i’m writting a shadow map shader in cg. Firstly i render the scene from the lights view.
I grab the depth texture and pass it off to the shader when rendering the scene from the main camera.
Now in this shader, i transform the vertices the same they are transformed when rendered from the lights camera this is so i know the distance from the light and should in theory be able to do a strait comparison between the depth to the light and the depth texture.
Problem is, that there is some difference between the texture and the transformed verts.
in my vertex shader:
float4 sceneDepth = mul(projection, float4(IN.position, 1.0));
OUT.zDepth = sceneDepth;
projection marix is light view * light projection (the mesh is using an identity matrix)
then in pixel shader:
float4 depth = tex2Dproj(shadow, IN.uvCoord);
OUT.colour = depth.r < (IN.zDepth.z/IN.zDepth.w) ? float4(1.0, 0.0, 0.0, 1.0) : float4(0.0, 0.0, 0.0, 1.0);
this is what i get: 
the light in both screenshots are at different distances, its like the texture depth map isnt in the same space as the space i transform the mesh to in my vertex shader.
Not that the camera transform and the light camera transform have the same perspective, aspect ratio’s and close/far clip planes.
any ideas on how to fix this?