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
|
/*
* ***** BEGIN GPL LICENSE BLOCK *****
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* as published by the Free Software Foundation; either version 2
* of the License, or (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
*
* The Original Code is Copyright (C) 2012 Blender Foundation.
* All rights reserved.
*
* The Original Code is: all of this file.
*
* Contributor(s): Peter Larabell.
*
* ***** END GPL LICENSE BLOCK *****
*/
/** \file raskter.c
* \ingroup RASKTER
*/
#include <stdlib.h>
#include "raskter.h"
//#define __PLX__FAKE_AA__
//#define __PLX_KD_TREE__
#ifdef __PLX_KD_TREE__
#include "kdtree.h"
#endif
/*
* Sort all the edges of the input polygon by Y, then by X, of the "first" vertex encountered.
* This will ensure we can scan convert the entire poly in one pass.
*
* Really the poly should be clipped to the frame buffer's dimensions here for speed of drawing
* just the poly. Since the DEM code could end up being coupled with this, we'll keep it separate
* for now.
*/
static void preprocess_all_edges(struct r_fill_context *ctx, struct poly_vert *verts, int num_verts, struct e_status *open_edge) {
int i;
int xbeg;
int ybeg;
int xend;
int yend;
int dx;
int dy;
int temp_pos;
int xdist;
struct e_status *e_new;
struct e_status *next_edge;
struct e_status **next_edge_ref;
struct poly_vert *v;
/* set up pointers */
v = verts;
ctx->all_edges = NULL;
/* initialize some boundaries */
ctx->rb.xmax = v[0].x;
ctx->rb.xmin = v[0].x;
ctx->rb.ymax = v[0].y;
ctx->rb.ymin = v[0].y;
/* loop all verts */
for(i = 0; i < num_verts; i++) {
/* determine beginnings and endings of edges, linking last vertex to first vertex */
xbeg = v[i].x;
ybeg = v[i].y;
/* keep track of our x and y bounds */
if(xbeg >= ctx->rb.xmax) {
ctx->rb.xmax = xbeg;
} else if(xbeg <= ctx->rb.xmin) {
ctx->rb.xmin = xbeg;
}
if(ybeg >= ctx->rb.ymax) {
ctx->rb.ymax= ybeg;
} else if(ybeg <= ctx->rb.ymin) {
ctx->rb.ymin=ybeg;
}
if(i) {
/* we're not at the last vert, so end of the edge is the previous vertex */
xend = v[i - 1].x;
yend = v[i - 1].y;
} else {
/* we're at the first vertex, so the "end" of this edge is the last vertex */
xend = v[num_verts - 1].x;
yend = v[num_verts - 1].y;
}
/* make sure our edges are facing the correct direction */
if(ybeg > yend) {
/* flip the Xs */
temp_pos = xbeg;
xbeg = xend;
xend = temp_pos;
/* flip the Ys */
temp_pos = ybeg;
ybeg = yend;
yend = temp_pos;
}
/* calculate y delta */
dy = yend - ybeg;
/* dont draw horizontal lines directly, they are scanned as part of the edges they connect, so skip em. :) */
if(dy) {
/* create the edge and determine it's slope (for incremental line drawing) */
e_new = open_edge++;
/* calculate x delta */
dx = xend - xbeg;
if(dx > 0) {
e_new->xdir = 1;
xdist = dx;
} else {
e_new->xdir = -1;
xdist = -dx;
}
e_new->x = xbeg;
e_new->ybeg = ybeg;
e_new->num = dy;
e_new->drift_dec = dy;
/* calculate deltas for incremental drawing */
if(dx >= 0) {
e_new->drift = 0;
} else {
e_new->drift = -dy + 1;
}
if(dy >= xdist) {
e_new->drift_inc = xdist;
e_new->xshift = 0;
} else {
e_new->drift_inc = xdist % dy;
e_new->xshift = (xdist / dy) * e_new->xdir;
}
next_edge_ref = &ctx->all_edges;
/* link in all the edges, in sorted order */
for(;;) {
next_edge = *next_edge_ref;
if(!next_edge || (next_edge->ybeg > ybeg) || ((next_edge->ybeg == ybeg) && (next_edge->x >= xbeg))) {
e_new->e_next = next_edge;
*next_edge_ref = e_new;
break;
}
next_edge_ref = &next_edge->e_next;
}
}
}
}
/*
* This function clips drawing to the frame buffer. That clipping will likely be moved into the preprocessor
* for speed, but waiting on final design choices for curve-data before eliminating data the DEM code will need
* if it ends up being coupled with this function.
*/
static int rast_scan_fill(struct r_fill_context *ctx, struct poly_vert *verts, int num_verts, float intensity) {
int x_curr; /* current pixel position in X */
int y_curr; /* current scan line being drawn */
int yp; /* y-pixel's position in frame buffer */
int swixd = 0; /* whether or not edges switched position in X */
float *cpxl; /* pixel pointers... */
float *mpxl;
float *spxl;
struct e_status *e_curr; /* edge pointers... */
struct e_status *e_temp;
struct e_status *edgbuf;
struct e_status **edgec;
/*
* If the number of verts specified to render as a polygon is less than 3,
* return immediately. Obviously we cant render a poly with sides < 3. The
* return for this we set to 1, simply so it can be distinguished from the
* next place we could return, /home/guest/blender-svn/soc-2011-tomato/intern/raskter/raskter.
* which is a failure to allocate memory.
*/
if(num_verts < 3) {
return(1);
}
/*
* Try to allocate an edge buffer in memory. needs to be the size of the edge tracking data
* multiplied by the number of edges, which is always equal to the number of verts in
* a 2D polygon. Here we return 0 to indicate a memory allocation failure, as opposed to a 1 for
* the preceeding error, which was a rasterization request on a 2D poly with less than
* 3 sides.
*/
if((edgbuf = (struct e_status *)(malloc(sizeof(struct e_status) * num_verts))) == NULL) {
return(0);
}
/*
* Do some preprocessing on all edges. This constructs a table structure in memory of all
* the edge properties and can "flip" some edges so sorting works correctly.
*/
preprocess_all_edges(ctx, verts, num_verts, edgbuf);
/* can happen with a zero area mask */
if (ctx->all_edges == NULL) {
free(edgbuf);
return(1);
}
/*
* Set the pointer for tracking the edges currently in processing to NULL to make sure
* we don't get some crazy value after initialization.
*/
ctx->possible_edges = NULL;
/*
* Loop through all scan lines to be drawn. Since we sorted by Y values during
* preprocess_all_edges(), we can already exact values for the lowest and
* highest Y values we could possibly need by induction. The preprocessing sorted
* out edges by Y position, we can cycle the current edge being processed once
* it runs out of Y pixels. When we have no more edges, meaning the current edge
* is NULL after setting the "current" edge to be the previous current edge's
* "next" edge in the Y sorted edge connection chain, we can stop looping Y values,
* since we can't possibly have more scan lines if we ran out of edges. :)
*
* TODO: This clips Y to the frame buffer, which should be done in the preprocessor, but for now is done here.
* Will get changed once DEM code gets in.
*/
for(y_curr = ctx->all_edges->ybeg; (ctx->all_edges || ctx->possible_edges); y_curr++) {
/*
* Link any edges that start on the current scan line into the list of
* edges currently needed to draw at least this, if not several, scan lines.
*/
/*
* Set the current edge to the beginning of the list of edges to be rasterized
* into this scan line.
*
* We could have lots of edge here, so iterate over all the edges needed. The
* preprocess_all_edges() function sorted edges by X within each chunk of Y sorting
* so we safely cycle edges to thier own "next" edges in order.
*
* At each iteration, make sure we still have a non-NULL edge.
*/
for(edgec = &ctx->possible_edges; ctx->all_edges && (ctx->all_edges->ybeg == y_curr);) {
x_curr = ctx->all_edges->x; /* Set current X position. */
for(;;) { /* Start looping edges. Will break when edges run out. */
e_curr = *edgec; /* Set up a current edge pointer. */
if(!e_curr || (e_curr->x >= x_curr)) { /* If we have an no edge, or we need to skip some X-span, */
e_temp = ctx->all_edges->e_next; /* set a temp "next" edge to test. */
*edgec = ctx->all_edges; /* Add this edge to the list to be scanned. */
ctx->all_edges->e_next = e_curr; /* Set up the next edge. */
edgec = &ctx->all_edges->e_next; /* Set our list to the next edge's location in memory. */
ctx->all_edges = e_temp; /* Skip the NULL or bad X edge, set pointer to next edge. */
break; /* Stop looping edges (since we ran out or hit empty X span. */
} else {
edgec = &e_curr->e_next; /* Set the pointer to the edge list the "next" edge. */
}
}
}
/*
* Determine the current scan line's offset in the pixel buffer based on its Y position.
* Basically we just multiply the current scan line's Y value by the number of pixels in each line.
*/
yp = y_curr * ctx->rb.sizex;
/*
* Set a "scan line pointer" in memory. The location of the buffer plus the row offset.
*/
spxl = ctx->rb.buf + (yp);
/*
* Set up the current edge to the first (in X) edge. The edges which could possibly be in this
* list were determined in the preceeding edge loop above. They were already sorted in X by the
* initial processing function.
*
* At each iteration, test for a NULL edge. Since we'll keep cycling edge's to their own "next" edge
* we will eventually hit a NULL when the list runs out.
*/
for(e_curr = ctx->possible_edges; e_curr; e_curr = e_curr->e_next) {
/*
* Calculate a span of pixels to fill on the current scan line.
*
* Set the current pixel pointer by adding the X offset to the scan line's start offset.
* Cycle the current edge the next edge.
* Set the max X value to draw to be one less than the next edge's first pixel. This way we are
* sure not to ever get into a situation where we have overdraw. (drawing the same pixel more than
* one time because it's on a vertex connecting two edges)
*
* Then blast through all the pixels in the span, advancing the pointer and setting the color to white.
*
* TODO: Here we clip to the scan line, this is not efficient, and should be done in the preprocessor,
* but for now it is done here until the DEM code comes in.
*/
/* set up xmin and xmax bounds on this scan line */
cpxl = spxl + MAX2(e_curr->x, 0);
e_curr = e_curr->e_next;
mpxl = spxl + MIN2(e_curr->x, ctx->rb.sizex) - 1;
if((y_curr >= 0) && (y_curr < ctx->rb.sizey)) {
/* draw the pixels. */
for(; cpxl <= mpxl; *cpxl++ += intensity);
}
}
/*
* Loop through all edges of polygon that could be hit by this scan line,
* and figure out their x-intersections with the next scan line.
*
* Either A.) we wont have any more edges to test, or B.) we just add on the
* slope delta computed in preprocessing step. Since this draws non-antialiased
* polygons, we dont have fractional positions, so we only move in x-direction
* when needed to get all the way to the next pixel over...
*/
for(edgec = &ctx->possible_edges; (e_curr = *edgec);) {
if(!(--(e_curr->num))) {
*edgec = e_curr->e_next;
} else {
e_curr->x += e_curr->xshift;
if((e_curr->drift += e_curr->drift_inc) > 0) {
e_curr->x += e_curr->xdir;
e_curr->drift -= e_curr->drift_dec;
}
edgec = &e_curr->e_next;
}
}
/*
* It's possible that some edges may have crossed during the last step, so we'll be sure
* that we ALWAYS intersect scan lines in order by shuffling if needed to make all edges
* sorted by x-intersection coordinate. We'll always scan through at least once to see if
* edges crossed, and if so, we set the 'swixd' flag. If 'swixd' gets set on the initial
* pass, then we know we need to sort by x, so then cycle through edges again and perform
* the sort.-
*/
if(ctx->possible_edges) {
for(edgec = &ctx->possible_edges; (e_curr = *edgec)->e_next; edgec = &(*edgec)->e_next) {
/* if the current edge hits scan line at greater X than the next edge, we need to exchange the edges */
if(e_curr->x > e_curr->e_next->x) {
*edgec = e_curr->e_next;
/* exchange the pointers */
e_temp = e_curr->e_next->e_next;
e_curr->e_next->e_next = e_curr;
e_curr->e_next = e_temp;
/* set flag that we had at least one switch */
swixd = 1;
}
}
/* if we did have a switch, look for more (there will more if there was one) */
for(;;) {
/* reset exchange flag so it's only set if we encounter another one */
swixd = 0;
for(edgec = &ctx->possible_edges; (e_curr = *edgec)->e_next; edgec = &(*edgec)->e_next) {
/* again, if current edge hits scan line at higher X than next edge, exchange the edges and set flag */
if(e_curr->x > e_curr->e_next->x) {
*edgec = e_curr->e_next;
/* exchange the pointers */
e_temp = e_curr->e_next->e_next;
e_curr->e_next->e_next = e_curr;
e_curr->e_next = e_temp;
/* flip the exchanged flag */
swixd = 1;
}
}
/* if we had no exchanges, we're done reshuffling the pointers */
if(!swixd) {
break;
}
}
}
}
free(edgbuf);
return 1;
}
int PLX_raskterize(float(*base_verts)[2], int num_base_verts,
float *buf, int buf_x, int buf_y, int do_mask_AA) {
int subdiv_AA = (do_mask_AA != 0)? 0:0;
int i; /* i: Loop counter. */
int sAx;
int sAy;
struct poly_vert *ply; /* ply: Pointer to a list of integer buffer-space vertex coordinates. */
struct r_fill_context ctx = {0};
const float buf_x_f = (float)(buf_x);
const float buf_y_f = (float)(buf_y);
float div_offset=(1.0f / (float)(subdiv_AA));
float div_offset_static = 0.5f * (float)(subdiv_AA) * div_offset;
/*
* Allocate enough memory for our poly_vert list. It'll be the size of the poly_vert
* data structure multiplied by the number of base_verts.
*
* In the event of a failure to allocate the memory, return 0, so this error can
* be distinguished as a memory allocation error.
*/
if((ply = (struct poly_vert *)(malloc(sizeof(struct poly_vert) * num_base_verts))) == NULL) {
return(0);
}
ctx.rb.buf = buf; /* Set the output buffer pointer. */
ctx.rb.sizex = buf_x; /* Set the output buffer size in X. (width) */
ctx.rb.sizey = buf_y; /* Set the output buffer size in Y. (height) */
/*
* Loop over all verts passed in to be rasterized. Each vertex's X and Y coordinates are
* then converted from normalized screen space (0.0 <= POS <= 1.0) to integer coordinates
* in the buffer-space coordinates passed in inside buf_x and buf_y.
*
* It's worth noting that this function ONLY outputs fully white pixels in a mask. Every pixel
* drawn will be 1.0f in value, there is no anti-aliasing.
*/
if(!subdiv_AA) {
for(i = 0; i < num_base_verts; i++) { /* Loop over all base_verts. */
ply[i].x = (int)((base_verts[i][0] * buf_x_f) + 0.5f); /* Range expand normalized X to integer buffer-space X. */
ply[i].y = (int)((base_verts[i][1] * buf_y_f) + 0.5f); /* Range expand normalized Y to integer buffer-space Y. */
}
i = rast_scan_fill(&ctx, ply, num_base_verts,1.0f); /* Call our rasterizer, passing in the integer coords for each vert. */
} else {
for(sAx=0; sAx < subdiv_AA; sAx++) {
for(sAy=0; sAy < subdiv_AA; sAy++) {
for(i=0; i < num_base_verts; i++) {
ply[i].x = (int)((base_verts[i][0]*buf_x_f)+0.5f - div_offset_static + (div_offset*(float)(sAx)));
ply[i].y = (int)((base_verts[i][1]*buf_y_f)+0.5f - div_offset_static + (div_offset*(float)(sAy)));
}
i = rast_scan_fill(&ctx, ply, num_base_verts,(1.0f / (float)(subdiv_AA*subdiv_AA)));
}
}
}
free(ply); /* Free the memory allocated for the integer coordinate table. */
return(i); /* Return the value returned by the rasterizer. */
}
/*
* This function clips drawing to the frame buffer. That clipping will likely be moved into the preprocessor
* for speed, but waiting on final design choices for curve-data before eliminating data the DEM code will need
* if it ends up being coupled with this function.
*/
static int rast_scan_feather(struct r_fill_context *ctx,
float(*base_verts_f)[2], int num_base_verts,
struct poly_vert *feather_verts, float(*feather_verts_f)[2], int num_feather_verts) {
int x_curr; /* current pixel position in X */
int y_curr; /* current scan line being drawn */
int yp; /* y-pixel's position in frame buffer */
int swixd = 0; /* whether or not edges switched position in X */
float *cpxl; /* pixel pointers... */
float *mpxl;
float *spxl;
struct e_status *e_curr; /* edge pointers... */
struct e_status *e_temp;
struct e_status *edgbuf;
struct e_status **edgec;
/* from dem */
int a; // a = temporary pixel index buffer loop counter
float fsz; // size of the frame
unsigned int rsl; // long used for finding fast 1.0/sqrt
float rsf; // float used for finding fast 1.0/sqrt
const float rsopf = 1.5f; // constant float used for finding fast 1.0/sqrt
//unsigned int gradientFillOffset;
float t;
float ud; // ud = unscaled edge distance
float dmin; // dmin = minimun edge distance
float odist; // odist = current outer edge distance
float idist; // idist = current inner edge distance
float dx; // dx = X-delta (used for distance proportion calculation)
float dy; // dy = Y-delta (used for distance proportion calculation)
float xpxw = (1.0f / (float)(ctx->rb.sizex)); // xpxw = normalized pixel width
float ypxh = (1.0f / (float)(ctx->rb.sizey)); // ypxh = normalized pixel height
#ifdef __PLX_KD_TREE__
void *res_kdi;
void *res_kdo;
float clup[2];
#endif
/*
* If the number of verts specified to render as a polygon is less than 3,
* return immediately. Obviously we cant render a poly with sides < 3. The
* return for this we set to 1, simply so it can be distinguished from the
* next place we could return,
* which is a failure to allocate memory.
*/
if(num_feather_verts < 3) {
return(1);
}
/*
* Try to allocate an edge buffer in memory. needs to be the size of the edge tracking data
* multiplied by the number of edges, which is always equal to the number of verts in
* a 2D polygon. Here we return 0 to indicate a memory allocation failure, as opposed to a 1 for
* the preceeding error, which was a rasterization request on a 2D poly with less than
* 3 sides.
*/
if((edgbuf = (struct e_status *)(malloc(sizeof(struct e_status) * num_feather_verts))) == NULL) {
return(0);
}
/*
* Do some preprocessing on all edges. This constructs a table structure in memory of all
* the edge properties and can "flip" some edges so sorting works correctly.
*/
preprocess_all_edges(ctx, feather_verts, num_feather_verts, edgbuf);
/* can happen with a zero area mask */
if (ctx->all_edges == NULL) {
free(edgbuf);
return(1);
}
/*
* Set the pointer for tracking the edges currently in processing to NULL to make sure
* we don't get some crazy value after initialization.
*/
ctx->possible_edges = NULL;
/*
* Loop through all scan lines to be drawn. Since we sorted by Y values during
* preprocess_all_edges(), we can already exact values for the lowest and
* highest Y values we could possibly need by induction. The preprocessing sorted
* out edges by Y position, we can cycle the current edge being processed once
* it runs out of Y pixels. When we have no more edges, meaning the current edge
* is NULL after setting the "current" edge to be the previous current edge's
* "next" edge in the Y sorted edge connection chain, we can stop looping Y values,
* since we can't possibly have more scan lines if we ran out of edges. :)
*
* TODO: This clips Y to the frame buffer, which should be done in the preprocessor, but for now is done here.
* Will get changed once DEM code gets in.
*/
for(y_curr = ctx->all_edges->ybeg; (ctx->all_edges || ctx->possible_edges); y_curr++) {
/*
* Link any edges that start on the current scan line into the list of
* edges currently needed to draw at least this, if not several, scan lines.
*/
/*
* Set the current edge to the beginning of the list of edges to be rasterized
* into this scan line.
*
* We could have lots of edge here, so iterate over all the edges needed. The
* preprocess_all_edges() function sorted edges by X within each chunk of Y sorting
* so we safely cycle edges to thier own "next" edges in order.
*
* At each iteration, make sure we still have a non-NULL edge.
*/
for(edgec = &ctx->possible_edges; ctx->all_edges && (ctx->all_edges->ybeg == y_curr);) {
x_curr = ctx->all_edges->x; /* Set current X position. */
for(;;) { /* Start looping edges. Will break when edges run out. */
e_curr = *edgec; /* Set up a current edge pointer. */
if(!e_curr || (e_curr->x >= x_curr)) { /* If we have an no edge, or we need to skip some X-span, */
e_temp = ctx->all_edges->e_next; /* set a temp "next" edge to test. */
*edgec = ctx->all_edges; /* Add this edge to the list to be scanned. */
ctx->all_edges->e_next = e_curr; /* Set up the next edge. */
edgec = &ctx->all_edges->e_next; /* Set our list to the next edge's location in memory. */
ctx->all_edges = e_temp; /* Skip the NULL or bad X edge, set pointer to next edge. */
break; /* Stop looping edges (since we ran out or hit empty X span. */
} else {
edgec = &e_curr->e_next; /* Set the pointer to the edge list the "next" edge. */
}
}
}
/*
* Determine the current scan line's offset in the pixel buffer based on its Y position.
* Basically we just multiply the current scan line's Y value by the number of pixels in each line.
*/
yp = y_curr * ctx->rb.sizex;
/*
* Set a "scan line pointer" in memory. The location of the buffer plus the row offset.
*/
spxl = ctx->rb.buf + (yp);
/*
* Set up the current edge to the first (in X) edge. The edges which could possibly be in this
* list were determined in the preceeding edge loop above. They were already sorted in X by the
* initial processing function.
*
* At each iteration, test for a NULL edge. Since we'll keep cycling edge's to their own "next" edge
* we will eventually hit a NULL when the list runs out.
*/
for(e_curr = ctx->possible_edges; e_curr; e_curr = e_curr->e_next) {
/*
* Calculate a span of pixels to fill on the current scan line.
*
* Set the current pixel pointer by adding the X offset to the scan line's start offset.
* Cycle the current edge the next edge.
* Set the max X value to draw to be one less than the next edge's first pixel. This way we are
* sure not to ever get into a situation where we have overdraw. (drawing the same pixel more than
* one time because it's on a vertex connecting two edges)
*
* Then blast through all the pixels in the span, advancing the pointer and setting the color to white.
*
* TODO: Here we clip to the scan line, this is not efficient, and should be done in the preprocessor,
* but for now it is done here until the DEM code comes in.
*/
/* set up xmin and xmax bounds on this scan line */
cpxl = spxl + MAX2(e_curr->x, 0);
e_curr = e_curr->e_next;
mpxl = spxl + MIN2(e_curr->x, ctx->rb.sizex) - 1;
if((y_curr >= 0) && (y_curr < ctx->rb.sizey)) {
t = ((float)((cpxl - spxl) % ctx->rb.sizex) + 0.5f) * xpxw;
fsz = ((float)(y_curr) + 0.5f) * ypxh;
/* draw the pixels. */
for(; cpxl <= mpxl; cpxl++, t += xpxw) {
//do feather check
// first check that pixel isn't already full, and only operate if it is not
if(*cpxl < 0.9999f) {
#ifndef __PLX_KD_TREE__
dmin = 2.0f; // reset min distance to edge pixel
for(a = 0; a < num_feather_verts; a++) { // loop through all outer edge buffer pixels
dx = t - feather_verts_f[a][0]; // set dx to gradient pixel column - outer edge pixel row
dy = fsz - feather_verts_f[a][1]; // set dy to gradient pixel row - outer edge pixel column
ud = dx * dx + dy * dy; // compute sum of squares
if(ud < dmin) { // if our new sum of squares is less than the current minimum
dmin = ud; // set a new minimum equal to the new lower value
}
}
odist = dmin; // cast outer min to a float
rsf = odist * 0.5f; //
rsl = *(unsigned int *)&odist; // use some peculiar properties of the way bits are stored
rsl = 0x5f3759df - (rsl >> 1); // in floats vs. unsigned ints to compute an approximate
odist = *(float *)&rsl; // reciprocal square root
odist = odist * (rsopf - (rsf * odist * odist)); // -- ** this line can be iterated for more accuracy ** --
odist = odist * (rsopf - (rsf * odist * odist));
dmin = 2.0f; // reset min distance to edge pixel
for(a = 0; a < num_base_verts; a++) { // loop through all inside edge pixels
dx = t - base_verts_f[a][0]; // compute delta in Y from gradient pixel to inside edge pixel
dy = fsz - base_verts_f[a][1]; // compute delta in X from gradient pixel to inside edge pixel
ud = dx * dx + dy * dy; // compute sum of squares
if(ud < dmin) { // if our new sum of squares is less than the current minimum we've found
dmin = ud; // set a new minimum equal to the new lower value
}
}
idist = dmin; // cast inner min to a float
rsf = idist * 0.5f; //
rsl = *(unsigned int *)&idist; //
rsl = 0x5f3759df - (rsl >> 1); // see notes above
idist = *(float *)&rsl; //
idist = idist * (rsopf - (rsf * idist * idist)); //
idist = idist * (rsopf - (rsf * idist * idist));
/*
* Note once again that since we are using reciprocals of distance values our
* proportion is already the correct intensity, and does not need to be
* subracted from 1.0 like it would have if we used real distances.
*/
#else
clup[0]=t;
clup[1]=fsz;
res_kdi=kd_nearestf(ctx->kdi,clup);
res_kdo=kd_nearestf(ctx->kdo,clup);
kd_res_itemf(res_kdi,clup);
dx=t-clup[0];
dy=fsz-clup[1];
idist=dx*dx+dy*dy;
rsf = idist * 0.5f; //
rsl = *(unsigned int *)&idist; //
rsl = 0x5f3759df - (rsl >> 1); // see notes above
idist = *(float *)&rsl; //
idist = idist * (rsopf - (rsf * idist * idist)); //
idist = idist * (rsopf - (rsf * idist * idist));
kd_res_itemf(res_kdo,clup);
dx=t-clup[0];
dy=fsz-clup[1];
odist=dx*dx+dy*dy;
rsf = odist * 0.5f; //
rsl = *(unsigned int *)&odist; // use some peculiar properties of the way bits are stored
rsl = 0x5f3759df - (rsl >> 1); // in floats vs. unsigned ints to compute an approximate
odist = *(float *)&rsl; // reciprocal square root
odist = odist * (rsopf - (rsf * odist * odist)); // -- ** this line can be iterated for more accuracy ** --
odist = odist * (rsopf - (rsf * odist * odist));
#endif
/* set intensity, do the += so overlapping gradients are additive */
*cpxl = (idist / (idist+odist));
}
}
}
}
/*
* Loop through all edges of polygon that could be hit by this scan line,
* and figure out their x-intersections with the next scan line.
*
* Either A.) we wont have any more edges to test, or B.) we just add on the
* slope delta computed in preprocessing step. Since this draws non-antialiased
* polygons, we dont have fractional positions, so we only move in x-direction
* when needed to get all the way to the next pixel over...
*/
for(edgec = &ctx->possible_edges; (e_curr = *edgec);) {
if(!(--(e_curr->num))) {
*edgec = e_curr->e_next;
} else {
e_curr->x += e_curr->xshift;
if((e_curr->drift += e_curr->drift_inc) > 0) {
e_curr->x += e_curr->xdir;
e_curr->drift -= e_curr->drift_dec;
}
edgec = &e_curr->e_next;
}
}
/*
* It's possible that some edges may have crossed during the last step, so we'll be sure
* that we ALWAYS intersect scan lines in order by shuffling if needed to make all edges
* sorted by x-intersection coordinate. We'll always scan through at least once to see if
* edges crossed, and if so, we set the 'swixd' flag. If 'swixd' gets set on the initial
* pass, then we know we need to sort by x, so then cycle through edges again and perform
* the sort.-
*/
if(ctx->possible_edges) {
for(edgec = &ctx->possible_edges; (e_curr = *edgec)->e_next; edgec = &(*edgec)->e_next) {
/* if the current edge hits scan line at greater X than the next edge, we need to exchange the edges */
if(e_curr->x > e_curr->e_next->x) {
*edgec = e_curr->e_next;
/* exchange the pointers */
e_temp = e_curr->e_next->e_next;
e_curr->e_next->e_next = e_curr;
e_curr->e_next = e_temp;
/* set flag that we had at least one switch */
swixd = 1;
}
}
/* if we did have a switch, look for more (there will more if there was one) */
for(;;) {
/* reset exchange flag so it's only set if we encounter another one */
swixd = 0;
for(edgec = &ctx->possible_edges; (e_curr = *edgec)->e_next; edgec = &(*edgec)->e_next) {
/* again, if current edge hits scan line at higher X than next edge,
* exchange the edges and set flag */
if(e_curr->x > e_curr->e_next->x) {
*edgec = e_curr->e_next;
/* exchange the pointers */
e_temp = e_curr->e_next->e_next;
e_curr->e_next->e_next = e_curr;
e_curr->e_next = e_temp;
/* flip the exchanged flag */
swixd = 1;
}
}
/* if we had no exchanges, we're done reshuffling the pointers */
if(!swixd) {
break;
}
}
}
}
free(edgbuf);
return 1;
}
int PLX_raskterize_feather(float(*base_verts)[2], int num_base_verts, float(*feather_verts)[2], int num_feather_verts,
float *buf, int buf_x, int buf_y) {
//void plx_floatsort(float(*f)[2], unsigned int n, int sortby);
int i; /* i: Loop counter. */
struct poly_vert *fe; /* fe: Pointer to a list of integer buffer-space feather vertex coords. */
struct r_fill_context ctx = {0};
/* for faster multiply */
const float buf_x_f = (float)buf_x;
const float buf_y_f = (float)buf_y;
#ifdef __PLX_KD_TREE__
ctx.kdi = kd_create(2);
ctx.kdo = kd_create(2);
#endif
/*
* Allocate enough memory for our poly_vert list. It'll be the size of the poly_vert
* data structure multiplied by the number of verts.
*
* In the event of a failure to allocate the memory, return 0, so this error can
* be distinguished as a memory allocation error.
*/
if((fe = (struct poly_vert *)(malloc(sizeof(struct poly_vert) * num_feather_verts))) == NULL) {
return(0);
}
/*
* Loop over all verts passed in to be rasterized. Each vertex's X and Y coordinates are
* then converted from normalized screen space (0.0 <= POS <= 1.0) to integer coordinates
* in the buffer-space coordinates passed in inside buf_x and buf_y.
*
* It's worth noting that this function ONLY outputs fully white pixels in a mask. Every pixel
* drawn will be 1.0f in value, there is no anti-aliasing.
*/
for(i = 0; i < num_feather_verts; i++) { /* Loop over all verts. */
fe[i].x = (int)((feather_verts[i][0] * buf_x_f) + 0.5f); /* Range expand normalized X to integer buffer-space X. */
fe[i].y = (int)((feather_verts[i][1] * buf_y_f) + 0.5f); /* Range expand normalized Y to integer buffer-space Y. */
#ifdef __PLX_KD_TREE__
kd_insertf(ctx.kdo,feather_verts[i],NULL);
}
for(i=0;i<num_base_verts;i++){
kd_insertf(ctx.kdi,base_verts[i],NULL);
#endif
}
ctx.rb.buf = buf; /* Set the output buffer pointer. */
ctx.rb.sizex = buf_x; /* Set the output buffer size in X. (width) */
ctx.rb.sizey = buf_y; /* Set the output buffer size in Y. (height) */
/* pre-sort the sets of edge verts on y */
//plx_floatsort(base_verts,num_base_verts,0);
//plx_floatsort(base_verts,num_base_verts,1);
//plx_floatsort(feather_verts,num_feather_verts,0);
//plx_floatsort(feather_verts,num_feather_verts,1);
/* Call our rasterizer, passing in the integer coords for each vert. */
i = rast_scan_feather(&ctx, base_verts, num_base_verts, fe, feather_verts, num_feather_verts);
free(fe);
return i; /* Return the value returned by the rasterizer. */
}
#ifndef __PLX__FAKE_AA__
int get_range_expanded_pixel_coord(float normalized_value, int max_value) {
return (int)((normalized_value * (float)(max_value)) + 0.5f);
}
float get_pixel_intensity(float *buf, int buf_x, int buf_y, int pos_x, int pos_y) {
if(pos_x < 0 || pos_x >= buf_x || pos_y < 0 || pos_y >= buf_y) {
return 0.0f;
}
return buf[(pos_y * buf_x) + pos_x];
}
float get_pixel_intensity_bilinear(float *buf, int buf_x, int buf_y, float u, float v) {
int a;
int b;
int a_plus_1;
int b_plus_1;
float prop_u;
float prop_v;
float inv_prop_u;
float inv_prop_v;
if(u<0.0f || u>1.0f || v<0.0f || v>1.0f) {
return 0.0f;
}
u = u * (float)(buf_x) - 0.5f;
v = v * (float)(buf_y) - 0.5f;
a = (int)(u);
b = (int)(v);
prop_u = u - (float)(a);
prop_v = v - (float)(b);
inv_prop_u = 1.0f - prop_u;
inv_prop_v = 1.0f - prop_v;
a_plus_1 = MIN2((buf_x-1),a+1);
b_plus_1 = MIN2((buf_y-1),b+1);
return (buf[(b * buf_x) + a] * inv_prop_u + buf[(b*buf_x)+(a_plus_1)] * prop_u)*inv_prop_v+(buf[((b_plus_1) * buf_x)+a] * inv_prop_u + buf[((b_plus_1)*buf_x)+(a_plus_1)] * prop_u) * prop_v;
}
void set_pixel_intensity(float *buf, int buf_x, int buf_y, int pos_x, int pos_y, float intensity) {
if(pos_x < 0 || pos_x >= buf_x || pos_y < 0 || pos_y >= buf_y) {
return;
}
buf[(pos_y * buf_x) + pos_x] = intensity;
}
#endif
int PLX_antialias_buffer(float *buf, int buf_x, int buf_y) {
#ifdef __PLX__FAKE_AA__
#ifdef __PLX_GREY_AA__
int i=0;
int sz = buf_x * buf_y;
for(i=0; i<sz; i++) {
buf[i] *= 0.5f;
}
#endif
(void)buf_x;
(void)buf_y;
(void)buf;
return 1;
#else
const float p0 = 1.0f;
const float p1 = 1.0f;
const float p2 = 1.0f;
const float p3 = 1.0f;
const float p4 = 1.0f;
const float p5 = 1.5f;
const float p6 = 2.0f;
const float p7 = 2.0f;
const float p8 = 2.0f;
const float p9 = 2.0f;
const float p10 = 4.0f;
const float p11 = 8.0f;
const float edge_threshold = 0.063f;
const float edge_threshold_min = 0.0312f;
const float quality_subpix = 1.0f;
float posM_x,posM_y;
float posB_x,posB_y;
float posN_x,posN_y;
float posP_x,posP_y;
float offNP_x,offNP_y;
float lumaM;
float lumaS;
float lumaE;
float lumaN;
float lumaW;
float lumaNW;
float lumaSE;
float lumaNE;
float lumaSW;
float lumaNS;
float lumaWE;
float lumaNESE;
float lumaNWNE;
float lumaNWSW;
float lumaSWSE;
float lumaNN;
float lumaSS;
float lumaEndN;
float lumaEndP;
float lumaMM;
float lumaMLTZero;
float subpixNWSWNESE;
float subpixRcpRange;
float subpixNSWE;
float maxSM;
float minSM;
float maxESM;
float minESM;
float maxWN;
float minWN;
float rangeMax;
float rangeMin;
float rangeMaxScaled;
float range;
float rangeMaxClamped;
float edgeHorz;
float edgeVert;
float edgeHorz1;
float edgeVert1;
float edgeHorz2;
float edgeVert2;
float edgeHorz3;
float edgeVert3;
float edgeHorz4;
float edgeVert4;
float lengthSign;
float subpixA;
float subpixB;
float subpixC;
float subpixD;
float subpixE;
float subpixF;
float subpixG;
float subpixH;
float gradientN;
float gradientS;
float gradient;
float gradientScaled;
float dstN;
float dstP;
float dst;
float spanLength;
float spanLengthRcp;
float pixelOffset;
float pixelOffsetGood;
float pixelOffsetSubpix;
float opx;
float opy;
int directionN;
int goodSpan;
int goodSpanN;
int goodSpanP;
int horzSpan;
int earlyExit;
int pairN;
int doneN;
int doneP;
int doneNP;
int curr_x=0;
int curr_y=0;
opx = (1.0f / (float)(buf_x));
opy = (1.0f / (float)(buf_y));
for(curr_y=0; curr_y < buf_y; curr_y++) {
for(curr_x=0; curr_x < buf_x; curr_x++) {
posM_x = ((float)(curr_x) + 0.5f) * opx;
posM_y = ((float)(curr_y) + 0.5f) * opy;
//#define __PLX_BILINEAR_INITIAL_SAMPLES__ 1
#ifdef __PLX_BILINEAR_INITIAL_SAMPLES__
lumaM = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x, posM_y);
lumaS = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x, posM_y + opy);
lumaE = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x + opx, posM_y);
lumaN = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x, posM_y - opy);
lumaW = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x - opx, posM_y);
#else
lumaM = get_pixel_intensity(buf, buf_x, buf_y, curr_x, curr_y);
lumaS = get_pixel_intensity(buf, buf_x, buf_y, curr_x, curr_y + 1);
lumaE = get_pixel_intensity(buf, buf_x, buf_y, curr_x + 1, curr_y);
lumaN = get_pixel_intensity(buf, buf_x, buf_y, curr_x, curr_y - 1);
lumaW = get_pixel_intensity(buf, buf_x, buf_y, curr_x - 1, curr_y);
#endif
maxSM = MAX2(lumaS, lumaM);
minSM = MIN2(lumaS, lumaM);
maxESM = MAX2(lumaE, maxSM);
minESM = MIN2(lumaE, minSM);
maxWN = MAX2(lumaN, lumaW);
minWN = MIN2(lumaN, lumaW);
rangeMax = MAX2(maxWN, maxESM);
rangeMin = MIN2(minWN, minESM);
rangeMaxScaled = rangeMax * edge_threshold;
range = rangeMax - rangeMin;
rangeMaxClamped = MAX2(edge_threshold_min, rangeMaxScaled);
earlyExit = range < rangeMaxClamped ? 1:0;
if(earlyExit) {
set_pixel_intensity(buf, buf_x, buf_y, curr_x, curr_y, lumaM);
} else {
lumaNW = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x - opx, posM_y - opy);
lumaSE = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x + opx, posM_y + opy);
lumaNE = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x + opx, posM_y - opy);
lumaSW = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x - opx, posM_y + opy);
lumaNS = lumaN + lumaS;
lumaWE = lumaW + lumaE;
subpixRcpRange = 1.0f/range;
subpixNSWE = lumaNS + lumaWE;
edgeHorz1 = (-2.0f * lumaM) + lumaNS;
edgeVert1 = (-2.0f * lumaM) + lumaWE;
lumaNESE = lumaNE + lumaSE;
lumaNWNE = lumaNW + lumaNE;
edgeHorz2 = (-2.0f * lumaE) + lumaNESE;
edgeVert2 = (-2.0f * lumaN) + lumaNWNE;
lumaNWSW = lumaNW + lumaSW;
lumaSWSE = lumaSW + lumaSE;
edgeHorz4 = (ABS(edgeHorz1) * 2.0f) + ABS(edgeHorz2);
edgeVert4 = (ABS(edgeVert1) * 2.0f) + ABS(edgeVert2);
edgeHorz3 = (-2.0f * lumaW) + lumaNWSW;
edgeVert3 = (-2.0f * lumaS) + lumaSWSE;
edgeHorz = ABS(edgeHorz3) + edgeHorz4;
edgeVert = ABS(edgeVert3) + edgeVert4;
subpixNWSWNESE = lumaNWSW + lumaNESE;
lengthSign = 1.0f / (float)(buf_x);
horzSpan = edgeHorz >= edgeVert ? 1:0;
subpixA = subpixNSWE * 2.0f + subpixNWSWNESE;
if(!horzSpan) {
lumaN = lumaW;
lumaS = lumaE;
} else {
lengthSign = 1.0f / (float)(buf_y);
}
subpixB = (subpixA * (1.0f/12.0f)) - lumaM;
gradientN = lumaN - lumaM;
gradientS = lumaS - lumaM;
lumaNN = lumaN + lumaM;
lumaSS = lumaS + lumaM;
pairN = (ABS(gradientN)) >= (ABS(gradientS)) ? 1:0;
gradient = MAX2(ABS(gradientN), ABS(gradientS));
if(pairN) {
lengthSign = -lengthSign;
}
subpixC = MAX2(MIN2(ABS(subpixB) * subpixRcpRange,1.0f),0.0f);
posB_x = posM_x;
posB_y = posM_y;
offNP_x = (!horzSpan) ? 0.0f:(1.0f / (float)(buf_x));
offNP_y = (horzSpan) ? 0.0f:(1.0f / (float)(buf_y));
if(!horzSpan) {
posB_x += lengthSign * 0.5f;
} else {
posB_y += lengthSign * 0.5f;
}
posN_x = posB_x - offNP_x * p0;
posN_y = posB_y - offNP_y * p0;
posP_x = posB_x + offNP_x * p0;
posP_y = posB_y + offNP_y * p0;
subpixD = ((-2.0f)*subpixC) + 3.0f;
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
subpixE = subpixC * subpixC;
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
if(!pairN) {
lumaNN = lumaSS;
}
gradientScaled = gradient * 1.0f/4.0f;
lumaMM =lumaM - lumaNN * 0.5f;
subpixF = subpixD * subpixE;
lumaMLTZero = lumaMM < 0.0f ? 1:0;
lumaEndN -= lumaNN * 0.5f;
lumaEndP -= lumaNN * 0.5f;
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p1;
posN_y -= offNP_y * p1;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p1;
posP_y += offNP_y * p1;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x, posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p2;
posN_y -= offNP_y * p2;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p2;
posP_y += offNP_y * p2;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p3;
posN_y -= offNP_y * p3;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p3;
posP_y += offNP_y * p3;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p4;
posN_y -= offNP_y * p4;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p4;
posP_y += offNP_y * p4;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p5;
posN_y -= offNP_y * p5;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p5;
posP_y += offNP_y * p5;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p6;
posN_y -= offNP_y * p6;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p6;
posP_y += offNP_y * p6;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p7;
posN_y -= offNP_y * p7;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p7;
posP_y += offNP_y * p7;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p8;
posN_y -= offNP_y * p8;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p8;
posP_y += offNP_y * p8;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p9;
posN_y -= offNP_y * p9;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p9;
posP_y += offNP_y * p9;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p10;
posN_y -= offNP_y * p10;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p10;
posP_y += offNP_y * p10;
}
if(doneNP) {
if(!doneN) {
lumaEndN = get_pixel_intensity_bilinear(buf, buf_x, buf_y,posN_x,posN_y);
}
if(!doneP) {
lumaEndP = get_pixel_intensity_bilinear(buf, buf_x, buf_y, posP_x,posP_y);
}
if(!doneN) {
lumaEndN = lumaEndN - lumaNN * 0.5;
}
if(!doneP) {
lumaEndP = lumaEndP - lumaNN * 0.5;
}
doneN = (ABS(lumaEndN)) >= gradientScaled ? 1:0;
doneP = (ABS(lumaEndP)) >= gradientScaled ? 1:0;
if(!doneN) {
posN_x -= offNP_x * p11;
posN_y -= offNP_y * p11;
}
doneNP = (!doneN) || (!doneP) ? 1:0;
if(!doneP) {
posP_x += offNP_x * p11;
posP_y += offNP_y * p11;
}
}
}
}
}
}
}
}
}
}
}
dstN = posM_x - posN_x;
dstP = posP_x - posM_x;
if(!horzSpan) {
dstN = posM_y - posN_y;
dstP = posP_y - posM_y;
}
goodSpanN = ((lumaEndN < 0.0f) ? 1:0) != lumaMLTZero ? 1:0;
spanLength = (dstP + dstN);
goodSpanP = ((lumaEndP < 0.0f) ? 1:0) != lumaMLTZero ? 1:0;
spanLengthRcp = 1.0f/spanLength;
directionN = dstN < dstP ? 1:0;
dst = MIN2(dstN, dstP);
goodSpan = (directionN==1) ? goodSpanN:goodSpanP;
subpixG = subpixF * subpixF;
pixelOffset = (dst * (-spanLengthRcp)) + 0.5f;
subpixH = subpixG * quality_subpix;
pixelOffsetGood = (goodSpan==1) ? pixelOffset : 0.0f;
pixelOffsetSubpix = MAX2(pixelOffsetGood, subpixH);
if(!horzSpan) {
posM_x += pixelOffsetSubpix * lengthSign;
} else {
posM_y += pixelOffsetSubpix * lengthSign;
}
//may need bilinear filtered get_pixel_intensity() here...
set_pixel_intensity(buf,buf_x,buf_y,curr_x,curr_y,get_pixel_intensity_bilinear(buf, buf_x, buf_y, posM_x,posM_y));
}
}
}
return 1;
#endif
}
#define SWAP_POLYVERT(a,b) ptemp[0]=(a)[0]; ptemp[1]=(a)[1]; (a)[0]=(b)[0]; (a)[1]=(b)[1]; (b)[0]=ptemp[0]; (b)[1]=ptemp[1];
#define __PLX_SMALL_COUNT__ 13
void plx_floatsort(float(*f)[2], unsigned int n, int sortby) {
unsigned int i;
unsigned int ir;
unsigned int j;
unsigned int k;
unsigned int l=1;
unsigned int istack[50];
int jstack=0;
float a[2];
float ptemp[2];
ir=n;
for(;;) {
if(ir-l < __PLX_SMALL_COUNT__) {
for(j=l+1; j<=ir; j++) {
a[1]=f[j][1];
a[0]=f[j][0];
for(i=j-1; i>=l; i--) {
if(f[i][sortby] <= a[sortby]) {
break;
}
f[i+1][1]=f[i][1];
f[i+1][0]=f[i][0];
}
f[i+1][1]=a[1];
f[i+1][0]=a[0];
}
if(jstack < 0) {
break;
}
ir=istack[jstack--];
l=istack[jstack--];
} else {
k=(l+ir) >> 1;
SWAP_POLYVERT(f[k],f[l+1])
if(f[l][sortby] > f[ir][sortby]) {
SWAP_POLYVERT(f[l],f[ir])
}
if(f[l+1][sortby] > f[ir][sortby]) {
SWAP_POLYVERT(f[l+1],f[ir])
}
if(f[l][sortby] > f[l+1][sortby]) {
SWAP_POLYVERT(f[l],f[l+1])
}
i=l+1;
j=ir;
a[0]=f[l+1][0];
a[1]=f[l+1][1];
for(;;) {
do i++;
while(f[i][sortby] < a[sortby]);
do j--;
while(f[j][sortby] > a[sortby]);
if(j < i) {
break;
}
SWAP_POLYVERT(f[i],f[j])
}
f[l+1][0]=f[j][0];
f[l+1][1]=f[j][1];
f[j][0]=a[0];
f[j][1]=a[1];
jstack+=2;
if(jstack > __PLX_SMALL_COUNT__) {
return;
}
if(ir-i+1 >= j-l) {
istack[jstack]=ir;
istack[jstack-1]=i;
ir=j-1;
} else {
istack[jstack]=j-1;
istack[jstack-1]=l;
l=i;
}
}
}
}
int plx_find_lower_bound(float v, float(*a)[2], int num_feather_verts) {
int x;
int l;
int r;
l=1;
r=num_feather_verts;
for(;;) {
// interpolation style search
//x=l+(v-a[l][1])*(r-l) / (a[r][1]-a[l][1]);
// binary search
x=(l+r) / 2;
if(v<a[x][1]) {
r=x-1;
} else {
l=x+1;
}
if((v>a[x-1][1] && v <= a[x][1]) || l>r) {
break;
}
}
if(v>a[x-1][1] && v <= a[x][1]) {
return x;
} else {
return num_feather_verts;
}
}
int plx_find_upper_bound(float v, float(*a)[2], int num_feather_verts) {
int x;
int l;
int r;
l=1;
r=num_feather_verts;
for(;;) {
// interpolation style search
//x=l+(v-a[l][1])*(r-l) / (a[r][1]-a[l][1]);
// binary search
x=(l+r) / 2;
if(v<a[x][1]) {
r=x-1;
} else {
l=x+1;
}
if((v>=a[x-1][1] && v < a[x][1]) || l>r) {
break;
}
}
if(v>=a[x-1][1] && v < a[x][1]) {
return x-1;
} else {
return num_feather_verts;
}
}
|