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
|
/*
* 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.
*/
#include "kernel/kernel_id_passes.h"
CCL_NAMESPACE_BEGIN
#ifdef __DENOISING_FEATURES__
ccl_device_inline void kernel_write_denoising_shadow(KernelGlobals *kg,
ccl_global float *buffer,
int sample,
float path_total,
float path_total_shaded)
{
if (kernel_data.film.pass_denoising_data == 0)
return;
buffer += (sample & 1) ? DENOISING_PASS_SHADOW_B : DENOISING_PASS_SHADOW_A;
path_total = ensure_finite(path_total);
path_total_shaded = ensure_finite(path_total_shaded);
kernel_write_pass_float(buffer, path_total);
kernel_write_pass_float(buffer + 1, path_total_shaded);
float value = path_total_shaded / max(path_total, 1e-7f);
kernel_write_pass_float(buffer + 2, value * value);
}
ccl_device_inline void kernel_update_denoising_features(KernelGlobals *kg,
ShaderData *sd,
ccl_addr_space PathState *state,
PathRadiance *L)
{
if (state->denoising_feature_weight == 0.0f) {
return;
}
L->denoising_depth += ensure_finite(state->denoising_feature_weight * sd->ray_length);
/* Skip implicitly transparent surfaces. */
if (sd->flag & SD_HAS_ONLY_VOLUME) {
return;
}
float3 normal = make_float3(0.0f, 0.0f, 0.0f);
float3 diffuse_albedo = make_float3(0.0f, 0.0f, 0.0f);
float3 specular_albedo = make_float3(0.0f, 0.0f, 0.0f);
float sum_weight = 0.0f, sum_nonspecular_weight = 0.0f;
for (int i = 0; i < sd->num_closure; i++) {
ShaderClosure *sc = &sd->closure[i];
if (!CLOSURE_IS_BSDF_OR_BSSRDF(sc->type))
continue;
/* All closures contribute to the normal feature, but only diffuse-like ones to the albedo. */
normal += sc->N * sc->sample_weight;
sum_weight += sc->sample_weight;
float3 closure_albedo = sc->weight;
/* Closures that include a Fresnel term typically have weights close to 1 even though their
* actual contribution is significantly lower.
* To account for this, we scale their weight by the average fresnel factor (the same is also
* done for the sample weight in the BSDF setup, so we don't need to scale that here). */
if (CLOSURE_IS_BSDF_MICROFACET_FRESNEL(sc->type)) {
MicrofacetBsdf *bsdf = (MicrofacetBsdf *)sc;
closure_albedo *= bsdf->extra->fresnel_color;
}
else if (sc->type == CLOSURE_BSDF_PRINCIPLED_SHEEN_ID) {
PrincipledSheenBsdf *bsdf = (PrincipledSheenBsdf *)sc;
closure_albedo *= bsdf->avg_value;
}
if (bsdf_get_specular_roughness_squared(sc) > sqr(0.075f)) {
diffuse_albedo += closure_albedo;
sum_nonspecular_weight += sc->sample_weight;
}
else {
specular_albedo += closure_albedo;
}
}
/* Wait for next bounce if 75% or more sample weight belongs to specular-like closures. */
if ((sum_weight == 0.0f) || (sum_nonspecular_weight * 4.0f > sum_weight)) {
if (sum_weight != 0.0f) {
normal /= sum_weight;
}
/* Transform normal into camera space. */
const Transform worldtocamera = kernel_data.cam.worldtocamera;
normal = transform_direction(&worldtocamera, normal);
L->denoising_normal += ensure_finite3(state->denoising_feature_weight * normal);
L->denoising_albedo += ensure_finite3(state->denoising_feature_weight *
state->denoising_feature_throughput * diffuse_albedo);
state->denoising_feature_weight = 0.0f;
}
else {
state->denoising_feature_throughput *= specular_albedo;
}
}
#endif /* __DENOISING_FEATURES__ */
#ifdef __KERNEL_DEBUG__
ccl_device_inline void kernel_write_debug_passes(KernelGlobals *kg,
ccl_global float *buffer,
PathRadiance *L)
{
int flag = kernel_data.film.pass_flag;
if (flag & PASSMASK(BVH_TRAVERSED_NODES)) {
kernel_write_pass_float(buffer + kernel_data.film.pass_bvh_traversed_nodes,
L->debug_data.num_bvh_traversed_nodes);
}
if (flag & PASSMASK(BVH_TRAVERSED_INSTANCES)) {
kernel_write_pass_float(buffer + kernel_data.film.pass_bvh_traversed_instances,
L->debug_data.num_bvh_traversed_instances);
}
if (flag & PASSMASK(BVH_INTERSECTIONS)) {
kernel_write_pass_float(buffer + kernel_data.film.pass_bvh_intersections,
L->debug_data.num_bvh_intersections);
}
if (flag & PASSMASK(RAY_BOUNCES)) {
kernel_write_pass_float(buffer + kernel_data.film.pass_ray_bounces,
L->debug_data.num_ray_bounces);
}
}
#endif /* __KERNEL_DEBUG__ */
#ifdef __KERNEL_CPU__
# define WRITE_ID_SLOT(buffer, depth, id, matte_weight, name) \
kernel_write_id_pass_cpu(buffer, depth * 2, id, matte_weight, kg->coverage_##name)
ccl_device_inline size_t kernel_write_id_pass_cpu(
float *buffer, size_t depth, float id, float matte_weight, CoverageMap *map)
{
if (map) {
(*map)[id] += matte_weight;
return 0;
}
#else /* __KERNEL_CPU__ */
# define WRITE_ID_SLOT(buffer, depth, id, matte_weight, name) \
kernel_write_id_slots_gpu(buffer, depth * 2, id, matte_weight)
ccl_device_inline size_t kernel_write_id_slots_gpu(ccl_global float *buffer,
size_t depth,
float id,
float matte_weight)
{
#endif /* __KERNEL_CPU__ */
kernel_write_id_slots(buffer, depth, id, matte_weight);
return depth * 2;
}
ccl_device_inline void kernel_write_data_passes(KernelGlobals *kg,
ccl_global float *buffer,
PathRadiance *L,
ShaderData *sd,
ccl_addr_space PathState *state,
float3 throughput)
{
#ifdef __PASSES__
int path_flag = state->flag;
if (!(path_flag & PATH_RAY_CAMERA))
return;
int flag = kernel_data.film.pass_flag;
int light_flag = kernel_data.film.light_pass_flag;
if (!((flag | light_flag) & PASS_ANY))
return;
if (!(path_flag & PATH_RAY_SINGLE_PASS_DONE)) {
if (!(sd->flag & SD_TRANSPARENT) || kernel_data.film.pass_alpha_threshold == 0.0f ||
average(shader_bsdf_alpha(kg, sd)) >= kernel_data.film.pass_alpha_threshold) {
if (state->sample == 0) {
if (flag & PASSMASK(DEPTH)) {
float depth = camera_distance(kg, sd->P);
kernel_write_pass_float(buffer + kernel_data.film.pass_depth, depth);
}
if (flag & PASSMASK(OBJECT_ID)) {
float id = object_pass_id(kg, sd->object);
kernel_write_pass_float(buffer + kernel_data.film.pass_object_id, id);
}
if (flag & PASSMASK(MATERIAL_ID)) {
float id = shader_pass_id(kg, sd);
kernel_write_pass_float(buffer + kernel_data.film.pass_material_id, id);
}
}
if (flag & PASSMASK(NORMAL)) {
float3 normal = shader_bsdf_average_normal(kg, sd);
kernel_write_pass_float3(buffer + kernel_data.film.pass_normal, normal);
}
if (flag & PASSMASK(UV)) {
float3 uv = primitive_uv(kg, sd);
kernel_write_pass_float3(buffer + kernel_data.film.pass_uv, uv);
}
if (flag & PASSMASK(MOTION)) {
float4 speed = primitive_motion_vector(kg, sd);
kernel_write_pass_float4(buffer + kernel_data.film.pass_motion, speed);
kernel_write_pass_float(buffer + kernel_data.film.pass_motion_weight, 1.0f);
}
state->flag |= PATH_RAY_SINGLE_PASS_DONE;
}
}
if (kernel_data.film.cryptomatte_passes) {
const float matte_weight = average(throughput) *
(1.0f - average(shader_bsdf_transparency(kg, sd)));
if (matte_weight > 0.0f) {
ccl_global float *cryptomatte_buffer = buffer + kernel_data.film.pass_cryptomatte;
if (kernel_data.film.cryptomatte_passes & CRYPT_OBJECT) {
float id = object_cryptomatte_id(kg, sd->object);
cryptomatte_buffer += WRITE_ID_SLOT(
cryptomatte_buffer, kernel_data.film.cryptomatte_depth, id, matte_weight, object);
}
if (kernel_data.film.cryptomatte_passes & CRYPT_MATERIAL) {
float id = shader_cryptomatte_id(kg, sd->shader);
cryptomatte_buffer += WRITE_ID_SLOT(
cryptomatte_buffer, kernel_data.film.cryptomatte_depth, id, matte_weight, material);
}
if (kernel_data.film.cryptomatte_passes & CRYPT_ASSET) {
float id = object_cryptomatte_asset_id(kg, sd->object);
cryptomatte_buffer += WRITE_ID_SLOT(
cryptomatte_buffer, kernel_data.film.cryptomatte_depth, id, matte_weight, asset);
}
}
}
if (light_flag & PASSMASK_COMPONENT(DIFFUSE))
L->color_diffuse += shader_bsdf_diffuse(kg, sd) * throughput;
if (light_flag & PASSMASK_COMPONENT(GLOSSY))
L->color_glossy += shader_bsdf_glossy(kg, sd) * throughput;
if (light_flag & PASSMASK_COMPONENT(TRANSMISSION))
L->color_transmission += shader_bsdf_transmission(kg, sd) * throughput;
if (light_flag & PASSMASK_COMPONENT(SUBSURFACE))
L->color_subsurface += shader_bsdf_subsurface(kg, sd) * throughput;
if (light_flag & PASSMASK(MIST)) {
/* bring depth into 0..1 range */
float mist_start = kernel_data.film.mist_start;
float mist_inv_depth = kernel_data.film.mist_inv_depth;
float depth = camera_distance(kg, sd->P);
float mist = saturate((depth - mist_start) * mist_inv_depth);
/* falloff */
float mist_falloff = kernel_data.film.mist_falloff;
if (mist_falloff == 1.0f)
;
else if (mist_falloff == 2.0f)
mist = mist * mist;
else if (mist_falloff == 0.5f)
mist = sqrtf(mist);
else
mist = powf(mist, mist_falloff);
/* modulate by transparency */
float3 alpha = shader_bsdf_alpha(kg, sd);
L->mist += (1.0f - mist) * average(throughput * alpha);
}
#endif
}
ccl_device_inline void kernel_write_light_passes(KernelGlobals *kg,
ccl_global float *buffer,
PathRadiance *L)
{
#ifdef __PASSES__
int light_flag = kernel_data.film.light_pass_flag;
if (!kernel_data.film.use_light_pass)
return;
if (light_flag & PASSMASK(DIFFUSE_INDIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_diffuse_indirect, L->indirect_diffuse);
if (light_flag & PASSMASK(GLOSSY_INDIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_glossy_indirect, L->indirect_glossy);
if (light_flag & PASSMASK(TRANSMISSION_INDIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_transmission_indirect,
L->indirect_transmission);
if (light_flag & PASSMASK(SUBSURFACE_INDIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_subsurface_indirect,
L->indirect_subsurface);
if (light_flag & PASSMASK(VOLUME_INDIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_volume_indirect, L->indirect_scatter);
if (light_flag & PASSMASK(DIFFUSE_DIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_diffuse_direct, L->direct_diffuse);
if (light_flag & PASSMASK(GLOSSY_DIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_glossy_direct, L->direct_glossy);
if (light_flag & PASSMASK(TRANSMISSION_DIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_transmission_direct,
L->direct_transmission);
if (light_flag & PASSMASK(SUBSURFACE_DIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_subsurface_direct,
L->direct_subsurface);
if (light_flag & PASSMASK(VOLUME_DIRECT))
kernel_write_pass_float3(buffer + kernel_data.film.pass_volume_direct, L->direct_scatter);
if (light_flag & PASSMASK(EMISSION))
kernel_write_pass_float3(buffer + kernel_data.film.pass_emission, L->emission);
if (light_flag & PASSMASK(BACKGROUND))
kernel_write_pass_float3(buffer + kernel_data.film.pass_background, L->background);
if (light_flag & PASSMASK(AO))
kernel_write_pass_float3(buffer + kernel_data.film.pass_ao, L->ao);
if (light_flag & PASSMASK(DIFFUSE_COLOR))
kernel_write_pass_float3(buffer + kernel_data.film.pass_diffuse_color, L->color_diffuse);
if (light_flag & PASSMASK(GLOSSY_COLOR))
kernel_write_pass_float3(buffer + kernel_data.film.pass_glossy_color, L->color_glossy);
if (light_flag & PASSMASK(TRANSMISSION_COLOR))
kernel_write_pass_float3(buffer + kernel_data.film.pass_transmission_color,
L->color_transmission);
if (light_flag & PASSMASK(SUBSURFACE_COLOR))
kernel_write_pass_float3(buffer + kernel_data.film.pass_subsurface_color, L->color_subsurface);
if (light_flag & PASSMASK(SHADOW)) {
float4 shadow = L->shadow;
shadow.w = kernel_data.film.pass_shadow_scale;
kernel_write_pass_float4(buffer + kernel_data.film.pass_shadow, shadow);
}
if (light_flag & PASSMASK(MIST))
kernel_write_pass_float(buffer + kernel_data.film.pass_mist, 1.0f - L->mist);
#endif
}
ccl_device_inline void kernel_write_result(KernelGlobals *kg,
ccl_global float *buffer,
int sample,
PathRadiance *L)
{
PROFILING_INIT(kg, PROFILING_WRITE_RESULT);
PROFILING_OBJECT(PRIM_NONE);
float alpha;
float3 L_sum = path_radiance_clamp_and_sum(kg, L, &alpha);
if (kernel_data.film.pass_flag & PASSMASK(COMBINED)) {
kernel_write_pass_float4(buffer, make_float4(L_sum.x, L_sum.y, L_sum.z, alpha));
}
kernel_write_light_passes(kg, buffer, L);
#ifdef __DENOISING_FEATURES__
if (kernel_data.film.pass_denoising_data) {
# ifdef __SHADOW_TRICKS__
kernel_write_denoising_shadow(kg,
buffer + kernel_data.film.pass_denoising_data,
sample,
average(L->path_total),
average(L->path_total_shaded));
# else
kernel_write_denoising_shadow(
kg, buffer + kernel_data.film.pass_denoising_data, sample, 0.0f, 0.0f);
# endif
if (kernel_data.film.pass_denoising_clean) {
float3 noisy, clean;
path_radiance_split_denoising(kg, L, &noisy, &clean);
kernel_write_pass_float3_variance(
buffer + kernel_data.film.pass_denoising_data + DENOISING_PASS_COLOR, noisy);
kernel_write_pass_float3_unaligned(buffer + kernel_data.film.pass_denoising_clean, clean);
}
else {
kernel_write_pass_float3_variance(buffer + kernel_data.film.pass_denoising_data +
DENOISING_PASS_COLOR,
ensure_finite3(L_sum));
}
kernel_write_pass_float3_variance(buffer + kernel_data.film.pass_denoising_data +
DENOISING_PASS_NORMAL,
L->denoising_normal);
kernel_write_pass_float3_variance(buffer + kernel_data.film.pass_denoising_data +
DENOISING_PASS_ALBEDO,
L->denoising_albedo);
kernel_write_pass_float_variance(
buffer + kernel_data.film.pass_denoising_data + DENOISING_PASS_DEPTH, L->denoising_depth);
}
#endif /* __DENOISING_FEATURES__ */
#ifdef __KERNEL_DEBUG__
kernel_write_debug_passes(kg, buffer, L);
#endif
}
CCL_NAMESPACE_END
|