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
|
/*
* Copyright 2014, 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.
*/
/* Triangle/Ray intersections.
*
* For BVH ray intersection we use a precomputed triangle storage to accelerate
* intersection at the cost of more memory usage.
*/
CCL_NAMESPACE_BEGIN
ccl_device_inline bool triangle_intersect(KernelGlobals *kg,
Intersection *isect,
float3 P,
float3 dir,
uint visibility,
int object,
int prim_addr)
{
const uint tri_vindex = kernel_tex_fetch(__prim_tri_index, prim_addr);
#if defined(__KERNEL_SSE2__) && defined(__KERNEL_SSE__)
const ssef *ssef_verts = (ssef*)&kg->__prim_tri_verts.data[tri_vindex];
#else
const float4 tri_a = kernel_tex_fetch(__prim_tri_verts, tri_vindex+0),
tri_b = kernel_tex_fetch(__prim_tri_verts, tri_vindex+1),
tri_c = kernel_tex_fetch(__prim_tri_verts, tri_vindex+2);
#endif
float t, u, v;
if(ray_triangle_intersect(P,
dir,
isect->t,
#if defined(__KERNEL_SSE2__) && defined(__KERNEL_SSE__)
ssef_verts,
#else
float4_to_float3(tri_a),
float4_to_float3(tri_b),
float4_to_float3(tri_c),
#endif
&u, &v, &t))
{
#ifdef __VISIBILITY_FLAG__
/* Visibility flag test. we do it here under the assumption
* that most triangles are culled by node flags.
*/
if(kernel_tex_fetch(__prim_visibility, prim_addr) & visibility)
#endif
{
isect->prim = prim_addr;
isect->object = object;
isect->type = PRIMITIVE_TRIANGLE;
isect->u = u;
isect->v = v;
isect->t = t;
return true;
}
}
return false;
}
/* Special ray intersection routines for subsurface scattering. In that case we
* only want to intersect with primitives in the same object, and if case of
* multiple hits we pick a single random primitive as the intersection point.
*/
#ifdef __SUBSURFACE__
ccl_device_inline void triangle_intersect_subsurface(
KernelGlobals *kg,
SubsurfaceIntersection *ss_isect,
float3 P,
float3 dir,
int object,
int prim_addr,
float tmax,
uint *lcg_state,
int max_hits)
{
const uint tri_vindex = kernel_tex_fetch(__prim_tri_index, prim_addr);
#if defined(__KERNEL_SSE2__) && defined(__KERNEL_SSE__)
const ssef *ssef_verts = (ssef*)&kg->__prim_tri_verts.data[tri_vindex];
#else
const float3 tri_a = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+0)),
tri_b = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+1)),
tri_c = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+2));
#endif
float t, u, v;
if(!ray_triangle_intersect(P,
dir,
tmax,
#if defined(__KERNEL_SSE2__) && defined(__KERNEL_SSE__)
ssef_verts,
#else
tri_a, tri_b, tri_c,
#endif
&u, &v, &t))
{
return;
}
for(int i = min(max_hits, ss_isect->num_hits) - 1; i >= 0; --i) {
if(ss_isect->hits[i].t == t) {
return;
}
}
ss_isect->num_hits++;
int hit;
if(ss_isect->num_hits <= max_hits) {
hit = ss_isect->num_hits - 1;
}
else {
/* reservoir sampling: if we are at the maximum number of
* hits, randomly replace element or skip it */
hit = lcg_step_uint(lcg_state) % ss_isect->num_hits;
if(hit >= max_hits)
return;
}
/* record intersection */
Intersection *isect = &ss_isect->hits[hit];
isect->prim = prim_addr;
isect->object = object;
isect->type = PRIMITIVE_TRIANGLE;
isect->u = u;
isect->v = v;
isect->t = t;
/* Record geometric normal. */
#if defined(__KERNEL_SSE2__) && defined(__KERNEL_SSE__)
const float3 tri_a = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+0)),
tri_b = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+1)),
tri_c = float4_to_float3(kernel_tex_fetch(__prim_tri_verts, tri_vindex+2));
#endif
ss_isect->Ng[hit] = normalize(cross(tri_b - tri_a, tri_c - tri_a));
}
#endif /* __SUBSURFACE__ */
/* Refine triangle intersection to more precise hit point. For rays that travel
* far the precision is often not so good, this reintersects the primitive from
* a closer distance. */
/* Reintersections uses the paper:
*
* Tomas Moeller
* Fast, minimum storage ray/triangle intersection
* http://www.cs.virginia.edu/~gfx/Courses/2003/ImageSynthesis/papers/Acceleration/Fast%20MinimumStorage%20RayTriangle%20Intersection.pdf
*/
ccl_device_inline float3 triangle_refine(KernelGlobals *kg,
ShaderData *sd,
const Intersection *isect,
const Ray *ray)
{
float3 P = ray->P;
float3 D = ray->D;
float t = isect->t;
#ifdef __INTERSECTION_REFINE__
if(isect->object != OBJECT_NONE) {
if(UNLIKELY(t == 0.0f)) {
return P;
}
# ifdef __OBJECT_MOTION__
Transform tfm = sd->ob_itfm;
# else
Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_INVERSE_TRANSFORM);
# endif
P = transform_point(&tfm, P);
D = transform_direction(&tfm, D*t);
D = normalize_len(D, &t);
}
P = P + D*t;
const uint tri_vindex = kernel_tex_fetch(__prim_tri_index, isect->prim);
const float4 tri_a = kernel_tex_fetch(__prim_tri_verts, tri_vindex+0),
tri_b = kernel_tex_fetch(__prim_tri_verts, tri_vindex+1),
tri_c = kernel_tex_fetch(__prim_tri_verts, tri_vindex+2);
float3 edge1 = make_float3(tri_a.x - tri_c.x, tri_a.y - tri_c.y, tri_a.z - tri_c.z);
float3 edge2 = make_float3(tri_b.x - tri_c.x, tri_b.y - tri_c.y, tri_b.z - tri_c.z);
float3 tvec = make_float3(P.x - tri_c.x, P.y - tri_c.y, P.z - tri_c.z);
float3 qvec = cross(tvec, edge1);
float3 pvec = cross(D, edge2);
float det = dot(edge1, pvec);
if(det != 0.0f) {
/* If determinant is zero it means ray lies in the plane of
* the triangle. It is possible in theory due to watertight
* nature of triangle intersection. For such cases we simply
* don't refine intersection hoping it'll go all fine.
*/
float rt = dot(edge2, qvec) / det;
P = P + D*rt;
}
if(isect->object != OBJECT_NONE) {
# ifdef __OBJECT_MOTION__
Transform tfm = sd->ob_tfm;
# else
Transform tfm = object_fetch_transform(kg, isect->object, OBJECT_TRANSFORM);
# endif
P = transform_point(&tfm, P);
}
return P;
#else
return P + D*t;
#endif
}
/* Same as above, except that isect->t is assumed to be in object space for
* instancing.
*/
ccl_device_inline float3 triangle_refine_subsurface(KernelGlobals *kg,
ShaderData *sd,
const Intersection *isect,
const Ray *ray)
{
float3 P = ray->P;
float3 D = ray->D;
float t = isect->t;
if(isect->object != OBJECT_NONE) {
#ifdef __OBJECT_MOTION__
Transform tfm = sd->ob_itfm;
#else
Transform tfm = object_fetch_transform(kg,
isect->object,
OBJECT_INVERSE_TRANSFORM);
#endif
P = transform_point(&tfm, P);
D = transform_direction(&tfm, D);
D = normalize(D);
}
P = P + D*t;
#ifdef __INTERSECTION_REFINE__
const uint tri_vindex = kernel_tex_fetch(__prim_tri_index, isect->prim);
const float4 tri_a = kernel_tex_fetch(__prim_tri_verts, tri_vindex+0),
tri_b = kernel_tex_fetch(__prim_tri_verts, tri_vindex+1),
tri_c = kernel_tex_fetch(__prim_tri_verts, tri_vindex+2);
float3 edge1 = make_float3(tri_a.x - tri_c.x, tri_a.y - tri_c.y, tri_a.z - tri_c.z);
float3 edge2 = make_float3(tri_b.x - tri_c.x, tri_b.y - tri_c.y, tri_b.z - tri_c.z);
float3 tvec = make_float3(P.x - tri_c.x, P.y - tri_c.y, P.z - tri_c.z);
float3 qvec = cross(tvec, edge1);
float3 pvec = cross(D, edge2);
float det = dot(edge1, pvec);
if(det != 0.0f) {
/* If determinant is zero it means ray lies in the plane of
* the triangle. It is possible in theory due to watertight
* nature of triangle intersection. For such cases we simply
* don't refine intersection hoping it'll go all fine.
*/
float rt = dot(edge2, qvec) / det;
P = P + D*rt;
}
#endif /* __INTERSECTION_REFINE__ */
if(isect->object != OBJECT_NONE) {
#ifdef __OBJECT_MOTION__
Transform tfm = sd->ob_tfm;
#else
Transform tfm = object_fetch_transform(kg,
isect->object,
OBJECT_TRANSFORM);
#endif
P = transform_point(&tfm, P);
}
return P;
}
CCL_NAMESPACE_END
|
