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
|
/*
* Copyright 2011-2013 Blender Foundation
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
CCL_NAMESPACE_BEGIN
/* Events for probalistic scattering */
typedef enum VolumeIntegrateResult {
VOLUME_PATH_SCATTERED = 0,
VOLUME_PATH_ATTENUATED = 1,
VOLUME_PATH_MISSED = 2
} VolumeIntegrateResult;
/* Volume shader properties
*
* extinction coefficient = absorption coefficient + scattering coefficient
* sigma_t = sigma_a + sigma_s */
typedef struct VolumeShaderCoefficients {
float3 sigma_a;
float3 sigma_s;
float3 emission;
} VolumeShaderCoefficients;
/* evaluate shader to get extinction coefficient at P */
ccl_device_inline bool volume_shader_extinction_sample(KernelGlobals *kg,
ShaderData *sd,
ccl_addr_space PathState *state,
float3 P,
float3 *extinction)
{
sd->P = P;
shader_eval_volume(kg, sd, state, state->volume_stack, PATH_RAY_SHADOW);
if(!(sd->flag & (SD_ABSORPTION|SD_SCATTER)))
return false;
float3 sigma_t = make_float3(0.0f, 0.0f, 0.0f);
for(int i = 0; i < sd->num_closure; i++) {
const ShaderClosure *sc = &sd->closure[i];
if(CLOSURE_IS_VOLUME(sc->type))
sigma_t += sc->weight;
}
*extinction = sigma_t;
return true;
}
/* evaluate shader to get absorption, scattering and emission at P */
ccl_device_inline bool volume_shader_sample(KernelGlobals *kg,
ShaderData *sd,
ccl_addr_space PathState *state,
float3 P,
VolumeShaderCoefficients *coeff)
{
sd->P = P;
shader_eval_volume(kg, sd, state, state->volume_stack, state->flag);
if(!(sd->flag & (SD_ABSORPTION|SD_SCATTER|SD_EMISSION)))
return false;
coeff->sigma_a = make_float3(0.0f, 0.0f, 0.0f);
coeff->sigma_s = make_float3(0.0f, 0.0f, 0.0f);
coeff->emission = make_float3(0.0f, 0.0f, 0.0f);
for(int i = 0; i < sd->num_closure; i++) {
const ShaderClosure *sc = &sd->closure[i];
if(sc->type == CLOSURE_VOLUME_ABSORPTION_ID)
coeff->sigma_a += sc->weight;
else if(sc->type == CLOSURE_EMISSION_ID)
coeff->emission += sc->weight;
else if(CLOSURE_IS_VOLUME(sc->type))
coeff->sigma_s += sc->weight;
}
/* when at the max number of bounces, treat scattering as absorption */
if(sd->flag & SD_SCATTER) {
if(state->volume_bounce >= kernel_data.integrator.max_volume_bounce) {
coeff->sigma_a += coeff->sigma_s;
coeff->sigma_s = make_float3(0.0f, 0.0f, 0.0f);
sd->flag &= ~SD_SCATTER;
sd->flag |= SD_ABSORPTION;
}
}
return true;
}
ccl_device float3 volume_color_transmittance(float3 sigma, float t)
{
return make_float3(expf(-sigma.x * t), expf(-sigma.y * t), expf(-sigma.z * t));
}
ccl_device float kernel_volume_channel_get(float3 value, int channel)
{
return (channel == 0)? value.x: ((channel == 1)? value.y: value.z);
}
ccl_device bool volume_stack_is_heterogeneous(KernelGlobals *kg, ccl_addr_space VolumeStack *stack)
{
for(int i = 0; stack[i].shader != SHADER_NONE; i++) {
int shader_flag = kernel_tex_fetch(__shader_flag, (stack[i].shader & SHADER_MASK)*SHADER_SIZE);
if(shader_flag & SD_HETEROGENEOUS_VOLUME)
return true;
}
return false;
}
ccl_device int volume_stack_sampling_method(KernelGlobals *kg, VolumeStack *stack)
{
if(kernel_data.integrator.num_all_lights == 0)
return 0;
int method = -1;
for(int i = 0; stack[i].shader != SHADER_NONE; i++) {
int shader_flag = kernel_tex_fetch(__shader_flag, (stack[i].shader & SHADER_MASK)*SHADER_SIZE);
if(shader_flag & SD_VOLUME_MIS) {
return SD_VOLUME_MIS;
}
else if(shader_flag & SD_VOLUME_EQUIANGULAR) {
if(method == 0)
return SD_VOLUME_MIS;
method = SD_VOLUME_EQUIANGULAR;
}
else {
if(method == SD_VOLUME_EQUIANGULAR)
return SD_VOLUME_MIS;
method = 0;
}
}
return method;
}
/* Volume Shadows
*
* These functions are used to attenuate shadow rays to lights. Both absorption
* and scattering will block light, represented by the extinction coefficient. */
/* homogeneous volume: assume shader evaluation at the starts gives
* the extinction coefficient for the entire line segment */
ccl_device void kernel_volume_shadow_homogeneous(KernelGlobals *kg,
ccl_addr_space PathState *state,
Ray *ray,
ShaderData *sd,
float3 *throughput)
{
float3 sigma_t;
if(volume_shader_extinction_sample(kg, sd, state, ray->P, &sigma_t))
*throughput *= volume_color_transmittance(sigma_t, ray->t);
}
/* heterogeneous volume: integrate stepping through the volume until we
* reach the end, get absorbed entirely, or run out of iterations */
ccl_device void kernel_volume_shadow_heterogeneous(KernelGlobals *kg,
ccl_addr_space PathState *state,
Ray *ray,
ShaderData *sd,
float3 *throughput)
{
float3 tp = *throughput;
const float tp_eps = 1e-6f; /* todo: this is likely not the right value */
/* prepare for stepping */
int max_steps = kernel_data.integrator.volume_max_steps;
float step = kernel_data.integrator.volume_step_size;
float random_jitter_offset = lcg_step_float_addrspace(&state->rng_congruential) * step;
/* compute extinction at the start */
float t = 0.0f;
float3 sum = make_float3(0.0f, 0.0f, 0.0f);
for(int i = 0; i < max_steps; i++) {
/* advance to new position */
float new_t = min(ray->t, (i+1) * step);
float dt = new_t - t;
/* use random position inside this segment to sample shader */
if(new_t == ray->t)
random_jitter_offset = lcg_step_float_addrspace(&state->rng_congruential) * dt;
float3 new_P = ray->P + ray->D * (t + random_jitter_offset);
float3 sigma_t;
/* compute attenuation over segment */
if(volume_shader_extinction_sample(kg, sd, state, new_P, &sigma_t)) {
/* Compute expf() only for every Nth step, to save some calculations
* because exp(a)*exp(b) = exp(a+b), also do a quick tp_eps check then. */
sum += (-sigma_t * (new_t - t));
if((i & 0x07) == 0) { /* ToDo: Other interval? */
tp = *throughput * make_float3(expf(sum.x), expf(sum.y), expf(sum.z));
/* stop if nearly all light is blocked */
if(tp.x < tp_eps && tp.y < tp_eps && tp.z < tp_eps)
break;
}
}
/* stop if at the end of the volume */
t = new_t;
if(t == ray->t) {
/* Update throughput in case we haven't done it above */
tp = *throughput * make_float3(expf(sum.x), expf(sum.y), expf(sum.z));
break;
}
}
*throughput = tp;
}
/* get the volume attenuation over line segment defined by ray, with the
* assumption that there are no surfaces blocking light between the endpoints */
ccl_device_noinline void kernel_volume_shadow(KernelGlobals *kg,
ShaderData *shadow_sd,
ccl_addr_space PathState *state,
Ray *ray,
float3 *throughput)
{
shader_setup_from_volume(kg, shadow_sd, ray);
if(volume_stack_is_heterogeneous(kg, state->volume_stack))
kernel_volume_shadow_heterogeneous(kg, state, ray, shadow_sd, throughput);
else
kernel_volume_shadow_homogeneous(kg, state, ray, shadow_sd, throughput);
}
/* Equi-angular sampling as in:
* "Importance Sampling Techniques for Path Tracing in Participating Media" */
ccl_device float kernel_volume_equiangular_sample(Ray *ray, float3 light_P, float xi, float *pdf)
{
float t = ray->t;
float delta = dot((light_P - ray->P) , ray->D);
float D = safe_sqrtf(len_squared(light_P - ray->P) - delta * delta);
if(UNLIKELY(D == 0.0f)) {
*pdf = 0.0f;
return 0.0f;
}
float theta_a = -atan2f(delta, D);
float theta_b = atan2f(t - delta, D);
float t_ = D * tanf((xi * theta_b) + (1 - xi) * theta_a);
if(UNLIKELY(theta_b == theta_a)) {
*pdf = 0.0f;
return 0.0f;
}
*pdf = D / ((theta_b - theta_a) * (D * D + t_ * t_));
return min(t, delta + t_); /* min is only for float precision errors */
}
ccl_device float kernel_volume_equiangular_pdf(Ray *ray, float3 light_P, float sample_t)
{
float delta = dot((light_P - ray->P) , ray->D);
float D = safe_sqrtf(len_squared(light_P - ray->P) - delta * delta);
if(UNLIKELY(D == 0.0f)) {
return 0.0f;
}
float t = ray->t;
float t_ = sample_t - delta;
float theta_a = -atan2f(delta, D);
float theta_b = atan2f(t - delta, D);
if(UNLIKELY(theta_b == theta_a)) {
return 0.0f;
}
float pdf = D / ((theta_b - theta_a) * (D * D + t_ * t_));
return pdf;
}
/* Distance sampling */
ccl_device float kernel_volume_distance_sample(float max_t, float3 sigma_t, int channel, float xi, float3 *transmittance, float3 *pdf)
{
/* xi is [0, 1[ so log(0) should never happen, division by zero is
* avoided because sample_sigma_t > 0 when SD_SCATTER is set */
float sample_sigma_t = kernel_volume_channel_get(sigma_t, channel);
float3 full_transmittance = volume_color_transmittance(sigma_t, max_t);
float sample_transmittance = kernel_volume_channel_get(full_transmittance, channel);
float sample_t = min(max_t, -logf(1.0f - xi*(1.0f - sample_transmittance))/sample_sigma_t);
*transmittance = volume_color_transmittance(sigma_t, sample_t);
*pdf = safe_divide_color(sigma_t * *transmittance, make_float3(1.0f, 1.0f, 1.0f) - full_transmittance);
/* todo: optimization: when taken together with hit/miss decision,
* the full_transmittance cancels out drops out and xi does not
* need to be remapped */
return sample_t;
}
ccl_device float3 kernel_volume_distance_pdf(float max_t, float3 sigma_t, float sample_t)
{
float3 full_transmittance = volume_color_transmittance(sigma_t, max_t);
float3 transmittance = volume_color_transmittance(sigma_t, sample_t);
return safe_divide_color(sigma_t * transmittance, make_float3(1.0f, 1.0f, 1.0f) - full_transmittance);
}
/* Emission */
ccl_device float3 kernel_volume_emission_integrate(VolumeShaderCoefficients *coeff, int closure_flag, float3 transmittance, float t)
{
/* integral E * exp(-sigma_t * t) from 0 to t = E * (1 - exp(-sigma_t * t))/sigma_t
* this goes to E * t as sigma_t goes to zero
*
* todo: we should use an epsilon to avoid precision issues near zero sigma_t */
float3 emission = coeff->emission;
if(closure_flag & SD_ABSORPTION) {
float3 sigma_t = coeff->sigma_a + coeff->sigma_s;
emission.x *= (sigma_t.x > 0.0f)? (1.0f - transmittance.x)/sigma_t.x: t;
emission.y *= (sigma_t.y > 0.0f)? (1.0f - transmittance.y)/sigma_t.y: t;
emission.z *= (sigma_t.z > 0.0f)? (1.0f - transmittance.z)/sigma_t.z: t;
}
else
emission *= t;
return emission;
}
/* Volume Path */
/* homogeneous volume: assume shader evaluation at the start gives
* the volume shading coefficient for the entire line segment */
ccl_device VolumeIntegrateResult kernel_volume_integrate_homogeneous(
KernelGlobals *kg,
ccl_addr_space PathState *state,
Ray *ray,
ShaderData *sd,
PathRadiance *L,
ccl_addr_space float3 *throughput,
bool probalistic_scatter)
{
VolumeShaderCoefficients coeff;
if(!volume_shader_sample(kg, sd, state, ray->P, &coeff))
return VOLUME_PATH_MISSED;
int closure_flag = sd->flag;
float t = ray->t;
float3 new_tp;
#ifdef __VOLUME_SCATTER__
/* randomly scatter, and if we do t is shortened */
if(closure_flag & SD_SCATTER) {
/* extinction coefficient */
float3 sigma_t = coeff.sigma_a + coeff.sigma_s;
/* pick random color channel, we use the Veach one-sample
* model with balance heuristic for the channels */
float rphase = path_state_rng_1D_for_decision(kg, state, PRNG_PHASE);
int channel = (int)(rphase*3.0f);
sd->randb_closure = rphase*3.0f - channel;
/* decide if we will hit or miss */
bool scatter = true;
float xi = path_state_rng_1D_for_decision(kg, state, PRNG_SCATTER_DISTANCE);
if(probalistic_scatter) {
float sample_sigma_t = kernel_volume_channel_get(sigma_t, channel);
float sample_transmittance = expf(-sample_sigma_t * t);
if(1.0f - xi >= sample_transmittance) {
scatter = true;
/* rescale random number so we can reuse it */
xi = 1.0f - (1.0f - xi - sample_transmittance)/(1.0f - sample_transmittance);
}
else
scatter = false;
}
if(scatter) {
/* scattering */
float3 pdf;
float3 transmittance;
float sample_t;
/* distance sampling */
sample_t = kernel_volume_distance_sample(ray->t, sigma_t, channel, xi, &transmittance, &pdf);
/* modify pdf for hit/miss decision */
if(probalistic_scatter)
pdf *= make_float3(1.0f, 1.0f, 1.0f) - volume_color_transmittance(sigma_t, t);
new_tp = *throughput * coeff.sigma_s * transmittance / average(pdf);
t = sample_t;
}
else {
/* no scattering */
float3 transmittance = volume_color_transmittance(sigma_t, t);
float pdf = average(transmittance);
new_tp = *throughput * transmittance / pdf;
}
}
else
#endif
if(closure_flag & SD_ABSORPTION) {
/* absorption only, no sampling needed */
float3 transmittance = volume_color_transmittance(coeff.sigma_a, t);
new_tp = *throughput * transmittance;
}
/* integrate emission attenuated by extinction */
if(L && (closure_flag & SD_EMISSION)) {
float3 sigma_t = coeff.sigma_a + coeff.sigma_s;
float3 transmittance = volume_color_transmittance(sigma_t, ray->t);
float3 emission = kernel_volume_emission_integrate(&coeff, closure_flag, transmittance, ray->t);
path_radiance_accum_emission(L, state, *throughput, emission);
}
/* modify throughput */
if(closure_flag & (SD_ABSORPTION|SD_SCATTER)) {
*throughput = new_tp;
/* prepare to scatter to new direction */
if(t < ray->t) {
/* adjust throughput and move to new location */
sd->P = ray->P + t*ray->D;
return VOLUME_PATH_SCATTERED;
}
}
return VOLUME_PATH_ATTENUATED;
}
/* heterogeneous volume distance sampling: integrate stepping through the
* volume until we reach the end, get absorbed entirely, or run out of
* iterations. this does probabilistically scatter or get transmitted through
* for path tracing where we don't want to branch. */
ccl_device VolumeIntegrateResult kernel_volume_integrate_heterogeneous_distance(
KernelGlobals *kg,
ccl_addr_space PathState *state,
Ray *ray,
ShaderData *sd,
PathRadiance *L,
ccl_addr_space float3 *throughput)
{
float3 tp = *throughput;
const float tp_eps = 1e-6f; /* todo: this is likely not the right value */
/* prepare for stepping */
int max_steps = kernel_data.integrator.volume_max_steps;
float step_size = kernel_data.integrator.volume_step_size;
float random_jitter_offset = lcg_step_float_addrspace(&state->rng_congruential) * step_size;
/* compute coefficients at the start */
float t = 0.0f;
float3 accum_transmittance = make_float3(1.0f, 1.0f, 1.0f);
/* pick random color channel, we use the Veach one-sample
* model with balance heuristic for the channels */
float xi = path_state_rng_1D_for_decision(kg, state, PRNG_SCATTER_DISTANCE);
float rphase = path_state_rng_1D_for_decision(kg, state, PRNG_PHASE);
int channel = (int)(rphase*3.0f);
sd->randb_closure = rphase*3.0f - channel;
bool has_scatter = false;
for(int i = 0; i < max_steps; i++) {
/* advance to new position */
float new_t = min(ray->t, (i+1) * step_size);
float dt = new_t - t;
/* use random position inside this segment to sample shader */
if(new_t == ray->t)
random_jitter_offset = lcg_step_float_addrspace(&state->rng_congruential) * dt;
float3 new_P = ray->P + ray->D * (t + random_jitter_offset);
VolumeShaderCoefficients coeff;
/* compute segment */
if(volume_shader_sample(kg, sd, state, new_P, &coeff)) {
int closure_flag = sd->flag;
float3 new_tp;
float3 transmittance;
bool scatter = false;
/* distance sampling */
#ifdef __VOLUME_SCATTER__
if((closure_flag & SD_SCATTER) || (has_scatter && (closure_flag & SD_ABSORPTION))) {
has_scatter = true;
float3 sigma_t = coeff.sigma_a + coeff.sigma_s;
float3 sigma_s = coeff.sigma_s;
/* compute transmittance over full step */
transmittance = volume_color_transmittance(sigma_t, dt);
/* decide if we will scatter or continue */
float sample_transmittance = kernel_volume_channel_get(transmittance, channel);
if(1.0f - xi >= sample_transmittance) {
/* compute sampling distance */
float sample_sigma_t = kernel_volume_channel_get(sigma_t, channel);
float new_dt = -logf(1.0f - xi)/sample_sigma_t;
new_t = t + new_dt;
/* transmittance and pdf */
float3 new_transmittance = volume_color_transmittance(sigma_t, new_dt);
float3 pdf = sigma_t * new_transmittance;
/* throughput */
new_tp = tp * sigma_s * new_transmittance / average(pdf);
scatter = true;
}
else {
/* throughput */
float pdf = average(transmittance);
new_tp = tp * transmittance / pdf;
/* remap xi so we can reuse it and keep thing stratified */
xi = 1.0f - (1.0f - xi)/sample_transmittance;
}
}
else
#endif
if(closure_flag & SD_ABSORPTION) {
/* absorption only, no sampling needed */
float3 sigma_a = coeff.sigma_a;
transmittance = volume_color_transmittance(sigma_a, dt);
new_tp = tp * transmittance;
}
/* integrate emission attenuated by absorption */
if(L && (closure_flag & SD_EMISSION)) {
float3 emission = kernel_volume_emission_integrate(&coeff, closure_flag, transmittance, dt);
path_radiance_accum_emission(L, state, tp, emission);
}
/* modify throughput */
if(closure_flag & (SD_ABSORPTION|SD_SCATTER)) {
tp = new_tp;
/* stop if nearly all light blocked */
if(tp.x < tp_eps && tp.y < tp_eps && tp.z < tp_eps) {
tp = make_float3(0.0f, 0.0f, 0.0f);
break;
}
}
/* prepare to scatter to new direction */
if(scatter) {
/* adjust throughput and move to new location */
sd->P = ray->P + new_t*ray->D;
*throughput = tp;
return VOLUME_PATH_SCATTERED;
}
else {
/* accumulate transmittance */
accum_transmittance *= transmittance;
}
}
/* stop if at the end of the volume */
t = new_t;
if(t == ray->t)
break;
}
*throughput = tp;
return VOLUME_PATH_ATTENUATED;
}
/* get the volume attenuation and emission over line segment defined by
* ray, with the assumption that there are no surfaces blocking light
* between the endpoints. distance sampling is used to decide if we will
* scatter or not. */
ccl_device_noinline VolumeIntegrateResult kernel_volume_integrate(
KernelGlobals *kg,
ccl_addr_space PathState *state,
ShaderData *sd,
Ray *ray,
PathRadiance *L,
ccl_addr_space float3 *throughput,
bool heterogeneous)
{
shader_setup_from_volume(kg, sd, ray);
if(heterogeneous)
return kernel_volume_integrate_heterogeneous_distance(kg, state, ray, sd, L, throughput);
else
return kernel_volume_integrate_homogeneous(kg, state, ray, sd, L, throughput, true);
}
#ifndef __SPLIT_KERNEL__
/* Decoupled Volume Sampling
*
* VolumeSegment is list of coefficients and transmittance stored at all steps
* through a volume. This can then later be used for decoupled sampling as in:
* "Importance Sampling Techniques for Path Tracing in Participating Media"
*
* On the GPU this is only supported (but currently not enabled)
* for homogeneous volumes (1 step), due to
* no support for malloc/free and too much stack usage with a fix size array. */
typedef struct VolumeStep {
float3 sigma_s; /* scatter coefficient */
float3 sigma_t; /* extinction coefficient */
float3 accum_transmittance; /* accumulated transmittance including this step */
float3 cdf_distance; /* cumulative density function for distance sampling */
float t; /* distance at end of this step */
float shade_t; /* jittered distance where shading was done in step */
int closure_flag; /* shader evaluation closure flags */
} VolumeStep;
typedef struct VolumeSegment {
VolumeStep stack_step; /* stack storage for homogeneous step, to avoid malloc */
VolumeStep *steps; /* recorded steps */
int numsteps; /* number of steps */
int closure_flag; /* accumulated closure flags from all steps */
float3 accum_emission; /* accumulated emission at end of segment */
float3 accum_transmittance; /* accumulated transmittance at end of segment */
int sampling_method; /* volume sampling method */
} VolumeSegment;
/* record volume steps to the end of the volume.
*
* it would be nice if we could only record up to the point that we need to scatter,
* but the entire segment is needed to do always scattering, rather than probabilistically
* hitting or missing the volume. if we don't know the transmittance at the end of the
* volume we can't generate stratified distance samples up to that transmittance */
#ifdef __VOLUME_DECOUPLED__
ccl_device void kernel_volume_decoupled_record(KernelGlobals *kg, PathState *state,
Ray *ray, ShaderData *sd, VolumeSegment *segment, bool heterogeneous)
{
const float tp_eps = 1e-6f; /* todo: this is likely not the right value */
/* prepare for volume stepping */
int max_steps;
float step_size, random_jitter_offset;
if(heterogeneous) {
const int global_max_steps = kernel_data.integrator.volume_max_steps;
step_size = kernel_data.integrator.volume_step_size;
/* compute exact steps in advance for malloc */
if(ray->t > global_max_steps*step_size) {
max_steps = global_max_steps;
step_size = ray->t / (float)max_steps;
}
else {
max_steps = max((int)ceilf(ray->t/step_size), 1);
}
#ifdef __KERNEL_CPU__
/* NOTE: For the branched path tracing it's possible to have direct
* and indirect light integration both having volume segments allocated.
* We detect this using index in the pre-allocated memory. Currently we
* only support two segments allocated at a time, if more needed some
* modifications to the KernelGlobals will be needed.
*
* This gives us restrictions that decoupled record should only happen
* in the stack manner, meaning if there's subsequent call of decoupled
* record it'll need to free memory before it's caller frees memory.
*/
const int index = kg->decoupled_volume_steps_index;
assert(index < sizeof(kg->decoupled_volume_steps) /
sizeof(*kg->decoupled_volume_steps));
if(kg->decoupled_volume_steps[index] == NULL) {
kg->decoupled_volume_steps[index] =
(VolumeStep*)malloc(sizeof(VolumeStep)*global_max_steps);
}
segment->steps = kg->decoupled_volume_steps[index];
++kg->decoupled_volume_steps_index;
#else
segment->steps = (VolumeStep*)malloc(sizeof(VolumeStep)*max_steps);
#endif
random_jitter_offset = lcg_step_float(&state->rng_congruential) * step_size;
}
else {
max_steps = 1;
step_size = ray->t;
random_jitter_offset = 0.0f;
segment->steps = &segment->stack_step;
}
/* init accumulation variables */
float3 accum_emission = make_float3(0.0f, 0.0f, 0.0f);
float3 accum_transmittance = make_float3(1.0f, 1.0f, 1.0f);
float3 cdf_distance = make_float3(0.0f, 0.0f, 0.0f);
float t = 0.0f;
segment->numsteps = 0;
segment->closure_flag = 0;
bool is_last_step_empty = false;
VolumeStep *step = segment->steps;
for(int i = 0; i < max_steps; i++, step++) {
/* advance to new position */
float new_t = min(ray->t, (i+1) * step_size);
float dt = new_t - t;
/* use random position inside this segment to sample shader */
if(heterogeneous && new_t == ray->t)
random_jitter_offset = lcg_step_float(&state->rng_congruential) * dt;
float3 new_P = ray->P + ray->D * (t + random_jitter_offset);
VolumeShaderCoefficients coeff;
/* compute segment */
if(volume_shader_sample(kg, sd, state, new_P, &coeff)) {
int closure_flag = sd->flag;
float3 sigma_t = coeff.sigma_a + coeff.sigma_s;
/* compute accumulated transmittance */
float3 transmittance = volume_color_transmittance(sigma_t, dt);
/* compute emission attenuated by absorption */
if(closure_flag & SD_EMISSION) {
float3 emission = kernel_volume_emission_integrate(&coeff, closure_flag, transmittance, dt);
accum_emission += accum_transmittance * emission;
}
accum_transmittance *= transmittance;
/* compute pdf for distance sampling */
float3 pdf_distance = dt * accum_transmittance * coeff.sigma_s;
cdf_distance = cdf_distance + pdf_distance;
/* write step data */
step->sigma_t = sigma_t;
step->sigma_s = coeff.sigma_s;
step->closure_flag = closure_flag;
segment->closure_flag |= closure_flag;
is_last_step_empty = false;
segment->numsteps++;
}
else {
if(is_last_step_empty) {
/* consecutive empty step, merge */
step--;
}
else {
/* store empty step */
step->sigma_t = make_float3(0.0f, 0.0f, 0.0f);
step->sigma_s = make_float3(0.0f, 0.0f, 0.0f);
step->closure_flag = 0;
segment->numsteps++;
is_last_step_empty = true;
}
}
step->accum_transmittance = accum_transmittance;
step->cdf_distance = cdf_distance;
step->t = new_t;
step->shade_t = t + random_jitter_offset;
/* stop if at the end of the volume */
t = new_t;
if(t == ray->t)
break;
/* stop if nearly all light blocked */
if(accum_transmittance.x < tp_eps && accum_transmittance.y < tp_eps && accum_transmittance.z < tp_eps)
break;
}
/* store total emission and transmittance */
segment->accum_emission = accum_emission;
segment->accum_transmittance = accum_transmittance;
/* normalize cumulative density function for distance sampling */
VolumeStep *last_step = segment->steps + segment->numsteps - 1;
if(!is_zero(last_step->cdf_distance)) {
VolumeStep *step = &segment->steps[0];
int numsteps = segment->numsteps;
float3 inv_cdf_distance_sum = safe_invert_color(last_step->cdf_distance);
for(int i = 0; i < numsteps; i++, step++)
step->cdf_distance *= inv_cdf_distance_sum;
}
}
ccl_device void kernel_volume_decoupled_free(KernelGlobals *kg, VolumeSegment *segment)
{
if(segment->steps != &segment->stack_step) {
#ifdef __KERNEL_CPU__
/* NOTE: We only allow free last allocated segment.
* No random order of alloc/free is supported.
*/
assert(kg->decoupled_volume_steps_index > 0);
assert(segment->steps == kg->decoupled_volume_steps[kg->decoupled_volume_steps_index - 1]);
--kg->decoupled_volume_steps_index;
#else
free(segment->steps);
#endif
}
}
#endif /* __VOLUME_DECOUPLED__ */
/* scattering for homogeneous and heterogeneous volumes, using decoupled ray
* marching.
*
* function is expected to return VOLUME_PATH_SCATTERED when probalistic_scatter is false */
ccl_device VolumeIntegrateResult kernel_volume_decoupled_scatter(
KernelGlobals *kg, PathState *state, Ray *ray, ShaderData *sd,
float3 *throughput, float rphase, float rscatter,
const VolumeSegment *segment, const float3 *light_P, bool probalistic_scatter)
{
kernel_assert(segment->closure_flag & SD_SCATTER);
/* pick random color channel, we use the Veach one-sample
* model with balance heuristic for the channels */
int channel = (int)(rphase*3.0f);
sd->randb_closure = rphase*3.0f - channel;
float xi = rscatter;
/* probabilistic scattering decision based on transmittance */
if(probalistic_scatter) {
float sample_transmittance = kernel_volume_channel_get(segment->accum_transmittance, channel);
if(1.0f - xi >= sample_transmittance) {
/* rescale random number so we can reuse it */
xi = 1.0f - (1.0f - xi - sample_transmittance)/(1.0f - sample_transmittance);
}
else {
*throughput /= sample_transmittance;
return VOLUME_PATH_MISSED;
}
}
VolumeStep *step;
float3 transmittance;
float pdf, sample_t;
float mis_weight = 1.0f;
bool distance_sample = true;
bool use_mis = false;
if(segment->sampling_method && light_P) {
if(segment->sampling_method == SD_VOLUME_MIS) {
/* multiple importance sample: randomly pick between
* equiangular and distance sampling strategy */
if(xi < 0.5f) {
xi *= 2.0f;
}
else {
xi = (xi - 0.5f)*2.0f;
distance_sample = false;
}
use_mis = true;
}
else {
/* only equiangular sampling */
distance_sample = false;
}
}
/* distance sampling */
if(distance_sample) {
/* find step in cdf */
step = segment->steps;
float prev_t = 0.0f;
float3 step_pdf_distance = make_float3(1.0f, 1.0f, 1.0f);
if(segment->numsteps > 1) {
float prev_cdf = 0.0f;
float step_cdf = 1.0f;
float3 prev_cdf_distance = make_float3(0.0f, 0.0f, 0.0f);
for(int i = 0; ; i++, step++) {
/* todo: optimize using binary search */
step_cdf = kernel_volume_channel_get(step->cdf_distance, channel);
if(xi < step_cdf || i == segment->numsteps-1)
break;
prev_cdf = step_cdf;
prev_t = step->t;
prev_cdf_distance = step->cdf_distance;
}
/* remap xi so we can reuse it */
xi = (xi - prev_cdf)/(step_cdf - prev_cdf);
/* pdf for picking step */
step_pdf_distance = step->cdf_distance - prev_cdf_distance;
}
/* determine range in which we will sample */
float step_t = step->t - prev_t;
/* sample distance and compute transmittance */
float3 distance_pdf;
sample_t = prev_t + kernel_volume_distance_sample(step_t, step->sigma_t, channel, xi, &transmittance, &distance_pdf);
/* modify pdf for hit/miss decision */
if(probalistic_scatter)
distance_pdf *= make_float3(1.0f, 1.0f, 1.0f) - segment->accum_transmittance;
pdf = average(distance_pdf * step_pdf_distance);
/* multiple importance sampling */
if(use_mis) {
float equi_pdf = kernel_volume_equiangular_pdf(ray, *light_P, sample_t);
mis_weight = 2.0f*power_heuristic(pdf, equi_pdf);
}
}
/* equi-angular sampling */
else {
/* sample distance */
sample_t = kernel_volume_equiangular_sample(ray, *light_P, xi, &pdf);
/* find step in which sampled distance is located */
step = segment->steps;
float prev_t = 0.0f;
float3 step_pdf_distance = make_float3(1.0f, 1.0f, 1.0f);
if(segment->numsteps > 1) {
float3 prev_cdf_distance = make_float3(0.0f, 0.0f, 0.0f);
int numsteps = segment->numsteps;
int high = numsteps - 1;
int low = 0;
int mid;
while(low < high) {
mid = (low + high) >> 1;
if(sample_t < step[mid].t)
high = mid;
else if(sample_t >= step[mid + 1].t)
low = mid + 1;
else {
/* found our interval in step[mid] .. step[mid+1] */
prev_t = step[mid].t;
prev_cdf_distance = step[mid].cdf_distance;
step += mid+1;
break;
}
}
if(low >= numsteps - 1) {
prev_t = step[numsteps - 1].t;
prev_cdf_distance = step[numsteps-1].cdf_distance;
step += numsteps - 1;
}
/* pdf for picking step with distance sampling */
step_pdf_distance = step->cdf_distance - prev_cdf_distance;
}
/* determine range in which we will sample */
float step_t = step->t - prev_t;
float step_sample_t = sample_t - prev_t;
/* compute transmittance */
transmittance = volume_color_transmittance(step->sigma_t, step_sample_t);
/* multiple importance sampling */
if(use_mis) {
float3 distance_pdf3 = kernel_volume_distance_pdf(step_t, step->sigma_t, step_sample_t);
float distance_pdf = average(distance_pdf3 * step_pdf_distance);
mis_weight = 2.0f*power_heuristic(pdf, distance_pdf);
}
}
if(sample_t < 0.0f || pdf == 0.0f) {
return VOLUME_PATH_MISSED;
}
/* compute transmittance up to this step */
if(step != segment->steps)
transmittance *= (step-1)->accum_transmittance;
/* modify throughput */
*throughput *= step->sigma_s * transmittance * (mis_weight / pdf);
/* evaluate shader to create closures at shading point */
if(segment->numsteps > 1) {
sd->P = ray->P + step->shade_t*ray->D;
VolumeShaderCoefficients coeff;
volume_shader_sample(kg, sd, state, sd->P, &coeff);
}
/* move to new position */
sd->P = ray->P + sample_t*ray->D;
return VOLUME_PATH_SCATTERED;
}
#endif /* __SPLIT_KERNEL */
/* decide if we need to use decoupled or not */
ccl_device bool kernel_volume_use_decoupled(KernelGlobals *kg, bool heterogeneous, bool direct, int sampling_method)
{
/* decoupled ray marching for heterogeneous volumes not supported on the GPU,
* which also means equiangular and multiple importance sampling is not
* support for that case */
#ifdef __KERNEL_GPU__
if(heterogeneous)
return false;
#endif
/* equiangular and multiple importance sampling only implemented for decoupled */
if(sampling_method != 0)
return true;
/* for all light sampling use decoupled, reusing shader evaluations is
* typically faster in that case */
if(direct)
return kernel_data.integrator.sample_all_lights_direct;
else
return kernel_data.integrator.sample_all_lights_indirect;
}
/* Volume Stack
*
* This is an array of object/shared ID's that the current segment of the path
* is inside of. */
ccl_device void kernel_volume_stack_init(KernelGlobals *kg,
ShaderData *stack_sd,
ccl_addr_space const PathState *state,
ccl_addr_space const Ray *ray,
ccl_addr_space VolumeStack *stack)
{
/* NULL ray happens in the baker, does it need proper initialization of
* camera in volume?
*/
if(!kernel_data.cam.is_inside_volume || ray == NULL) {
/* Camera is guaranteed to be in the air, only take background volume
* into account in this case.
*/
if(kernel_data.background.volume_shader != SHADER_NONE) {
stack[0].shader = kernel_data.background.volume_shader;
stack[0].object = PRIM_NONE;
stack[1].shader = SHADER_NONE;
}
else {
stack[0].shader = SHADER_NONE;
}
return;
}
kernel_assert(state->flag & PATH_RAY_CAMERA);
Ray volume_ray = *ray;
volume_ray.t = FLT_MAX;
const uint visibility = (state->flag & PATH_RAY_ALL_VISIBILITY);
int stack_index = 0, enclosed_index = 0;
#ifdef __VOLUME_RECORD_ALL__
Intersection hits[2*VOLUME_STACK_SIZE + 1];
uint num_hits = scene_intersect_volume_all(kg,
&volume_ray,
hits,
2*VOLUME_STACK_SIZE,
visibility);
if(num_hits > 0) {
int enclosed_volumes[VOLUME_STACK_SIZE];
Intersection *isect = hits;
qsort(hits, num_hits, sizeof(Intersection), intersections_compare);
for(uint hit = 0; hit < num_hits; ++hit, ++isect) {
shader_setup_from_ray(kg, stack_sd, isect, &volume_ray);
if(stack_sd->flag & SD_BACKFACING) {
bool need_add = true;
for(int i = 0; i < enclosed_index && need_add; ++i) {
/* If ray exited the volume and never entered to that volume
* it means that camera is inside such a volume.
*/
if(enclosed_volumes[i] == stack_sd->object) {
need_add = false;
}
}
for(int i = 0; i < stack_index && need_add; ++i) {
/* Don't add intersections twice. */
if(stack[i].object == stack_sd->object) {
need_add = false;
break;
}
}
if(need_add) {
stack[stack_index].object = stack_sd->object;
stack[stack_index].shader = stack_sd->shader;
++stack_index;
}
}
else {
/* If ray from camera enters the volume, this volume shouldn't
* be added to the stack on exit.
*/
enclosed_volumes[enclosed_index++] = stack_sd->object;
}
}
}
#else
int enclosed_volumes[VOLUME_STACK_SIZE];
int step = 0;
while(stack_index < VOLUME_STACK_SIZE - 1 &&
enclosed_index < VOLUME_STACK_SIZE - 1 &&
step < 2 * VOLUME_STACK_SIZE)
{
Intersection isect;
if(!scene_intersect_volume(kg, &volume_ray, &isect, visibility)) {
break;
}
shader_setup_from_ray(kg, stack_sd, &isect, &volume_ray);
if(stack_sd->flag & SD_BACKFACING) {
/* If ray exited the volume and never entered to that volume
* it means that camera is inside such a volume.
*/
bool need_add = true;
for(int i = 0; i < enclosed_index && need_add; ++i) {
/* If ray exited the volume and never entered to that volume
* it means that camera is inside such a volume.
*/
if(enclosed_volumes[i] == stack_sd->object) {
need_add = false;
}
}
for(int i = 0; i < stack_index && need_add; ++i) {
/* Don't add intersections twice. */
if(stack[i].object == stack_sd->object) {
need_add = false;
break;
}
}
if(need_add) {
stack[stack_index].object = stack_sd->object;
stack[stack_index].shader = stack_sd->shader;
++stack_index;
}
}
else {
/* If ray from camera enters the volume, this volume shouldn't
* be added to the stack on exit.
*/
enclosed_volumes[enclosed_index++] = stack_sd->object;
}
/* Move ray forward. */
volume_ray.P = ray_offset(stack_sd->P, -stack_sd->Ng);
++step;
}
#endif
/* stack_index of 0 means quick checks outside of the kernel gave false
* positive, nothing to worry about, just we've wasted quite a few of
* ticks just to come into conclusion that camera is in the air.
*
* In this case we're doing the same above -- check whether background has
* volume.
*/
if(stack_index == 0 && kernel_data.background.volume_shader == SHADER_NONE) {
stack[0].shader = kernel_data.background.volume_shader;
stack[0].object = PRIM_NONE;
stack[1].shader = SHADER_NONE;
}
else {
stack[stack_index].shader = SHADER_NONE;
}
}
ccl_device void kernel_volume_stack_enter_exit(KernelGlobals *kg, ShaderData *sd, ccl_addr_space VolumeStack *stack)
{
/* todo: we should have some way for objects to indicate if they want the
* world shader to work inside them. excluding it by default is problematic
* because non-volume objects can't be assumed to be closed manifolds */
if(!(sd->flag & SD_HAS_VOLUME))
return;
if(sd->flag & SD_BACKFACING) {
/* exit volume object: remove from stack */
for(int i = 0; stack[i].shader != SHADER_NONE; i++) {
if(stack[i].object == sd->object) {
/* shift back next stack entries */
do {
stack[i] = stack[i+1];
i++;
}
while(stack[i].shader != SHADER_NONE);
return;
}
}
}
else {
/* enter volume object: add to stack */
int i;
for(i = 0; stack[i].shader != SHADER_NONE; i++) {
/* already in the stack? then we have nothing to do */
if(stack[i].object == sd->object)
return;
}
/* if we exceed the stack limit, ignore */
if(i >= VOLUME_STACK_SIZE-1)
return;
/* add to the end of the stack */
stack[i].shader = sd->shader;
stack[i].object = sd->object;
stack[i+1].shader = SHADER_NONE;
}
}
#ifdef __SUBSURFACE__
ccl_device void kernel_volume_stack_update_for_subsurface(KernelGlobals *kg,
ShaderData *stack_sd,
Ray *ray,
ccl_addr_space VolumeStack *stack)
{
kernel_assert(kernel_data.integrator.use_volumes);
Ray volume_ray = *ray;
# ifdef __VOLUME_RECORD_ALL__
Intersection hits[2*VOLUME_STACK_SIZE + 1];
uint num_hits = scene_intersect_volume_all(kg,
&volume_ray,
hits,
2*VOLUME_STACK_SIZE,
PATH_RAY_ALL_VISIBILITY);
if(num_hits > 0) {
Intersection *isect = hits;
qsort(hits, num_hits, sizeof(Intersection), intersections_compare);
for(uint hit = 0; hit < num_hits; ++hit, ++isect) {
shader_setup_from_ray(kg, stack_sd, isect, &volume_ray);
kernel_volume_stack_enter_exit(kg, stack_sd, stack);
}
}
# else
Intersection isect;
int step = 0;
float3 Pend = ray->P + ray->D*ray->t;
while(step < 2 * VOLUME_STACK_SIZE &&
scene_intersect_volume(kg,
&volume_ray,
&isect,
PATH_RAY_ALL_VISIBILITY))
{
shader_setup_from_ray(kg, stack_sd, &isect, &volume_ray);
kernel_volume_stack_enter_exit(kg, stack_sd, stack);
/* Move ray forward. */
volume_ray.P = ray_offset(stack_sd->P, -stack_sd->Ng);
if(volume_ray.t != FLT_MAX) {
volume_ray.D = normalize_len(Pend - volume_ray.P, &volume_ray.t);
}
++step;
}
# endif
}
#endif
/* Clean stack after the last bounce.
*
* It is expected that all volumes are closed manifolds, so at the time when ray
* hits nothing (for example, it is a last bounce which goes to environment) the
* only expected volume in the stack is the world's one. All the rest volume
* entries should have been exited already.
*
* This isn't always true because of ray intersection precision issues, which
* could lead us to an infinite non-world volume in the stack, causing render
* artifacts.
*
* Use this function after the last bounce to get rid of all volumes apart from
* the world's one after the last bounce to avoid render artifacts.
*/
ccl_device_inline void kernel_volume_clean_stack(KernelGlobals *kg,
ccl_addr_space VolumeStack *volume_stack)
{
if(kernel_data.background.volume_shader != SHADER_NONE) {
/* Keep the world's volume in stack. */
volume_stack[1].shader = SHADER_NONE;
}
else {
volume_stack[0].shader = SHADER_NONE;
}
}
CCL_NAMESPACE_END
|