Shadow Mapping in Ogre

Hamilton Chong
Aug 2006

Abstract
A discussion of shadow mapping: the algorithm, variants, theory and analysis,
and implementation.

1 Introduction to the Shadow Mapping Algorithm

Shadow mapping, an algorithm introduced by Lance Williams [9] and now prevalent
in real-time and off-line rendering, is based on a simple idea: First, a snapshot of the
scene is taken from the viewpoint of the light. Then, when creating an image from the
perspective of the camera, the lights snapshot is used to determine visibility. Parts of
the scene seen by both the light and the camera must be lit (by the light in question).
Parts of the scene visible only to the camera must be shadowed. We do not care about
parts of the scene seen only by the light.
In practice, the snapshot from the viewpoint of the light is stored as a floating point
depth buffer. It is important to use a format that supports enough precision to avoid
shadow acne (z-fighting) on lit surfaces. In Ogre, we can specify the depth format to
use; in the example code, we will choose the 32-bit format.
Once shadow determination has occurred (whether a fragment is in shadow or not),
Ogre provides two different ways to render the shadows into the final image. The
modulative technique will uniformly darken regions of the image determined to be in
shadow. This is a cheaper and less accurate lighting model. For instance, specular
highlights in shadow will appear as darkened specular highlights. The other technique
is additive light masking. This technique builds up contributions from each light in nonshadowed areas and adds them together to create the final image. The code in section
4 will use additive light masking, but could just as easily be adapted for modulative
shadows.

1.1

Formalism

Mathematically, the process can be represented as follows: Let Pl and Pc be the projection matrices for the light and camera respectively. Let Ml and Mc be the modelview
matrices for the light and camera coordinate systems. Let ~x = [x1 , x2 , x3 , 1]t be a point
in object space, ~y = [y1 , y2 , y3 , 1]t the screen space coordinates, and ~u = [u1 , u2 , u3 , 1]t
the shadow map coordinates.

u1 wl
u2 wl

u3 wl = Pl Ml
wl

x1
x2

x3
1

(1)

y1 wc

y2 wc
= Pc Mc

y3 wc
wc

x1
x2

x3
1

(2)

Figure 1: Shadow map sample must use one float to represent a range of possible depth values.
A depth sample is chosen in the middle. Any camera image point in between the two camera rays
will see the geometry, and depending on distance from light will report differently on shadowed
versus lit. However, every such point should be lit.
These equations can be written more concisely as: ~uwl = Pl Ml~x and ~ywc = Pc Mc~x.
Division of~uwl and~ywc by their respective homogeneous coordinates yields the Euclidean representations ~u and ~y.
Note that while Pc and Mc are completely determined by the camera image we want
to produce, we have some ambiguity in the Pl and Ml chosen for shadow mapping. The
degrees of freedom here are later exploited to combat the aliasing issue.

1.2

Depth Biasing

Due to the finite precision of floating point representations and inherent inability of one
number to represent a range of values, it is often necessary to add a little bias to the
depth values stored in a shadow map. One does not simply store the u3 value. Figure 1
illustrates the issue. Here we have used blue dots on the lights image plane to represent
boundaries between shadow texels. The interval in between the dots then represents
a shadow map sample for which a single depth value (float) is stored. For the sample
whose boundary rays are shown, the red dots depth is saved. However, note that from
the cameras perspective, any (camera) image point in between the two drawn camera
rays will hit the scene geometry within the shadow map samples interval. Hence, the
same shadow map sample depth will be used to determine visibility for all such camera
pixels. Camera pixels whose rays fall to the right of the red dot will be marked as
shadowed, while pixels whose rays fall to the left of the red dot will be marked as lit.
This is not the right behavior because clearly all the pixels should be marked as lit. As
we can see, a depth bias is needed. By pushing the shadow map samples depth farther
(to the 2nd red dot), we can achieve correct shadow determination.
One could approach the depth bias issue in a completely ad hoc manner, but it is
possible to do better. One would ideally compute a bias that depends on how depth (u3 )

changes between shadow map samples. The change in depth as one moves a unit step
(to the next shadow map sample) represents the ambiguity of the depth value. Such a
value may seem intractable to compute, but calculus and linear algebra save the day.
From calculus, we learn that the derivative provides the best linear approximation to
any function (u3 = u3 (u1 , u2 ) in particular). In multiple dimensions, this role is played
by the Jacobian (matrix of partial derivatives). In other words, we want to compute
du3
du3
du1 and du2 , where we have treated u3 as a function of u1 and u2 . Once these values
are computed, it makes sense to then add some weighted combination of these to the
stored depth value (e.g., some scale of the Jacobians Frobenius norm).
But even if the light is staring at a plane straight on (view direciton lines up with
du3
3
and du
planes normal), making du
du2 both zero, we would still need a slight offset
1
because rounding due to the floats finite representation may still cause shadow acne.
In this case, wed like to offset the depth by a small value that pushes it beyond rounding
ambiguity. While one could use an arbitrary constant offset, this is unsatisfactory since
the constant in light image space corresponds to varying amounts of offset in light
space (pre-projection Euclidean space with lights position at origin). Let us instead
choose a constant offset in the z direction of light space and compute what the offset
for a particular sample should be in light image space. In Ogres example code, the
small constant offset in light space is chosen to be 1 unit. If 1 is not a small amount in
your engines chosen scale, you can easily change this choice. At any rate, the relevant
quantity is Xu3 where ~X = Ml~x.
3
The choices here closely mirror what OpenGL implements through glPolygonOffset. The second adjustment is slightly different since OpenGL chooses a vendor specific fudge factor.
Equations for computing the stated quantities are provided below. One need not
wade through these to use the depth biasing code. Understanding what the relevant
parameters explained above are (in case adjustment is needed) is sufficient.

(~xql )
= i-th column of Ml1 Pl1Vl1
ui

(3)

where Vl is the viewport matrix for the light and i = 1, 2, 3. ql turns out to be 1/wl .


~x
ql
1 (~xql )
=
~x
ui ql
ui
ui
1 


~x
~x
du3
~n
= ~n
du j
u3
uj

(4)
(5)

where ~n is the normal at point ~x and j = 1, 2. Note that (5) is exactly the set of values
needed for the first part.

(~uwl )
= 3rd column of Pl
X3


~u
wl
1 (~uwl )
=
~u
X3 wl
X3
X3
4

(6)
(7)

Note that (7) is the quantity needed for the second bias term. This is also the term
to scale for different choices of small offset in light space. If 0.01 units is the small
offset, scale this value by 0.01.

1.3

Percentage Closest Filtering

As widely known, shadow mapping can exhibit significant aliasing. When this happens during texture mapping we apply filtering. Wed like to apply a similar principle
with shadow maps, but filtering depth values is categorically the wrong thing to do. As
described in [7], one should instead filter depth test results. This is termed percentage closest filtering. Ideally this would be a filtering technique much like anisotropic
texture filtering, but for simplicity and efficiency, Ogres example code implements the
bilinear analogue.

2 Variants
There are many shadow mapping variants. Enumerating (much less describing) all
of them would take us too far afield in this article. We instead defer to the provided
references and google for such coverage. The many variants can, however, be broken
up into three broad categories: (1) Those that store additional information beyond a
single float, (2) those that divide up shadow frusta into multiple frusta to be handled
separately, and (3) those that propose less naive Pl and Ml to use and thereby affect the
sampling distribution. Algorithms in each category usually work quite independently
and so many hybrid approaches are easily conceivable.

2.1

Storing Additional Info

One example of this is Deep Shadow Maps [5]. In this work, instead of storing a single
depth value and treating visibility as a binary value, a transfer function is stored and
visibility is continuous. This algorithm is important in offline movie rendering, but also
relevant to the Variance Shadow Mapping algorithm elucidated by the game developer
community [3].
While variance shadow maps are motivated by statistical considerations, it is perhaps more properly understood in the Deep Shadow Maps framework. Analyzing it in
terms of distributions is flawed for two reasons: (1) the inequality considered is valid
only for unimodal distributions whereas depth values are often discontinuous in regions
that matter; (2) the inequality is treated as equality. The equations are justified with a
very specific example in which two planes are viewed straight on. In practice there
are very noticeable halo effects around objects, which makes more heuristic tweaks
necessary.
Recasting this into the framework of deep shadow maps, we see that the proposed
equality is simply a particular functional approximation to the transfer function. Variance shadow maps proposes a two-parameter family of approximation functions whose
parameters are linearly interpolated in the usual way. This viewpoint allows for analy-

Figure 2: Region I is defined as the set of all points along rays between the light and a point on
the plane of interest in the cameras view. Everything in region I is shadowed and self-shadowed
properly. Objects in region II are not self-shadowed properly.
sis and also suggests the possibility of getting improvements via other approximating
functional forms.

2.2

Breaking up Shadow Frusta

Adaptive Shadow Maps [4] are an example of this. It is still largely considered too
expensive for real-time rendering, but continued research and growing GPU power may
make some variant worthwhile.

2.3

Playing with Projection Matrices

There are various heuristic approaches for choosing Pl and Ml , but here we will focus
on one method, the Plane Optimal algorithm [1], that provides a particular guarantee.
For this algorithm, we specify a plane of interest (e.g., ground plane, wall, table top)
for which we want perfect shadowing no matter the configuration of light and camera
in the scene (even dueling frusta). The algorithm will then compute Pl and Ml so that
the mapping between camera image and light image is the identity when restricted to
the plane. If the shadow map matches the resolution of the screen, then each pixel
gets exactly one shadow sample. Shadows off the plane of interest have no guarantees.
One limitation of the method is shown in Figure 2. Only region I will be shadowed and
self-shadowed properly, with points on the plane being shadowed perfectly (alias-free).
This makes the method perhaps most useful for games where the view is top-down or
isometric (like RTS games). It is also useful for cases like dueling frusta (where just
about all other methods fail).

3 Theory and Analysis

A full discussion of shadow map analysis is beyond the scope of this article. For those
interested, the references [2] and [1] are good (in my extremely biased opinion). Note
that as research papers, they are quite concise. Unfortunately there dont seem to more
step-by-step expositions available at this moment.
There has been a lot of academic and industry research on improving shadow maps.
However, analyses presented on shadow maps often do not say what people claim they
say. These faulty conclusions usually come from considering very special cases and
assuming the general case is very similar. For clarification, we explore some of these
misconceptions here.

3.1

(Non)Optimality of Logarithmic Shadow Maps

We start with one heuristic that has gained quite a bit of traction: the idea of using some
logarithmic mapping between light space and light image space instead of a projective
transform. A number of algorithms based on this idea have been proposed, and even
some hardware changes. Much of this work seems to be motivated by the incorrect
assumption that logarithmic mappings are optimal.
The very special motivating case is this: The camera looks down the z axis. Directional light illuminates the scene perpendicular to the z axis. An angled piece of a
plane is viewed by the camera. As the angled piece of plane is pulled along the camera
ray direction, using a logarithmic shadow map gives us constant shadow quality on this
geometric piece. But unless were rendering translucent dust particles along a camera
ray, this analysis is irrelevant. If the dust particles are not translucent, we only care
about shadow determination on the first one, not a whole line of them. If we are rendering continuous surfaces (resp. curves), we care about the quality as one moves in
the tangent plane (resp. tangent) direction because this is the best linear approximation
to the surface (resp. curve), not the camera ray direction.
In fact, in the case of a chosen plane of interest for example, we know we can
get completely alias free shadow mapping using a projective transform (section 2.3).
Logarithmic shadow maps may be an interesting heuristic to try out, but certainly not
worth changing hardware over in my opinion. If youre going to change hardware,
might as well aim for true optimality.

3.2

Sampling Aliasing versus Depth Precision Aliasing

Sometimes people tend to conflate these two sources of aliasing. They note that after applying some sort of custom projective transform, the depth values are warped as
well. This problem can be completely overcome via the depth replacement method
prescribed in Trapezoidal Shadow Maps [6]. So this is a completely orthogonal issue. Depth precision can be just as good as normal shadow maps, no matter the
perspective warp used to affect sampling.

3.3

Projective versus Perspective Aliasing

The terms perspective and projective aliasing appeared in the Perspective Shadow Maps
[8] paper and has since been used extensively by those who work on improving shadow
heuristics. Often it is claimed that methods ameliorate perspective aliasing while projective aliasing is either unavoidable or must be addressed via completely separate
means. However, the distinction between the two is somewhat artificial. Both result
from not allocating enough shadow map samples to regions that matter to the viewer.
As the Plane Optimal algorithm demonstrates, it is possible to completely remove projective aliasing (as well as perspective aliasing) in certain scenes. In general, there
should be one combined measure of aliasing and algorithms must minimize this quantity. See [2] for a unified notion of aliasing.

4 Implementation
Ogre provides a powerful framework that allows us to do a lot of shadow map customization. In Ogre, we turn on custom shadow mapping through the scene manager
(here, sceneMgr). It is recommended that this happen early as it may affect how certain
resources are loaded.
// Use Ogres custom shadow mapping ability
sceneMgr->setShadowTexturePixelFormat(PF_FLOAT32_R);
sceneMgr->setShadowTechnique( SHADOWTYPE_TEXTURE_ADDITIVE );
sceneMgr->setShadowTextureCasterMaterial("CustomShadows/ShadowCaster");
sceneMgr->setShadowTextureReceiverMaterial("CustomShadows/ShadowReceiver");
sceneMgr->setShadowTextureSelfShadow(true);
sceneMgr->setShadowTextureSize(512);
The setShadowTechnique call is all that is required for Ogres default shadow mapping. In the code above, we have told Ogre to use the R channel of a floating point
texture to store depth values. This tends to be a very portable method (over graphics
cards and APIs). The sample sticks to using Ogres default of 512x512 shadow maps.
Self-shadowing is turned on, but be warned that this will only work properly if appropriate depth biasing is also used. The example code will manually account for depth
biasing via the method described above in section 1.2. The shadow caster and shadow
receiver materials are defined in a materials script. They tell Ogre which shaders to use
when rendering shadow casters into the shadow map and rendering shadow receivers
during shadow determination.
The CustomShadows.material material script is given below:
// Shadow Caster __________________________________________________

vertex_program CustomShadows/ShadowCasterVP/Cg cg
{
source customshadowcastervp.cg

entry_point main
profiles arbvp1 vs_2_0
default_params
{
param_named_auto uModelViewProjection worldviewproj_matrix
}
}
fragment_program CustomShadows/ShadowCasterFP/Cg cg
{
source customshadowcasterfp.cg
entry_point main
profiles arbfp1 ps_2_0
default_params
{
param_named
param_named
param_named
param_named_auto
param_named_auto
}

uDepthOffset float
uSTexWidth float
uSTexHeight float
uInvModelViewProjection
uProjection

1.0
512.0
512.0
inverse_worldviewproj_matrix
projection_matrix

}
vertex_program CustomShadows/ShadowCasterVP/GLSL glsl
{
source customshadowcastervp.vert
default_params
{
param_named_auto uModelViewProjection worldviewproj_matrix
}
}
fragment_program CustomShadows/ShadowCasterFP/GLSL glsl
{
source customshadowcasterfp.frag
default_params
{
param_named
param_named
param_named
param_named_auto
param_named_auto

uDepthOffset float
uSTexWidth float
uSTexHeight float
uInvModelViewProjection
uProjection
9

1.0
512.0
512.0
inverse_worldviewproj_matrix
projection_matrix

}
}
vertex_program CustomShadows/ShadowCasterVP/HLSL hlsl
{
source customshadowcastervp.hlsl
entry_point main
target vs_2_0
default_params
{
param_named_auto uModelViewProjection worldviewproj_matrix
}
}
fragment_program CustomShadows/ShadowCasterFP/HLSL hlsl
{
source customshadowcasterfp.hlsl
entry_point main
target ps_2_0
default_params
{
param_named
param_named
param_named
param_named_auto
param_named_auto
}

uDepthOffset float
uSTexWidth float
uSTexHeight float
uInvModelViewProjection
uProjection

1.0
512.0
512.0
inverse_worldviewproj_matrix
projection_matrix

}
material CustomShadows/ShadowCaster
{
technique glsl
{
// Z-write only pass
pass Z-write
{
vertex_program_ref CustomShadows/ShadowCasterVP/GLSL
{
}
fragment_program_ref CustomShadows/ShadowCasterFP/GLSL
{
}
}
}
10

technique hlsl
{
// Z-write only pass
pass Z-write
{
//Instead of using depth_bias, well be implementing it manually
vertex_program_ref CustomShadows/ShadowCasterVP/HLSL
{
}
fragment_program_ref CustomShadows/ShadowCasterFP/HLSL
{
}
}
}
technique cg
{
// Z-write only pass
pass Z-write
{
//Instead of using depth_bias, well be implementing it manually
vertex_program_ref CustomShadows/ShadowCasterVP/Cg
{
}
fragment_program_ref CustomShadows/ShadowCasterFP/Cg
{
}
}
}
}

// Shadow Receiver ________________________________________________

vertex_program CustomShadows/ShadowReceiverVP/Cg cg
{
source customshadowreceivervp.cg
entry_point main
profiles arbvp1 vs_2_0
default_params
{
11

param_named_auto
param_named_auto
param_named_auto
param_named_auto

uModelViewProjection
uLightPosition
uModel
uTextureViewProjection

worldviewproj_matrix
light_position_object_space 0
world_matrix
texture_viewproj_matrix

}
}
fragment_program CustomShadows/ShadowReceiverFP/Cg cg
{
source customshadowreceiverfp.cg
entry_point main
profiles arbfp1 ps_2_x
default_params
{
param_named uSTexWidth float 512.0
param_named uSTexHeight float 512.0
}
}
vertex_program CustomShadows/ShadowReceiverVP/GLSL glsl
{
source customshadowreceiver.vert
default_params
{
param_named_auto
param_named_auto
param_named_auto
param_named_auto
}

uModelViewProjection
uLightPosition
uModel
uTextureViewProjection

worldviewproj_matrix
light_position_object_space 0
world_matrix
texture_viewproj_matrix

}
fragment_program CustomShadows/ShadowReceiverFP/GLSL glsl
{
source customshadowreceiver.frag
default_params
{
param_named uSTexWidth float 512.0
param_named uSTexHeight float 512.0
}
}
vertex_program CustomShadows/ShadowReceiverVP/HLSL hlsl
{
12

source customshadowreceivervp.hlsl
entry_point main
target vs_2_0
default_params
{
param_named_auto
param_named_auto
param_named_auto
param_named_auto
}

uModelViewProjection
uLightPosition
uModel
uTextureViewProjection

worldviewproj_matrix
light_position_object_space 0
world_matrix
texture_viewproj_matrix

}
fragment_program CustomShadows/ShadowReceiverFP/HLSL hlsl
{
source customshadowreceiverfp.hlsl
entry_point main
target ps_3_0
default_params
{
param_named uSTexWidth float 512.0
param_named uSTexHeight float 512.0
}
}
material CustomShadows/ShadowReceiver
{
technique glsl
{
pass lighting
{
vertex_program_ref CustomShadows/ShadowReceiverVP/GLSL
{
}
fragment_program_ref CustomShadows/ShadowReceiverFP/GLSL
{
param_named uShadowMap int 0
}
texture_unit ShadowMap
{
tex_address_mode clamp
filtering none
}
13

}
}
technique hlsl
{
pass lighting
{
vertex_program_ref CustomShadows/ShadowReceiverVP/HLSL
{
}
fragment_program_ref CustomShadows/ShadowReceiverFP/HLSL
{
}
// we wont rely on hardware specific filtering of z-tests
texture_unit ShadowMap
{
tex_address_mode clamp
filtering none
}
}
}
technique cg
{
pass lighting
{
vertex_program_ref CustomShadows/ShadowReceiverVP/Cg
{
}
fragment_program_ref CustomShadows/ShadowReceiverFP/Cg
{
}
// we wont rely on hardware specific filtering of z-tests
texture_unit ShadowMap
{
tex_address_mode clamp
filtering none
}
}
}
}

14

Three techniques are presented, one for GLSL, one for HLSL, and one for Cg.
Well present the GLSL code below. Note that while most of the shader files are direct translations of each other, DirectX HLSL shaders must handle percentage closest
filtering slightly differently from OpenGL. OpenGL chooses the convention of having
integers index sample centers whereas DirectX chooses integers to index sample corners. Also note the variable names in the shaders presented below are slightly different
from those presented earlier in this document. This is due in part to the awkwardness
of expressing subscripts in variable names and also in part because u3 is less evocative of depth than z, etc. With minimal effort one can match the shader equations with
those presented earlier. The code is presented here mostly to demonstrate how things
fit together.
//////////////////////////////////////////////////////////////////
//
// shadowcastervp.vert
//
// This is an example vertex shader for shadow caster objects.
//
//////////////////////////////////////////////////////////////////

// I N P U T

V A R I A B L E S /////////////////////////////////

uniform mat4 uModelViewProjection;

// O U T P U T

// modelview projection matrix

V A R I A B L E S ///////////////////////////////

varying vec4 pPosition;

varying vec4 pNormal;
varying vec4 pModelPos;

// post projection position coordinates

// normal in object space (to be interpolated)
// position in object space (to be interpolated)

// M A I N ///////////////////////////////////////////////////////
void main()
{
// Transform vertex position into post projective (homogenous screen) space.
gl_Position = uModelViewProjection * gl_Vertex;
pPosition
= uModelViewProjection * gl_Vertex;
// copy over data to interpolate using perspective correct interpolation
pNormal
= vec4(gl_Normal.x, gl_Normal.y, gl_Normal.z, 0.0);
pModelPos = gl_Vertex;
}
This is a pretty standard vertex shader.
/////////////////////////////////////////////////////////////////////////////////
15

//
// shadowcasterfp.frag
//
// This is an example fragment shader for shadow caster objects.
//
/////////////////////////////////////////////////////////////////////////////////
// I N P U T

V A R I A B L E S ////////////////////////////////////////////////

// uniform constants
uniform float uDepthOffset;
//
uniform float uSTexWidth;
//
uniform float uSTexHeight;
//
uniform mat4 uInvModelViewProjection;//
uniform mat4 uProjection;
//
// per fragment inputs
varying vec4 pPosition;
varying vec4 pNormal;
varying vec4 pModelPos;

offset amount (constant in eye space)

shadow map texture width
shadow map texture height
inverse model-view-projection matrix
projection matrix

// position of fragment (in homogeneous coordinates)

// un-normalized normal in object space
// coordinates of model in object space at this point

// M A I N //////////////////////////////////////////////////////////////////////
void main(void)
{
// compute the "normalized device coordinates" (no viewport applied yet)
vec4 postProj = pPosition / pPosition.w;
// get the normalized normal of the geometry seen at this point
vec4 normal = normalize(pNormal);

// -- Computing Depth Bias Quantities ----------------------------// We want to compute the "depth slope" of the polygon.
// This is the change in z value that accompanies a change in x or y on screen
// such that the coordinates stay on the triangle.
// The depth slope, dzlen below, is a measure of the uncertainty in our z value
// Roughly, these equations come from re-arrangement of the product rule:
// d(uq) = d(u)q + u d(q) --> d(u) = 1/q * (d(uq) - u d(q))
vec4 duqdx = uInvModelViewProjection * vec4(1.0/uSTexWidth,0.0,0.0,0.0);
vec4 dudx = pPosition.w * (duqdx - (pModelPos * duqdx.w));
vec4 duqdy = uInvModelViewProjection * vec4(0.0,1.0/uSTexHeight,0.0,0.0);
vec4 dudy = pPosition.w * (duqdy - (pModelPos * duqdy.w));
vec4 duqdz = uInvModelViewProjection * vec4(0.0,0.0,1.0,0.0);
16

vec4 dudz = pPosition.w * (duqdz - (pModelPos * duqdz.w));

// The next relations come from the requirement dot(normal, displacement) = 0
float denom = 1.0 / dot(normal.xyz, dudz.xyz);
vec2 dz = - vec2( dot(normal.xyz, dudx.xyz) * denom ,
dot(normal.xyz, dudy.xyz) * denom );
float dzlen = max(abs(dz.x), abs(dz.y));

// We now compute the change in z that would signify a push in the z direction
// by 1 unit in eye space. Note that eye space z is related in a nonlinear way to
// screen space z, so this is not just a constant.
// ddepth below is how much screen space z at this point would change for that push.
// NOTE: computation of ddepth likely differs from OpenGLs glPolygonOffset "unit"
// computation, which is allowed to be vendor specific.
vec4 dpwdz = uProjection * vec4(0.0, 0.0, 1.0, 0.0);
vec4 dpdz = (dpwdz - (postProj * dpwdz.w)) / pPosition.w;
float ddepth = abs(dpdz.z);
// -- End depth bias helper section -------------------------------// We now compute the depth of the fragment. This is the actual depth value plus
// our depth bias. The depth bias depends on how uncertain we are about the z value
// plus some constant push in the z direction. The exact coefficients to use are
// up to you, but at least it should be somewhat intuitive now what the tradeoffs are.
float depthval = postProj.z + (0.5 * dzlen)+ (uDepthOffset * ddepth);
depthval = (0.5 * depthval) + 0.5; // put into [0,1] range instead of [-1,1]
gl_FragColor = vec4(depthval, depthval, depthval, 0.0);
}
This shader computes the two depth bias pieces described in section 1.2. These are
used to offset the stored depth value. This is where the notation differs from above, but
the translation is quite straightforward.
//////////////////////////////////////////////////////////////////
//
// shadowreceiver.vert
//
//////////////////////////////////////////////////////////////////

// I N P U T

V A R I A B L E S /////////////////////////////////

uniform mat4 uModelViewProjection; // modelview projection matrix

uniform mat4 uModel;
// model matrix
uniform mat4 uTextureViewProjection;
// shadow maps view projection matrix
17

uniform vec4 uLightPosition;

// O U T P U T
varying vec4
varying float

// light position in object space

V A R I A B L E S ///////////////////////////////
pShadowCoord;
pDiffuse;

// vertex position in shadow map coordinates

// diffuse shading value

// M A I N ///////////////////////////////////////////////////////
void main()
{
// compute diffuse shading
vec3 lightDirection = normalize(uLightPosition.xyz - gl_Vertex.xyz);
pDiffuse = dot(gl_Normal.xyz, lightDirection);
// compute shadow map lookup coordinates
pShadowCoord = uTextureViewProjection * (uModel * gl_Vertex);
// compute vertexs homogenous screen-space coordinates
// Use following line if other passes use shaders
//gl_Position = uModelViewProjection * gl_Vertex;
gl_Position = ftransform(); // uncomment if other passes use fixed function pipeline
}
This is a pretty standard vertex shader as well. The ftransform() function guarantees
the output matches the fixed function pipeline. If the objects you render use shaders
instead of fixed function, then you should do so here as well.
/////////////////////////////////////////////////////////////////////////////////
//
// shadowreceiver.frag
//
/////////////////////////////////////////////////////////////////////////////////
// I N P U T

V A R I A B L E S ////////////////////////////////////////////////

// uniform constants
uniform sampler2D
uShadowMap;
uniform float
uSTexWidth;
uniform float
uSTexHeight;
// per fragment inputs
varying vec4
pShadowCoord;
varying float pDiffuse;

// vertex position in shadow map coordinates

// diffuse shading value

18

// M A I N //////////////////////////////////////////////////////////////////////
void main(void)
{
// compute the shadow coordinates for texture lookup
// NOTE: texture_viewproj_matrix maps z into [0,1] range, not [-1,1], so
// have to make sure shadow caster stores depth values with same convention.
vec4 scoord = pShadowCoord / pShadowCoord.w;

// -- "Percentage Closest Filtering" ----------------------------------------//

//
//
//
//
//
//
//

One could use scoord.xy to look up the shadow map for depth testing, but
well be implementing a simple "percentage closest filtering" algorithm instead.
This mimics the behavior of turning on bilinear filtering on NVIDIA hardware
when also performing shadow comparisons. This causes bilinear filtering of
depth tests. Note that this is NOT the same as bilinear filtering the depth
values and then doing the depth comparison. The two operations are not
commutative. PCF is explicitly about filtering the test values since
testing filtered z values is often meaningless.

// Real percentage closest filtering should sample from the entire footprint
// on the shadow map, not just seek the closest four sample points. Such
// an improvement is for future work.

// NOTE: Assuming OpenGL convention for texture lookups with integers in centers.
// DX convention is to have integers mark sample corners
vec2 tcoord;
tcoord.x = (scoord.x * uSTexWidth) - 0.5;
tcoord.y = (scoord.y * uSTexHeight) - 0.5;
float x0 = floor(tcoord.x);
float x1 = ceil(tcoord.x);
float fracx = fract(tcoord.x);
float y0 = floor(tcoord.y);
float y1 = ceil(tcoord.y);
float fracy = fract(tcoord.y);
// sample coordinates in [0,1]^2 domain
vec2 t00, t01, t10, t11;
float invWidth = 1.0 / uSTexWidth;
float invHeight = 1.0 / uSTexHeight;
t00 = float2((x0+0.5) * invWidth, (y0+0.5)
t10 = float2((x1+0.5) * invWidth, (y0+0.5)
t01 = float2((x0+0.5) * invWidth, (y1+0.5)
t11 = float2((x1+0.5) * invWidth, (y1+0.5)
19

*
*
*
*

invHeight);
invHeight);
invHeight);
invHeight);

// grab the
float z00 =
float viz00
float z01 =
float viz01
float z10 =
float viz10
float z11 =
float viz11

samples
texture2D(uShadowMap,
= (z00 <= scoord.z) ?
texture2D(uShadowMap,
= (z01 <= scoord.z) ?
texture2D(uShadowMap,
= (z10 <= scoord.z) ?
texture2D(uShadowMap,
= (z11 <= scoord.z) ?

// determine that all

viz00 = ((abs(t00.x viz01 = ((abs(t01.x viz10 = ((abs(t10.x viz11 = ((abs(t11.x -

t00).x;
0.0 : 1.0;
t01).x;
0.0 : 1.0;
t10).x;
0.0 : 1.0;
t11).x;
0.0 : 1.0;

geometry outside the shadow test frustum is lit

0.5) <= 0.5) && (abs(t00.y - 0.5) <= 0.5)) ? viz00
0.5) <= 0.5) && (abs(t01.y - 0.5) <= 0.5)) ? viz01
0.5) <= 0.5) && (abs(t10.y - 0.5) <= 0.5)) ? viz10
0.5) <= 0.5) && (abs(t11.y - 0.5) <= 0.5)) ? viz11

:
:
:
:

1.0;
1.0;
1.0;
1.0;

// bilinear filter test results

float v0 = (1.0 - fracx) * viz00 + fracx * viz10;
float v1 = (1.0 - fracx) * viz01 + fracx * viz11;
float visibility = (1.0 - fracy) * v0 + fracy * v1;
// -----------------------------------------------------------------------------// Non-PCF code (comment out above section and uncomment the following three lines)
//float zvalue = texture2D(uShadowMap, scoord.xy).x;
//float visibility = (zvalue <= scoord.z) ? 0.0 : 1.0;
//visibility = ((abs(scoord.x - 0.5) <= 0.5) && (abs(scoord.y - 0.5) <= 0.5))
//
? visibility : 1.0;
// -----------------------------------------------------------------------------visibility *= pDiffuse;
gl_FragColor = vec4(visibility, visibility, visibility, 0.0);
}
This file implements percentage closest filtering. To use unfiltered shadow mapping,
comment out the PCF block as noted and uncomment the Non-PCF block. Note that
after doing this, the uSTexWidth and uSTexHeight variables are likely to be optimized
away and so you should uncomment these variables in the materials script as well.
The following shows how to activate plane optimal shadow mapping given some
pointer to a MovablePlane and a pointer to a light.
PlaneOptimalShadowCameraSetup *planeOptShadowCamera =
new PlaneOptimalShadowCameraSetup(movablePlane);
20

Entity *movablePlaneEntity = sceneMgr->createEntity( "movablePlane", "plane.mesh" );

SceneNode *movablePlaneNode =
sceneMgr->getRootSceneNode()->createChildSceneNode("MovablePlaneNode");
movablePlaneNode->attachObject(movablePlaneEntity);
SharedPtr<ShadowCameraSetup> planeOptPtr(planeOptShadowCamera);
light->setCustomShadowCameraSetup(planeOptPtr);

References
[1] Hamilton Y. Chong and Steven J. Gortler. A lixel for every pixel. In Proceedings
of the Eurographics Symposium on Rendering. Eurographics Association, 2004.
[2] Hamilton Y. Chong and Steven J. Gortler. Scene optimized shadow maps. In
Harvard Technical Report TR-11-06, 2006.
[3] William Donnelly and Andrew Lauritzen. Variance shadow maps. In SI3D 06:
Proceedings of the 2006 symposium on Interactive 3D graphics and games, pages
161165, New York, NY, USA, 2006. ACM Press.
[4] Randima Fernando, Sebastian Fernandez, Kavita Bala, and Donald P. Greenberg.
Adaptive shadow maps. In SIGGRAPH 01: Proceedings of the 28th annual conference on Computer graphics and interactive techniques, pages 387390, New
York, NY, USA, 2001. ACM Press.
[5] Tom Lokovic and Eric Veach. Deep shadow maps. In SIGGRAPH 00: Proceedings of the 27th annual conference on Computer graphics and interactive techniques, New York, NY, USA, 2000. ACM Press.
[6] Tobias Martin and Tiow-Seng Tan. Anti-aliasing and continuity with trapezoidal
shadow maps. In Proceedings of the Eurographics Symposium on Rendering, pages
153160. Eurographics Association, 2004.
[7] William T. Reeves, David H. Salesin, and Robert L. Cook. Rendering antialiased
shadows with depth maps. In SIGGRAPH 87: Proceedings of the 14th annual
conference on Computer graphics and interactive techniques, pages 283291, New
York, NY, USA, 1987. ACM Press.
[8] Marc Stamminger and George Drettakis. Perspective shadow maps. In SIGGRAPH
02: Proceedings of the 29th annual conference on Computer graphics and interactive techniques, pages 557562, New York, NY, USA, 2002. ACM Press.
[9] Lance Williams. Casting curved shadows on curved surfaces. In SIGGRAPH 78:
Proceedings of the 5th annual conference on Computer graphics and interactive
techniques, pages 270274, New York, NY, USA, 1978. ACM Press.

21