R2ShadowVariance.h File Reference

Functions for variance shadows. More...

#include "R2LogDepth.h"

Include dependency graph for R2ShadowVariance.h:

This graph shows which files directly or indirectly include this file:

Go to the source code of this file.

Data Structures
struct	R2_shadow_variance_t
	A variance shadow. More...

Functions
float	R2_varianceChebyshevUpperBound (const R2_shadow_variance_t s, const vec2 moments, const float depth)
float	R2_varianceLinearStep (const float min, const float max, const float x)
float	R2_varianceLightBleedReduction (const R2_shadow_variance_t s, const float p_max)
float	R2_varianceShadowFactor (const R2_shadow_variance_t s, const R2_light_projective_vectors_t v)

Detailed Description

Functions for variance shadows.

Definition in file R2ShadowVariance.h.

Function Documentation

R2_varianceChebyshevUpperBound()

float R2_varianceChebyshevUpperBound	(	const R2_shadow_variance_t	s,
		const vec2	moments,
		const float	depth
	)

Compute an upper bound on the probability that the position at depth is in shadow.

Parameters

s	The shadow
moments	The sampled shadow map depth moments
depth	The current depth

Returns

A value between 0.0 and 1.0, where 1.0 indicates that the area is certainly in shadow.

Definition at line 38 of file R2ShadowVariance.h.

{
  float p        = float (depth <= moments.x);
  float variance = max (s.variance_minimum, moments.y - (moments.x * moments.x));
  float delta    = depth - moments.x;
  float p_max    = variance / (variance + (delta * delta));
  return max (p, p_max);
}

R2_varianceLightBleedReduction()

float R2_varianceLightBleedReduction	(	const R2_shadow_variance_t	s,
		const float	p_max
	)

Apply light bleed reduction to the given shadow probability.

Parameters

s	The shadow
p	The shadow probability

Returns

p with light bleed reduction applied

Definition at line 79 of file R2ShadowVariance.h.

{
  return R2_varianceLinearStep (s.bleed_reduction, 1.0, p_max);
}

R2_varianceLinearStep()

float R2_varianceLinearStep	(	const float	min,
		const float	max,
		const float	x
	)

A linear step function that maps all values less than min to min, all values greater than max to max, and linearly scales all values in [min, max] to [min, max].

Parameters

min	The minimum value
max	The maximum value
x	The value to scale

Returns

A linearly scaled value

Definition at line 61 of file R2ShadowVariance.h.

{
  float vsm = x - min;
  float msm = max - min;
  return clamp (vsm / msm, 0.0, 1.0);
}

R2_varianceShadowFactor()

float R2_varianceShadowFactor	(	const R2_shadow_variance_t	s,
		const R2_light_projective_vectors_t	v
	)

Calculate a shadow factor for the given position.

Parameters

s	The shadow
v	The projective light vectors, including the position in various light coordinate spaces

Returns

s.factor_minimum if the point is fully in shadow, or 1.0 if the point is

Definition at line 95 of file R2ShadowVariance.h.

{
  float pos_light_depth =
    R2_logDepthEncodePartial(
      R2_logDepthPrepareEyeZ (v.surface_light_eye.z),
      s.depth_coefficient);

  // Sample the variance map for the depth distribution
  vec2 moments =
    texture (s.map, v.surface_light_uv.xy).xy;

  // Calculate the probability that the point is in shadow.
  float p_max =
    R2_varianceChebyshevUpperBound (s, moments, pos_light_depth);

  // Apply light bleeding reduction
  float p_reduced =
    R2_varianceLightBleedReduction (s, p_max);

  return max (p_reduced, s.factor_minimum);
}