I’m looking at the official “The Matrix Awakens” project to implement the Fake 3D geometry wtih parallax, and I’m confused by the “front-back depth map” capture. I don’t know how to make a depth map like the official one.Is there anyone who can help me solve this problem or provide a tutorial on it?
Latest Progress：
I can’t create a corresponding depth map texture, but I can do it in a roundabout way, by rendering the target to create a depth material of the corresponding object instead.
Is there any way to create static textures from R32f format render targets?
While I wish I could help you here, I myself would love to learn exactly how they pulled their false interiors trick off. It does appear to be Parallax Occlusion mapping through a heightfield generated with stencils of given objects, jittered to create the look of 3D meshes, and a cubemap as the background.
If anyone, a dev or someone familar with the technique could elaborate, I would also be extremely grateful.
I could have sworn I found a developer stream that actually went exactly into how they accomplished the interiors- the materials, how they raymarched the stacked textures, all of it. I just can’t remember where it is.
It’s actually not all that difficult, it’s just straight parallax mapping, doing depth intersection tests against two depth maps. Hammersley is used to dither the slices, I don’t know why hammersley specifically was chosen though.
What I find much more confusing about it is how on earth they got the built-in interior cubemapping function to work with non-cube rooms. That whole node is black magic to me, I can’t tell what the hell it is doing, and of course… it’s not explained here.
