ok i found the error and I saw that it isn't necessary to create the SRV in every frame...

I'm not sure if i have to create a new shader resource view every time...

I created a ID3D10Texture2D object and a ID3D10RenderTargetView. on rendering I'm doing the following: pDevice->OMGetRenderTargets(1, &pFirstRenderTargetView, NULL);pDevice->OMSetRenderTargets(1, &pSecondRenderTargetView, NULL);pDevice->ClearRenderTargets(pSecondRenderTarget, D3DXCOLOR(0.0f, 0.0f, 0.0f, ...

can anybody tell me how I'm able to render to a texture instead of using the back buffer? thx!

when I'm describing the d3d behavior with the help of DXGI_SWAP_CHAIN_DESC I have to set the buffer usage. what is the difference between DXGI_USAGE_RENDER_TARGET_OUTPUT und DXGI_USAGE_BACK_BUFFER?

melvil: will you be at the games convention?

for texture splatting i used an rgb-map. each channel holds 8 bit for every pixel. with that information i'm able to get a percentage for each pixel and each texture (max: 4 textures). with a 2d graphics application you are able to draw the image for each channel.

i didn't think about using ray tracing to cull geometry yet. i would assume that you have to create a bunch of rays pointing to the direction the camera is looking at (sorry, my english is crappy).  after that you check if one ray intersects your bounding box. if so, you could check against the triangles too. but first of all you have ...

i would recommend to implement a lod system in order to optimize the quad tree. another possibility is to wait for occlusion culling in XNA 2.0.

try to draw the geometry with PrimitiveType.PointList so you are able to see if the vertex positions and indices are correct... melvil:will you publish the source? I'm interesseted in the lod-code...