My latest attempt on SSAO. In a next step I’m going to implement SSDO and this demo is supposed to serve as a basis for that.
It uses a linear depth buffer which is packed into a A8R8G8B8 render target. Check this article for details. Position reconstruction is done via the frustum corner method as described here. The amount of occlusion contribution of taken samples is based on this tutorial. However I replaced the sampling pattern with a basic spherical filter kernel. No blurring is done on the occlusion term. The final image is antialiased with Nvidia’s FXAA filter.
Images below are done with RenderMonkey.