surface_shader.h

Go to the documentation of this file.

/* SPDX-FileCopyrightText: 2011-2022 Blender Foundation
 *
 * SPDX-License-Identifier: Apache-2.0 */
 
/* Functions to evaluate shaders. */
 
#pragma once
 
#include "kernel/closure/bsdf.h"
#include "kernel/closure/emissive.h"
 
#include "kernel/film/light_passes.h"
 
#include "kernel/integrator/guiding.h"
 
#ifdef __SVM__
#  include "kernel/svm/svm.h"
#endif
#ifdef __OSL__
#  include "kernel/osl/osl.h"
#endif
 
CCL_NAMESPACE_BEGIN
 
/* Guiding */
 
#ifdef __PATH_GUIDING__
 
ccl_device float surface_shader_average_sample_weight_squared_roughness(
    const ccl_private ShaderData *sd)
{
  float avg_squared_roughness = 0.0f;
  float sum_sample_weight = 0.0f;
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (!CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
      continue;
    }
    avg_squared_roughness += sc->sample_weight * bsdf_get_specular_roughness_squared(sc);
    sum_sample_weight += sc->sample_weight;
  }
 
  avg_squared_roughness = avg_squared_roughness > 0.0f ?
                              avg_squared_roughness / sum_sample_weight :
                              0.0f;
  return avg_squared_roughness;
}
 
ccl_device_inline void surface_shader_prepare_guiding(KernelGlobals kg,
                                                      IntegratorState state,
                                                      ccl_private ShaderData *sd,
                                                      const ccl_private RNGState *rng_state)
{
  /* Have any BSDF to guide? */
  if (!(kernel_data.integrator.use_surface_guiding && (sd->flag & SD_BSDF_HAS_EVAL))) {
    state->guiding.use_surface_guiding = false;
    return;
  }
 
  const float surface_guiding_probability = kernel_data.integrator.surface_guiding_probability;
  const int guiding_directional_sampling_type =
      kernel_data.integrator.guiding_directional_sampling_type;
  const float guiding_roughness_threshold = kernel_data.integrator.guiding_roughness_threshold;
  float rand_bsdf_guiding = path_state_rng_1D(kg, rng_state, PRNG_SURFACE_BSDF_GUIDING);
 
  /* Compute proportion of diffuse BSDF and BSSRDFs. */
  float diffuse_sampling_fraction = 0.0f;
  float bssrdf_sampling_fraction = 0.0f;
  float bsdf_bssrdf_sampling_sum = 0.0f;
 
  bool fully_opaque = true;
 
  for (int i = 0; i < sd->num_closure; i++) {
    ShaderClosure *sc = &sd->closure[i];
    if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
      const float sweight = sc->sample_weight;
      kernel_assert(sweight >= 0.0f);
 
      bsdf_bssrdf_sampling_sum += sweight;
      if (CLOSURE_IS_BSDF_DIFFUSE(sc->type) && sc->type < CLOSURE_BSDF_TRANSLUCENT_ID) {
        diffuse_sampling_fraction += sweight;
      }
      if (CLOSURE_IS_BSSRDF(sc->type)) {
        bssrdf_sampling_fraction += sweight;
      }
 
      if (CLOSURE_IS_BSDF_TRANSPARENT(sc->type) || CLOSURE_IS_BSDF_TRANSMISSION(sc->type)) {
        fully_opaque = false;
      }
    }
  }
 
  if (bsdf_bssrdf_sampling_sum > 0.0f) {
    diffuse_sampling_fraction /= bsdf_bssrdf_sampling_sum;
    bssrdf_sampling_fraction /= bsdf_bssrdf_sampling_sum;
  }
 
  /* Initial guiding */
  /* The roughness because the function returns `alpha.x * alpha.y`.
   * In addition alpha is squared again. */
  float avg_roughness = surface_shader_average_sample_weight_squared_roughness(sd);
  avg_roughness = safe_sqrtf(avg_roughness);
  if (!fully_opaque || avg_roughness < guiding_roughness_threshold ||
      ((guiding_directional_sampling_type == GUIDING_DIRECTIONAL_SAMPLING_TYPE_PRODUCT_MIS) &&
       (diffuse_sampling_fraction <= 0.0f)) ||
      !guiding_bsdf_init(kg, state, sd->P, sd->N, rand_bsdf_guiding))
  {
    state->guiding.use_surface_guiding = false;
    state->guiding.surface_guiding_sampling_prob = 0.0f;
    return;
  }
 
  state->guiding.use_surface_guiding = true;
  if (kernel_data.integrator.guiding_directional_sampling_type ==
      GUIDING_DIRECTIONAL_SAMPLING_TYPE_PRODUCT_MIS)
  {
    state->guiding.surface_guiding_sampling_prob = surface_guiding_probability *
                                                   diffuse_sampling_fraction;
  }
  else if (kernel_data.integrator.guiding_directional_sampling_type ==
           GUIDING_DIRECTIONAL_SAMPLING_TYPE_RIS)
  {
    state->guiding.surface_guiding_sampling_prob = surface_guiding_probability;
  }
  else {  // GUIDING_DIRECTIONAL_SAMPLING_TYPE_ROUGHNESS
    state->guiding.surface_guiding_sampling_prob = surface_guiding_probability * avg_roughness;
  }
  state->guiding.bssrdf_sampling_prob = bssrdf_sampling_fraction;
  state->guiding.sample_surface_guiding_rand = rand_bsdf_guiding;
 
  kernel_assert(state->guiding.surface_guiding_sampling_prob > 0.0f &&
                state->guiding.surface_guiding_sampling_prob <= 1.0f);
}
#endif
 
ccl_device_inline void surface_shader_prepare_closures(KernelGlobals kg,
                                                       ConstIntegratorState state,
                                                       ccl_private ShaderData *sd,
                                                       const uint32_t path_flag)
{
  /* Filter out closures. */
  if (kernel_data.integrator.filter_closures) {
    const int filter_closures = kernel_data.integrator.filter_closures;
    if (filter_closures & FILTER_CLOSURE_EMISSION) {
      sd->closure_emission_background = zero_spectrum();
    }
 
    if (path_flag & PATH_RAY_CAMERA) {
      if (filter_closures & FILTER_CLOSURE_DIRECT_LIGHT) {
        sd->flag &= ~SD_BSDF_HAS_EVAL;
      }
 
      for (int i = 0; i < sd->num_closure; i++) {
        ccl_private ShaderClosure *sc = &sd->closure[i];
 
        const bool filter_diffuse = (filter_closures & FILTER_CLOSURE_DIFFUSE);
        const bool filter_glossy = (filter_closures & FILTER_CLOSURE_GLOSSY);
        const bool filter_transmission = (filter_closures & FILTER_CLOSURE_TRANSMISSION);
        const bool filter_glass = filter_glossy && filter_transmission;
        if ((CLOSURE_IS_BSDF_DIFFUSE(sc->type) && filter_diffuse) ||
            (CLOSURE_IS_BSDF_GLOSSY(sc->type) && filter_glossy) ||
            (CLOSURE_IS_BSDF_TRANSMISSION(sc->type) && filter_transmission) ||
            (CLOSURE_IS_GLASS(sc->type) && filter_glass))
        {
          sc->type = CLOSURE_NONE_ID;
          sc->sample_weight = 0.0f;
        }
        else if ((CLOSURE_IS_BSDF_TRANSPARENT(sc->type) &&
                  (filter_closures & FILTER_CLOSURE_TRANSPARENT)))
        {
          sc->type = CLOSURE_HOLDOUT_ID;
          sc->sample_weight = 0.0f;
          sd->flag |= SD_HOLDOUT;
        }
      }
    }
  }
 
  /* Filter glossy.
   *
   * Blurring of bsdf after bounces, for rays that have a small likelihood
   * of following this particular path (diffuse, rough glossy) */
  if (kernel_data.integrator.filter_glossy != FLT_MAX
#ifdef __MNEE__
      && !(INTEGRATOR_STATE(state, path, mnee) & PATH_MNEE_VALID)
#endif
  )
  {
    const float blur_pdf = kernel_data.integrator.filter_glossy *
                           INTEGRATOR_STATE(state, path, min_ray_pdf);
 
    if (blur_pdf < 1.0f) {
      const float blur_roughness = sqrtf(1.0f - blur_pdf) * 0.5f;
 
      for (int i = 0; i < sd->num_closure; i++) {
        ccl_private ShaderClosure *sc = &sd->closure[i];
        if (CLOSURE_IS_BSDF(sc->type)) {
          bsdf_blur(kg, sc, blur_roughness);
        }
      }
 
      /* NOTE: this is a sufficient condition. If `blur_roughness < THRESH < original_roughness`
       * then the flag was already set. */
      if (sqr(blur_roughness) > BSDF_ROUGHNESS_SQ_THRESH) {
        sd->flag |= SD_BSDF_HAS_EVAL;
      }
    }
  }
}
 
/* BSDF */
#ifdef WITH_CYCLES_DEBUG
ccl_device_inline void surface_shader_validate_bsdf_sample(const KernelGlobals kg,
                                                           const ccl_private ShaderClosure *sc,
                                                           const float3 wo,
                                                           const int org_label,
                                                           const float2 org_roughness,
                                                           const float org_eta)
{
  /* Validate the #bsdf_label and #bsdf_roughness_eta functions
   * by estimating the values after a BSDF sample. */
  kernel_assert(org_label == bsdf_label(kg, sc, wo));
 
  float2 comp_roughness;
  float comp_eta;
  bsdf_roughness_eta(kg, sc, wo, &comp_roughness, &comp_eta);
  kernel_assert(org_eta == comp_eta);
  kernel_assert(org_roughness.x == comp_roughness.x);
  kernel_assert(org_roughness.y == comp_roughness.y);
}
#endif
 
ccl_device_forceinline bool _surface_shader_exclude(ClosureType type,
                                                    const uint light_shader_flags)
{
  if (!(light_shader_flags & SHADER_EXCLUDE_ANY)) {
    return false;
  }
  if (light_shader_flags & SHADER_EXCLUDE_DIFFUSE) {
    if (CLOSURE_IS_BSDF_DIFFUSE(type)) {
      return true;
    }
  }
  if (light_shader_flags & SHADER_EXCLUDE_GLOSSY) {
    if (CLOSURE_IS_BSDF_GLOSSY(type)) {
      return true;
    }
  }
  if (light_shader_flags & SHADER_EXCLUDE_TRANSMIT) {
    if (CLOSURE_IS_BSDF_TRANSMISSION(type)) {
      return true;
    }
  }
  /* Glass closures are both glossy and transmissive, so only exclude them if both are filtered. */
  const uint exclude_glass = SHADER_EXCLUDE_TRANSMIT | SHADER_EXCLUDE_GLOSSY;
  if ((light_shader_flags & exclude_glass) == exclude_glass) {
    if (CLOSURE_IS_GLASS(type)) {
      return true;
    }
  }
  return false;
}
 
ccl_device_inline float _surface_shader_bsdf_eval_mis(KernelGlobals kg,
                                                      ccl_private ShaderData *sd,
                                                      const float3 wo,
                                                      const ccl_private ShaderClosure *skip_sc,
                                                      ccl_private BsdfEval *result_eval,
                                                      float sum_pdf,
                                                      float sum_sample_weight,
                                                      const uint light_shader_flags)
{
  /* This is the veach one-sample model with balance heuristic,
   * some PDF factors drop out when using balance heuristic weighting. */
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (sc == skip_sc) {
      continue;
    }
 
    if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
      if (CLOSURE_IS_BSDF(sc->type) && !_surface_shader_exclude(sc->type, light_shader_flags)) {
        float bsdf_pdf = 0.0f;
        const Spectrum eval = bsdf_eval(kg, sd, sc, wo, &bsdf_pdf);
 
        if (bsdf_pdf != 0.0f) {
          bsdf_eval_accum(result_eval, sc, wo, eval * sc->weight);
          sum_pdf += bsdf_pdf * sc->sample_weight;
        }
      }
 
      sum_sample_weight += sc->sample_weight;
    }
  }
 
  return (sum_sample_weight > 0.0f) ? sum_pdf / sum_sample_weight : 0.0f;
}
 
ccl_device_inline float surface_shader_bsdf_eval_pdfs(const KernelGlobals kg,
                                                      ccl_private ShaderData *sd,
                                                      const float3 wo,
                                                      ccl_private BsdfEval *result_eval,
                                                      ccl_private float *pdfs,
                                                      const uint light_shader_flags)
{
  /* This is the veach one-sample model with balance heuristic, some pdf
   * factors drop out when using balance heuristic weighting. */
  float sum_pdf = 0.0f;
  float sum_sample_weight = 0.0f;
  bsdf_eval_init(result_eval, zero_spectrum());
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
      if (CLOSURE_IS_BSDF(sc->type) && !_surface_shader_exclude(sc->type, light_shader_flags)) {
        float bsdf_pdf = 0.0f;
        const Spectrum eval = bsdf_eval(kg, sd, sc, wo, &bsdf_pdf);
        kernel_assert(bsdf_pdf >= 0.0f);
        if (bsdf_pdf != 0.0f) {
          bsdf_eval_accum(result_eval, sc, wo, eval * sc->weight);
          sum_pdf += bsdf_pdf * sc->sample_weight;
          kernel_assert(bsdf_pdf * sc->sample_weight >= 0.0f);
          pdfs[i] = bsdf_pdf * sc->sample_weight;
        }
        else {
          pdfs[i] = 0.0f;
        }
      }
      else {
        pdfs[i] = 0.0f;
      }
 
      sum_sample_weight += sc->sample_weight;
    }
    else {
      pdfs[i] = 0.0f;
    }
  }
  if (sum_pdf > 0.0f) {
    for (int i = 0; i < sd->num_closure; i++) {
      pdfs[i] /= sum_pdf;
    }
  }
 
  return (sum_sample_weight > 0.0f) ? sum_pdf / sum_sample_weight : 0.0f;
}
 
#ifndef __KERNEL_CUDA__
ccl_device
#else
ccl_device_inline
#endif
    float
    surface_shader_bsdf_eval(KernelGlobals kg,
                             IntegratorState state,
                             ccl_private ShaderData *sd,
                             const float3 wo,
                             ccl_private BsdfEval *bsdf_eval,
                             const uint light_shader_flags)
{
  bsdf_eval_init(bsdf_eval, zero_spectrum());
 
  float pdf = _surface_shader_bsdf_eval_mis(
      kg, sd, wo, nullptr, bsdf_eval, 0.0f, 0.0f, light_shader_flags);
 
  /* If the light does not use MIS, then it is only sampled via NEE, so the probability of hitting
   * the light using BSDF sampling is zero. */
  if (!(light_shader_flags & SHADER_USE_MIS)) {
    pdf = 0.0f;
  }
 
#if defined(__PATH_GUIDING__) && PATH_GUIDING_LEVEL >= 4
  if (pdf > 0.0f && state->guiding.use_surface_guiding) {
    const float guiding_sampling_prob = state->guiding.surface_guiding_sampling_prob;
    const float bssrdf_sampling_prob = state->guiding.bssrdf_sampling_prob;
    const float guide_pdf = guiding_bsdf_pdf(kg, state, wo);
 
    if (kernel_data.integrator.guiding_directional_sampling_type ==
        GUIDING_DIRECTIONAL_SAMPLING_TYPE_RIS)
    {
      pdf = (0.5f * guide_pdf * (1.0f - bssrdf_sampling_prob)) + 0.5f * pdf;
    }
    else {
      pdf = (guiding_sampling_prob * guide_pdf * (1.0f - bssrdf_sampling_prob)) +
            (1.0f - guiding_sampling_prob) * pdf;
    }
  }
#endif
 
  return pdf;
}
 
/* Randomly sample a BSSRDF or BSDF proportional to ShaderClosure.sample_weight. */
const ccl_device_inline ccl_private ShaderClosure *surface_shader_bsdf_bssrdf_pick(
    const ccl_private ShaderData *ccl_restrict sd, ccl_private float3 *rand_bsdf)
{
  int sampled = 0;
 
  if (sd->num_closure > 1) {
    /* Pick a BSDF or based on sample weights. */
    float sum = 0.0f;
 
    for (int i = 0; i < sd->num_closure; i++) {
      const ccl_private ShaderClosure *sc = &sd->closure[i];
 
      if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
        sum += sc->sample_weight;
      }
    }
 
    const float r = (*rand_bsdf).z * sum;
    float partial_sum = 0.0f;
 
    for (int i = 0; i < sd->num_closure; i++) {
      const ccl_private ShaderClosure *sc = &sd->closure[i];
 
      if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
        const float next_sum = partial_sum + sc->sample_weight;
 
        if (r < next_sum) {
          sampled = i;
 
          /* Rescale to reuse for direction sample, to better preserve stratification. */
          (*rand_bsdf).z = (r - partial_sum) / sc->sample_weight;
          break;
        }
 
        partial_sum = next_sum;
      }
    }
  }
 
  return &sd->closure[sampled];
}
 
/* Return weight for picked BSSRDF. */
ccl_device_inline Spectrum
surface_shader_bssrdf_sample_weight(const ccl_private ShaderData *ccl_restrict sd,
                                    const ccl_private ShaderClosure *ccl_restrict bssrdf_sc)
{
  Spectrum weight = bssrdf_sc->weight;
 
  if (sd->num_closure > 1) {
    float sum = 0.0f;
    for (int i = 0; i < sd->num_closure; i++) {
      const ccl_private ShaderClosure *sc = &sd->closure[i];
 
      if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
        sum += sc->sample_weight;
      }
    }
    weight *= sum / bssrdf_sc->sample_weight;
  }
 
  return weight;
}
 
#ifdef __PATH_GUIDING__
/* Sample direction for picked BSDF, and return evaluation and pdf for all
 * BSDFs combined using MIS. */
 
ccl_device int surface_shader_bsdf_guided_sample_closure_mis(KernelGlobals kg,
                                                             IntegratorState state,
                                                             ccl_private ShaderData *sd,
                                                             const ccl_private ShaderClosure *sc,
                                                             const float3 rand_bsdf,
                                                             ccl_private BsdfEval *bsdf_eval,
                                                             ccl_private float3 *wo,
                                                             ccl_private float *bsdf_pdf,
                                                             ccl_private float *unguided_bsdf_pdf,
                                                             ccl_private float2 *sampled_roughness,
                                                             ccl_private float *eta)
{
  /* BSSRDF should already have been handled elsewhere. */
  kernel_assert(CLOSURE_IS_BSDF(sc->type));
 
  const bool use_surface_guiding = state->guiding.use_surface_guiding;
  const float guiding_sampling_prob = state->guiding.surface_guiding_sampling_prob;
  const float bssrdf_sampling_prob = state->guiding.bssrdf_sampling_prob;
 
  /* Decide between sampling guiding distribution and BSDF. */
  bool sample_guiding = false;
  float rand_bsdf_guiding = state->guiding.sample_surface_guiding_rand;
 
  if (use_surface_guiding && rand_bsdf_guiding < guiding_sampling_prob) {
    sample_guiding = true;
    rand_bsdf_guiding /= guiding_sampling_prob;
  }
  else {
    rand_bsdf_guiding -= guiding_sampling_prob;
    rand_bsdf_guiding /= (1.0f - guiding_sampling_prob);
  }
 
  /* Initialize to zero. */
  int label = LABEL_NONE;
  Spectrum eval = zero_spectrum();
  bsdf_eval_init(bsdf_eval, eval);
 
  *unguided_bsdf_pdf = 0.0f;
  float guide_pdf = 0.0f;
 
  if (sample_guiding) {
    /* Sample guiding distribution. */
    guide_pdf = guiding_bsdf_sample(kg, state, make_float2(rand_bsdf), wo);
    *bsdf_pdf = 0.0f;
 
    if (guide_pdf != 0.0f) {
      float unguided_bsdf_pdfs[MAX_CLOSURE];
 
      *unguided_bsdf_pdf = surface_shader_bsdf_eval_pdfs(
          kg, sd, *wo, bsdf_eval, unguided_bsdf_pdfs, 0);
      *bsdf_pdf = (guiding_sampling_prob * guide_pdf * (1.0f - bssrdf_sampling_prob)) +
                  ((1.0f - guiding_sampling_prob) * (*unguided_bsdf_pdf));
      float sum_pdfs = 0.0f;
 
      if (*unguided_bsdf_pdf > 0.0f) {
        int idx = -1;
        for (int i = 0; i < sd->num_closure; i++) {
          sum_pdfs += unguided_bsdf_pdfs[i];
          if (rand_bsdf_guiding <= sum_pdfs) {
            idx = i;
            break;
          }
        }
 
        kernel_assert(idx >= 0);
        /* Set the default idx to the last in the list.
         * in case of numerical problems and rand_bsdf_guiding is just >=1.0f and
         * the sum of all unguided_bsdf_pdfs is just < 1.0f. */
        idx = (rand_bsdf_guiding > sum_pdfs) ? sd->num_closure - 1 : idx;
 
        label = bsdf_label(kg, &sd->closure[idx], *wo);
      }
      else {
        *bsdf_pdf = 0.0f;
        *unguided_bsdf_pdf = 0.0f;
      }
    }
 
    kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
 
    *sampled_roughness = make_float2(1.0f, 1.0f);
    *eta = 1.0f;
  }
  else {
    /* Sample BSDF. */
    *bsdf_pdf = 0.0f;
    label = bsdf_sample(kg,
                        sd,
                        sc,
                        INTEGRATOR_STATE(state, path, flag),
                        rand_bsdf,
                        &eval,
                        wo,
                        unguided_bsdf_pdf,
                        sampled_roughness,
                        eta);
#  ifdef WITH_CYCLES_DEBUG
    /* Code path to validate the estimation of the label, sampled roughness and eta. This should be
     * activated from time to time when the BSDFs change to check if everything is still working
     * correctly. */
    if (*unguided_bsdf_pdf > 0.0f) {
      surface_shader_validate_bsdf_sample(kg, sc, *wo, label, *sampled_roughness, *eta);
    }
#  endif
 
    if (*unguided_bsdf_pdf != 0.0f) {
      bsdf_eval_init(bsdf_eval, sc, *wo, eval * sc->weight);
 
      kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
 
      if (sd->num_closure > 1) {
        const float sweight = sc->sample_weight;
        *unguided_bsdf_pdf = _surface_shader_bsdf_eval_mis(
            kg, sd, *wo, sc, bsdf_eval, (*unguided_bsdf_pdf) * sweight, sweight, 0);
        kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
      }
      *bsdf_pdf = *unguided_bsdf_pdf;
 
      if (use_surface_guiding) {
        guide_pdf = guiding_bsdf_pdf(kg, state, *wo);
        *bsdf_pdf *= 1.0f - guiding_sampling_prob;
        *bsdf_pdf += guiding_sampling_prob * guide_pdf * (1.0f - bssrdf_sampling_prob);
      }
    }
 
    kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
  }
 
  return label;
}
 
ccl_device int surface_shader_bsdf_guided_sample_closure_ris(KernelGlobals kg,
                                                             IntegratorState state,
                                                             ccl_private ShaderData *sd,
                                                             const ccl_private ShaderClosure *sc,
                                                             const float3 rand_bsdf,
                                                             const ccl_private RNGState *rng_state,
                                                             ccl_private BsdfEval *bsdf_eval,
                                                             ccl_private float3 *wo,
                                                             ccl_private float *bsdf_pdf,
                                                             ccl_private float *mis_pdf,
                                                             ccl_private float *unguided_bsdf_pdf,
                                                             ccl_private float2 *sampled_roughness,
                                                             ccl_private float *eta)
{
  /* BSSRDF should already have been handled elsewhere. */
  kernel_assert(CLOSURE_IS_BSDF(sc->type));
 
  const bool use_surface_guiding = state->guiding.use_surface_guiding;
  const float guiding_sampling_prob = state->guiding.surface_guiding_sampling_prob;
  const float bssrdf_sampling_prob = state->guiding.bssrdf_sampling_prob;
 
  /* Decide between sampling guiding distribution and BSDF. */
  const float rand_bsdf_guiding = state->guiding.sample_surface_guiding_rand;
 
  /* Initialize to zero. */
  int label = LABEL_NONE;
  Spectrum eval = zero_spectrum();
  bsdf_eval_init(bsdf_eval, eval);
 
  *unguided_bsdf_pdf = 0.0f;
  float guide_pdf = 0.0f;
 
  if (use_surface_guiding && guiding_sampling_prob > 0.0f) {
    /* Performing guided sampling using RIS */
 
    // selected RIS candidate
    int ris_idx = 0;
 
    // meta data for the two RIS candidates
    GuidingRISSample ris_samples[2];
    ris_samples[0].rand = rand_bsdf;
    ris_samples[1].rand = path_state_rng_3D(kg, rng_state, PRNG_SURFACE_RIS_GUIDING_0);
 
    // ----------------------------------------------------
    // generate the first RIS candidate using a BSDF sample
    // ----------------------------------------------------
    ris_samples[0].label = bsdf_sample(kg,
                                       sd,
                                       sc,
                                       INTEGRATOR_STATE(state, path, flag),
                                       ris_samples[0].rand,
                                       &ris_samples[0].eval,
                                       &ris_samples[0].wo,
                                       &ris_samples[0].bsdf_pdf,
                                       &ris_samples[0].sampled_roughness,
                                       &ris_samples[0].eta);
 
    bsdf_eval_init(
        &ris_samples[0].bsdf_eval, sc, ris_samples[0].wo, ris_samples[0].eval * sc->weight);
    if (ris_samples[0].bsdf_pdf > 0.0f) {
      if (sd->num_closure > 1) {
        const float sweight = sc->sample_weight;
        ris_samples[0].bsdf_pdf = _surface_shader_bsdf_eval_mis(kg,
                                                                sd,
                                                                ris_samples[0].wo,
                                                                sc,
                                                                &ris_samples[0].bsdf_eval,
                                                                (ris_samples[0].bsdf_pdf) *
                                                                    sweight,
                                                                sweight,
                                                                0);
        kernel_assert(reduce_min(bsdf_eval_sum(&ris_samples[0].bsdf_eval)) >= 0.0f);
      }
      ris_samples[0].avg_bsdf_eval = average(ris_samples[0].bsdf_eval.sum);
      ris_samples[0].guide_pdf = guiding_bsdf_pdf(kg, state, ris_samples[0].wo);
      ris_samples[0].guide_pdf *= (1.0f - bssrdf_sampling_prob);
      ris_samples[0].incoming_radiance_pdf = guiding_surface_incoming_radiance_pdf(
          kg, state, ris_samples[0].wo);
      ris_samples[0].bsdf_pdf = max(0.0f, ris_samples[0].bsdf_pdf);
    }
 
    // ------------------------------------------------------------------------------
    // generate the second RIS candidate using a sample from the guiding distribution
    // ------------------------------------------------------------------------------
    float unguided_bsdf_pdfs[MAX_CLOSURE];
    bsdf_eval_init(&ris_samples[1].bsdf_eval, eval);
    ris_samples[1].guide_pdf = guiding_bsdf_sample(
        kg, state, make_float2(ris_samples[1].rand), &ris_samples[1].wo);
    ris_samples[1].guide_pdf *= (1.0f - bssrdf_sampling_prob);
    ris_samples[1].incoming_radiance_pdf = guiding_surface_incoming_radiance_pdf(
        kg, state, ris_samples[1].wo);
    ris_samples[1].bsdf_pdf = surface_shader_bsdf_eval_pdfs(
        kg, sd, ris_samples[1].wo, &ris_samples[1].bsdf_eval, unguided_bsdf_pdfs, 0);
    ris_samples[1].label = ris_samples[0].label;
    ris_samples[1].avg_bsdf_eval = average(ris_samples[1].bsdf_eval.sum);
    ris_samples[1].bsdf_pdf = max(0.0f, ris_samples[1].bsdf_pdf);
 
    // ------------------------------------------------------------------------------
    // calculate the RIS target functions for each RIS candidate
    // ------------------------------------------------------------------------------
    int num_ris_candidates = 0;
    float sum_ris_weights = 0.0f;
    if (calculate_ris_target(&ris_samples[0], guiding_sampling_prob)) {
      sum_ris_weights += ris_samples[0].ris_weight;
      num_ris_candidates++;
    }
    kernel_assert(ris_samples[0].ris_weight >= 0.0f);
    kernel_assert(sum_ris_weights >= 0.0f);
 
    if (calculate_ris_target(&ris_samples[1], guiding_sampling_prob)) {
      sum_ris_weights += ris_samples[1].ris_weight;
      num_ris_candidates++;
    }
    kernel_assert(ris_samples[1].ris_weight >= 0.0f);
    kernel_assert(sum_ris_weights >= 0.0f);
 
    // ------------------------------------------------------------------------------
    // Sample/Select a sample from the RIS candidates proportional to the target
    // ------------------------------------------------------------------------------
    if (num_ris_candidates == 0 || !(sum_ris_weights > 1e-10f)) {
      *bsdf_pdf = 0.0f;
      *mis_pdf = 0.0f;
      return label;
    }
 
    const float rand_ris_select = rand_bsdf_guiding * sum_ris_weights;
 
    float sum_ris = 0.0f;
    for (int i = 0; i < 2; i++) {
      sum_ris += ris_samples[i].ris_weight;
      if (rand_ris_select <= sum_ris) {
        ris_idx = i;
        break;
      }
    }
 
    kernel_assert(sum_ris >= 0.0f);
    kernel_assert(ris_idx < 2);
 
    // ------------------------------------------------------------------------------
    // Fill in the sample data for the selected RIS candidate
    // ------------------------------------------------------------------------------
    guide_pdf = ris_samples[ris_idx].ris_target * (2.0f / sum_ris_weights);
    *unguided_bsdf_pdf = ris_samples[ris_idx].bsdf_pdf;
    *mis_pdf = 0.5f * (ris_samples[ris_idx].bsdf_pdf + ris_samples[ris_idx].guide_pdf);
    *bsdf_pdf = guide_pdf;
 
    *wo = ris_samples[ris_idx].wo;
    label = ris_samples[ris_idx].label;
 
    *sampled_roughness = ris_samples[ris_idx].sampled_roughness;
    *eta = ris_samples[ris_idx].eta;
    *bsdf_eval = ris_samples[ris_idx].bsdf_eval;
 
    kernel_assert(isfinite_safe(guide_pdf));
    kernel_assert(isfinite_safe(*bsdf_pdf));
 
    if (!(*bsdf_pdf > 1e-10f)) {
      *bsdf_pdf = 0.0f;
      *mis_pdf = 0.0f;
      return label;
    }
 
    kernel_assert(*bsdf_pdf > 0.0f);
    kernel_assert(*bsdf_pdf >= 1e-20f);
    kernel_assert(guide_pdf >= 0.0f);

    if (ris_idx == 1 && ris_samples[1].bsdf_pdf > 0.0f) {
 
      const float rnd = path_state_rng_1D(kg, rng_state, PRNG_SURFACE_RIS_GUIDING_1);
 
      float sum_pdfs = 0.0f;
      int idx = -1;
      for (int i = 0; i < sd->num_closure; i++) {
        sum_pdfs += unguided_bsdf_pdfs[i];
        if (rnd <= sum_pdfs) {
          idx = i;
          break;
        }
      }
      // kernel_assert(idx >= 0);
      /* Set the default idx to the last in the list.
       * in case of numerical problems and rand_bsdf_guiding is just >=1.0f and
       * the sum of all unguided_bsdf_pdfs is just < 1.0f. */
      idx = (rnd > sum_pdfs) ? sd->num_closure - 1 : idx;
 
      label = bsdf_label(kg, &sd->closure[idx], *wo);
      bsdf_roughness_eta(kg, &sd->closure[idx], *wo, sampled_roughness, eta);
    }
 
    kernel_assert(isfinite_safe(*bsdf_pdf));
    kernel_assert(*bsdf_pdf >= 0.0f);
    kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
  }
  else {
    /* Sample BSDF. */
    *bsdf_pdf = 0.0f;
    label = bsdf_sample(kg,
                        sd,
                        sc,
                        INTEGRATOR_STATE(state, path, flag),
                        rand_bsdf,
                        &eval,
                        wo,
                        unguided_bsdf_pdf,
                        sampled_roughness,
                        eta);
#  ifdef WITH_CYCLES_DEBUG
    // Code path to validate the estimation of the label, sampled roughness and eta
    // This should be activated from time to time when the BSDFs change to check if everything
    // is still working correctly.
    if (*unguided_bsdf_pdf > 0.0f) {
      surface_shader_validate_bsdf_sample(kg, sc, *wo, label, *sampled_roughness, *eta);
    }
#  endif
 
    if (*unguided_bsdf_pdf != 0.0f) {
      bsdf_eval_init(bsdf_eval, sc, *wo, eval * sc->weight);
 
      kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
 
      if (sd->num_closure > 1) {
        const float sweight = sc->sample_weight;
        *unguided_bsdf_pdf = _surface_shader_bsdf_eval_mis(
            kg, sd, *wo, sc, bsdf_eval, (*unguided_bsdf_pdf) * sweight, sweight, 0);
        kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
      }
      *bsdf_pdf = *unguided_bsdf_pdf;
      *mis_pdf = *bsdf_pdf;
    }
 
    kernel_assert(reduce_min(bsdf_eval_sum(bsdf_eval)) >= 0.0f);
  }
 
  return label;
}
 
ccl_device int surface_shader_bsdf_guided_sample_closure(KernelGlobals kg,
                                                         IntegratorState state,
                                                         ccl_private ShaderData *sd,
                                                         const ccl_private ShaderClosure *sc,
                                                         const float3 rand_bsdf,
                                                         ccl_private BsdfEval *bsdf_eval,
                                                         ccl_private float3 *wo,
                                                         ccl_private float *bsdf_pdf,
                                                         ccl_private float *mis_pdf,
                                                         ccl_private float *unguided_bsdf_pdf,
                                                         ccl_private float2 *sampled_roughness,
                                                         ccl_private float *eta,
                                                         const ccl_private RNGState *rng_state)
{
  int label = LABEL_NONE;
  if (kernel_data.integrator.guiding_directional_sampling_type ==
          GUIDING_DIRECTIONAL_SAMPLING_TYPE_PRODUCT_MIS ||
      kernel_data.integrator.guiding_directional_sampling_type ==
          GUIDING_DIRECTIONAL_SAMPLING_TYPE_ROUGHNESS)
  {
    label = surface_shader_bsdf_guided_sample_closure_mis(kg,
                                                          state,
                                                          sd,
                                                          sc,
                                                          rand_bsdf,
                                                          bsdf_eval,
                                                          wo,
                                                          bsdf_pdf,
                                                          unguided_bsdf_pdf,
                                                          sampled_roughness,
                                                          eta);
    *mis_pdf = (*unguided_bsdf_pdf > 0.0f) ? *bsdf_pdf : 0.0f;
  }
  else if (kernel_data.integrator.guiding_directional_sampling_type ==
           GUIDING_DIRECTIONAL_SAMPLING_TYPE_RIS)
  {
    label = surface_shader_bsdf_guided_sample_closure_ris(kg,
                                                          state,
                                                          sd,
                                                          sc,
                                                          rand_bsdf,
                                                          rng_state,
                                                          bsdf_eval,
                                                          wo,
                                                          bsdf_pdf,
                                                          mis_pdf,
                                                          unguided_bsdf_pdf,
                                                          sampled_roughness,
                                                          eta);
  }
  if (!(*unguided_bsdf_pdf > 0.0f)) {
    *bsdf_pdf = 0.0f;
    *mis_pdf = 0.0f;
  }
  return label;
}
 
#endif
 
/* Sample direction for picked BSDF, and return evaluation and pdf for all
 * BSDFs combined using MIS. */
ccl_device int surface_shader_bsdf_sample_closure(KernelGlobals kg,
                                                  ccl_private ShaderData *sd,
                                                  const ccl_private ShaderClosure *sc,
                                                  const int path_flag,
                                                  const float3 rand_bsdf,
                                                  ccl_private BsdfEval *bsdf_eval,
                                                  ccl_private float3 *wo,
                                                  ccl_private float *pdf,
                                                  ccl_private float2 *sampled_roughness,
                                                  ccl_private float *eta)
{
  /* BSSRDF should already have been handled elsewhere. */
  kernel_assert(CLOSURE_IS_BSDF(sc->type));
 
  int label;
  Spectrum eval = zero_spectrum();
 
  *pdf = 0.0f;
  label = bsdf_sample(kg, sd, sc, path_flag, rand_bsdf, &eval, wo, pdf, sampled_roughness, eta);
 
  if (*pdf != 0.0f) {
    bsdf_eval_init(bsdf_eval, sc, *wo, eval * sc->weight);
 
    if (sd->num_closure > 1) {
      const float sweight = sc->sample_weight;
      *pdf = _surface_shader_bsdf_eval_mis(kg, sd, *wo, sc, bsdf_eval, *pdf * sweight, sweight, 0);
    }
  }
  else {
    bsdf_eval_init(bsdf_eval, zero_spectrum());
  }
 
  return label;
}
 
ccl_device float surface_shader_average_roughness(const ccl_private ShaderData *sd)
{
  float roughness = 0.0f;
  float sum_weight = 0.0f;
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF(sc->type)) {
      /* `sqrt()` once to undo the squaring from multiplying roughness on the
       * two axes, and once for the squared roughness convention. */
      const float value = bsdf_get_roughness_pass_squared(sc);
      if (value >= 0.0f) {
        const float weight = fabsf(average(sc->weight));
        roughness += weight * sqrtf(sqrtf(value));
        sum_weight += weight;
      }
    }
  }
 
  return (sum_weight > 0.0f) ? roughness / sum_weight : 1.0f;
}
 
ccl_device Spectrum surface_shader_transparency(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  if (sd->flag & SD_HAS_ONLY_VOLUME) {
    return one_spectrum();
  }
  if (sd->flag & (SD_TRANSPARENT | SD_RAY_PORTAL)) {
    return sd->closure_transparent_extinction;
  }
  return zero_spectrum();
}
 
ccl_device Spectrum surface_shader_alpha(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  Spectrum alpha = one_spectrum() - surface_shader_transparency(kg, sd);
 
  alpha = saturate(alpha);
 
  return alpha;
}
 
ccl_device Spectrum surface_shader_diffuse(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  Spectrum eval = zero_spectrum();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF_DIFFUSE(sc->type) || CLOSURE_IS_BSSRDF(sc->type)) {
      eval += bsdf_albedo(kg, sd, sc, true, true);
    }
  }
 
  return eval;
}
 
ccl_device Spectrum surface_shader_glossy(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  Spectrum eval = zero_spectrum();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF_GLOSSY(sc->type) || CLOSURE_IS_GLASS(sc->type)) {
      eval += bsdf_albedo(kg, sd, sc, true, false);
    }
  }
 
  return eval;
}
 
ccl_device Spectrum surface_shader_transmission(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  Spectrum eval = zero_spectrum();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF_TRANSMISSION(sc->type) || CLOSURE_IS_GLASS(sc->type)) {
      eval += bsdf_albedo(kg, sd, sc, false, true);
    }
  }
 
  return eval;
}
 
ccl_device float3 surface_shader_average_normal(KernelGlobals kg, const ccl_private ShaderData *sd)
{
  float3 N = zero_float3();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
    if (CLOSURE_IS_BSDF_OR_BSSRDF(sc->type)) {
      N += sc->N * fabsf(average(sc->weight));
    }
  }
 
  return (is_zero(N)) ? sd->N : normalize(N);
}
 
ccl_device Spectrum surface_shader_ao(KernelGlobals kg,
                                      const ccl_private ShaderData *sd,
                                      const float ao_factor,
                                      ccl_private float3 *N_)
{
  Spectrum eval = zero_spectrum();
  float3 N = zero_float3();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSDF_DIFFUSE(sc->type)) {
      const ccl_private DiffuseBsdf *bsdf = (const ccl_private DiffuseBsdf *)sc;
      eval += sc->weight * ao_factor;
      N += bsdf->N * fabsf(average(sc->weight));
    }
  }
 
  *N_ = (is_zero(N)) ? sd->N : normalize(N);
  return eval;
}
 
#ifdef __SUBSURFACE__
ccl_device float3 surface_shader_bssrdf_normal(const ccl_private ShaderData *sd)
{
  float3 N = zero_float3();
 
  for (int i = 0; i < sd->num_closure; i++) {
    const ccl_private ShaderClosure *sc = &sd->closure[i];
 
    if (CLOSURE_IS_BSSRDF(sc->type)) {
      const ccl_private Bssrdf *bssrdf = (const ccl_private Bssrdf *)sc;
      const float avg_weight = fabsf(average(sc->weight));
 
      N += bssrdf->N * avg_weight;
    }
  }
 
  return (is_zero(N)) ? sd->N : normalize(N);
}
#endif /* __SUBSURFACE__ */
 
/* Constant emission optimization */
 
ccl_device bool surface_shader_constant_emission(KernelGlobals kg,
                                                 const int shader,
                                                 ccl_private Spectrum *eval)
{
  const int shader_index = shader & SHADER_MASK;
  const int shader_flag = kernel_data_fetch(shaders, shader_index).flags;
 
  if (shader_flag & SD_HAS_CONSTANT_EMISSION) {
    const float3 emission_rgb = make_float3(
        kernel_data_fetch(shaders, shader_index).constant_emission[0],
        kernel_data_fetch(shaders, shader_index).constant_emission[1],
        kernel_data_fetch(shaders, shader_index).constant_emission[2]);
    *eval = rgb_to_spectrum(emission_rgb);
 
    return true;
  }
 
  return false;
}
 
/* Background */
 
ccl_device Spectrum surface_shader_background(const ccl_private ShaderData *sd)
{
  if (sd->flag & SD_EMISSION) {
    return sd->closure_emission_background;
  }
  return zero_spectrum();
}
 
/* Emission */
 
ccl_device Spectrum surface_shader_emission(const ccl_private ShaderData *sd)
{
  if (sd->flag & SD_EMISSION) {
    return emissive_simple_eval(sd->Ng, sd->wi) * sd->closure_emission_background;
  }
  return zero_spectrum();
}
 
/* Holdout */
 
ccl_device Spectrum surface_shader_apply_holdout(KernelGlobals kg, ccl_private ShaderData *sd)
{
  Spectrum weight = zero_spectrum();
 
  /* For objects marked as holdout, preserve transparency and remove all other
   * closures, replacing them with a holdout weight. */
  if (sd->object_flag & SD_OBJECT_HOLDOUT_MASK) {
    if ((sd->flag & SD_TRANSPARENT) && !(sd->flag & SD_HAS_ONLY_VOLUME)) {
      weight = one_spectrum() - sd->closure_transparent_extinction;
 
      for (int i = 0; i < sd->num_closure; i++) {
        ccl_private ShaderClosure *sc = &sd->closure[i];
        if (!CLOSURE_IS_BSDF_TRANSPARENT(sc->type)) {
          sc->type = NBUILTIN_CLOSURES;
        }
      }
 
      sd->flag &= ~(SD_CLOSURE_FLAGS - (SD_TRANSPARENT | SD_BSDF));
    }
    else {
      weight = one_spectrum();
    }
  }
  else {
    for (int i = 0; i < sd->num_closure; i++) {
      const ccl_private ShaderClosure *sc = &sd->closure[i];
      if (CLOSURE_IS_HOLDOUT(sc->type)) {
        weight += sc->weight;
      }
    }
  }
 
  return weight;
}
 
/* Surface Evaluation */
 
template<uint node_feature_mask, typename ConstIntegratorGenericState>
ccl_device void surface_shader_eval(KernelGlobals kg,
                                    ConstIntegratorGenericState state,
                                    ccl_private ShaderData *ccl_restrict sd,
                                    ccl_global float *ccl_restrict buffer,
                                    const uint32_t path_flag,
                                    bool use_caustics_storage = false)
{
  /* If path is being terminated, we are tracing a shadow ray or evaluating
   * emission, then we don't need to store closures. The emission and shadow
   * shader data also do not have a closure array to save GPU memory. */
  int max_closures;
  if (path_flag & (PATH_RAY_TERMINATE | PATH_RAY_SHADOW | PATH_RAY_EMISSION)) {
    max_closures = 0;
  }
  else {
    max_closures = use_caustics_storage ? CAUSTICS_MAX_CLOSURE : kernel_data.max_closures;
  }
 
  sd->num_closure = 0;
  sd->num_closure_left = max_closures;
  sd->closure_transparent_extinction = zero_spectrum();
 
#ifdef __OSL__
  if (kernel_data.kernel_features & KERNEL_FEATURE_OSL_SHADING) {
    osl_eval_nodes<SHADER_TYPE_SURFACE>(kg, state, sd, path_flag);
  }
  else
#endif
  {
#ifdef __SVM__
    svm_eval_nodes<node_feature_mask, SHADER_TYPE_SURFACE>(kg, state, sd, buffer, path_flag);
#else
    if (sd->object == OBJECT_NONE) {
      sd->closure_emission_background = make_spectrum(0.8f);
      sd->flag |= SD_EMISSION;
    }
    else {
      ccl_private DiffuseBsdf *bsdf = (ccl_private DiffuseBsdf *)bsdf_alloc(
          sd, sizeof(DiffuseBsdf), make_spectrum(0.8f));
      if (bsdf != nullptr) {
        bsdf->N = sd->N;
        sd->flag |= bsdf_diffuse_setup(bsdf);
      }
    }
#endif
  }
}
 
CCL_NAMESPACE_END