
This is the first of a sequence of changes to support compiling Cycles kernels as MSL (Metal Shading Language) in preparation for a Metal GPU device implementation. MSL requires that all pointer types be declared with explicit address space attributes (device, thread, etc...). There is already precedent for this with Cycles' address space macros (ccl_global, ccl_private, etc...), therefore the first step of MSL-enablement is to apply these consistently. Line-for-line this represents the largest change required to enable MSL. Applying this change first will simplify future patches as well as offering the emergent benefit of enhanced descriptiveness. The vast majority of deltas in this patch fall into one of two cases: - Ensuring ccl_private is specified for thread-local pointer types - Ensuring ccl_global is specified for device-wide pointer types Additionally, the ccl_addr_space qualifier can be removed. Prior to Cycles X, ccl_addr_space was used as a context-dependent address space qualifier, but now it is either redundant (e.g. in struct typedefs), or can be replaced by ccl_global in the case of pointer types. Associated function variants (e.g. lcg_step_float_addrspace) are also redundant. In cases where address space qualifiers are chained with "const", this patch places the address space qualifier first. The rationale for this is that the choice of address space is likely to have the greater impact on runtime performance and overall architecture. The final part of this patch is the addition of a metal/compat.h header. This is partially complete and will be extended in future patches, paving the way for the full Metal implementation. Ref T92212 Reviewed By: brecht Maniphest Tasks: T92212 Differential Revision: https://developer.blender.org/D12864
247 lines
8.6 KiB
C
247 lines
8.6 KiB
C
/*
|
|
* Copyright 2011-2014 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.
|
|
*/
|
|
|
|
/*
|
|
* ASHIKHMIN SHIRLEY BSDF
|
|
*
|
|
* Implementation of
|
|
* Michael Ashikhmin and Peter Shirley: "An Anisotropic Phong BRDF Model" (2000)
|
|
*
|
|
* The Fresnel factor is missing to get a separable bsdf (intensity*color), as is
|
|
* the case with all other microfacet-based BSDF implementations in Cycles.
|
|
*
|
|
* Other than that, the implementation directly follows the paper.
|
|
*/
|
|
|
|
#pragma once
|
|
|
|
CCL_NAMESPACE_BEGIN
|
|
|
|
ccl_device int bsdf_ashikhmin_shirley_setup(ccl_private MicrofacetBsdf *bsdf)
|
|
{
|
|
bsdf->alpha_x = clamp(bsdf->alpha_x, 1e-4f, 1.0f);
|
|
bsdf->alpha_y = clamp(bsdf->alpha_y, 1e-4f, 1.0f);
|
|
|
|
bsdf->type = CLOSURE_BSDF_ASHIKHMIN_SHIRLEY_ID;
|
|
return SD_BSDF | SD_BSDF_HAS_EVAL;
|
|
}
|
|
|
|
ccl_device void bsdf_ashikhmin_shirley_blur(ccl_private ShaderClosure *sc, float roughness)
|
|
{
|
|
ccl_private MicrofacetBsdf *bsdf = (ccl_private MicrofacetBsdf *)sc;
|
|
|
|
bsdf->alpha_x = fmaxf(roughness, bsdf->alpha_x);
|
|
bsdf->alpha_y = fmaxf(roughness, bsdf->alpha_y);
|
|
}
|
|
|
|
ccl_device_inline float bsdf_ashikhmin_shirley_roughness_to_exponent(float roughness)
|
|
{
|
|
return 2.0f / (roughness * roughness) - 2.0f;
|
|
}
|
|
|
|
ccl_device_forceinline float3
|
|
bsdf_ashikhmin_shirley_eval_reflect(ccl_private const ShaderClosure *sc,
|
|
const float3 I,
|
|
const float3 omega_in,
|
|
ccl_private float *pdf)
|
|
{
|
|
ccl_private const MicrofacetBsdf *bsdf = (ccl_private const MicrofacetBsdf *)sc;
|
|
float3 N = bsdf->N;
|
|
|
|
float NdotI = dot(N, I); /* in Cycles/OSL convention I is omega_out */
|
|
float NdotO = dot(N, omega_in); /* and consequently we use for O omaga_in ;) */
|
|
|
|
float out = 0.0f;
|
|
|
|
if (fmaxf(bsdf->alpha_x, bsdf->alpha_y) <= 1e-4f)
|
|
return make_float3(0.0f, 0.0f, 0.0f);
|
|
|
|
if (NdotI > 0.0f && NdotO > 0.0f) {
|
|
NdotI = fmaxf(NdotI, 1e-6f);
|
|
NdotO = fmaxf(NdotO, 1e-6f);
|
|
float3 H = normalize(omega_in + I);
|
|
float HdotI = fmaxf(fabsf(dot(H, I)), 1e-6f);
|
|
float HdotN = fmaxf(dot(H, N), 1e-6f);
|
|
|
|
/* pump from original paper
|
|
* (first derivative disc., but cancels the HdotI in the pdf nicely) */
|
|
float pump = 1.0f / fmaxf(1e-6f, (HdotI * fmaxf(NdotO, NdotI)));
|
|
/* pump from d-brdf paper */
|
|
/*float pump = 1.0f / fmaxf(1e-4f, ((NdotO + NdotI) * (NdotO*NdotI))); */
|
|
|
|
float n_x = bsdf_ashikhmin_shirley_roughness_to_exponent(bsdf->alpha_x);
|
|
float n_y = bsdf_ashikhmin_shirley_roughness_to_exponent(bsdf->alpha_y);
|
|
|
|
if (n_x == n_y) {
|
|
/* isotropic */
|
|
float e = n_x;
|
|
float lobe = powf(HdotN, e);
|
|
float norm = (n_x + 1.0f) / (8.0f * M_PI_F);
|
|
|
|
out = NdotO * norm * lobe * pump;
|
|
/* this is p_h / 4(H.I) (conversion from 'wh measure' to 'wi measure', eq. 8 in paper). */
|
|
*pdf = norm * lobe / HdotI;
|
|
}
|
|
else {
|
|
/* anisotropic */
|
|
float3 X, Y;
|
|
make_orthonormals_tangent(N, bsdf->T, &X, &Y);
|
|
|
|
float HdotX = dot(H, X);
|
|
float HdotY = dot(H, Y);
|
|
float lobe;
|
|
if (HdotN < 1.0f) {
|
|
float e = (n_x * HdotX * HdotX + n_y * HdotY * HdotY) / (1.0f - HdotN * HdotN);
|
|
lobe = powf(HdotN, e);
|
|
}
|
|
else {
|
|
lobe = 1.0f;
|
|
}
|
|
float norm = sqrtf((n_x + 1.0f) * (n_y + 1.0f)) / (8.0f * M_PI_F);
|
|
|
|
out = NdotO * norm * lobe * pump;
|
|
*pdf = norm * lobe / HdotI;
|
|
}
|
|
}
|
|
|
|
return make_float3(out, out, out);
|
|
}
|
|
|
|
ccl_device float3 bsdf_ashikhmin_shirley_eval_transmit(ccl_private const ShaderClosure *sc,
|
|
const float3 I,
|
|
const float3 omega_in,
|
|
ccl_private float *pdf)
|
|
{
|
|
return make_float3(0.0f, 0.0f, 0.0f);
|
|
}
|
|
|
|
ccl_device_inline void bsdf_ashikhmin_shirley_sample_first_quadrant(float n_x,
|
|
float n_y,
|
|
float randu,
|
|
float randv,
|
|
ccl_private float *phi,
|
|
ccl_private float *cos_theta)
|
|
{
|
|
*phi = atanf(sqrtf((n_x + 1.0f) / (n_y + 1.0f)) * tanf(M_PI_2_F * randu));
|
|
float cos_phi = cosf(*phi);
|
|
float sin_phi = sinf(*phi);
|
|
*cos_theta = powf(randv, 1.0f / (n_x * cos_phi * cos_phi + n_y * sin_phi * sin_phi + 1.0f));
|
|
}
|
|
|
|
ccl_device int bsdf_ashikhmin_shirley_sample(ccl_private const ShaderClosure *sc,
|
|
float3 Ng,
|
|
float3 I,
|
|
float3 dIdx,
|
|
float3 dIdy,
|
|
float randu,
|
|
float randv,
|
|
ccl_private float3 *eval,
|
|
ccl_private float3 *omega_in,
|
|
ccl_private float3 *domega_in_dx,
|
|
ccl_private float3 *domega_in_dy,
|
|
ccl_private float *pdf)
|
|
{
|
|
ccl_private const MicrofacetBsdf *bsdf = (ccl_private const MicrofacetBsdf *)sc;
|
|
float3 N = bsdf->N;
|
|
int label = LABEL_REFLECT | LABEL_GLOSSY;
|
|
|
|
float NdotI = dot(N, I);
|
|
if (NdotI > 0.0f) {
|
|
|
|
float n_x = bsdf_ashikhmin_shirley_roughness_to_exponent(bsdf->alpha_x);
|
|
float n_y = bsdf_ashikhmin_shirley_roughness_to_exponent(bsdf->alpha_y);
|
|
|
|
/* get x,y basis on the surface for anisotropy */
|
|
float3 X, Y;
|
|
|
|
if (n_x == n_y)
|
|
make_orthonormals(N, &X, &Y);
|
|
else
|
|
make_orthonormals_tangent(N, bsdf->T, &X, &Y);
|
|
|
|
/* sample spherical coords for h in tangent space */
|
|
float phi;
|
|
float cos_theta;
|
|
if (n_x == n_y) {
|
|
/* isotropic sampling */
|
|
phi = M_2PI_F * randu;
|
|
cos_theta = powf(randv, 1.0f / (n_x + 1.0f));
|
|
}
|
|
else {
|
|
/* anisotropic sampling */
|
|
if (randu < 0.25f) { /* first quadrant */
|
|
float remapped_randu = 4.0f * randu;
|
|
bsdf_ashikhmin_shirley_sample_first_quadrant(
|
|
n_x, n_y, remapped_randu, randv, &phi, &cos_theta);
|
|
}
|
|
else if (randu < 0.5f) { /* second quadrant */
|
|
float remapped_randu = 4.0f * (.5f - randu);
|
|
bsdf_ashikhmin_shirley_sample_first_quadrant(
|
|
n_x, n_y, remapped_randu, randv, &phi, &cos_theta);
|
|
phi = M_PI_F - phi;
|
|
}
|
|
else if (randu < 0.75f) { /* third quadrant */
|
|
float remapped_randu = 4.0f * (randu - 0.5f);
|
|
bsdf_ashikhmin_shirley_sample_first_quadrant(
|
|
n_x, n_y, remapped_randu, randv, &phi, &cos_theta);
|
|
phi = M_PI_F + phi;
|
|
}
|
|
else { /* fourth quadrant */
|
|
float remapped_randu = 4.0f * (1.0f - randu);
|
|
bsdf_ashikhmin_shirley_sample_first_quadrant(
|
|
n_x, n_y, remapped_randu, randv, &phi, &cos_theta);
|
|
phi = 2.0f * M_PI_F - phi;
|
|
}
|
|
}
|
|
|
|
/* get half vector in tangent space */
|
|
float sin_theta = sqrtf(fmaxf(0.0f, 1.0f - cos_theta * cos_theta));
|
|
float cos_phi = cosf(phi);
|
|
float sin_phi = sinf(phi); /* no sqrt(1-cos^2) here b/c it causes artifacts */
|
|
float3 h = make_float3(sin_theta * cos_phi, sin_theta * sin_phi, cos_theta);
|
|
|
|
/* half vector to world space */
|
|
float3 H = h.x * X + h.y * Y + h.z * N;
|
|
float HdotI = dot(H, I);
|
|
if (HdotI < 0.0f)
|
|
H = -H;
|
|
|
|
/* reflect I on H to get omega_in */
|
|
*omega_in = -I + (2.0f * HdotI) * H;
|
|
|
|
if (fmaxf(bsdf->alpha_x, bsdf->alpha_y) <= 1e-4f) {
|
|
/* Some high number for MIS. */
|
|
*pdf = 1e6f;
|
|
*eval = make_float3(1e6f, 1e6f, 1e6f);
|
|
label = LABEL_REFLECT | LABEL_SINGULAR;
|
|
}
|
|
else {
|
|
/* leave the rest to eval_reflect */
|
|
*eval = bsdf_ashikhmin_shirley_eval_reflect(sc, I, *omega_in, pdf);
|
|
}
|
|
|
|
#ifdef __RAY_DIFFERENTIALS__
|
|
/* just do the reflection thing for now */
|
|
*domega_in_dx = (2.0f * dot(N, dIdx)) * N - dIdx;
|
|
*domega_in_dy = (2.0f * dot(N, dIdy)) * N - dIdy;
|
|
#endif
|
|
}
|
|
|
|
return label;
|
|
}
|
|
|
|
CCL_NAMESPACE_END
|