I have read some articles on shadow maps, and looked at some demos. Most Demos focus on spotlights, and I have found a couple that have implementations for directional lights (NVidia and ATI) The problem is that for the implementations there are tons of magic numbers being used, and it is unclear to me how the ortho projection for the directional light is being built. I was hoping somebody could either explain the "vanilla" ortho projection derivation for directional lights to me or point me to some good resources. I am not interested on advanced techniques (PSM, TSM, etc, etc), just the basics. Some examples of the code I have found: // ATI shadow map sample // Projection for directional light D3DXMatrixOrthoLH(&matProj, 5.0f, 5.0f, 0.0f, DEPTH_RANGE); D3DXMATRIX matShadowModelMVP, matShadowFloorMVP; D3DXMatrixMultiply(&mat, &matView, &matProj); D3DXMatrixMultiply(&matShadowModelMVP, &matWorldModel, &mat); // Texture adjustment matrix D3DXMATRIX matTexAdj(/*magic numbers here*/); D3DXMatrixMultiplyTranspose(&m_matShadowModelTex, &matShadowModelMVP, &matTexAdj); For example, I do not understand why the value 5.0 is used for width and height for the projection, or why they have a adjustment matrix.

Quote:Original post by SimmerD
Nothing comes to mind. Which part are you stuck on?

I actually took out the orthogonal projection for directional lights out of my engine to simplify the code. Since I am doing a separate shadow map for each shadow caster, I can treat a directional light as a distant point light.

Humm...why not still make the seperate shadow maps orthogonal though? Just to kind of unify the shadowing system?

I got the directional shadow map samples from both the NVidia and the ATI sites:

http://www.ati.com/developer/samples/dx9/ShadowMap.html
http://download.developer.nvidia.com/developer/SDK/Individual_Samples/samples.html

The Nvidia sample if for prespective shadow mapping, but it also has modes for directional lights with plain shadow maps.