Isn't the 3D just rendering two images, then sending one to each of your eyes through some screen positioning black magic? It should be doable to replicate these 2 images with a capture software, or even just emulating the 3D with the streamed image, though you wouldn't actually be able to...