flipcode | Daily Game Development | Programming 3D Games and Graphics | Message Center - Shadow map in Pixel shader texRECT(samlperRECT tex, float3 sz)

March 29, 2005, 06:53 AM

I m learning how to program shaders in Cg. I 'm working on shadow mapping. There is a function called texRECT(samlperRECT tex, float3 sz) that I use to perform a depth test. I think I have set up the texture correctly in openGl to perform the depth test. What I don't understand is the use of z (in the function texRECT). In Cg toolkit manual they say "z indicates a depth comparison value for shadow map lookups". What I am suppose to give to the function ???

And then the result of this function is a float4 but where is the result of the depth test ?

In fact I would like to do the same thing that in assembly code

TEX R0.x, fragment.position, texture[0], SHADOWRECT;

Reedbeta

March 29, 2005, 11:14 AM

I don't know CG, but there is probably a function called something like texRECTShadow(samplerRECT, float3) to do the shadow comparison. You put the Z value that you want to compare against into the Z-component of the texture coordinate, if this Z value is greater than the value stored in the texture it returns a vector of all 0's, otherwise a vector of all 1's.

SirKnight

April 14, 2005, 11:06 PM

The tex coords you use when doing the shadow map lookup should be the world space vertex transformed by a matrix that is the combination of light view matrix, light projection matrix and a scale/bias matrix to convert the range from [-1,1] to [0,1]. If you are not going to use a proj version of the texture lookup function then you will have to do the divide by w yourself to the tex coords before using them in the shadow map lookup. You should use texRECTproj so it does the divide by w for you.

Here is a snippet of a cg vertex program that I use for shadow mapping:

 
vertexOut vout;
        
mv  = glstate.matrix.modelview[0];
mvp = glstate.matrix.mvp;
        
float4 epos = mul( mv, vert.pos );      // Eye-space vertex position
float4 wpos = mul( cvi, epos    );      // Transform back to world space
float4 proj = mul( tex, wpos    );  // Transform from world space to projected texture space
 
vout.hpos       = mul( mvp, vert.pos );
vout.projCoords = proj; 
vout.tex0       = vert.tex0;
vout.normal     = vert.normal;
vout.v2light    = float4( lightPos.xyz - vert.pos.xyz, 1 );
vout.view2v     = float4( vert.pos.xyz - eyePos.xyz, 1 );
 
return vout;
 

BTW, cvi is the inverse camera view matrix and tex is the matrix that has the matrix products: [Scale/Bias]*[Light Proj]*[Light View].

And here is the fragment program (note I use 2D texures):

 
float4 finalColor;
 
float  shadow   = ( vert.projCoords.w < 0 ) ? 0 : tex2Dproj( shadowMap, vert.projCoords ).r;    
float4 lightTex = ( vert.projCoords.w < 0 ) ? 0 : tex2Dproj( projLightTexture, vert.projCoords );       
 
float2 tex0 = vert.tex0.xy;
 
float4 Kd       = tex2D( rgbTexture,  tex0 );
float4 Ks       = tex2D( specTexture, tex0 );
 
float3 normal  = normalize( vert.normal.xyz  );
float3 v2light = normalize( vert.v2light.xyz );
float3 view2v  = normalize( vert.view2v.xyz  );
        
float3 l2vert = -v2light;
        
float diffuse, specular;
 
float3 R = reflect( -v2light, normal );  
 
float4 lighting = float4( 0, 0, 0, 0 );
        
lighting = lit( dot( normal, v2light ), dot( R, -view2v ), specExp );
 
diffuse  = lighting.y;
specular = lighting.z;
 
float3 finalShadow = shadow * lightTex.xyz;
        
finalColor  = float4( ( Kd.xyz * ambient ) + ( Kd.xyz * diffuse + Ks.xyz * specular ) * finalShadow, 1 );
 
return finalColor;

-SirKnight

killah

May 25, 2005, 06:06 AM

If I'm using OpenGL and, once configured the camera, I want to transform "Camera View Matrix" to "Inverse Camera View Matrix" (your "cvi" variable) to send it to the vertex shader, which matrix I have to use? (MODELVIEW, PROJECTION...), at first I thought that I had to use MODELVIEW but I'm not sure at all...

Cheers

This thread contains 4 messages.