I am currently trying to get the Depth Buffer as a texture to use it for edge detection algorithms. Since XNA4 does not allow direct access to the Depth Buffer as texture i have to render it into a texture. The effect i currently use looks like this:
float4x4 World;
float4x4 View;
float4x4 Projection;
struct DEPTH_VS_OUTPUT
{
float4 Position : POSITION;
float2 Depth : TEXCOORD;
};
DEPTH_VS_OUTPUT DepthTextureVS(float4 Position : POSITION)
{
DEPTH_VS_OUTPUT output;
float4 worldPosition = mul(Position, World);
float4 viewPosition = mul(worldPosition, View);
output.Position = mul(viewPosition, Projection);
output.Depth = output.Position.zw;
return output;
}
float4 DepthTexturePS(DEPTH_VS_OUTPUT input) : COLOR0
{
float4 result;
result.rgb = input.Depth.x / input.Depth.y;
result.a = 1.0;
return result;
}
technique DepthTextureRender
{
pass P0
{
CullMode = NONE;
ZEnable = TRUE;
ZWriteEnable = TRUE;
AlphaBlendEnable = FALSE;
VertexShader = compile vs_2_0 DepthTextureVS();
PixelShader = compile ps_2_0 DepthTexturePS();
}
}
It looks like this:
The cubes in this scene are 1x1x1 the projection matrix is made like this:
var projection = Matrix.CreatePerspectiveFieldOfView(MathHelper.PiOver4, 16f / 9f, 0.1f, 1000f);
As far as I understood the Depth Buffer concept z/w should give me the distance from the camera position to the pixel as value between 0-1 where 0 would be the Near-plane and 1 the Far-plane but since the cubes in this scene are only about 5 Units away they shouldn't be white.
