate = r;
    *distortion = d;

842
    x->e_mbd.mode_info_context->mbmi.uv_mode = mode_selected;
John Koleszar's avatar
John Koleszar committed
843
844
845
846
847
848
849
}

int vp8_cost_mv_ref(MB_PREDICTION_MODE m, const int near_mv_ref_ct[4])
{
    vp8_prob p [VP8_MVREFS-1];
    assert(NEARESTMV <= m  &&  m <= SPLITMV);
    vp8_mv_ref_probs(p, near_mv_ref_ct);
850
851
    return vp8_cost_token(vp8_mv_ref_tree, p,
                          vp8_mv_ref_encoding_array - NEARESTMV + m);
John Koleszar's avatar
John Koleszar committed
852
853
854
855
856
857
}

void vp8_set_mbmode_and_mvs(MACROBLOCK *x, MB_PREDICTION_MODE mb, MV *mv)
{
    int i;

858
859
860
    x->e_mbd.mode_info_context->mbmi.mode = mb;
    x->e_mbd.mode_info_context->mbmi.mv.as_mv.row = mv->row;
    x->e_mbd.mode_info_context->mbmi.mv.as_mv.col = mv->col;
John Koleszar's avatar
John Koleszar committed
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948

    for (i = 0; i < 16; i++)
    {
        B_MODE_INFO *bmi = &x->e_mbd.block[i].bmi;
        bmi->mode = (B_PREDICTION_MODE) mb;
        bmi->mv.as_mv.row = mv->row;
        bmi->mv.as_mv.col = mv->col;
    }
}

static int labels2mode(
    MACROBLOCK *x,
    int const *labelings, int which_label,
    B_PREDICTION_MODE this_mode,
    MV *this_mv, MV *best_ref_mv,
    int *mvcost[2]
)
{
    MACROBLOCKD *const xd = & x->e_mbd;
    MODE_INFO *const mic = xd->mode_info_context;
    const int mis = xd->mode_info_stride;

    int cost = 0;
    int thismvcost = 0;

    /* We have to be careful retrieving previously-encoded motion vectors.
       Ones from this macroblock have to be pulled from the BLOCKD array
       as they have not yet made it to the bmi array in our MB_MODE_INFO. */

    int i = 0;

    do
    {
        BLOCKD *const d = xd->block + i;
        const int row = i >> 2,  col = i & 3;

        B_PREDICTION_MODE m;

        if (labelings[i] != which_label)
            continue;

        if (col  &&  labelings[i] == labelings[i-1])
            m = LEFT4X4;
        else if (row  &&  labelings[i] == labelings[i-4])
            m = ABOVE4X4;
        else
        {
            // the only time we should do costing for new motion vector or mode
            // is when we are on a new label  (jbb May 08, 2007)
            switch (m = this_mode)
            {
            case NEW4X4 :
                thismvcost  = vp8_mv_bit_cost(this_mv, best_ref_mv, mvcost, 102);
                break;
            case LEFT4X4:
                *this_mv = col ? d[-1].bmi.mv.as_mv : vp8_left_bmi(mic, i)->mv.as_mv;
                break;
            case ABOVE4X4:
                *this_mv = row ? d[-4].bmi.mv.as_mv : vp8_above_bmi(mic, i, mis)->mv.as_mv;
                break;
            case ZERO4X4:
                this_mv->row = this_mv->col = 0;
                break;
            default:
                break;
            }

            if (m == ABOVE4X4)  // replace above with left if same
            {
                const MV mv = col ? d[-1].bmi.mv.as_mv : vp8_left_bmi(mic, i)->mv.as_mv;

                if (mv.row == this_mv->row  &&  mv.col == this_mv->col)
                    m = LEFT4X4;
            }

            cost = x->inter_bmode_costs[ m];
        }

        d->bmi.mode = m;
        d->bmi.mv.as_mv = *this_mv;

    }
    while (++i < 16);

    cost += thismvcost ;
    return cost;
}

949
950
951
static int rdcost_mbsegment_y(MACROBLOCK *mb, const int *labels,
                              int which_label, ENTROPY_CONTEXT *ta,
                              ENTROPY_CONTEXT *tl)
John Koleszar's avatar
John Koleszar committed
952
953
954
955
956
957
958
{
    int cost = 0;
    int b;
    MACROBLOCKD *x = &mb->e_mbd;

    for (b = 0; b < 16; b++)
        if (labels[ b] == which_label)
Scott LaVarnway's avatar
Scott LaVarnway committed
959
            cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_WITH_DC,
960
961
                                ta + vp8_block2above[b],
                                tl + vp8_block2left[b]);
John Koleszar's avatar
John Koleszar committed
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980

    return cost;

}
static unsigned int vp8_encode_inter_mb_segment(MACROBLOCK *x, int const *labels, int which_label, const vp8_encodemb_rtcd_vtable_t *rtcd)
{
    int i;
    unsigned int distortion = 0;

    for (i = 0; i < 16; i++)
    {
        if (labels[i] == which_label)
        {
            BLOCKD *bd = &x->e_mbd.block[i];
            BLOCK *be = &x->block[i];


            vp8_build_inter_predictors_b(bd, 16, x->e_mbd.subpixel_predict);
            ENCODEMB_INVOKE(rtcd, subb)(be, bd, 16);
Yaowu Xu's avatar
Yaowu Xu committed
981
            x->vp8_short_fdct4x4(be->src_diff, be->coeff, 32);
John Koleszar's avatar
John Koleszar committed
982
983
984

            // set to 0 no way to account for 2nd order DC so discount
            //be->coeff[0] = 0;
985
            x->quantize_b(be, bd);
John Koleszar's avatar
John Koleszar committed
986
987
988
989
990
991
992
993

            distortion += ENCODEMB_INVOKE(rtcd, berr)(be->coeff, bd->dqcoeff);
        }
    }

    return distortion;
}

994
995
996
997
998

static const unsigned int segmentation_to_sseshift[4] = {3, 3, 2, 0};


typedef struct
John Koleszar's avatar
John Koleszar committed
999
{
1000
  MV *ref_mv;
1001
  MV *mvp;
1002
1003
1004
1005
1006
1007
1008

  int segment_rd;
  int segment_num;
  int r;
  int d;
  int segment_yrate;
  B_PREDICTION_MODE modes[16];
1009
  int_mv mvs[16];
1010
1011
1012
1013
1014
  unsigned char eobs[16];

  int mvthresh;
  int *mdcounts;

1015
1016
1017
  MV sv_mvp[4];     // save 4 mvp from 8x8
  int sv_istep[2];  // save 2 initial step_param for 16x8/8x16

1018
1019
1020
} BEST_SEG_INFO;


1021
1022
static void rd_check_segment(VP8_COMP *cpi, MACROBLOCK *x,
                             BEST_SEG_INFO *bsi, unsigned int segmentation)
1023
1024
{
    int i;
John Koleszar's avatar
John Koleszar committed
1025
1026
1027
    int const *labels;
    int br = 0;
    int bd = 0;
1028
    B_PREDICTION_MODE this_mode;
John Koleszar's avatar
John Koleszar committed
1029

1030

1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
    int label_count;
    int this_segment_rd = 0;
    int label_mv_thresh;
    int rate = 0;
    int sbr = 0;
    int sbd = 0;
    int segmentyrate = 0;

    vp8_variance_fn_ptr_t *v_fn_ptr;

    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;
    ENTROPY_CONTEXT_PLANES t_above_b, t_left_b;
    ENTROPY_CONTEXT *ta_b;
    ENTROPY_CONTEXT *tl_b;
John Koleszar's avatar
John Koleszar committed
1047

1048
1049
    vpx_memcpy(&t_above, x->e_mbd.above_context, sizeof(ENTROPY_CONTEXT_PLANES));
    vpx_memcpy(&t_left, x->e_mbd.left_context, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
1050

1051
1052
1053
1054
1055
1056
1057
    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
    ta_b = (ENTROPY_CONTEXT *)&t_above_b;
    tl_b = (ENTROPY_CONTEXT *)&t_left_b;

    br = 0;
    bd = 0;
1058

1059
1060
1061
    v_fn_ptr = &cpi->fn_ptr[segmentation];
    labels = vp8_mbsplits[segmentation];
    label_count = vp8_mbsplit_count[segmentation];
1062

1063
1064
1065
1066
1067
1068
1069
1070
1071
    // 64 makes this threshold really big effectively
    // making it so that we very rarely check mvs on
    // segments.   setting this to 1 would make mv thresh
    // roughly equal to what it is for macroblocks
    label_mv_thresh = 1 * bsi->mvthresh / label_count ;

    // Segmentation method overheads
    rate = vp8_cost_token(vp8_mbsplit_tree, vp8_mbsplit_probs, vp8_mbsplit_encodings + segmentation);
    rate += vp8_cost_mv_ref(SPLITMV, bsi->mdcounts);
1072
    this_segment_rd += RDCOST(x->rdmult, x->rddiv, rate, 0);
1073
1074
1075
    br += rate;

    for (i = 0; i < label_count; i++)
John Koleszar's avatar
John Koleszar committed
1076
    {
1077
1078
1079
1080
        MV mode_mv[B_MODE_COUNT];
        int best_label_rd = INT_MAX;
        B_PREDICTION_MODE mode_selected = ZERO4X4;
        int bestlabelyrate = 0;
John Koleszar's avatar
John Koleszar committed
1081

1082
1083
1084
1085
1086
1087
1088
1089
1090
        // search for the best motion vector on this segment
        for (this_mode = LEFT4X4; this_mode <= NEW4X4 ; this_mode ++)
        {
            int this_rd;
            int distortion;
            int labelyrate;
            ENTROPY_CONTEXT_PLANES t_above_s, t_left_s;
            ENTROPY_CONTEXT *ta_s;
            ENTROPY_CONTEXT *tl_s;
John Koleszar's avatar
John Koleszar committed
1091

1092
1093
            vpx_memcpy(&t_above_s, &t_above, sizeof(ENTROPY_CONTEXT_PLANES));
            vpx_memcpy(&t_left_s, &t_left, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
1094

1095
1096
            ta_s = (ENTROPY_CONTEXT *)&t_above_s;
            tl_s = (ENTROPY_CONTEXT *)&t_left_s;
John Koleszar's avatar
John Koleszar committed
1097

1098
            if (this_mode == NEW4X4)
John Koleszar's avatar
John Koleszar committed
1099
            {
1100
                int sseshift;
John Koleszar's avatar
John Koleszar committed
1101
                int num00;
1102
                int step_param = 0;
1103
                int further_steps;
1104
1105
1106
1107
1108
1109
                int n;
                int thissme;
                int bestsme = INT_MAX;
                MV  temp_mv;
                BLOCK *c;
                BLOCKD *e;
1110

1111
1112
1113
                // Is the best so far sufficiently good that we cant justify doing and new motion search.
                if (best_label_rd < label_mv_thresh)
                    break;
John Koleszar's avatar
John Koleszar committed
1114

1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
                if(cpi->compressor_speed)
                {
                    if (segmentation == BLOCK_8X16 || segmentation == BLOCK_16X8)
                    {
                        bsi->mvp = &bsi->sv_mvp[i];
                        if (i==1 && segmentation == BLOCK_16X8) bsi->mvp = &bsi->sv_mvp[2];

                        step_param = bsi->sv_istep[i];
                    }

                    // use previous block's result as next block's MV predictor.
                    if (segmentation == BLOCK_4X4 && i>0)
                    {
                        bsi->mvp = &(x->e_mbd.block[i-1].bmi.mv.as_mv);
                        if (i==4 || i==8 || i==12) bsi->mvp = &(x->e_mbd.block[i-4].bmi.mv.as_mv);
                        step_param = 2;
                    }
                }

                further_steps = (MAX_MVSEARCH_STEPS - 1) - step_param;

John Koleszar's avatar
John Koleszar committed
1136
                {
1137
                    int sadpb = x->sadperbit4;
John Koleszar's avatar
John Koleszar committed
1138

1139
                    // find first label
1140
                    n = vp8_mbsplit_offset[segmentation][i];
John Koleszar's avatar
John Koleszar committed
1141

1142
1143
                    c = &x->block[n];
                    e = &x->e_mbd.block[n];
John Koleszar's avatar
John Koleszar committed
1144

1145
1146
                    if (cpi->sf.search_method == HEX)
                        bestsme = vp8_hex_search(x, c, e, bsi->ref_mv,
1147
                                                 &mode_mv[NEW4X4], step_param, sadpb, &num00, v_fn_ptr, x->mvsadcost, x->mvcost, bsi->ref_mv);
John Koleszar's avatar
John Koleszar committed
1148

1149
1150
                    else
                    {
1151
                        bestsme = cpi->diamond_search_sad(x, c, e, bsi->mvp,
1152
                                                          &mode_mv[NEW4X4], step_param,
1153
                                                          sadpb / 2, &num00, v_fn_ptr, x->mvcost, bsi->ref_mv);
John Koleszar's avatar
John Koleszar committed
1154

1155
1156
                        n = num00;
                        num00 = 0;
John Koleszar's avatar
John Koleszar committed
1157

1158
                        while (n < further_steps)
John Koleszar's avatar
John Koleszar committed
1159
                        {
1160
                            n++;
John Koleszar's avatar
John Koleszar committed
1161

1162
1163
                            if (num00)
                                num00--;
John Koleszar's avatar
John Koleszar committed
1164
1165
                            else
                            {
1166
                                thissme = cpi->diamond_search_sad(x, c, e, bsi->mvp,
1167
                                                                  &temp_mv, step_param + n,
1168
                                                                  sadpb / 2, &num00, v_fn_ptr, x->mvcost, bsi->ref_mv);
1169
1170
1171
1172
1173
1174
1175

                                if (thissme < bestsme)
                                {
                                    bestsme = thissme;
                                    mode_mv[NEW4X4].row = temp_mv.row;
                                    mode_mv[NEW4X4].col = temp_mv.col;
                                }
John Koleszar's avatar
John Koleszar committed
1176
1177
1178
1179
                            }
                        }
                    }

1180
1181
1182
1183
                    sseshift = segmentation_to_sseshift[segmentation];

                    // Should we do a full search (best quality only)
                    if ((cpi->compressor_speed == 0) && (bestsme >> sseshift) > 4000)
John Koleszar's avatar
John Koleszar committed
1184
                    {
Yunqing Wang's avatar
Yunqing Wang committed
1185
1186
1187
1188
1189
1190
                        MV full_mvp;

                        full_mvp.row = bsi->mvp->row >>3;
                        full_mvp.col = bsi->mvp->col >>3;

                        thissme = cpi->full_search_sad(x, c, e, &full_mvp,
1191
                                                       sadpb / 4, 16, v_fn_ptr, x->mvcost, bsi->ref_mv);
1192
1193
1194
1195
1196
1197

                        if (thissme < bestsme)
                        {
                            bestsme = thissme;
                            mode_mv[NEW4X4] = e->bmi.mv.as_mv;
                        }
John Koleszar's avatar
John Koleszar committed
1198
                        else
1199
1200
1201
1202
                        {
                            // The full search result is actually worse so re-instate the previous best vector
                            e->bmi.mv.as_mv = mode_mv[NEW4X4];
                        }
John Koleszar's avatar
John Koleszar committed
1203
1204
1205
                    }
                }

1206
                if (bestsme < INT_MAX)
John Koleszar's avatar
John Koleszar committed
1207
                {
1208
                    int distortion;
1209
                    unsigned int sse;
1210

1211
1212
                    if (!cpi->common.full_pixel)
                        cpi->find_fractional_mv_step(x, c, e, &mode_mv[NEW4X4],
1213
                                                     bsi->ref_mv, x->errorperbit / 2, v_fn_ptr, x->mvcost, &distortion, &sse);
1214
1215
                    else
                        vp8_skip_fractional_mv_step(x, c, e, &mode_mv[NEW4X4],
1216
                                                    bsi->ref_mv, x->errorperbit, v_fn_ptr, x->mvcost, &distortion, &sse);
John Koleszar's avatar
John Koleszar committed
1217
                }
1218
1219
1220
1221
            } /* NEW4X4 */

            rate = labels2mode(x, labels, i, this_mode, &mode_mv[this_mode],
                               bsi->ref_mv, x->mvcost);
John Koleszar's avatar
John Koleszar committed
1222

1223
1224
1225
1226
1227
1228
            // Trap vectors that reach beyond the UMV borders
            if (((mode_mv[this_mode].row >> 3) < x->mv_row_min) || ((mode_mv[this_mode].row >> 3) > x->mv_row_max) ||
                ((mode_mv[this_mode].col >> 3) < x->mv_col_min) || ((mode_mv[this_mode].col >> 3) > x->mv_col_max))
            {
                continue;
            }
John Koleszar's avatar
John Koleszar committed
1229

1230
            distortion = vp8_encode_inter_mb_segment(x, labels, i, IF_RTCD(&cpi->rtcd.encodemb)) / 4;
John Koleszar's avatar
John Koleszar committed
1231

1232
1233
            labelyrate = rdcost_mbsegment_y(x, labels, i, ta_s, tl_s);
            rate += labelyrate;
John Koleszar's avatar
John Koleszar committed
1234

1235
            this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
1236

1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
            if (this_rd < best_label_rd)
            {
                sbr = rate;
                sbd = distortion;
                bestlabelyrate = labelyrate;
                mode_selected = this_mode;
                best_label_rd = this_rd;

                vpx_memcpy(ta_b, ta_s, sizeof(ENTROPY_CONTEXT_PLANES));
                vpx_memcpy(tl_b, tl_s, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
1247
1248

            }
1249
        } /*for each 4x4 mode*/
John Koleszar's avatar
John Koleszar committed
1250

1251
1252
        vpx_memcpy(ta, ta_b, sizeof(ENTROPY_CONTEXT_PLANES));
        vpx_memcpy(tl, tl_b, sizeof(ENTROPY_CONTEXT_PLANES));
John Koleszar's avatar
John Koleszar committed
1253

1254
1255
        labels2mode(x, labels, i, mode_selected, &mode_mv[mode_selected],
                    bsi->ref_mv, x->mvcost);
John Koleszar's avatar
John Koleszar committed
1256

1257
1258
1259
1260
        br += sbr;
        bd += sbd;
        segmentyrate += bestlabelyrate;
        this_segment_rd += best_label_rd;
John Koleszar's avatar
John Koleszar committed
1261

1262
        if (this_segment_rd >= bsi->segment_rd)
1263
            break;
1264

1265
    } /* for each label */
John Koleszar's avatar
John Koleszar committed
1266

1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
    if (this_segment_rd < bsi->segment_rd)
    {
        bsi->r = br;
        bsi->d = bd;
        bsi->segment_yrate = segmentyrate;
        bsi->segment_rd = this_segment_rd;
        bsi->segment_num = segmentation;

        // store everything needed to come back to this!!
        for (i = 0; i < 16; i++)
John Koleszar's avatar
John Koleszar committed
1277
        {
1278
            BLOCKD *bd = &x->e_mbd.block[i];
John Koleszar's avatar
John Koleszar committed
1279

1280
            bsi->mvs[i].as_mv = bd->bmi.mv.as_mv;
1281
1282
            bsi->modes[i] = bd->bmi.mode;
            bsi->eobs[i] = bd->eob;
John Koleszar's avatar
John Koleszar committed
1283
1284
        }
    }
1285
}
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300

static __inline
void vp8_cal_step_param(int sr, int *sp)
{
    int step = 0;

    if (sr > MAX_FIRST_STEP) sr = MAX_FIRST_STEP;
    else if (sr < 1) sr = 1;

    while (sr>>=1)
        step++;

    *sp = MAX_MVSEARCH_STEPS - 1 - step;
}

1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
static int vp8_rd_pick_best_mbsegmentation(VP8_COMP *cpi, MACROBLOCK *x,
                                           MV *best_ref_mv, int best_rd,
                                           int *mdcounts, int *returntotrate,
                                           int *returnyrate, int *returndistortion,
                                           int mvthresh)
{
    int i;
    BEST_SEG_INFO bsi;

    vpx_memset(&bsi, 0, sizeof(bsi));

    bsi.segment_rd = best_rd;
    bsi.ref_mv = best_ref_mv;
1314
    bsi.mvp = best_ref_mv;
1315
1316
1317
1318
1319
1320
1321
    bsi.mvthresh = mvthresh;
    bsi.mdcounts = mdcounts;

    for(i = 0; i < 16; i++)
    {
        bsi.modes[i] = ZERO4X4;
    }
1322

1323
1324
1325
1326
    if(cpi->compressor_speed == 0)
    {
        /* for now, we will keep the original segmentation order
           when in best quality mode */
1327
1328
1329
1330
        rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
        rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
        rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
        rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
1331
1332
1333
    }
    else
    {
1334
1335
        int sr;

1336
        rd_check_segment(cpi, x, &bsi, BLOCK_8X8);
1337

1338
1339
        if (bsi.segment_rd < best_rd)
        {
Yunqing Wang's avatar
Yunqing Wang committed
1340
1341
1342
1343
            int col_min = (best_ref_mv->col - MAX_FULL_PEL_VAL) >>3;
            int col_max = (best_ref_mv->col + MAX_FULL_PEL_VAL) >>3;
            int row_min = (best_ref_mv->row - MAX_FULL_PEL_VAL) >>3;
            int row_max = (best_ref_mv->row + MAX_FULL_PEL_VAL) >>3;
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360

            int tmp_col_min = x->mv_col_min;
            int tmp_col_max = x->mv_col_max;
            int tmp_row_min = x->mv_row_min;
            int tmp_row_max = x->mv_row_max;

            /* Get intersection of UMV window and valid MV window to reduce # of checks in diamond search. */
            if (x->mv_col_min < col_min )
                x->mv_col_min = col_min;
            if (x->mv_col_max > col_max )
                x->mv_col_max = col_max;
            if (x->mv_row_min < row_min )
                x->mv_row_min = row_min;
            if (x->mv_row_max > row_max )
                x->mv_row_max = row_max;

            /* Get 8x8 result */
1361
1362
1363
1364
1365
            bsi.sv_mvp[0] = bsi.mvs[0].as_mv;
            bsi.sv_mvp[1] = bsi.mvs[2].as_mv;
            bsi.sv_mvp[2] = bsi.mvs[8].as_mv;
            bsi.sv_mvp[3] = bsi.mvs[10].as_mv;

1366
1367
            /* Use 8x8 result as 16x8/8x16's predictor MV. Adjust search range according to the closeness of 2 MV. */
            /* block 8X16 */
1368
1369
1370
1371
1372
1373
1374
            {
                sr = MAXF((abs(bsi.sv_mvp[0].row - bsi.sv_mvp[2].row))>>3, (abs(bsi.sv_mvp[0].col - bsi.sv_mvp[2].col))>>3);
                vp8_cal_step_param(sr, &bsi.sv_istep[0]);

                sr = MAXF((abs(bsi.sv_mvp[1].row - bsi.sv_mvp[3].row))>>3, (abs(bsi.sv_mvp[1].col - bsi.sv_mvp[3].col))>>3);
                vp8_cal_step_param(sr, &bsi.sv_istep[1]);

1375
                rd_check_segment(cpi, x, &bsi, BLOCK_8X16);
1376
1377
            }

1378
            /* block 16X8 */
1379
1380
1381
1382
1383
1384
1385
            {
                sr = MAXF((abs(bsi.sv_mvp[0].row - bsi.sv_mvp[1].row))>>3, (abs(bsi.sv_mvp[0].col - bsi.sv_mvp[1].col))>>3);
                vp8_cal_step_param(sr, &bsi.sv_istep[0]);

                sr = MAXF((abs(bsi.sv_mvp[2].row - bsi.sv_mvp[3].row))>>3, (abs(bsi.sv_mvp[2].col - bsi.sv_mvp[3].col))>>3);
                vp8_cal_step_param(sr, &bsi.sv_istep[1]);

1386
                rd_check_segment(cpi, x, &bsi, BLOCK_16X8);
1387
1388
            }

1389
            /* If 8x8 is better than 16x8/8x16, then do 4x4 search */
1390
1391
            /* Not skip 4x4 if speed=0 (good quality) */
            if (cpi->sf.no_skip_block4x4_search || bsi.segment_num == BLOCK_8X8)  /* || (sv_segment_rd8x8-bsi.segment_rd) < sv_segment_rd8x8>>5) */
1392
1393
            {
                bsi.mvp = &bsi.sv_mvp[0];
1394
                rd_check_segment(cpi, x, &bsi, BLOCK_4X4);
1395
            }
1396
1397
1398
1399
1400
1401

            /* restore UMV window */
            x->mv_col_min = tmp_col_min;
            x->mv_col_max = tmp_col_max;
            x->mv_row_min = tmp_row_min;
            x->mv_row_max = tmp_row_max;
1402
1403
        }
    }
1404
1405

    /* set it to the best */
John Koleszar's avatar
John Koleszar committed
1406
1407
1408
1409
    for (i = 0; i < 16; i++)
    {
        BLOCKD *bd = &x->e_mbd.block[i];

1410
        bd->bmi.mv.as_mv = bsi.mvs[i].as_mv;
1411
1412
        bd->bmi.mode = bsi.modes[i];
        bd->eob = bsi.eobs[i];
John Koleszar's avatar
John Koleszar committed
1413
1414
    }

1415
1416
1417
    *returntotrate = bsi.r;
    *returndistortion = bsi.d;
    *returnyrate = bsi.segment_yrate;
John Koleszar's avatar
John Koleszar committed
1418

1419
1420
1421
    /* save partitions */
    x->e_mbd.mode_info_context->mbmi.partitioning = bsi.segment_num;
    x->partition_info->count = vp8_mbsplit_count[bsi.segment_num];
John Koleszar's avatar
John Koleszar committed
1422

1423
    for (i = 0; i < x->partition_info->count; i++)
John Koleszar's avatar
John Koleszar committed
1424
1425
1426
    {
        int j;

1427
        j = vp8_mbsplit_offset[bsi.segment_num][i];
John Koleszar's avatar
John Koleszar committed
1428

1429
1430
        x->partition_info->bmi[i].mode = x->e_mbd.block[j].bmi.mode;
        x->partition_info->bmi[i].mv.as_mv = x->e_mbd.block[j].bmi.mv.as_mv;
John Koleszar's avatar
John Koleszar committed
1431
1432
    }

1433
    return bsi.segment_rd;
John Koleszar's avatar
John Koleszar committed
1434
1435
}

1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
static void swap(int *x,int *y)
{
   int tmp;

   tmp = *x;
   *x = *y;
   *y = tmp;
}

static void quicksortmv(int arr[],int left, int right)
{
   int lidx,ridx,pivot;

   lidx = left;
   ridx = right;

   if( left < right)
   {
      pivot = (left + right)/2;

      while(lidx <=pivot && ridx >=pivot)
      {
          while(arr[lidx] < arr[pivot] && lidx <= pivot)
              lidx++;
          while(arr[ridx] > arr[pivot] && ridx >= pivot)
              ridx--;
          swap(&arr[lidx], &arr[ridx]);
          lidx++;
          ridx--;
          if(lidx-1 == pivot)
          {
              ridx++;
              pivot = ridx;
          }
          else if(ridx+1 == pivot)
          {
              lidx--;
              pivot = lidx;
          }
      }
      quicksortmv(arr, left, pivot - 1);
      quicksortmv(arr, pivot + 1, right);
   }
}

static void quicksortsad(int arr[],int idx[], int left, int right)
{
   int lidx,ridx,pivot;

   lidx = left;
   ridx = right;

   if( left < right)
   {
      pivot = (left + right)/2;

      while(lidx <=pivot && ridx >=pivot)
      {
          while(arr[lidx] < arr[pivot] && lidx <= pivot)
              lidx++;
          while(arr[ridx] > arr[pivot] && ridx >= pivot)
              ridx--;
          swap(&arr[lidx], &arr[ridx]);
          swap(&idx[lidx], &idx[ridx]);
          lidx++;
          ridx--;
          if(lidx-1 == pivot)
          {
              ridx++;
              pivot = ridx;
          }
          else if(ridx+1 == pivot)
          {
              lidx--;
              pivot = lidx;
          }
      }
      quicksortsad(arr, idx, left, pivot - 1);
      quicksortsad(arr, idx, pivot + 1, right);
   }
}

//The improved MV prediction
Yunqing Wang's avatar
Yunqing Wang committed
1519
void vp8_mv_pred
Yunqing Wang's avatar