As you all know that 3D Rad generates shadows using a method called shadow volumes, which works by casting a 3D mesh on other meshes. Now the problem with this method is it's really outdated, slow, works horribly with animated meshes and generates artefacts when using high-resolution models. I wondering if it would be possible to implement a different method called shadow mapping, which works by getting the z depth pass of the camera and using it as a reference to generate a texture. This texture is cast on other meshes to generate shadows. Not only this method solves the above problems but also is used some highly popular engines like Unity and Unreal Engine 4.
The only that may be available is through shaders HLSL but don't ask me how as I have no idea about shaders. But I do remember some basic discussions about it back on the original forum and I don't think anything was resolved then.