rdopt.c 133 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"
Christian Duvivier's avatar
Christian Duvivier committed
28
#include "vp8/common/quant_common.h"
John Koleszar's avatar
John Koleszar committed
29
30
#include "encodemb.h"
#include "quantize.h"
John Koleszar's avatar
John Koleszar committed
31
32
#include "vp8/common/idct.h"
#include "vp8/common/g_common.h"
John Koleszar's avatar
John Koleszar committed
33
34
#include "variance.h"
#include "mcomp.h"
Yunqing Wang's avatar
Yunqing Wang committed
35
#include "rdopt.h"
Paul Wilkins's avatar
Paul Wilkins committed
36
#include "ratectrl.h"
John Koleszar's avatar
John Koleszar committed
37
38
#include "vpx_mem/vpx_mem.h"
#include "dct.h"
John Koleszar's avatar
John Koleszar committed
39
#include "vp8/common/systemdependent.h"
John Koleszar's avatar
John Koleszar committed
40

41
#include "vp8/common/seg_common.h"
42
#include "vp8/common/pred_common.h"
43

John Koleszar's avatar
John Koleszar committed
44
45
46
47
48
49
#if CONFIG_RUNTIME_CPU_DETECT
#define IF_RTCD(x)  (x)
#else
#define IF_RTCD(x)  NULL
#endif

Scott LaVarnway's avatar
Scott LaVarnway committed
50
51
52
extern void vp8cx_mb_init_quantizer(VP8_COMP *cpi, MACROBLOCK *x);
extern void vp8_update_zbin_extra(VP8_COMP *cpi, MACROBLOCK *x);

53
54
55
56
57
58
#if CONFIG_HIGH_PRECISION_MV
#define XMVCOST (x->e_mbd.allow_high_precision_mv?x->mvcost_hp:x->mvcost)
#else
#define XMVCOST (x->mvcost)
#endif

John Koleszar's avatar
John Koleszar committed
59
60
#define MAXF(a,b)            (((a) > (b)) ? (a) : (b))

61
62
#define INVALID_MV 0x80008000

63
static const int auto_speed_thresh[17] =
John Koleszar's avatar
John Koleszar committed
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
{
    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,
114
    I8X8_PRED,
115

116
    /* compound prediction modes */
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
    ZEROMV,
    NEARESTMV,
    NEARMV,

    ZEROMV,
    NEARESTMV,
    NEARMV,

    ZEROMV,
    NEARESTMV,
    NEARMV,

    NEWMV,
    NEWMV,
    NEWMV,
132
133
134
135

    SPLITMV,
    SPLITMV,
    SPLITMV,
John Koleszar's avatar
John Koleszar committed
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
};

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,
168
    INTRA_FRAME,
169

170
    /* compound prediction modes */
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
    LAST_FRAME,
    LAST_FRAME,
    LAST_FRAME,

    ALTREF_FRAME,
    ALTREF_FRAME,
    ALTREF_FRAME,

    GOLDEN_FRAME,
    GOLDEN_FRAME,
    GOLDEN_FRAME,

    LAST_FRAME,
    ALTREF_FRAME,
    GOLDEN_FRAME,
186
187
188
189

    LAST_FRAME,
    ALTREF_FRAME,
    GOLDEN_FRAME,
190
191
192
193
194
195
196
197
};

const MV_REFERENCE_FRAME vp8_second_ref_frame_order[MAX_MODES] =
{
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
    0,

198
    /* compound prediction modes */
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
    GOLDEN_FRAME,
    GOLDEN_FRAME,
    GOLDEN_FRAME,

    LAST_FRAME,
    LAST_FRAME,
    LAST_FRAME,

    ALTREF_FRAME,
    ALTREF_FRAME,
    ALTREF_FRAME,

    GOLDEN_FRAME,
    LAST_FRAME,
    ALTREF_FRAME,
214
215
216
217

    GOLDEN_FRAME,
    LAST_FRAME,
    ALTREF_FRAME,
John Koleszar's avatar
John Koleszar committed
218
219
220
};

static void fill_token_costs(
221
222
    unsigned int c      [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [MAX_ENTROPY_TOKENS],
    const vp8_prob p    [BLOCK_TYPES] [COEF_BANDS] [PREV_COEF_CONTEXTS] [ENTROPY_NODES]
John Koleszar's avatar
John Koleszar committed
223
224
225
226
227
228
229
230
)
{
    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++)
231
232
233
234
235
236
237
238
            {

                if(k == 0 && ((j > 0 && i > 0) || (j > 1 && i == 0)))
                    vp8_cost_tokens_skip((int *)(c [i][j][k]), p [i][j][k], vp8_coef_tree);
                else

                    vp8_cost_tokens((int *)(c [i][j][k]), p [i][j][k], vp8_coef_tree);
            }
John Koleszar's avatar
John Koleszar committed
239
240
241

}

Paul Wilkins's avatar
Paul Wilkins committed
242
static int rd_iifactor [ 32 ] =  {    4,   4,   3,   2,   1,   0,   0,   0,
John Koleszar's avatar
John Koleszar committed
243
244
245
246
247
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                      0,   0,   0,   0,   0,   0,   0,   0,
                                 };

248
// 3* dc_qlookup[Q]*dc_qlookup[Q];
249

250
/* values are now correlated to quantizer */
Paul Wilkins's avatar
Paul Wilkins committed
251
252
253
254
static int sad_per_bit16lut[QINDEX_RANGE];
static int sad_per_bit4lut[QINDEX_RANGE];

void vp8_init_me_luts()
John Koleszar's avatar
John Koleszar committed
255
{
Paul Wilkins's avatar
Paul Wilkins committed
256
257
258
259
260
261
262
263
264
265
266
267
    int i;

    // Initialize the sad lut tables using a formulaic calculation for now
    // This is to make it easier to resolve the impact of experimental changes
    // to the quantizer tables.
    for ( i = 0; i < QINDEX_RANGE; i++ )
    {
        sad_per_bit16lut[i] =
            (int)((0.0418*vp8_convert_qindex_to_q(i)) + 2.4107);
        sad_per_bit4lut[i] = (int)((0.063*vp8_convert_qindex_to_q(i)) + 2.742);
    }
}
John Koleszar's avatar
John Koleszar committed
268

269
270
271
272
273
int compute_rd_mult( int qindex )
{
    int q;

    q = vp8_dc_quant(qindex,0);
274
    return (11 * q * q) >> 6;
275
276
}

John Koleszar's avatar
John Koleszar committed
277
278
void vp8cx_initialize_me_consts(VP8_COMP *cpi, int QIndex)
{
Yaowu Xu's avatar
Yaowu Xu committed
279
280
    cpi->mb.sadperbit16 =  sad_per_bit16lut[QIndex];
    cpi->mb.sadperbit4  =  sad_per_bit4lut[QIndex];
John Koleszar's avatar
John Koleszar committed
281
282
}

283
284

void vp8_initialize_rd_consts(VP8_COMP *cpi, int QIndex)
John Koleszar's avatar
John Koleszar committed
285
286
287
288
289
290
{
    int q;
    int i;

    vp8_clear_system_state();  //__asm emms;

291
292
293
294
    // 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)
Paul Wilkins's avatar
Paul Wilkins committed
295
    QIndex=(QIndex<0)? 0 : ((QIndex>MAXQ)?MAXQ : QIndex);
296
297

    cpi->RDMULT = compute_rd_mult(QIndex);
John Koleszar's avatar
John Koleszar committed
298

299
300
301
302
303
304
305
306
    // Extend rate multiplier along side quantizer zbin increases
    if (cpi->zbin_over_quant  > 0)
    {
        double oq_factor;

        // 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);
307
        cpi->RDMULT = (int)((double)cpi->RDMULT * oq_factor * oq_factor);
308
    }
John Koleszar's avatar
John Koleszar committed
309

Paul Wilkins's avatar
Paul Wilkins committed
310
    if (cpi->pass == 2 && (cpi->common.frame_type != KEY_FRAME))
John Koleszar's avatar
John Koleszar committed
311
    {
312
        if (cpi->twopass.next_iiratio > 31)
Paul Wilkins's avatar
Paul Wilkins committed
313
            cpi->RDMULT += (cpi->RDMULT * rd_iifactor[31]) >> 4;
John Koleszar's avatar
John Koleszar committed
314
        else
315
316
            cpi->RDMULT +=
                (cpi->RDMULT * rd_iifactor[cpi->twopass.next_iiratio]) >> 4;
John Koleszar's avatar
John Koleszar committed
317
318
    }

319
320
    if (cpi->RDMULT < 7)
        cpi->RDMULT = 7;
321

322
    cpi->mb.errorperbit = (cpi->RDMULT / 110);
323
324
    cpi->mb.errorperbit += (cpi->mb.errorperbit==0);

John Koleszar's avatar
John Koleszar committed
325
326
    vp8_set_speed_features(cpi);

327
328
    q = (int)pow(vp8_dc_quant(QIndex,0)>>2, 1.25);
    q = q << 2;
329
    cpi->RDMULT = cpi->RDMULT << 4;
Paul Wilkins's avatar
Paul Wilkins committed
330

331
332
    if (q < 8)
        q = 8;
333

John Koleszar's avatar
John Koleszar committed
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
    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
    );

377
378
379
380
    fill_token_costs(
        cpi->mb.token_costs_8x8,
        (const vp8_prob( *)[8][3][11]) cpi->common.fc.coef_probs_8x8
    );
Yaowu Xu's avatar
Yaowu Xu committed
381
    /*rough estimate for costing*/
382
    cpi->common.kf_ymode_probs_index = cpi->common.base_qindex>>4;
John Koleszar's avatar
John Koleszar committed
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
    vp8_init_mode_costs(cpi);

}

void vp8_auto_select_speed(VP8_COMP *cpi)
{
    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
Johann's avatar
Johann committed
409
410
411
412
    if( cpi->oxcf.cpu_used > 16)
        cpi->oxcf.cpu_used = 16;
    if( cpi->oxcf.cpu_used < -16)
        cpi->oxcf.cpu_used = -16;
John Koleszar's avatar
John Koleszar committed
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
    */

    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;
                }
            }

435
            if (milliseconds_for_compress * 100 > cpi->avg_encode_time * auto_speed_thresh[cpi->Speed])
John Koleszar's avatar
John Koleszar committed
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
            {
                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;
532
533
    int mv_row = x->e_mbd.mode_info_context->mbmi.mv.as_mv.row;
    int mv_col = x->e_mbd.mode_info_context->mbmi.mv.as_mv.col;
John Koleszar's avatar
John Koleszar committed
534
535
536
    int offset;
    int pre_stride = x->e_mbd.block[16].pre_stride;

537
538
539
540
541
542
543
544
545
546
547
548
    if (mv_row < 0)
        mv_row -= 1;
    else
        mv_row += 1;

    if (mv_col < 0)
        mv_col -= 1;
    else
        mv_col += 1;

    mv_row /= 2;
    mv_col /= 2;
John Koleszar's avatar
John Koleszar committed
549
550
551
552
553
554
555

    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)
    {
556
557
558
559
560
561
#if CONFIG_SIXTEENTH_SUBPEL_UV
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(uptr, pre_stride,
            (mv_col & 7)<<1, (mv_row & 7)<<1, upred_ptr, uv_stride, &sse2);
        VARIANCE_INVOKE(rtcd, subpixvar8x8)(vptr, pre_stride,
            (mv_col & 7)<<1, (mv_row & 7)<<1, vpred_ptr, uv_stride, &sse1);
#else
562
563
564
565
        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);
566
#endif
John Koleszar's avatar
John Koleszar committed
567
568
569
570
        sse2 += sse1;
    }
    else
    {
571
572
573
574
        VARIANCE_INVOKE(rtcd, var8x8)(uptr, pre_stride,
            upred_ptr, uv_stride, &sse2);
        VARIANCE_INVOKE(rtcd, var8x8)(vptr, pre_stride,
            vpred_ptr, uv_stride, &sse1);
John Koleszar's avatar
John Koleszar committed
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
        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
613
static int vp8_rdcost_mby(MACROBLOCK *mb)
John Koleszar's avatar
John Koleszar committed
614
615
616
617
{
    int cost = 0;
    int b;
    MACROBLOCKD *x = &mb->e_mbd;
618
619
620
621
622
623
    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
624

625
626
    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
John Koleszar's avatar
John Koleszar committed
627
628

    for (b = 0; b < 16; b++)
Scott LaVarnway's avatar
Scott LaVarnway committed
629
        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_Y_NO_DC,
630
                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
John Koleszar's avatar
John Koleszar committed
631

Scott LaVarnway's avatar
Scott LaVarnway committed
632
    cost += cost_coeffs(mb, x->block + 24, PLANE_TYPE_Y2,
Scott LaVarnway's avatar
Scott LaVarnway committed
633
                ta + vp8_block2above[24], tl + vp8_block2left[24]);
John Koleszar's avatar
John Koleszar committed
634
635
636
637

    return cost;
}

638
639
640
static void macro_block_yrd( MACROBLOCK *mb,
                             int *Rate,
                             int *Distortion,
641
                             const VP8_ENCODER_RTCD *rtcd)
642
643
644
645
646
647
648
649
650
{
    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;

651
652
653
654
655
    ENCODEMB_INVOKE(&rtcd->encodemb, submby)(
        mb->src_diff,
        *(mb->block[0].base_src),
        mb->e_mbd.predictor,
        mb->block[0].src_stride );
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677

    // 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
678
    d = ENCODEMB_INVOKE(&rtcd->encodemb, mberr)(mb, 1);
679

680
    d += ENCODEMB_INVOKE(&rtcd->encodemb, berr)(mb_y2->coeff, x_y2->dqcoeff);
681

682
    *Distortion = (d >> 2);
683
684
685
    // rate
    *Rate = vp8_rdcost_mby(mb);
}
John Koleszar's avatar
John Koleszar committed
686

687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724

static int cost_coeffs_2x2(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);
    assert(eob<=4);

# define QC2X2( I)  ( qcoeff_ptr [vp8_default_zig_zag1d[I]] )

    for (; c < eob; c++)
    {
        int v = QC2X2(c);
        int t = vp8_dct_value_tokens_ptr[v].Token;
        cost += mb->token_costs_8x8[type] [vp8_coef_bands[c]] [pt] [t];
        cost += vp8_dct_value_cost_ptr[v];
        pt = vp8_prev_token_class[t];
    }

# undef QC2X2
    if (c < 4)
        cost += mb->token_costs_8x8 [type][vp8_coef_bands[c]]
                                    [pt] [DCT_EOB_TOKEN];

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


static int cost_coeffs_8x8(MACROBLOCK *mb,
                           BLOCKD *b, int type,
725
                           ENTROPY_CONTEXT *a, ENTROPY_CONTEXT *l)
726
727
728
729
730
731
732
{
    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;

733
    VP8_COMBINEENTROPYCONTEXTS(pt, *a, *l);
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771

# define QC8X8( I)  ( qcoeff_ptr [vp8_default_zig_zag1d_8x8[I]] )

    for (; c < eob; c++)
    {
        int v = QC8X8(c);
        int t = vp8_dct_value_tokens_ptr[v].Token;
        cost += mb->token_costs_8x8[type] [vp8_coef_bands_8x8[c]] [pt] [t];
        cost += vp8_dct_value_cost_ptr[v];
        pt = vp8_prev_token_class[t];
    }

# undef QC8X8
    if (c < 64)
        cost += mb->token_costs_8x8 [type][vp8_coef_bands_8x8[c]]
                                    [pt] [DCT_EOB_TOKEN];

    pt = (c != !type); // is eob first coefficient;
    *a = *l = pt;
    return cost;
}
static int vp8_rdcost_mby_8x8(MACROBLOCK *mb)
{
    int cost = 0;
    int b;
    MACROBLOCKD *x = &mb->e_mbd;
    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));

    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;

    for (b = 0; b < 16; b+=4)
        cost += cost_coeffs_8x8(mb, x->block + b, PLANE_TYPE_Y_NO_DC,
772
                    ta + vp8_block2above_8x8[b], tl + vp8_block2left_8x8[b]);
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789

    cost += cost_coeffs_2x2(mb, x->block + 24, PLANE_TYPE_Y2,
                ta + vp8_block2above[24], tl + vp8_block2left[24]);
    return cost;
}

static void macro_block_yrd_8x8( MACROBLOCK *mb,
                             int *Rate,
                             int *Distortion,
                             const VP8_ENCODER_RTCD *rtcd)
{
    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;
    int d;

790
791
792
793
794
    ENCODEMB_INVOKE(&rtcd->encodemb, submby)(
        mb->src_diff,
        *(mb->block[0].base_src),
        mb->e_mbd.predictor,
        mb->block[0].src_stride );
795
796
797
798
799
800
801
802
803
804
805
806
807
808

    vp8_transform_mby_8x8(mb);
    vp8_quantize_mby_8x8(mb);

    /* remove 1st order dc to properly combine 1st/2nd order distortion */
    mb->coeff[0] = 0;
    mb->coeff[64] = 0;
    mb->coeff[128] = 0;
    mb->coeff[192] = 0;
    mb->e_mbd.dqcoeff[0] = 0;
    mb->e_mbd.dqcoeff[64] = 0;
    mb->e_mbd.dqcoeff[128] = 0;
    mb->e_mbd.dqcoeff[192] = 0;

809
810
    d = ENCODEMB_INVOKE(&rtcd->encodemb, mberr)(mb, 0);
    d += ENCODEMB_INVOKE(&rtcd->encodemb, berr)(mb_y2->coeff, x_y2->dqcoeff);
811

812
    *Distortion = (d >> 2);
813
814
815
    // rate
    *Rate = vp8_rdcost_mby_8x8(mb);
}
816

817
static void copy_predictor(unsigned char *dst, const unsigned char *predictor)
818
{
819
820
821
822
823
824
    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];
825
}
Yaowu Xu's avatar
Yaowu Xu committed
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848

static void copy_predictor_8x8(unsigned char *dst, const unsigned char *predictor)
{
    const unsigned int *p = (const unsigned int *)predictor;
    unsigned int *d = (unsigned int *)dst;
    d[0] = p[0];
    d[1] = p[1];
    d[4] = p[4];
    d[5] = p[5];
    d[8] = p[8];
    d[9] = p[9];
    d[12] = p[12];
    d[13] = p[13];
    d[16] = p[16];
    d[17] = p[17];
    d[20] = p[20];
    d[21] = p[21];
    d[24] = p[24];
    d[25] = p[25];
    d[28] = p[28];
    d[29] = p[29];
}

849
static int rd_pick_intra4x4block(
John Koleszar's avatar
John Koleszar committed
850
851
852
853
854
    VP8_COMP *cpi,
    MACROBLOCK *x,
    BLOCK *be,
    BLOCKD *b,
    B_PREDICTION_MODE *best_mode,
855
856
#if CONFIG_COMP_INTRA_PRED
    B_PREDICTION_MODE *best_second_mode,
857
    int allow_comp,
858
#endif
859
    unsigned int *bmode_costs,
John Koleszar's avatar
John Koleszar committed
860
861
862
863
864
    ENTROPY_CONTEXT *a,
    ENTROPY_CONTEXT *l,

    int *bestrate,
    int *bestratey,
865
    int *bestdistortion)
John Koleszar's avatar
John Koleszar committed
866
867
{
    B_PREDICTION_MODE mode;
868
869
870
#if CONFIG_COMP_INTRA_PRED
    B_PREDICTION_MODE mode2;
#endif
871
    int best_rd = INT_MAX;
John Koleszar's avatar
John Koleszar committed
872
873
874
875
876
    int rate = 0;
    int distortion;

    ENTROPY_CONTEXT ta = *a, tempa = *a;
    ENTROPY_CONTEXT tl = *l, templ = *l;
877
878
879
880
881
882
883
    /*
     * 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
884
885
886

    for (mode = B_DC_PRED; mode <= B_HU_PRED; mode++)
    {
887
#if CONFIG_COMP_INTRA_PRED
888
        for (mode2 = (allow_comp ? 0 : (B_DC_PRED - 1)); mode2 != (allow_comp ? (mode + 1) : 0); mode2++)
889
890
        {
#endif
John Koleszar's avatar
John Koleszar committed
891
892
893
        int this_rd;
        int ratey;

Adrian Grange's avatar
Adrian Grange committed
894
895
896
897
898
899
        // TODO Temporarily ignore modes that need the above-right data. SB
        // encoding means this data is not available for the bottom right MB
        // Do we need to do this for mode2 also?
        if (mode==B_LD_PRED || mode==B_VL_PRED)
            continue;

900
901
        rate = bmode_costs[mode];

902
903
904
905
#if CONFIG_COMP_INTRA_PRED
            if (mode2 == (B_PREDICTION_MODE) (B_DC_PRED - 1))
            {
#endif
906
907
        RECON_INVOKE(&cpi->rtcd.common->recon, intra4x4_predict)
                     (b, mode, b->predictor);
908
909
910
911
912
913
914
915
916
#if CONFIG_COMP_INTRA_PRED
            }
            else
            {
                RECON_INVOKE(&cpi->rtcd.common->recon, comp_intra4x4_predict)
                    (b, mode, mode2, b->predictor);
                rate += bmode_costs[mode2];
            }
#endif
917
918
919
        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
920
921
922
923

        tempa = ta;
        templ = tl;

Scott LaVarnway's avatar
Scott LaVarnway committed
924
        ratey = cost_coeffs(x, b, PLANE_TYPE_Y_WITH_DC, &tempa, &templ);
John Koleszar's avatar
John Koleszar committed
925
        rate += ratey;
Yaowu Xu's avatar
Yaowu Xu committed
926
927
        distortion = ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)(
            be->coeff, b->dqcoeff) >> 2;
John Koleszar's avatar
John Koleszar committed
928
929
930
931
932
933
934
935
936
937

        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;
938
939
940
#if CONFIG_COMP_INTRA_PRED
            *best_second_mode = mode2;
#endif
John Koleszar's avatar
John Koleszar committed
941
942
            *a = tempa;
            *l = templ;
943
944
            copy_predictor(best_predictor, b->predictor);
            vpx_memcpy(best_dqcoeff, b->dqcoeff, 32);
945
946
947
#if CONFIG_COMP_INTRA_PRED
        }
#endif
John Koleszar's avatar
John Koleszar committed
948
949
        }
    }
950
951
952
953
    b->bmi.as_mode.first = (B_PREDICTION_MODE)(*best_mode);
#if CONFIG_COMP_INTRA_PRED
    b->bmi.as_mode.second = (B_PREDICTION_MODE)(*best_second_mode);
#endif
954

955
956
    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
957

958
    return best_rd;
John Koleszar's avatar
John Koleszar committed
959
960
}

961
static int rd_pick_intra4x4mby_modes(VP8_COMP *cpi, MACROBLOCK *mb, int *Rate,
962
                                     int *rate_y, int *Distortion, int best_rd,
963
964
965
966
#if CONFIG_COMP_INTRA_PRED
                                     int allow_comp,
#endif
                                     int update_contexts)
John Koleszar's avatar
John Koleszar committed
967
968
{
    int i;
Adrian Grange's avatar
Adrian Grange committed
969
    MACROBLOCKD *const xd = &mb->e_mbd;
John Koleszar's avatar
John Koleszar committed
970
971
972
    int cost = mb->mbmode_cost [xd->frame_type] [B_PRED];
    int distortion = 0;
    int tot_rate_y = 0;
973
    int64_t total_rd = 0;
974
975
976
    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;
977
    unsigned int *bmode_costs;
978

Adrian Grange's avatar
Adrian Grange committed
979
980
981
982
983
984
985
986
987
988
989
    if (update_contexts)
    {
        ta = (ENTROPY_CONTEXT *)mb->e_mbd.above_context;
        tl = (ENTROPY_CONTEXT *)mb->e_mbd.left_context;
    }
    else
    {
        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));
990

Adrian Grange's avatar
Adrian Grange committed
991
992
993
        ta = (ENTROPY_CONTEXT *)&t_above;
        tl = (ENTROPY_CONTEXT *)&t_left;
    }
John Koleszar's avatar
John Koleszar committed
994

Adrian Grange's avatar
Adrian Grange committed
995
#if 0
John Koleszar's avatar
John Koleszar committed
996
    vp8_intra_prediction_down_copy(xd);
Adrian Grange's avatar
Adrian Grange committed
997
#endif
John Koleszar's avatar
John Koleszar committed
998

999
1000
    bmode_costs = mb->inter_bmode_costs;

John Koleszar's avatar
John Koleszar committed
1001
1002
1003
1004
1005
    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);
1006
1007
1008
#if CONFIG_COMP_INTRA_PRED
        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_second_mode);
#endif
John Koleszar's avatar
John Koleszar committed
1009
1010
        int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(ry), UNINITIALIZED_IS_SAFE(d);

1011
1012
        if (mb->e_mbd.frame_type == KEY_FRAME)
        {
Scott LaVarnway's avatar
Scott LaVarnway committed
1013
1014
            const B_PREDICTION_MODE A = above_block_mode(mic, i, mis);
            const B_PREDICTION_MODE L = left_block_mode(mic, i);
1015
1016
1017
1018

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

1019
        total_rd += rd_pick_intra4x4block(
1020
1021
            cpi, mb, mb->block + i, xd->block + i, &best_mode,
#if CONFIG_COMP_INTRA_PRED
1022
            &best_second_mode, allow_comp,
1023
1024
#endif
            bmode_costs, ta + vp8_block2above[i],
1025
            tl + vp8_block2left[i], &r, &ry, &d);
John Koleszar's avatar
John Koleszar committed
1026
1027
1028
1029

        cost += r;
        distortion += d;
        tot_rate_y += ry;
Scott LaVarnway's avatar
Scott LaVarnway committed
1030

1031
1032
1033
1034
        mic->bmi[i].as_mode.first = best_mode;
#if CONFIG_COMP_INTRA_PRED
        mic->bmi[i].as_mode.second = best_second_mode;
#endif
1035

1036
        if(total_rd >= (int64_t)best_rd)
1037
            break;
John Koleszar's avatar
John Koleszar committed
1038
1039
    }

1040
    if(total_rd >= (int64_t)best_rd)
1041
        return INT_MAX;
1042

1043
1044
1045
1046
#if CONFIG_COMP_INTRA_PRED
    cost += vp8_cost_bit(128, allow_comp);
#endif
    *Rate = cost;
John Koleszar's avatar
John Koleszar committed
1047
1048
1049
1050
1051
    *rate_y += tot_rate_y;
    *Distortion = distortion;

    return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
}
1052
1053
1054
1055
1056
1057
1058


static int rd_pick_intra16x16mby_mode(VP8_COMP *cpi,
                                      MACROBLOCK *x,
                                      int *Rate,
                                      int *rate_y,
                                      int *Distortion)
John Koleszar's avatar
John Koleszar committed
1059
1060
1061
{
    MB_PREDICTION_MODE mode;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
1062
1063
1064
1065
#if CONFIG_COMP_INTRA_PRED
    MB_PREDICTION_MODE mode2;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode2_selected);
#endif
John Koleszar's avatar
John Koleszar committed
1066
    int rate, ratey;
1067
    int distortion;
John Koleszar's avatar
John Koleszar committed
1068
    int best_rd = INT_MAX;
1069
    int this_rd;
John Koleszar's avatar
John Koleszar committed
1070
1071
1072
1073

    //Y Search for 16x16 intra prediction mode
    for (mode = DC_PRED; mode <= TM_PRED; mode++)
    {
1074
        x->e_mbd.mode_info_context->mbmi.mode = mode;
1075
1076
1077
1078
1079
1080
1081
#if CONFIG_COMP_INTRA_PRED
        for (mode2 = DC_PRED - 1; mode2 != TM_PRED + 1; mode2++)
        {
            x->e_mbd.mode_info_context->mbmi.second_mode = mode2;
            if (mode2 == (MB_PREDICTION_MODE) (DC_PRED - 1))
            {
#endif
1082
1083
        RECON_INVOKE(&cpi->common.rtcd.recon, build_intra_predictors_mby)
            (&x->e_mbd);
1084
1085
1086
1087
1088
1089
1090
1091
#if CONFIG_COMP_INTRA_PRED
            }
            else
            {
                continue; // i.e. disable for now
                RECON_INVOKE(&cpi->common.rtcd.recon, build_comp_intra_predictors_mby)(&x->e_mbd);
            }
#endif
John Koleszar's avatar
John Koleszar committed
1092

1093
        macro_block_yrd(x, &ratey, &distortion, IF_RTCD(&cpi->rtcd));
1094
1095
            // FIXME add compoundmode cost
            // FIXME add rate for mode2
1096
1097
        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
1098
1099
1100
1101
1102
1103

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

        if (this_rd < best_rd)
        {
            mode_selected = mode;
1104
1105
1106
#if CONFIG_COMP_INTRA_PRED
            mode2_selected = mode2;
#endif
John Koleszar's avatar
John Koleszar committed
1107
1108
1109
            best_rd = this_rd;
            *Rate = rate;
            *rate_y = ratey;
1110
            *Distortion = distortion;
John Koleszar's avatar
John Koleszar committed
1111
        }
1112
1113
1114
#if CONFIG_COMP_INTRA_PRED
        }
#endif
John Koleszar's avatar
John Koleszar committed
1115
1116
    }

1117
    x->e_mbd.mode_info_context->mbmi.mode = mode_selected;
1118
1119
1120
#if CONFIG_COMP_INTRA_PRED
    x->e_mbd.mode_info_context->mbmi.second_mode = mode2_selected;
#endif
John Koleszar's avatar
John Koleszar committed
1121
1122
    return best_rd;
}
Yaowu Xu's avatar
Yaowu Xu committed
1123
1124
1125
1126
1127
static int rd_pick_intra8x8block(
    VP8_COMP *cpi,
    MACROBLOCK *x,
    int ib,
    B_PREDICTION_MODE *best_mode,
1128
1129
1130
#if CONFIG_COMP_INTRA_PRED
    B_PREDICTION_MODE *best_second_mode,
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1131
1132
1133
1134
1135
1136
1137
1138
    unsigned int *mode_costs,
    ENTROPY_CONTEXT *a,
    ENTROPY_CONTEXT *l,
    int *bestrate,
    int *bestratey,
    int *bestdistortion)
{
    MB_PREDICTION_MODE mode;
1139
1140
1141
#if CONFIG_COMP_INTRA_PRED
    MB_PREDICTION_MODE mode2;
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
    MACROBLOCKD *xd = &x->e_mbd;
    int best_rd = INT_MAX;
    int rate = 0;
    int distortion;
    BLOCK  *be=x->block + ib;
    BLOCKD *b=x->e_mbd.block + ib;
    ENTROPY_CONTEXT ta0, ta1, besta0, besta1;
    ENTROPY_CONTEXT tl0, tl1, bestl0, bestl1;


    /*
     * 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 8x8 block
     * */

    DECLARE_ALIGNED_ARRAY(16, unsigned char,  best_predictor, 16*8);
    DECLARE_ALIGNED_ARRAY(16, short, best_dqcoeff, 16*4);

    for (mode = DC_PRED; mode <= TM_PRED; mode++)
    {
1163
1164
1165
1166
#if CONFIG_COMP_INTRA_PRED
        for (mode2 = DC_PRED - 1; mode2 != TM_PRED + 1; mode2++)
        {
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1167
1168
1169
        int this_rd;
        int rate_t;

1170
        // FIXME rate for compound mode and second intrapred mode
Yaowu Xu's avatar
Yaowu Xu committed
1171
1172
        rate = mode_costs[mode];

1173
1174
1175
1176
#if CONFIG_COMP_INTRA_PRED
            if (mode2 == (MB_PREDICTION_MODE) (DC_PRED - 1))
            {
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1177
1178
        RECON_INVOKE(&cpi->rtcd.common->recon, intra8x8_predict)
                     (b, mode, b->predictor);
1179
1180
1181
1182
1183
1184
1185
1186
1187
#if CONFIG_COMP_INTRA_PRED
            }
            else
            {
                continue; // i.e. disable for now
                RECON_INVOKE(&cpi->rtcd.common->recon, comp_intra8x8_predict)
                    (b, mode, mode2, b->predictor);
            }
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232

        vp8_subtract_4b_c(be, b, 16);

        x->vp8_short_fdct8x4(be->src_diff, be->coeff, 32);
        x->vp8_short_fdct8x4(be->src_diff + 64, be->coeff + 64, 32);

        x->quantize_b_pair(x->block+ib, x->block+ib+1,
                            xd->block+ib, xd->block+ib+1);
        x->quantize_b_pair(x->block+ib+4, x->block+ib+5,
                            xd->block+ib+4, xd->block+ib+5);

        distortion = ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)
            ((x->block+ib)->coeff,(xd->block+ib)->dqcoeff)>>2;
        distortion += ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)
            ((x->block+ib+1)->coeff,(xd->block+ib+1)->dqcoeff)>>2;
        distortion += ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)
            ((x->block+ib+4)->coeff,(xd->block+ib+4)->dqcoeff)>>2;
        distortion += ENCODEMB_INVOKE(IF_RTCD(&cpi->rtcd.encodemb), berr)
            ((x->block+ib+5)->coeff,(xd->block+ib+5)->dqcoeff)>>2;

        ta0 = *(a + vp8_block2above[ib]);
        ta1 = *(a + vp8_block2above[ib+1]);
        tl0 = *(l + vp8_block2above[ib]);
        tl1 = *(l + vp8_block2above[ib+4]);
        rate_t = cost_coeffs(x, xd->block+ib, PLANE_TYPE_Y_WITH_DC,
            &ta0, &tl0);
        rate_t += cost_coeffs(x, xd->block+ib+1, PLANE_TYPE_Y_WITH_DC,
            &ta1, &tl0);
        rate_t += cost_coeffs(x, xd->block+ib+4, PLANE_TYPE_Y_WITH_DC,
            &ta0, &tl1);
        rate_t += cost_coeffs(x, xd->block+ib+5, PLANE_TYPE_Y_WITH_DC,
            &ta1, &tl1);
        rate += rate_t;
        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
        if (this_rd < best_rd)
        {
            *bestrate = rate;
            *bestratey = rate_t;
            *bestdistortion = distortion;
            besta0 = ta0;
            besta1 = ta1;
            bestl0 = tl0;
            bestl1 = tl1;
            best_rd = this_rd;
            *best_mode = mode;
1233
1234
1235
#if CONFIG_COMP_INTRA_PRED
            *best_second_mode = mode2;
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1236
1237
1238
            copy_predictor_8x8(best_predictor, b->predictor);
            vpx_memcpy(best_dqcoeff, b->dqcoeff, 64);
            vpx_memcpy(best_dqcoeff+32, b->dqcoeff+64, 64);
1239
1240
1241
#if CONFIG_COMP_INTRA_PRED
        }
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1242
1243
        }
    }
1244
1245
1246
1247
    b->bmi.as_mode.first = (*best_mode);
#if CONFIG_COMP_INTRA_PRED
    b->bmi.as_mode.second = (*best_second_mode);
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
    vp8_encode_intra8x8 (IF_RTCD(&cpi->rtcd), x, ib);
    *(a + vp8_block2above[ib])   = besta0;
    *(a + vp8_block2above[ib+1]) = besta1;
    *(l + vp8_block2above[ib])   = bestl0;
    *(l + vp8_block2above[ib+4]) = bestl1;
    return best_rd;
}

const int vp8_i8x8_block[4]={0, 2, 8, 10};
int rd_pick_intra8x8mby_modes(VP8_COMP *cpi,
                                      MACROBLOCK *mb,
                                      int *Rate,
                                      int *rate_y,
                                      int *Distortion,
                                      int best_rd)
{
    MACROBLOCKD *const xd = &mb->e_mbd;
    int i,ib;
    int cost = mb->mbmode_cost [xd->frame_type] [I8X8_PRED];
    int distortion = 0;
    int tot_rate_y = 0;
    long long total_rd = 0;
    ENTROPY_CONTEXT_PLANES t_above, t_left;
    ENTROPY_CONTEXT *ta;
    ENTROPY_CONTEXT *tl;
    unsigned int *i8x8mode_costs;

    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;

    i8x8mode_costs  = mb->i8x8_mode_costs;

    for (i = 0; i < 4; i++)
    {
        MODE_INFO *const mic = xd->mode_info_context;
        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_mode);
1287
1288
1289
#if CONFIG_COMP_INTRA_PRED
        B_PREDICTION_MODE UNINITIALIZED_IS_SAFE(best_second_mode);
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1290
1291
1292
1293
        int UNINITIALIZED_IS_SAFE(r), UNINITIALIZED_IS_SAFE(ry), UNINITIALIZED_IS_SAFE(d);

        ib = vp8_i8x8_block[i];
        total_rd += rd_pick_intra8x8block(
1294
1295
1296
1297
1298
            cpi, mb, ib, &best_mode,
#if CONFIG_COMP_INTRA_PRED
            &best_second_mode,
#endif
            i8x8mode_costs, ta, tl, &r, &ry, &d);
Yaowu Xu's avatar
Yaowu Xu committed
1299
1300
1301
        cost += r;
        distortion += d;
        tot_rate_y += ry;
1302
1303
1304
1305
        mic->bmi[ib].as_mode.first = best_mode;
#if CONFIG_COMP_INTRA_PRED
        mic->bmi[ib].as_mode.second = best_second_mode;
#endif
Yaowu Xu's avatar
Yaowu Xu committed
1306
1307
1308
1309
1310
1311
    }
    *Rate = cost;
    *rate_y += tot_rate_y;
    *Distortion = distortion;
    return RDCOST(mb->rdmult, mb->rddiv, cost, distortion);
}
John Koleszar's avatar
John Koleszar committed
1312
1313
1314
1315
1316
1317

static int rd_cost_mbuv(MACROBLOCK *mb)
{
    int b;
    int cost = 0;
    MACROBLOCKD *x = &mb->e_mbd;
1318
1319
1320
1321
1322
1323
    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
1324

1325
1326
    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;
John Koleszar's avatar
John Koleszar committed
1327

Scott LaVarnway's avatar
Scott LaVarnway committed
1328
1329
    for (b = 16; b < 24; b++)
        cost += cost_coeffs(mb, x->block + b, PLANE_TYPE_UV,
1330
                    ta + vp8_block2above[b], tl + vp8_block2left[b]);
John Koleszar's avatar
John Koleszar committed
1331
1332
1333
1334
1335

    return cost;
}


1336
1337
static int rd_inter16x16_uv(VP8_COMP *cpi, MACROBLOCK *x, int *rate,
                            int *distortion, int fullpixel)
John Koleszar's avatar
John Koleszar committed
1338
{
1339
1340
    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);
John Koleszar's avatar
John Koleszar committed
1341

1342
1343
1344
1345
1346
1347
1348
1349
    vp8_transform_mbuv(x);
    vp8_quantize_mbuv(x);

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

    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
}
Yaowu Xu's avatar
Yaowu Xu committed
1350

1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
static int rd_cost_mbuv_8x8(MACROBLOCK *mb)
{
    int b;
    int cost = 0;
    MACROBLOCKD *x = &mb->e_mbd;
    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));

    ta = (ENTROPY_CONTEXT *)&t_above;
    tl = (ENTROPY_CONTEXT *)&t_left;

    for (b = 16; b < 24; b+=4)
        cost += cost_coeffs_8x8(mb, x->block + b, PLANE_TYPE_UV,
1368
1369
                                ta + vp8_block2above_8x8[b],
                                tl + vp8_block2left_8x8[b]);
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389

    return cost;
}


static int rd_inter16x16_uv_8x8(VP8_COMP *cpi, MACROBLOCK *x, int *rate,
                            int *distortion, int fullpixel)
{
    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_8x8(x);

    vp8_quantize_mbuv_8x8(x);

    *rate       = rd_cost_mbuv_8x8(x);
    *distortion = ENCODEMB_INVOKE(&cpi->rtcd.encodemb, mbuverr)(x) / 4;

    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
}
Yaowu Xu's avatar
Yaowu Xu committed
1390

1391
1392
1393
1394
1395
1396
1397
1398
1399
1400

static int rd_inter4x4_uv(VP8_COMP *cpi, MACROBLOCK *x, int *rate,
                          int *distortion, int fullpixel)
{
    vp8_build_inter4x4_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
1401
1402
1403
1404

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

1405
    return RDCOST(x->rdmult, x->rddiv, *rate, *distortion);
John Koleszar's avatar
John Koleszar committed
1406
1407
}

1408
1409
1410
1411
1412
static void rd_pick_intra_mbuv_mode(VP8_COMP *cpi,
                                    MACROBLOCK *x,
                                    int *rate,
                                    int *rate_tokenonly,
                                    int *distortion)
John Koleszar's avatar
John Koleszar committed
1413
1414
1415
{
    MB_PREDICTION_MODE mode;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode_selected);
1416
1417
1418
1419
#if CONFIG_COMP_INTRA_PRED
    MB_PREDICTION_MODE mode2;
    MB_PREDICTION_MODE UNINITIALIZED_IS_SAFE(mode2_selected);
#endif
John Koleszar's avatar
John Koleszar committed
1420
1421
1422
1423
1424
1425
    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++)
    {
1426
1427
1428
1429
#if CONFIG_COMP_INTRA_PRED
        for (mode2 = DC_PRED - 1; mode2 != TM_PRED + 1; mode2++)
        {
#endif
John Koleszar's avatar
John Koleszar committed
1430
1431
1432
1433
        int rate;
        int distortion;
        int this_rd;

1434
        x->e_mbd.mode_info_context->mbmi.uv_mode = mode;
1435
1436
1437
1438
1439
#if CONFIG_COMP_INTRA_PRED
            x->e_mbd.mode_info_context->mbmi.second_uv_mode = mode2;
            if (mode2 == (MB_PREDICTION_MODE) (DC_PRED - 1))
            {
#endif
1440
1441
        RECON_INVOKE(&cpi->rtcd.common->recon, build_intra_predictors_mbuv)
                     (&x->e_mbd);
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
#if CONFIG_COMP_INTRA_PRED
            }
            else
            {
                continue;
                RECON_INVOKE(&cpi->rtcd.common->recon, build_comp_intra_predictors_mbuv)
                    (&x->e_mbd);
            }
#endif

1452
1453
1454
1455
1456
        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
1457
1458

        rate_to = rd_cost_mbuv(x);
Yaowu Xu's avatar
Yaowu Xu committed
1459
1460
1461
        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
1462

1463
        distortion = ENCODEMB_INVOKE(&cpi->rtcd.encodemb, mbuverr)(x) / 4;
John Koleszar's avatar
John Koleszar committed
1464

1465
        this_rd = RDCOST(x->rdmult, x->rddiv, rate, distortion);
John Koleszar's avatar
John Koleszar committed
1466
1467
1468
1469
1470
1471
1472
1473

        if (this_rd < best_rd)
        {
            best_rd = this_rd;
            d = distortion;
            r = rate;
            *rate_tokenonly = rate_to;
            mode_selected = mode;
1474
1475
1476
1477
#if CONFIG_COMP_INTRA_PRED
            mode2_selected = mode2;
        }
#endif
John Koleszar's avatar
John Koleszar committed
1478
1479
1480
1481
1482
1483
        }
    }

    *rate = r;
    *distortion = d;

1484
    x->e_mbd.mode_info_context->mbmi.uv_mode = mode_selected;
1485
1486
1487
#if CONFIG_COMP_INTRA_PRED