summaryrefslogtreecommitdiffstats
path: root/third_party/aom/av1/common/reconintra.c
blob: 71a52e73e517443995edc972de97a9fcac0d2e0d (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
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
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
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
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
725
726
727
728
729
730
731
732
733
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
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
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
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
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
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
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
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
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
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
/*
 * Copyright (c) 2016, Alliance for Open Media. All rights reserved
 *
 * This source code is subject to the terms of the BSD 2 Clause License and
 * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
 * was not distributed with this source code in the LICENSE file, you can
 * obtain it at www.aomedia.org/license/software. If the Alliance for Open
 * Media Patent License 1.0 was not distributed with this source code in the
 * PATENTS file, you can obtain it at www.aomedia.org/license/patent.
 */

#include <math.h>

#include "config/aom_config.h"
#include "config/aom_dsp_rtcd.h"
#include "config/av1_rtcd.h"

#include "aom_dsp/aom_dsp_common.h"
#include "aom_mem/aom_mem.h"
#include "aom_ports/aom_once.h"
#include "aom_ports/mem.h"
#include "aom_ports/system_state.h"
#include "av1/common/reconintra.h"
#include "av1/common/onyxc_int.h"
#include "av1/common/cfl.h"

enum {
  NEED_LEFT = 1 << 1,
  NEED_ABOVE = 1 << 2,
  NEED_ABOVERIGHT = 1 << 3,
  NEED_ABOVELEFT = 1 << 4,
  NEED_BOTTOMLEFT = 1 << 5,
};

#define INTRA_EDGE_FILT 3
#define INTRA_EDGE_TAPS 5
#define MAX_UPSAMPLE_SZ 16

static const uint8_t extend_modes[INTRA_MODES] = {
  NEED_ABOVE | NEED_LEFT,                   // DC
  NEED_ABOVE,                               // V
  NEED_LEFT,                                // H
  NEED_ABOVE | NEED_ABOVERIGHT,             // D45
  NEED_LEFT | NEED_ABOVE | NEED_ABOVELEFT,  // D135
  NEED_LEFT | NEED_ABOVE | NEED_ABOVELEFT,  // D113
  NEED_LEFT | NEED_ABOVE | NEED_ABOVELEFT,  // D157
  NEED_LEFT | NEED_BOTTOMLEFT,              // D203
  NEED_ABOVE | NEED_ABOVERIGHT,             // D67
  NEED_LEFT | NEED_ABOVE,                   // SMOOTH
  NEED_LEFT | NEED_ABOVE,                   // SMOOTH_V
  NEED_LEFT | NEED_ABOVE,                   // SMOOTH_H
  NEED_LEFT | NEED_ABOVE | NEED_ABOVELEFT,  // PAETH
};

// Tables to store if the top-right reference pixels are available. The flags
// are represented with bits, packed into 8-bit integers. E.g., for the 32x32
// blocks in a 128x128 superblock, the index of the "o" block is 10 (in raster
// order), so its flag is stored at the 3rd bit of the 2nd entry in the table,
// i.e. (table[10 / 8] >> (10 % 8)) & 1.
//       . . . .
//       . . . .
//       . . o .
//       . . . .
static uint8_t has_tr_4x4[128] = {
  255, 255, 255, 255, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  127, 127, 127, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  255, 127, 255, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  127, 127, 127, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  255, 255, 255, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  127, 127, 127, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  255, 127, 255, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
  127, 127, 127, 127, 85, 85, 85, 85, 119, 119, 119, 119, 85, 85, 85, 85,
};
static uint8_t has_tr_4x8[64] = {
  255, 255, 255, 255, 119, 119, 119, 119, 127, 127, 127, 127, 119,
  119, 119, 119, 255, 127, 255, 127, 119, 119, 119, 119, 127, 127,
  127, 127, 119, 119, 119, 119, 255, 255, 255, 127, 119, 119, 119,
  119, 127, 127, 127, 127, 119, 119, 119, 119, 255, 127, 255, 127,
  119, 119, 119, 119, 127, 127, 127, 127, 119, 119, 119, 119,
};
static uint8_t has_tr_8x4[64] = {
  255, 255, 0, 0, 85, 85, 0, 0, 119, 119, 0, 0, 85, 85, 0, 0,
  127, 127, 0, 0, 85, 85, 0, 0, 119, 119, 0, 0, 85, 85, 0, 0,
  255, 127, 0, 0, 85, 85, 0, 0, 119, 119, 0, 0, 85, 85, 0, 0,
  127, 127, 0, 0, 85, 85, 0, 0, 119, 119, 0, 0, 85, 85, 0, 0,
};
static uint8_t has_tr_8x8[32] = {
  255, 255, 85, 85, 119, 119, 85, 85, 127, 127, 85, 85, 119, 119, 85, 85,
  255, 127, 85, 85, 119, 119, 85, 85, 127, 127, 85, 85, 119, 119, 85, 85,
};
static uint8_t has_tr_8x16[16] = {
  255, 255, 119, 119, 127, 127, 119, 119,
  255, 127, 119, 119, 127, 127, 119, 119,
};
static uint8_t has_tr_16x8[16] = {
  255, 0, 85, 0, 119, 0, 85, 0, 127, 0, 85, 0, 119, 0, 85, 0,
};
static uint8_t has_tr_16x16[8] = {
  255, 85, 119, 85, 127, 85, 119, 85,
};
static uint8_t has_tr_16x32[4] = { 255, 119, 127, 119 };
static uint8_t has_tr_32x16[4] = { 15, 5, 7, 5 };
static uint8_t has_tr_32x32[2] = { 95, 87 };
static uint8_t has_tr_32x64[1] = { 127 };
static uint8_t has_tr_64x32[1] = { 19 };
static uint8_t has_tr_64x64[1] = { 7 };
static uint8_t has_tr_64x128[1] = { 3 };
static uint8_t has_tr_128x64[1] = { 1 };
static uint8_t has_tr_128x128[1] = { 1 };
static uint8_t has_tr_4x16[32] = {
  255, 255, 255, 255, 127, 127, 127, 127, 255, 127, 255,
  127, 127, 127, 127, 127, 255, 255, 255, 127, 127, 127,
  127, 127, 255, 127, 255, 127, 127, 127, 127, 127,
};
static uint8_t has_tr_16x4[32] = {
  255, 0, 0, 0, 85, 0, 0, 0, 119, 0, 0, 0, 85, 0, 0, 0,
  127, 0, 0, 0, 85, 0, 0, 0, 119, 0, 0, 0, 85, 0, 0, 0,
};
static uint8_t has_tr_8x32[8] = {
  255, 255, 127, 127, 255, 127, 127, 127,
};
static uint8_t has_tr_32x8[8] = {
  15, 0, 5, 0, 7, 0, 5, 0,
};
static uint8_t has_tr_16x64[2] = { 255, 127 };
static uint8_t has_tr_64x16[2] = { 3, 1 };

static const uint8_t *const has_tr_tables[BLOCK_SIZES_ALL] = {
  // 4X4
  has_tr_4x4,
  // 4X8,       8X4,            8X8
  has_tr_4x8, has_tr_8x4, has_tr_8x8,
  // 8X16,      16X8,           16X16
  has_tr_8x16, has_tr_16x8, has_tr_16x16,
  // 16X32,     32X16,          32X32
  has_tr_16x32, has_tr_32x16, has_tr_32x32,
  // 32X64,     64X32,          64X64
  has_tr_32x64, has_tr_64x32, has_tr_64x64,
  // 64x128,    128x64,         128x128
  has_tr_64x128, has_tr_128x64, has_tr_128x128,
  // 4x16,      16x4,            8x32
  has_tr_4x16, has_tr_16x4, has_tr_8x32,
  // 32x8,      16x64,           64x16
  has_tr_32x8, has_tr_16x64, has_tr_64x16
};

static uint8_t has_tr_vert_8x8[32] = {
  255, 255, 0, 0, 119, 119, 0, 0, 127, 127, 0, 0, 119, 119, 0, 0,
  255, 127, 0, 0, 119, 119, 0, 0, 127, 127, 0, 0, 119, 119, 0, 0,
};
static uint8_t has_tr_vert_16x16[8] = {
  255, 0, 119, 0, 127, 0, 119, 0,
};
static uint8_t has_tr_vert_32x32[2] = { 15, 7 };
static uint8_t has_tr_vert_64x64[1] = { 3 };

// The _vert_* tables are like the ordinary tables above, but describe the
// order we visit square blocks when doing a PARTITION_VERT_A or
// PARTITION_VERT_B. This is the same order as normal except for on the last
// split where we go vertically (TL, BL, TR, BR). We treat the rectangular block
// as a pair of squares, which means that these tables work correctly for both
// mixed vertical partition types.
//
// There are tables for each of the square sizes. Vertical rectangles (like
// BLOCK_16X32) use their respective "non-vert" table
static const uint8_t *const has_tr_vert_tables[BLOCK_SIZES] = {
  // 4X4
  NULL,
  // 4X8,      8X4,         8X8
  has_tr_4x8, NULL, has_tr_vert_8x8,
  // 8X16,     16X8,        16X16
  has_tr_8x16, NULL, has_tr_vert_16x16,
  // 16X32,    32X16,       32X32
  has_tr_16x32, NULL, has_tr_vert_32x32,
  // 32X64,    64X32,       64X64
  has_tr_32x64, NULL, has_tr_vert_64x64,
  // 64x128,   128x64,      128x128
  has_tr_64x128, NULL, has_tr_128x128
};

static const uint8_t *get_has_tr_table(PARTITION_TYPE partition,
                                       BLOCK_SIZE bsize) {
  const uint8_t *ret = NULL;
  // If this is a mixed vertical partition, look up bsize in orders_vert.
  if (partition == PARTITION_VERT_A || partition == PARTITION_VERT_B) {
    assert(bsize < BLOCK_SIZES);
    ret = has_tr_vert_tables[bsize];
  } else {
    ret = has_tr_tables[bsize];
  }
  assert(ret);
  return ret;
}

static int has_top_right(const AV1_COMMON *cm, BLOCK_SIZE bsize, int mi_row,
                         int mi_col, int top_available, int right_available,
                         PARTITION_TYPE partition, TX_SIZE txsz, int row_off,
                         int col_off, int ss_x, int ss_y) {
  if (!top_available || !right_available) return 0;

  const int bw_unit = block_size_wide[bsize] >> tx_size_wide_log2[0];
  const int plane_bw_unit = AOMMAX(bw_unit >> ss_x, 1);
  const int top_right_count_unit = tx_size_wide_unit[txsz];

  if (row_off > 0) {  // Just need to check if enough pixels on the right.
    if (block_size_wide[bsize] > block_size_wide[BLOCK_64X64]) {
      // Special case: For 128x128 blocks, the transform unit whose
      // top-right corner is at the center of the block does in fact have
      // pixels available at its top-right corner.
      if (row_off == mi_size_high[BLOCK_64X64] >> ss_y &&
          col_off + top_right_count_unit == mi_size_wide[BLOCK_64X64] >> ss_x) {
        return 1;
      }
      const int plane_bw_unit_64 = mi_size_wide[BLOCK_64X64] >> ss_x;
      const int col_off_64 = col_off % plane_bw_unit_64;
      return col_off_64 + top_right_count_unit < plane_bw_unit_64;
    }
    return col_off + top_right_count_unit < plane_bw_unit;
  } else {
    // All top-right pixels are in the block above, which is already available.
    if (col_off + top_right_count_unit < plane_bw_unit) return 1;

    const int bw_in_mi_log2 = mi_size_wide_log2[bsize];
    const int bh_in_mi_log2 = mi_size_high_log2[bsize];
    const int sb_mi_size = mi_size_high[cm->seq_params.sb_size];
    const int blk_row_in_sb = (mi_row & (sb_mi_size - 1)) >> bh_in_mi_log2;
    const int blk_col_in_sb = (mi_col & (sb_mi_size - 1)) >> bw_in_mi_log2;

    // Top row of superblock: so top-right pixels are in the top and/or
    // top-right superblocks, both of which are already available.
    if (blk_row_in_sb == 0) return 1;

    // Rightmost column of superblock (and not the top row): so top-right pixels
    // fall in the right superblock, which is not available yet.
    if (((blk_col_in_sb + 1) << bw_in_mi_log2) >= sb_mi_size) {
      return 0;
    }

    // General case (neither top row nor rightmost column): check if the
    // top-right block is coded before the current block.
    const int this_blk_index =
        ((blk_row_in_sb + 0) << (MAX_MIB_SIZE_LOG2 - bw_in_mi_log2)) +
        blk_col_in_sb + 0;
    const int idx1 = this_blk_index / 8;
    const int idx2 = this_blk_index % 8;
    const uint8_t *has_tr_table = get_has_tr_table(partition, bsize);
    return (has_tr_table[idx1] >> idx2) & 1;
  }
}

// Similar to the has_tr_* tables, but store if the bottom-left reference
// pixels are available.
static uint8_t has_bl_4x4[128] = {
  84, 85, 85, 85, 16, 17, 17, 17, 84, 85, 85, 85, 0,  1,  1,  1,  84, 85, 85,
  85, 16, 17, 17, 17, 84, 85, 85, 85, 0,  0,  1,  0,  84, 85, 85, 85, 16, 17,
  17, 17, 84, 85, 85, 85, 0,  1,  1,  1,  84, 85, 85, 85, 16, 17, 17, 17, 84,
  85, 85, 85, 0,  0,  0,  0,  84, 85, 85, 85, 16, 17, 17, 17, 84, 85, 85, 85,
  0,  1,  1,  1,  84, 85, 85, 85, 16, 17, 17, 17, 84, 85, 85, 85, 0,  0,  1,
  0,  84, 85, 85, 85, 16, 17, 17, 17, 84, 85, 85, 85, 0,  1,  1,  1,  84, 85,
  85, 85, 16, 17, 17, 17, 84, 85, 85, 85, 0,  0,  0,  0,
};
static uint8_t has_bl_4x8[64] = {
  16, 17, 17, 17, 0, 1, 1, 1, 16, 17, 17, 17, 0, 0, 1, 0,
  16, 17, 17, 17, 0, 1, 1, 1, 16, 17, 17, 17, 0, 0, 0, 0,
  16, 17, 17, 17, 0, 1, 1, 1, 16, 17, 17, 17, 0, 0, 1, 0,
  16, 17, 17, 17, 0, 1, 1, 1, 16, 17, 17, 17, 0, 0, 0, 0,
};
static uint8_t has_bl_8x4[64] = {
  254, 255, 84, 85, 254, 255, 16, 17, 254, 255, 84, 85, 254, 255, 0, 1,
  254, 255, 84, 85, 254, 255, 16, 17, 254, 255, 84, 85, 254, 255, 0, 0,
  254, 255, 84, 85, 254, 255, 16, 17, 254, 255, 84, 85, 254, 255, 0, 1,
  254, 255, 84, 85, 254, 255, 16, 17, 254, 255, 84, 85, 254, 255, 0, 0,
};
static uint8_t has_bl_8x8[32] = {
  84, 85, 16, 17, 84, 85, 0, 1, 84, 85, 16, 17, 84, 85, 0, 0,
  84, 85, 16, 17, 84, 85, 0, 1, 84, 85, 16, 17, 84, 85, 0, 0,
};
static uint8_t has_bl_8x16[16] = {
  16, 17, 0, 1, 16, 17, 0, 0, 16, 17, 0, 1, 16, 17, 0, 0,
};
static uint8_t has_bl_16x8[16] = {
  254, 84, 254, 16, 254, 84, 254, 0, 254, 84, 254, 16, 254, 84, 254, 0,
};
static uint8_t has_bl_16x16[8] = {
  84, 16, 84, 0, 84, 16, 84, 0,
};
static uint8_t has_bl_16x32[4] = { 16, 0, 16, 0 };
static uint8_t has_bl_32x16[4] = { 78, 14, 78, 14 };
static uint8_t has_bl_32x32[2] = { 4, 4 };
static uint8_t has_bl_32x64[1] = { 0 };
static uint8_t has_bl_64x32[1] = { 34 };
static uint8_t has_bl_64x64[1] = { 0 };
static uint8_t has_bl_64x128[1] = { 0 };
static uint8_t has_bl_128x64[1] = { 0 };
static uint8_t has_bl_128x128[1] = { 0 };
static uint8_t has_bl_4x16[32] = {
  0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0,
  0, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 0, 0,
};
static uint8_t has_bl_16x4[32] = {
  254, 254, 254, 84, 254, 254, 254, 16, 254, 254, 254, 84, 254, 254, 254, 0,
  254, 254, 254, 84, 254, 254, 254, 16, 254, 254, 254, 84, 254, 254, 254, 0,
};
static uint8_t has_bl_8x32[8] = {
  0, 1, 0, 0, 0, 1, 0, 0,
};
static uint8_t has_bl_32x8[8] = {
  238, 78, 238, 14, 238, 78, 238, 14,
};
static uint8_t has_bl_16x64[2] = { 0, 0 };
static uint8_t has_bl_64x16[2] = { 42, 42 };

static const uint8_t *const has_bl_tables[BLOCK_SIZES_ALL] = {
  // 4X4
  has_bl_4x4,
  // 4X8,         8X4,         8X8
  has_bl_4x8, has_bl_8x4, has_bl_8x8,
  // 8X16,        16X8,        16X16
  has_bl_8x16, has_bl_16x8, has_bl_16x16,
  // 16X32,       32X16,       32X32
  has_bl_16x32, has_bl_32x16, has_bl_32x32,
  // 32X64,       64X32,       64X64
  has_bl_32x64, has_bl_64x32, has_bl_64x64,
  // 64x128,      128x64,      128x128
  has_bl_64x128, has_bl_128x64, has_bl_128x128,
  // 4x16,        16x4,        8x32
  has_bl_4x16, has_bl_16x4, has_bl_8x32,
  // 32x8,        16x64,       64x16
  has_bl_32x8, has_bl_16x64, has_bl_64x16
};

static uint8_t has_bl_vert_8x8[32] = {
  254, 255, 16, 17, 254, 255, 0, 1, 254, 255, 16, 17, 254, 255, 0, 0,
  254, 255, 16, 17, 254, 255, 0, 1, 254, 255, 16, 17, 254, 255, 0, 0,
};
static uint8_t has_bl_vert_16x16[8] = {
  254, 16, 254, 0, 254, 16, 254, 0,
};
static uint8_t has_bl_vert_32x32[2] = { 14, 14 };
static uint8_t has_bl_vert_64x64[1] = { 2 };

// The _vert_* tables are like the ordinary tables above, but describe the
// order we visit square blocks when doing a PARTITION_VERT_A or
// PARTITION_VERT_B. This is the same order as normal except for on the last
// split where we go vertically (TL, BL, TR, BR). We treat the rectangular block
// as a pair of squares, which means that these tables work correctly for both
// mixed vertical partition types.
//
// There are tables for each of the square sizes. Vertical rectangles (like
// BLOCK_16X32) use their respective "non-vert" table
static const uint8_t *const has_bl_vert_tables[BLOCK_SIZES] = {
  // 4X4
  NULL,
  // 4X8,     8X4,         8X8
  has_bl_4x8, NULL, has_bl_vert_8x8,
  // 8X16,    16X8,        16X16
  has_bl_8x16, NULL, has_bl_vert_16x16,
  // 16X32,   32X16,       32X32
  has_bl_16x32, NULL, has_bl_vert_32x32,
  // 32X64,   64X32,       64X64
  has_bl_32x64, NULL, has_bl_vert_64x64,
  // 64x128,  128x64,      128x128
  has_bl_64x128, NULL, has_bl_128x128
};

static const uint8_t *get_has_bl_table(PARTITION_TYPE partition,
                                       BLOCK_SIZE bsize) {
  const uint8_t *ret = NULL;
  // If this is a mixed vertical partition, look up bsize in orders_vert.
  if (partition == PARTITION_VERT_A || partition == PARTITION_VERT_B) {
    assert(bsize < BLOCK_SIZES);
    ret = has_bl_vert_tables[bsize];
  } else {
    ret = has_bl_tables[bsize];
  }
  assert(ret);
  return ret;
}

static int has_bottom_left(const AV1_COMMON *cm, BLOCK_SIZE bsize, int mi_row,
                           int mi_col, int bottom_available, int left_available,
                           PARTITION_TYPE partition, TX_SIZE txsz, int row_off,
                           int col_off, int ss_x, int ss_y) {
  if (!bottom_available || !left_available) return 0;

  // Special case for 128x* blocks, when col_off is half the block width.
  // This is needed because 128x* superblocks are divided into 64x* blocks in
  // raster order
  if (block_size_wide[bsize] > block_size_wide[BLOCK_64X64] && col_off > 0) {
    const int plane_bw_unit_64 = mi_size_wide[BLOCK_64X64] >> ss_x;
    const int col_off_64 = col_off % plane_bw_unit_64;
    if (col_off_64 == 0) {
      // We are at the left edge of top-right or bottom-right 64x* block.
      const int plane_bh_unit_64 = mi_size_high[BLOCK_64X64] >> ss_y;
      const int row_off_64 = row_off % plane_bh_unit_64;
      const int plane_bh_unit =
          AOMMIN(mi_size_high[bsize] >> ss_y, plane_bh_unit_64);
      // Check if all bottom-left pixels are in the left 64x* block (which is
      // already coded).
      return row_off_64 + tx_size_high_unit[txsz] < plane_bh_unit;
    }
  }

  if (col_off > 0) {
    // Bottom-left pixels are in the bottom-left block, which is not available.
    return 0;
  } else {
    const int bh_unit = block_size_high[bsize] >> tx_size_high_log2[0];
    const int plane_bh_unit = AOMMAX(bh_unit >> ss_y, 1);
    const int bottom_left_count_unit = tx_size_high_unit[txsz];

    // All bottom-left pixels are in the left block, which is already available.
    if (row_off + bottom_left_count_unit < plane_bh_unit) return 1;

    const int bw_in_mi_log2 = mi_size_wide_log2[bsize];
    const int bh_in_mi_log2 = mi_size_high_log2[bsize];
    const int sb_mi_size = mi_size_high[cm->seq_params.sb_size];
    const int blk_row_in_sb = (mi_row & (sb_mi_size - 1)) >> bh_in_mi_log2;
    const int blk_col_in_sb = (mi_col & (sb_mi_size - 1)) >> bw_in_mi_log2;

    // Leftmost column of superblock: so bottom-left pixels maybe in the left
    // and/or bottom-left superblocks. But only the left superblock is
    // available, so check if all required pixels fall in that superblock.
    if (blk_col_in_sb == 0) {
      const int blk_start_row_off = blk_row_in_sb
                                        << (bh_in_mi_log2 + MI_SIZE_LOG2 -
                                            tx_size_wide_log2[0]) >>
                                    ss_y;
      const int row_off_in_sb = blk_start_row_off + row_off;
      const int sb_height_unit = sb_mi_size >> ss_y;
      return row_off_in_sb + bottom_left_count_unit < sb_height_unit;
    }

    // Bottom row of superblock (and not the leftmost column): so bottom-left
    // pixels fall in the bottom superblock, which is not available yet.
    if (((blk_row_in_sb + 1) << bh_in_mi_log2) >= sb_mi_size) return 0;

    // General case (neither leftmost column nor bottom row): check if the
    // bottom-left block is coded before the current block.
    const int this_blk_index =
        ((blk_row_in_sb + 0) << (MAX_MIB_SIZE_LOG2 - bw_in_mi_log2)) +
        blk_col_in_sb + 0;
    const int idx1 = this_blk_index / 8;
    const int idx2 = this_blk_index % 8;
    const uint8_t *has_bl_table = get_has_bl_table(partition, bsize);
    return (has_bl_table[idx1] >> idx2) & 1;
  }
}

typedef void (*intra_pred_fn)(uint8_t *dst, ptrdiff_t stride,
                              const uint8_t *above, const uint8_t *left);

static intra_pred_fn pred[INTRA_MODES][TX_SIZES_ALL];
static intra_pred_fn dc_pred[2][2][TX_SIZES_ALL];

typedef void (*intra_high_pred_fn)(uint16_t *dst, ptrdiff_t stride,
                                   const uint16_t *above, const uint16_t *left,
                                   int bd);
static intra_high_pred_fn pred_high[INTRA_MODES][TX_SIZES_ALL];
static intra_high_pred_fn dc_pred_high[2][2][TX_SIZES_ALL];

static void init_intra_predictors_internal(void) {
  assert(NELEMENTS(mode_to_angle_map) == INTRA_MODES);

#define INIT_RECTANGULAR(p, type)             \
  p[TX_4X8] = aom_##type##_predictor_4x8;     \
  p[TX_8X4] = aom_##type##_predictor_8x4;     \
  p[TX_8X16] = aom_##type##_predictor_8x16;   \
  p[TX_16X8] = aom_##type##_predictor_16x8;   \
  p[TX_16X32] = aom_##type##_predictor_16x32; \
  p[TX_32X16] = aom_##type##_predictor_32x16; \
  p[TX_32X64] = aom_##type##_predictor_32x64; \
  p[TX_64X32] = aom_##type##_predictor_64x32; \
  p[TX_4X16] = aom_##type##_predictor_4x16;   \
  p[TX_16X4] = aom_##type##_predictor_16x4;   \
  p[TX_8X32] = aom_##type##_predictor_8x32;   \
  p[TX_32X8] = aom_##type##_predictor_32x8;   \
  p[TX_16X64] = aom_##type##_predictor_16x64; \
  p[TX_64X16] = aom_##type##_predictor_64x16;

#define INIT_NO_4X4(p, type)                  \
  p[TX_8X8] = aom_##type##_predictor_8x8;     \
  p[TX_16X16] = aom_##type##_predictor_16x16; \
  p[TX_32X32] = aom_##type##_predictor_32x32; \
  p[TX_64X64] = aom_##type##_predictor_64x64; \
  INIT_RECTANGULAR(p, type)

#define INIT_ALL_SIZES(p, type)           \
  p[TX_4X4] = aom_##type##_predictor_4x4; \
  INIT_NO_4X4(p, type)

  INIT_ALL_SIZES(pred[V_PRED], v);
  INIT_ALL_SIZES(pred[H_PRED], h);
  INIT_ALL_SIZES(pred[PAETH_PRED], paeth);
  INIT_ALL_SIZES(pred[SMOOTH_PRED], smooth);
  INIT_ALL_SIZES(pred[SMOOTH_V_PRED], smooth_v);
  INIT_ALL_SIZES(pred[SMOOTH_H_PRED], smooth_h);
  INIT_ALL_SIZES(dc_pred[0][0], dc_128);
  INIT_ALL_SIZES(dc_pred[0][1], dc_top);
  INIT_ALL_SIZES(dc_pred[1][0], dc_left);
  INIT_ALL_SIZES(dc_pred[1][1], dc);

  INIT_ALL_SIZES(pred_high[V_PRED], highbd_v);
  INIT_ALL_SIZES(pred_high[H_PRED], highbd_h);
  INIT_ALL_SIZES(pred_high[PAETH_PRED], highbd_paeth);
  INIT_ALL_SIZES(pred_high[SMOOTH_PRED], highbd_smooth);
  INIT_ALL_SIZES(pred_high[SMOOTH_V_PRED], highbd_smooth_v);
  INIT_ALL_SIZES(pred_high[SMOOTH_H_PRED], highbd_smooth_h);
  INIT_ALL_SIZES(dc_pred_high[0][0], highbd_dc_128);
  INIT_ALL_SIZES(dc_pred_high[0][1], highbd_dc_top);
  INIT_ALL_SIZES(dc_pred_high[1][0], highbd_dc_left);
  INIT_ALL_SIZES(dc_pred_high[1][1], highbd_dc);
#undef intra_pred_allsizes
}

// Directional prediction, zone 1: 0 < angle < 90
void av1_dr_prediction_z1_c(uint8_t *dst, ptrdiff_t stride, int bw, int bh,
                            const uint8_t *above, const uint8_t *left,
                            int upsample_above, int dx, int dy) {
  int r, c, x, base, shift, val;

  (void)left;
  (void)dy;
  assert(dy == 1);
  assert(dx > 0);

  const int max_base_x = ((bw + bh) - 1) << upsample_above;
  const int frac_bits = 6 - upsample_above;
  const int base_inc = 1 << upsample_above;
  x = dx;
  for (r = 0; r < bh; ++r, dst += stride, x += dx) {
    base = x >> frac_bits;
    shift = ((x << upsample_above) & 0x3F) >> 1;

    if (base >= max_base_x) {
      for (int i = r; i < bh; ++i) {
        memset(dst, above[max_base_x], bw * sizeof(dst[0]));
        dst += stride;
      }
      return;
    }

    for (c = 0; c < bw; ++c, base += base_inc) {
      if (base < max_base_x) {
        val = above[base] * (32 - shift) + above[base + 1] * shift;
        dst[c] = ROUND_POWER_OF_TWO(val, 5);
      } else {
        dst[c] = above[max_base_x];
      }
    }
  }
}

// Directional prediction, zone 2: 90 < angle < 180
void av1_dr_prediction_z2_c(uint8_t *dst, ptrdiff_t stride, int bw, int bh,
                            const uint8_t *above, const uint8_t *left,
                            int upsample_above, int upsample_left, int dx,
                            int dy) {
  int r, c, x, y, shift1, shift2, val, base1, base2;

  assert(dx > 0);
  assert(dy > 0);

  const int min_base_x = -(1 << upsample_above);
  const int frac_bits_x = 6 - upsample_above;
  const int frac_bits_y = 6 - upsample_left;
  const int base_inc_x = 1 << upsample_above;
  x = -dx;
  for (r = 0; r < bh; ++r, x -= dx, dst += stride) {
    base1 = x >> frac_bits_x;
    y = (r << 6) - dy;
    for (c = 0; c < bw; ++c, base1 += base_inc_x, y -= dy) {
      if (base1 >= min_base_x) {
        shift1 = ((x * (1 << upsample_above)) & 0x3F) >> 1;
        val = above[base1] * (32 - shift1) + above[base1 + 1] * shift1;
        val = ROUND_POWER_OF_TWO(val, 5);
      } else {
        base2 = y >> frac_bits_y;
        assert(base2 >= -(1 << upsample_left));
        shift2 = ((y * (1 << upsample_left)) & 0x3F) >> 1;
        val = left[base2] * (32 - shift2) + left[base2 + 1] * shift2;
        val = ROUND_POWER_OF_TWO(val, 5);
      }
      dst[c] = val;
    }
  }
}

// Directional prediction, zone 3: 180 < angle < 270
void av1_dr_prediction_z3_c(uint8_t *dst, ptrdiff_t stride, int bw, int bh,
                            const uint8_t *above, const uint8_t *left,
                            int upsample_left, int dx, int dy) {
  int r, c, y, base, shift, val;

  (void)above;
  (void)dx;

  assert(dx == 1);
  assert(dy > 0);

  const int max_base_y = (bw + bh - 1) << upsample_left;
  const int frac_bits = 6 - upsample_left;
  const int base_inc = 1 << upsample_left;
  y = dy;
  for (c = 0; c < bw; ++c, y += dy) {
    base = y >> frac_bits;
    shift = ((y << upsample_left) & 0x3F) >> 1;

    for (r = 0; r < bh; ++r, base += base_inc) {
      if (base < max_base_y) {
        val = left[base] * (32 - shift) + left[base + 1] * shift;
        dst[r * stride + c] = val = ROUND_POWER_OF_TWO(val, 5);
      } else {
        for (; r < bh; ++r) dst[r * stride + c] = left[max_base_y];
        break;
      }
    }
  }
}

static void dr_predictor(uint8_t *dst, ptrdiff_t stride, TX_SIZE tx_size,
                         const uint8_t *above, const uint8_t *left,
                         int upsample_above, int upsample_left, int angle) {
  const int dx = av1_get_dx(angle);
  const int dy = av1_get_dy(angle);
  const int bw = tx_size_wide[tx_size];
  const int bh = tx_size_high[tx_size];
  assert(angle > 0 && angle < 270);

  if (angle > 0 && angle < 90) {
    av1_dr_prediction_z1(dst, stride, bw, bh, above, left, upsample_above, dx,
                         dy);
  } else if (angle > 90 && angle < 180) {
    av1_dr_prediction_z2(dst, stride, bw, bh, above, left, upsample_above,
                         upsample_left, dx, dy);
  } else if (angle > 180 && angle < 270) {
    av1_dr_prediction_z3(dst, stride, bw, bh, above, left, upsample_left, dx,
                         dy);
  } else if (angle == 90) {
    pred[V_PRED][tx_size](dst, stride, above, left);
  } else if (angle == 180) {
    pred[H_PRED][tx_size](dst, stride, above, left);
  }
}

// Directional prediction, zone 1: 0 < angle < 90
void av1_highbd_dr_prediction_z1_c(uint16_t *dst, ptrdiff_t stride, int bw,
                                   int bh, const uint16_t *above,
                                   const uint16_t *left, int upsample_above,
                                   int dx, int dy, int bd) {
  int r, c, x, base, shift, val;

  (void)left;
  (void)dy;
  (void)bd;
  assert(dy == 1);
  assert(dx > 0);

  const int max_base_x = ((bw + bh) - 1) << upsample_above;
  const int frac_bits = 6 - upsample_above;
  const int base_inc = 1 << upsample_above;
  x = dx;
  for (r = 0; r < bh; ++r, dst += stride, x += dx) {
    base = x >> frac_bits;
    shift = ((x << upsample_above) & 0x3F) >> 1;

    if (base >= max_base_x) {
      for (int i = r; i < bh; ++i) {
        aom_memset16(dst, above[max_base_x], bw);
        dst += stride;
      }
      return;
    }

    for (c = 0; c < bw; ++c, base += base_inc) {
      if (base < max_base_x) {
        val = above[base] * (32 - shift) + above[base + 1] * shift;
        dst[c] = ROUND_POWER_OF_TWO(val, 5);
      } else {
        dst[c] = above[max_base_x];
      }
    }
  }
}

// Directional prediction, zone 2: 90 < angle < 180
void av1_highbd_dr_prediction_z2_c(uint16_t *dst, ptrdiff_t stride, int bw,
                                   int bh, const uint16_t *above,
                                   const uint16_t *left, int upsample_above,
                                   int upsample_left, int dx, int dy, int bd) {
  int r, c, x, y, shift, val, base;

  (void)bd;
  assert(dx > 0);
  assert(dy > 0);

  const int min_base_x = -(1 << upsample_above);
  const int frac_bits_x = 6 - upsample_above;
  const int frac_bits_y = 6 - upsample_left;
  for (r = 0; r < bh; ++r) {
    for (c = 0; c < bw; ++c) {
      y = r + 1;
      x = (c << 6) - y * dx;
      base = x >> frac_bits_x;
      if (base >= min_base_x) {
        shift = ((x * (1 << upsample_above)) & 0x3F) >> 1;
        val = above[base] * (32 - shift) + above[base + 1] * shift;
        val = ROUND_POWER_OF_TWO(val, 5);
      } else {
        x = c + 1;
        y = (r << 6) - x * dy;
        base = y >> frac_bits_y;
        shift = ((y * (1 << upsample_left)) & 0x3F) >> 1;
        val = left[base] * (32 - shift) + left[base + 1] * shift;
        val = ROUND_POWER_OF_TWO(val, 5);
      }
      dst[c] = val;
    }
    dst += stride;
  }
}

// Directional prediction, zone 3: 180 < angle < 270
void av1_highbd_dr_prediction_z3_c(uint16_t *dst, ptrdiff_t stride, int bw,
                                   int bh, const uint16_t *above,
                                   const uint16_t *left, int upsample_left,
                                   int dx, int dy, int bd) {
  int r, c, y, base, shift, val;

  (void)above;
  (void)dx;
  (void)bd;
  assert(dx == 1);
  assert(dy > 0);

  const int max_base_y = (bw + bh - 1) << upsample_left;
  const int frac_bits = 6 - upsample_left;
  const int base_inc = 1 << upsample_left;
  y = dy;
  for (c = 0; c < bw; ++c, y += dy) {
    base = y >> frac_bits;
    shift = ((y << upsample_left) & 0x3F) >> 1;

    for (r = 0; r < bh; ++r, base += base_inc) {
      if (base < max_base_y) {
        val = left[base] * (32 - shift) + left[base + 1] * shift;
        dst[r * stride + c] = ROUND_POWER_OF_TWO(val, 5);
      } else {
        for (; r < bh; ++r) dst[r * stride + c] = left[max_base_y];
        break;
      }
    }
  }
}

static void highbd_dr_predictor(uint16_t *dst, ptrdiff_t stride,
                                TX_SIZE tx_size, const uint16_t *above,
                                const uint16_t *left, int upsample_above,
                                int upsample_left, int angle, int bd) {
  const int dx = av1_get_dx(angle);
  const int dy = av1_get_dy(angle);
  const int bw = tx_size_wide[tx_size];
  const int bh = tx_size_high[tx_size];
  assert(angle > 0 && angle < 270);

  if (angle > 0 && angle < 90) {
    av1_highbd_dr_prediction_z1(dst, stride, bw, bh, above, left,
                                upsample_above, dx, dy, bd);
  } else if (angle > 90 && angle < 180) {
    av1_highbd_dr_prediction_z2(dst, stride, bw, bh, above, left,
                                upsample_above, upsample_left, dx, dy, bd);
  } else if (angle > 180 && angle < 270) {
    av1_highbd_dr_prediction_z3(dst, stride, bw, bh, above, left, upsample_left,
                                dx, dy, bd);
  } else if (angle == 90) {
    pred_high[V_PRED][tx_size](dst, stride, above, left, bd);
  } else if (angle == 180) {
    pred_high[H_PRED][tx_size](dst, stride, above, left, bd);
  }
}

DECLARE_ALIGNED(16, const int8_t,
                av1_filter_intra_taps[FILTER_INTRA_MODES][8][8]) = {
  {
      { -6, 10, 0, 0, 0, 12, 0, 0 },
      { -5, 2, 10, 0, 0, 9, 0, 0 },
      { -3, 1, 1, 10, 0, 7, 0, 0 },
      { -3, 1, 1, 2, 10, 5, 0, 0 },
      { -4, 6, 0, 0, 0, 2, 12, 0 },
      { -3, 2, 6, 0, 0, 2, 9, 0 },
      { -3, 2, 2, 6, 0, 2, 7, 0 },
      { -3, 1, 2, 2, 6, 3, 5, 0 },
  },
  {
      { -10, 16, 0, 0, 0, 10, 0, 0 },
      { -6, 0, 16, 0, 0, 6, 0, 0 },
      { -4, 0, 0, 16, 0, 4, 0, 0 },
      { -2, 0, 0, 0, 16, 2, 0, 0 },
      { -10, 16, 0, 0, 0, 0, 10, 0 },
      { -6, 0, 16, 0, 0, 0, 6, 0 },
      { -4, 0, 0, 16, 0, 0, 4, 0 },
      { -2, 0, 0, 0, 16, 0, 2, 0 },
  },
  {
      { -8, 8, 0, 0, 0, 16, 0, 0 },
      { -8, 0, 8, 0, 0, 16, 0, 0 },
      { -8, 0, 0, 8, 0, 16, 0, 0 },
      { -8, 0, 0, 0, 8, 16, 0, 0 },
      { -4, 4, 0, 0, 0, 0, 16, 0 },
      { -4, 0, 4, 0, 0, 0, 16, 0 },
      { -4, 0, 0, 4, 0, 0, 16, 0 },
      { -4, 0, 0, 0, 4, 0, 16, 0 },
  },
  {
      { -2, 8, 0, 0, 0, 10, 0, 0 },
      { -1, 3, 8, 0, 0, 6, 0, 0 },
      { -1, 2, 3, 8, 0, 4, 0, 0 },
      { 0, 1, 2, 3, 8, 2, 0, 0 },
      { -1, 4, 0, 0, 0, 3, 10, 0 },
      { -1, 3, 4, 0, 0, 4, 6, 0 },
      { -1, 2, 3, 4, 0, 4, 4, 0 },
      { -1, 2, 2, 3, 4, 3, 3, 0 },
  },
  {
      { -12, 14, 0, 0, 0, 14, 0, 0 },
      { -10, 0, 14, 0, 0, 12, 0, 0 },
      { -9, 0, 0, 14, 0, 11, 0, 0 },
      { -8, 0, 0, 0, 14, 10, 0, 0 },
      { -10, 12, 0, 0, 0, 0, 14, 0 },
      { -9, 1, 12, 0, 0, 0, 12, 0 },
      { -8, 0, 0, 12, 0, 1, 11, 0 },
      { -7, 0, 0, 1, 12, 1, 9, 0 },
  },
};

void av1_filter_intra_predictor_c(uint8_t *dst, ptrdiff_t stride,
                                  TX_SIZE tx_size, const uint8_t *above,
                                  const uint8_t *left, int mode) {
  int r, c;
  uint8_t buffer[33][33];
  const int bw = tx_size_wide[tx_size];
  const int bh = tx_size_high[tx_size];

  assert(bw <= 32 && bh <= 32);

  // The initialization is just for silencing Jenkins static analysis warnings
  for (r = 0; r < bh + 1; ++r)
    memset(buffer[r], 0, (bw + 1) * sizeof(buffer[0][0]));

  for (r = 0; r < bh; ++r) buffer[r + 1][0] = left[r];
  memcpy(buffer[0], &above[-1], (bw + 1) * sizeof(uint8_t));

  for (r = 1; r < bh + 1; r += 2)
    for (c = 1; c < bw + 1; c += 4) {
      const uint8_t p0 = buffer[r - 1][c - 1];
      const uint8_t p1 = buffer[r - 1][c];
      const uint8_t p2 = buffer[r - 1][c + 1];
      const uint8_t p3 = buffer[r - 1][c + 2];
      const uint8_t p4 = buffer[r - 1][c + 3];
      const uint8_t p5 = buffer[r][c - 1];
      const uint8_t p6 = buffer[r + 1][c - 1];
      for (int k = 0; k < 8; ++k) {
        int r_offset = k >> 2;
        int c_offset = k & 0x03;
        buffer[r + r_offset][c + c_offset] =
            clip_pixel(ROUND_POWER_OF_TWO_SIGNED(
                av1_filter_intra_taps[mode][k][0] * p0 +
                    av1_filter_intra_taps[mode][k][1] * p1 +
                    av1_filter_intra_taps[mode][k][2] * p2 +
                    av1_filter_intra_taps[mode][k][3] * p3 +
                    av1_filter_intra_taps[mode][k][4] * p4 +
                    av1_filter_intra_taps[mode][k][5] * p5 +
                    av1_filter_intra_taps[mode][k][6] * p6,
                FILTER_INTRA_SCALE_BITS));
      }
    }

  for (r = 0; r < bh; ++r) {
    memcpy(dst, &buffer[r + 1][1], bw * sizeof(uint8_t));
    dst += stride;
  }
}

static void highbd_filter_intra_predictor(uint16_t *dst, ptrdiff_t stride,
                                          TX_SIZE tx_size,
                                          const uint16_t *above,
                                          const uint16_t *left, int mode,
                                          int bd) {
  int r, c;
  uint16_t buffer[33][33];
  const int bw = tx_size_wide[tx_size];
  const int bh = tx_size_high[tx_size];

  assert(bw <= 32 && bh <= 32);

  // The initialization is just for silencing Jenkins static analysis warnings
  for (r = 0; r < bh + 1; ++r)
    memset(buffer[r], 0, (bw + 1) * sizeof(buffer[0][0]));

  for (r = 0; r < bh; ++r) buffer[r + 1][0] = left[r];
  memcpy(buffer[0], &above[-1], (bw + 1) * sizeof(buffer[0][0]));

  for (r = 1; r < bh + 1; r += 2)
    for (c = 1; c < bw + 1; c += 4) {
      const uint16_t p0 = buffer[r - 1][c - 1];
      const uint16_t p1 = buffer[r - 1][c];
      const uint16_t p2 = buffer[r - 1][c + 1];
      const uint16_t p3 = buffer[r - 1][c + 2];
      const uint16_t p4 = buffer[r - 1][c + 3];
      const uint16_t p5 = buffer[r][c - 1];
      const uint16_t p6 = buffer[r + 1][c - 1];
      for (int k = 0; k < 8; ++k) {
        int r_offset = k >> 2;
        int c_offset = k & 0x03;
        buffer[r + r_offset][c + c_offset] =
            clip_pixel_highbd(ROUND_POWER_OF_TWO_SIGNED(
                                  av1_filter_intra_taps[mode][k][0] * p0 +
                                      av1_filter_intra_taps[mode][k][1] * p1 +
                                      av1_filter_intra_taps[mode][k][2] * p2 +
                                      av1_filter_intra_taps[mode][k][3] * p3 +
                                      av1_filter_intra_taps[mode][k][4] * p4 +
                                      av1_filter_intra_taps[mode][k][5] * p5 +
                                      av1_filter_intra_taps[mode][k][6] * p6,
                                  FILTER_INTRA_SCALE_BITS),
                              bd);
      }
    }

  for (r = 0; r < bh; ++r) {
    memcpy(dst, &buffer[r + 1][1], bw * sizeof(dst[0]));
    dst += stride;
  }
}

static int is_smooth(const MB_MODE_INFO *mbmi, int plane) {
  if (plane == 0) {
    const PREDICTION_MODE mode = mbmi->mode;
    return (mode == SMOOTH_PRED || mode == SMOOTH_V_PRED ||
            mode == SMOOTH_H_PRED);
  } else {
    // uv_mode is not set for inter blocks, so need to explicitly
    // detect that case.
    if (is_inter_block(mbmi)) return 0;

    const UV_PREDICTION_MODE uv_mode = mbmi->uv_mode;
    return (uv_mode == UV_SMOOTH_PRED || uv_mode == UV_SMOOTH_V_PRED ||
            uv_mode == UV_SMOOTH_H_PRED);
  }
}

static int get_filt_type(const MACROBLOCKD *xd, int plane) {
  int ab_sm, le_sm;

  if (plane == 0) {
    const MB_MODE_INFO *ab = xd->above_mbmi;
    const MB_MODE_INFO *le = xd->left_mbmi;
    ab_sm = ab ? is_smooth(ab, plane) : 0;
    le_sm = le ? is_smooth(le, plane) : 0;
  } else {
    const MB_MODE_INFO *ab = xd->chroma_above_mbmi;
    const MB_MODE_INFO *le = xd->chroma_left_mbmi;
    ab_sm = ab ? is_smooth(ab, plane) : 0;
    le_sm = le ? is_smooth(le, plane) : 0;
  }

  return (ab_sm || le_sm) ? 1 : 0;
}

static int intra_edge_filter_strength(int bs0, int bs1, int delta, int type) {
  const int d = abs(delta);
  int strength = 0;

  const int blk_wh = bs0 + bs1;
  if (type == 0) {
    if (blk_wh <= 8) {
      if (d >= 56) strength = 1;
    } else if (blk_wh <= 12) {
      if (d >= 40) strength = 1;
    } else if (blk_wh <= 16) {
      if (d >= 40) strength = 1;
    } else if (blk_wh <= 24) {
      if (d >= 8) strength = 1;
      if (d >= 16) strength = 2;
      if (d >= 32) strength = 3;
    } else if (blk_wh <= 32) {
      if (d >= 1) strength = 1;
      if (d >= 4) strength = 2;
      if (d >= 32) strength = 3;
    } else {
      if (d >= 1) strength = 3;
    }
  } else {
    if (blk_wh <= 8) {
      if (d >= 40) strength = 1;
      if (d >= 64) strength = 2;
    } else if (blk_wh <= 16) {
      if (d >= 20) strength = 1;
      if (d >= 48) strength = 2;
    } else if (blk_wh <= 24) {
      if (d >= 4) strength = 3;
    } else {
      if (d >= 1) strength = 3;
    }
  }
  return strength;
}

void av1_filter_intra_edge_c(uint8_t *p, int sz, int strength) {
  if (!strength) return;

  const int kernel[INTRA_EDGE_FILT][INTRA_EDGE_TAPS] = {
    { 0, 4, 8, 4, 0 }, { 0, 5, 6, 5, 0 }, { 2, 4, 4, 4, 2 }
  };
  const int filt = strength - 1;
  uint8_t edge[129];

  memcpy(edge, p, sz * sizeof(*p));
  for (int i = 1; i < sz; i++) {
    int s = 0;
    for (int j = 0; j < INTRA_EDGE_TAPS; j++) {
      int k = i - 2 + j;
      k = (k < 0) ? 0 : k;
      k = (k > sz - 1) ? sz - 1 : k;
      s += edge[k] * kernel[filt][j];
    }
    s = (s + 8) >> 4;
    p[i] = s;
  }
}

static void filter_intra_edge_corner(uint8_t *p_above, uint8_t *p_left) {
  const int kernel[3] = { 5, 6, 5 };

  int s = (p_left[0] * kernel[0]) + (p_above[-1] * kernel[1]) +
          (p_above[0] * kernel[2]);
  s = (s + 8) >> 4;
  p_above[-1] = s;
  p_left[-1] = s;
}

void av1_filter_intra_edge_high_c(uint16_t *p, int sz, int strength) {
  if (!strength) return;

  const int kernel[INTRA_EDGE_FILT][INTRA_EDGE_TAPS] = {
    { 0, 4, 8, 4, 0 }, { 0, 5, 6, 5, 0 }, { 2, 4, 4, 4, 2 }
  };
  const int filt = strength - 1;
  uint16_t edge[129];

  memcpy(edge, p, sz * sizeof(*p));
  for (int i = 1; i < sz; i++) {
    int s = 0;
    for (int j = 0; j < INTRA_EDGE_TAPS; j++) {
      int k = i - 2 + j;
      k = (k < 0) ? 0 : k;
      k = (k > sz - 1) ? sz - 1 : k;
      s += edge[k] * kernel[filt][j];
    }
    s = (s + 8) >> 4;
    p[i] = s;
  }
}

static void filter_intra_edge_corner_high(uint16_t *p_above, uint16_t *p_left) {
  const int kernel[3] = { 5, 6, 5 };

  int s = (p_left[0] * kernel[0]) + (p_above[-1] * kernel[1]) +
          (p_above[0] * kernel[2]);
  s = (s + 8) >> 4;
  p_above[-1] = s;
  p_left[-1] = s;
}

void av1_upsample_intra_edge_c(uint8_t *p, int sz) {
  // interpolate half-sample positions
  assert(sz <= MAX_UPSAMPLE_SZ);

  uint8_t in[MAX_UPSAMPLE_SZ + 3];
  // copy p[-1..(sz-1)] and extend first and last samples
  in[0] = p[-1];
  in[1] = p[-1];
  for (int i = 0; i < sz; i++) {
    in[i + 2] = p[i];
  }
  in[sz + 2] = p[sz - 1];

  // interpolate half-sample edge positions
  p[-2] = in[0];
  for (int i = 0; i < sz; i++) {
    int s = -in[i] + (9 * in[i + 1]) + (9 * in[i + 2]) - in[i + 3];
    s = clip_pixel((s + 8) >> 4);
    p[2 * i - 1] = s;
    p[2 * i] = in[i + 2];
  }
}

void av1_upsample_intra_edge_high_c(uint16_t *p, int sz, int bd) {
  // interpolate half-sample positions
  assert(sz <= MAX_UPSAMPLE_SZ);

  uint16_t in[MAX_UPSAMPLE_SZ + 3];
  // copy p[-1..(sz-1)] and extend first and last samples
  in[0] = p[-1];
  in[1] = p[-1];
  for (int i = 0; i < sz; i++) {
    in[i + 2] = p[i];
  }
  in[sz + 2] = p[sz - 1];

  // interpolate half-sample edge positions
  p[-2] = in[0];
  for (int i = 0; i < sz; i++) {
    int s = -in[i] + (9 * in[i + 1]) + (9 * in[i + 2]) - in[i + 3];
    s = (s + 8) >> 4;
    s = clip_pixel_highbd(s, bd);
    p[2 * i - 1] = s;
    p[2 * i] = in[i + 2];
  }
}

static void build_intra_predictors_high(
    const MACROBLOCKD *xd, const uint8_t *ref8, int ref_stride, uint8_t *dst8,
    int dst_stride, PREDICTION_MODE mode, int angle_delta,
    FILTER_INTRA_MODE filter_intra_mode, TX_SIZE tx_size,
    int disable_edge_filter, int n_top_px, int n_topright_px, int n_left_px,
    int n_bottomleft_px, int plane) {
  int i;
  uint16_t *dst = CONVERT_TO_SHORTPTR(dst8);
  uint16_t *ref = CONVERT_TO_SHORTPTR(ref8);
  DECLARE_ALIGNED(16, uint16_t, left_data[MAX_TX_SIZE * 2 + 32]);
  DECLARE_ALIGNED(16, uint16_t, above_data[MAX_TX_SIZE * 2 + 32]);
  uint16_t *const above_row = above_data + 16;
  uint16_t *const left_col = left_data + 16;
  const int txwpx = tx_size_wide[tx_size];
  const int txhpx = tx_size_high[tx_size];
  int need_left = extend_modes[mode] & NEED_LEFT;
  int need_above = extend_modes[mode] & NEED_ABOVE;
  int need_above_left = extend_modes[mode] & NEED_ABOVELEFT;
  const uint16_t *above_ref = ref - ref_stride;
  const uint16_t *left_ref = ref - 1;
  int p_angle = 0;
  const int is_dr_mode = av1_is_directional_mode(mode);
  const int use_filter_intra = filter_intra_mode != FILTER_INTRA_MODES;
  int base = 128 << (xd->bd - 8);

  // The default values if ref pixels are not available:
  // base-1 base-1 base-1 .. base-1 base-1 base-1 base-1 base-1 base-1
  // base+1   A      B  ..     Y      Z
  // base+1   C      D  ..     W      X
  // base+1   E      F  ..     U      V
  // base+1   G      H  ..     S      T      T      T      T      T

  if (is_dr_mode) {
    p_angle = mode_to_angle_map[mode] + angle_delta;
    if (p_angle <= 90)
      need_above = 1, need_left = 0, need_above_left = 1;
    else if (p_angle < 180)
      need_above = 1, need_left = 1, need_above_left = 1;
    else
      need_above = 0, need_left = 1, need_above_left = 1;
  }
  if (use_filter_intra) need_left = need_above = need_above_left = 1;

  assert(n_top_px >= 0);
  assert(n_topright_px >= 0);
  assert(n_left_px >= 0);
  assert(n_bottomleft_px >= 0);

  if ((!need_above && n_left_px == 0) || (!need_left && n_top_px == 0)) {
    int val;
    if (need_left) {
      val = (n_top_px > 0) ? above_ref[0] : base + 1;
    } else {
      val = (n_left_px > 0) ? left_ref[0] : base - 1;
    }
    for (i = 0; i < txhpx; ++i) {
      aom_memset16(dst, val, txwpx);
      dst += dst_stride;
    }
    return;
  }

  // NEED_LEFT
  if (need_left) {
    int need_bottom = !!(extend_modes[mode] & NEED_BOTTOMLEFT);
    if (use_filter_intra) need_bottom = 0;
    if (is_dr_mode) need_bottom = p_angle > 180;
    const int num_left_pixels_needed = txhpx + (need_bottom ? txwpx : 0);
    i = 0;
    if (n_left_px > 0) {
      for (; i < n_left_px; i++) left_col[i] = left_ref[i * ref_stride];
      if (need_bottom && n_bottomleft_px > 0) {
        assert(i == txhpx);
        for (; i < txhpx + n_bottomleft_px; i++)
          left_col[i] = left_ref[i * ref_stride];
      }
      if (i < num_left_pixels_needed)
        aom_memset16(&left_col[i], left_col[i - 1], num_left_pixels_needed - i);
    } else {
      if (n_top_px > 0) {
        aom_memset16(left_col, above_ref[0], num_left_pixels_needed);
      } else {
        aom_memset16(left_col, base + 1, num_left_pixels_needed);
      }
    }
  }

  // NEED_ABOVE
  if (need_above) {
    int need_right = !!(extend_modes[mode] & NEED_ABOVERIGHT);
    if (use_filter_intra) need_right = 0;
    if (is_dr_mode) need_right = p_angle < 90;
    const int num_top_pixels_needed = txwpx + (need_right ? txhpx : 0);
    if (n_top_px > 0) {
      memcpy(above_row, above_ref, n_top_px * sizeof(above_ref[0]));
      i = n_top_px;
      if (need_right && n_topright_px > 0) {
        assert(n_top_px == txwpx);
        memcpy(above_row + txwpx, above_ref + txwpx,
               n_topright_px * sizeof(above_ref[0]));
        i += n_topright_px;
      }
      if (i < num_top_pixels_needed)
        aom_memset16(&above_row[i], above_row[i - 1],
                     num_top_pixels_needed - i);
    } else {
      if (n_left_px > 0) {
        aom_memset16(above_row, left_ref[0], num_top_pixels_needed);
      } else {
        aom_memset16(above_row, base - 1, num_top_pixels_needed);
      }
    }
  }

  if (need_above_left) {
    if (n_top_px > 0 && n_left_px > 0) {
      above_row[-1] = above_ref[-1];
    } else if (n_top_px > 0) {
      above_row[-1] = above_ref[0];
    } else if (n_left_px > 0) {
      above_row[-1] = left_ref[0];
    } else {
      above_row[-1] = base;
    }
    left_col[-1] = above_row[-1];
  }

  if (use_filter_intra) {
    highbd_filter_intra_predictor(dst, dst_stride, tx_size, above_row, left_col,
                                  filter_intra_mode, xd->bd);
    return;
  }

  if (is_dr_mode) {
    int upsample_above = 0;
    int upsample_left = 0;
    if (!disable_edge_filter) {
      const int need_right = p_angle < 90;
      const int need_bottom = p_angle > 180;
      const int filt_type = get_filt_type(xd, plane);
      if (p_angle != 90 && p_angle != 180) {
        const int ab_le = need_above_left ? 1 : 0;
        if (need_above && need_left && (txwpx + txhpx >= 24)) {
          filter_intra_edge_corner_high(above_row, left_col);
        }
        if (need_above && n_top_px > 0) {
          const int strength =
              intra_edge_filter_strength(txwpx, txhpx, p_angle - 90, filt_type);
          const int n_px = n_top_px + ab_le + (need_right ? txhpx : 0);
          av1_filter_intra_edge_high(above_row - ab_le, n_px, strength);
        }
        if (need_left && n_left_px > 0) {
          const int strength = intra_edge_filter_strength(
              txhpx, txwpx, p_angle - 180, filt_type);
          const int n_px = n_left_px + ab_le + (need_bottom ? txwpx : 0);
          av1_filter_intra_edge_high(left_col - ab_le, n_px, strength);
        }
      }
      upsample_above =
          av1_use_intra_edge_upsample(txwpx, txhpx, p_angle - 90, filt_type);
      if (need_above && upsample_above) {
        const int n_px = txwpx + (need_right ? txhpx : 0);
        av1_upsample_intra_edge_high(above_row, n_px, xd->bd);
      }
      upsample_left =
          av1_use_intra_edge_upsample(txhpx, txwpx, p_angle - 180, filt_type);
      if (need_left && upsample_left) {
        const int n_px = txhpx + (need_bottom ? txwpx : 0);
        av1_upsample_intra_edge_high(left_col, n_px, xd->bd);
      }
    }
    highbd_dr_predictor(dst, dst_stride, tx_size, above_row, left_col,
                        upsample_above, upsample_left, p_angle, xd->bd);
    return;
  }

  // predict
  if (mode == DC_PRED) {
    dc_pred_high[n_left_px > 0][n_top_px > 0][tx_size](
        dst, dst_stride, above_row, left_col, xd->bd);
  } else {
    pred_high[mode][tx_size](dst, dst_stride, above_row, left_col, xd->bd);
  }
}

static void build_intra_predictors(const MACROBLOCKD *xd, const uint8_t *ref,
                                   int ref_stride, uint8_t *dst, int dst_stride,
                                   PREDICTION_MODE mode, int angle_delta,
                                   FILTER_INTRA_MODE filter_intra_mode,
                                   TX_SIZE tx_size, int disable_edge_filter,
                                   int n_top_px, int n_topright_px,
                                   int n_left_px, int n_bottomleft_px,
                                   int plane) {
  int i;
  const uint8_t *above_ref = ref - ref_stride;
  const uint8_t *left_ref = ref - 1;
  DECLARE_ALIGNED(16, uint8_t, left_data[MAX_TX_SIZE * 2 + 32]);
  DECLARE_ALIGNED(16, uint8_t, above_data[MAX_TX_SIZE * 2 + 32]);
  uint8_t *const above_row = above_data + 16;
  uint8_t *const left_col = left_data + 16;
  const int txwpx = tx_size_wide[tx_size];
  const int txhpx = tx_size_high[tx_size];
  int need_left = extend_modes[mode] & NEED_LEFT;
  int need_above = extend_modes[mode] & NEED_ABOVE;
  int need_above_left = extend_modes[mode] & NEED_ABOVELEFT;
  int p_angle = 0;
  const int is_dr_mode = av1_is_directional_mode(mode);
  const int use_filter_intra = filter_intra_mode != FILTER_INTRA_MODES;

  // The default values if ref pixels are not available:
  // 127 127 127 .. 127 127 127 127 127 127
  // 129  A   B  ..  Y   Z
  // 129  C   D  ..  W   X
  // 129  E   F  ..  U   V
  // 129  G   H  ..  S   T   T   T   T   T
  // ..

  if (is_dr_mode) {
    p_angle = mode_to_angle_map[mode] + angle_delta;
    if (p_angle <= 90)
      need_above = 1, need_left = 0, need_above_left = 1;
    else if (p_angle < 180)
      need_above = 1, need_left = 1, need_above_left = 1;
    else
      need_above = 0, need_left = 1, need_above_left = 1;
  }
  if (use_filter_intra) need_left = need_above = need_above_left = 1;

  assert(n_top_px >= 0);
  assert(n_topright_px >= 0);
  assert(n_left_px >= 0);
  assert(n_bottomleft_px >= 0);

  if ((!need_above && n_left_px == 0) || (!need_left && n_top_px == 0)) {
    int val;
    if (need_left) {
      val = (n_top_px > 0) ? above_ref[0] : 129;
    } else {
      val = (n_left_px > 0) ? left_ref[0] : 127;
    }
    for (i = 0; i < txhpx; ++i) {
      memset(dst, val, txwpx);
      dst += dst_stride;
    }
    return;
  }

  // NEED_LEFT
  if (need_left) {
    int need_bottom = !!(extend_modes[mode] & NEED_BOTTOMLEFT);
    if (use_filter_intra) need_bottom = 0;
    if (is_dr_mode) need_bottom = p_angle > 180;
    const int num_left_pixels_needed = txhpx + (need_bottom ? txwpx : 0);
    i = 0;
    if (n_left_px > 0) {
      for (; i < n_left_px; i++) left_col[i] = left_ref[i * ref_stride];
      if (need_bottom && n_bottomleft_px > 0) {
        assert(i == txhpx);
        for (; i < txhpx + n_bottomleft_px; i++)
          left_col[i] = left_ref[i * ref_stride];
      }
      if (i < num_left_pixels_needed)
        memset(&left_col[i], left_col[i - 1], num_left_pixels_needed - i);
    } else {
      if (n_top_px > 0) {
        memset(left_col, above_ref[0], num_left_pixels_needed);
      } else {
        memset(left_col, 129, num_left_pixels_needed);
      }
    }
  }

  // NEED_ABOVE
  if (need_above) {
    int need_right = !!(extend_modes[mode] & NEED_ABOVERIGHT);
    if (use_filter_intra) need_right = 0;
    if (is_dr_mode) need_right = p_angle < 90;
    const int num_top_pixels_needed = txwpx + (need_right ? txhpx : 0);
    if (n_top_px > 0) {
      memcpy(above_row, above_ref, n_top_px);
      i = n_top_px;
      if (need_right && n_topright_px > 0) {
        assert(n_top_px == txwpx);
        memcpy(above_row + txwpx, above_ref + txwpx, n_topright_px);
        i += n_topright_px;
      }
      if (i < num_top_pixels_needed)
        memset(&above_row[i], above_row[i - 1], num_top_pixels_needed - i);
    } else {
      if (n_left_px > 0) {
        memset(above_row, left_ref[0], num_top_pixels_needed);
      } else {
        memset(above_row, 127, num_top_pixels_needed);
      }
    }
  }

  if (need_above_left) {
    if (n_top_px > 0 && n_left_px > 0) {
      above_row[-1] = above_ref[-1];
    } else if (n_top_px > 0) {
      above_row[-1] = above_ref[0];
    } else if (n_left_px > 0) {
      above_row[-1] = left_ref[0];
    } else {
      above_row[-1] = 128;
    }
    left_col[-1] = above_row[-1];
  }

  if (use_filter_intra) {
    av1_filter_intra_predictor(dst, dst_stride, tx_size, above_row, left_col,
                               filter_intra_mode);
    return;
  }

  if (is_dr_mode) {
    int upsample_above = 0;
    int upsample_left = 0;
    if (!disable_edge_filter) {
      const int need_right = p_angle < 90;
      const int need_bottom = p_angle > 180;
      const int filt_type = get_filt_type(xd, plane);
      if (p_angle != 90 && p_angle != 180) {
        const int ab_le = need_above_left ? 1 : 0;
        if (need_above && need_left && (txwpx + txhpx >= 24)) {
          filter_intra_edge_corner(above_row, left_col);
        }
        if (need_above && n_top_px > 0) {
          const int strength =
              intra_edge_filter_strength(txwpx, txhpx, p_angle - 90, filt_type);
          const int n_px = n_top_px + ab_le + (need_right ? txhpx : 0);
          av1_filter_intra_edge(above_row - ab_le, n_px, strength);
        }
        if (need_left && n_left_px > 0) {
          const int strength = intra_edge_filter_strength(
              txhpx, txwpx, p_angle - 180, filt_type);
          const int n_px = n_left_px + ab_le + (need_bottom ? txwpx : 0);
          av1_filter_intra_edge(left_col - ab_le, n_px, strength);
        }
      }
      upsample_above =
          av1_use_intra_edge_upsample(txwpx, txhpx, p_angle - 90, filt_type);
      if (need_above && upsample_above) {
        const int n_px = txwpx + (need_right ? txhpx : 0);
        av1_upsample_intra_edge(above_row, n_px);
      }
      upsample_left =
          av1_use_intra_edge_upsample(txhpx, txwpx, p_angle - 180, filt_type);
      if (need_left && upsample_left) {
        const int n_px = txhpx + (need_bottom ? txwpx : 0);
        av1_upsample_intra_edge(left_col, n_px);
      }
    }
    dr_predictor(dst, dst_stride, tx_size, above_row, left_col, upsample_above,
                 upsample_left, p_angle);
    return;
  }

  // predict
  if (mode == DC_PRED) {
    dc_pred[n_left_px > 0][n_top_px > 0][tx_size](dst, dst_stride, above_row,
                                                  left_col);
  } else {
    pred[mode][tx_size](dst, dst_stride, above_row, left_col);
  }
}

void av1_predict_intra_block(
    const AV1_COMMON *cm, const MACROBLOCKD *xd, int wpx, int hpx,
    TX_SIZE tx_size, PREDICTION_MODE mode, int angle_delta, int use_palette,
    FILTER_INTRA_MODE filter_intra_mode, const uint8_t *ref, int ref_stride,
    uint8_t *dst, int dst_stride, int col_off, int row_off, int plane) {
  const MB_MODE_INFO *const mbmi = xd->mi[0];
  const int txwpx = tx_size_wide[tx_size];
  const int txhpx = tx_size_high[tx_size];
  const int x = col_off << tx_size_wide_log2[0];
  const int y = row_off << tx_size_high_log2[0];

  if (use_palette) {
    int r, c;
    const uint8_t *const map = xd->plane[plane != 0].color_index_map +
                               xd->color_index_map_offset[plane != 0];
    const uint16_t *const palette =
        mbmi->palette_mode_info.palette_colors + plane * PALETTE_MAX_SIZE;
    if (xd->cur_buf->flags & YV12_FLAG_HIGHBITDEPTH) {
      uint16_t *dst16 = CONVERT_TO_SHORTPTR(dst);
      for (r = 0; r < txhpx; ++r) {
        for (c = 0; c < txwpx; ++c) {
          dst16[r * dst_stride + c] = palette[map[(r + y) * wpx + c + x]];
        }
      }
    } else {
      for (r = 0; r < txhpx; ++r) {
        for (c = 0; c < txwpx; ++c) {
          dst[r * dst_stride + c] =
              (uint8_t)palette[map[(r + y) * wpx + c + x]];
        }
      }
    }
    return;
  }

  BLOCK_SIZE bsize = mbmi->sb_type;
  const struct macroblockd_plane *const pd = &xd->plane[plane];
  const int txw = tx_size_wide_unit[tx_size];
  const int txh = tx_size_high_unit[tx_size];
  const int have_top = row_off || (pd->subsampling_y ? xd->chroma_up_available
                                                     : xd->up_available);
  const int have_left =
      col_off ||
      (pd->subsampling_x ? xd->chroma_left_available : xd->left_available);
  const int mi_row = -xd->mb_to_top_edge >> (3 + MI_SIZE_LOG2);
  const int mi_col = -xd->mb_to_left_edge >> (3 + MI_SIZE_LOG2);
  const int xr_chr_offset = 0;
  const int yd_chr_offset = 0;

  // Distance between the right edge of this prediction block to
  // the frame right edge
  const int xr = (xd->mb_to_right_edge >> (3 + pd->subsampling_x)) +
                 (wpx - x - txwpx) - xr_chr_offset;
  // Distance between the bottom edge of this prediction block to
  // the frame bottom edge
  const int yd = (xd->mb_to_bottom_edge >> (3 + pd->subsampling_y)) +
                 (hpx - y - txhpx) - yd_chr_offset;
  const int right_available =
      mi_col + ((col_off + txw) << pd->subsampling_x) < xd->tile.mi_col_end;
  const int bottom_available =
      (yd > 0) &&
      (mi_row + ((row_off + txh) << pd->subsampling_y) < xd->tile.mi_row_end);

  const PARTITION_TYPE partition = mbmi->partition;

  // force 4x4 chroma component block size.
  bsize = scale_chroma_bsize(bsize, pd->subsampling_x, pd->subsampling_y);

  const int have_top_right = has_top_right(
      cm, bsize, mi_row, mi_col, have_top, right_available, partition, tx_size,
      row_off, col_off, pd->subsampling_x, pd->subsampling_y);
  const int have_bottom_left = has_bottom_left(
      cm, bsize, mi_row, mi_col, bottom_available, have_left, partition,
      tx_size, row_off, col_off, pd->subsampling_x, pd->subsampling_y);

  const int disable_edge_filter = !cm->seq_params.enable_intra_edge_filter;
  if (xd->cur_buf->flags & YV12_FLAG_HIGHBITDEPTH) {
    build_intra_predictors_high(
        xd, ref, ref_stride, dst, dst_stride, mode, angle_delta,
        filter_intra_mode, tx_size, disable_edge_filter,
        have_top ? AOMMIN(txwpx, xr + txwpx) : 0,
        have_top_right ? AOMMIN(txwpx, xr) : 0,
        have_left ? AOMMIN(txhpx, yd + txhpx) : 0,
        have_bottom_left ? AOMMIN(txhpx, yd) : 0, plane);
    return;
  }

  build_intra_predictors(xd, ref, ref_stride, dst, dst_stride, mode,
                         angle_delta, filter_intra_mode, tx_size,
                         disable_edge_filter,
                         have_top ? AOMMIN(txwpx, xr + txwpx) : 0,
                         have_top_right ? AOMMIN(txwpx, xr) : 0,
                         have_left ? AOMMIN(txhpx, yd + txhpx) : 0,
                         have_bottom_left ? AOMMIN(txhpx, yd) : 0, plane);
}

void av1_predict_intra_block_facade(const AV1_COMMON *cm, MACROBLOCKD *xd,
                                    int plane, int blk_col, int blk_row,
                                    TX_SIZE tx_size) {
  const MB_MODE_INFO *const mbmi = xd->mi[0];
  struct macroblockd_plane *const pd = &xd->plane[plane];
  const int dst_stride = pd->dst.stride;
  uint8_t *dst =
      &pd->dst.buf[(blk_row * dst_stride + blk_col) << tx_size_wide_log2[0]];
  const PREDICTION_MODE mode =
      (plane == AOM_PLANE_Y) ? mbmi->mode : get_uv_mode(mbmi->uv_mode);
  const int use_palette = mbmi->palette_mode_info.palette_size[plane != 0] > 0;
  const FILTER_INTRA_MODE filter_intra_mode =
      (plane == AOM_PLANE_Y && mbmi->filter_intra_mode_info.use_filter_intra)
          ? mbmi->filter_intra_mode_info.filter_intra_mode
          : FILTER_INTRA_MODES;
  const int angle_delta = mbmi->angle_delta[plane != AOM_PLANE_Y] * ANGLE_STEP;

  if (plane != AOM_PLANE_Y && mbmi->uv_mode == UV_CFL_PRED) {
#if CONFIG_DEBUG
    assert(is_cfl_allowed(xd));
    const BLOCK_SIZE plane_bsize = get_plane_block_size(
        mbmi->sb_type, pd->subsampling_x, pd->subsampling_y);
    (void)plane_bsize;
    assert(plane_bsize < BLOCK_SIZES_ALL);
    if (!xd->lossless[mbmi->segment_id]) {
      assert(blk_col == 0);
      assert(blk_row == 0);
      assert(block_size_wide[plane_bsize] == tx_size_wide[tx_size]);
      assert(block_size_high[plane_bsize] == tx_size_high[tx_size]);
    }
#endif
    CFL_CTX *const cfl = &xd->cfl;
    CFL_PRED_TYPE pred_plane = get_cfl_pred_type(plane);
    if (cfl->dc_pred_is_cached[pred_plane] == 0) {
      av1_predict_intra_block(cm, xd, pd->width, pd->height, tx_size, mode,
                              angle_delta, use_palette, filter_intra_mode, dst,
                              dst_stride, dst, dst_stride, blk_col, blk_row,
                              plane);
      if (cfl->use_dc_pred_cache) {
        cfl_store_dc_pred(xd, dst, pred_plane, tx_size_wide[tx_size]);
        cfl->dc_pred_is_cached[pred_plane] = 1;
      }
    } else {
      cfl_load_dc_pred(xd, dst, dst_stride, tx_size, pred_plane);
    }
    cfl_predict_block(xd, dst, dst_stride, tx_size, plane);
    return;
  }
  av1_predict_intra_block(cm, xd, pd->width, pd->height, tx_size, mode,
                          angle_delta, use_palette, filter_intra_mode, dst,
                          dst_stride, dst, dst_stride, blk_col, blk_row, plane);
}

void av1_init_intra_predictors(void) {
  aom_once(init_intra_predictors_internal);
}