So a Dedicated buffer is no longer needed
While we're at it we need a place to put Roughness, since we currently store that inside the Position buffers alpha channel.
We can compress Normal down to 2 floats, so Rename the Normal buffer to Surface, and pack like so:
Normal into X and Y
Roughness into Z
Metalness into W
Position buffer stores the View Position of each pixel But we can calculate that anywhere with:
So a Dedicated buffer is no longer needed While we're at it we need a place to put Roughness, since we currently store that inside the Position buffers alpha channel.
We can compress Normal down to 2 floats, so Rename the Normal buffer to Surface, and pack like so: Normal into X and Y Roughness into Z Metalness into W