Question Help me understand Position node in Fullscreen material mode (shader graph)

Unity 2023.1.7f1

When hooking up a Position node (using Object space) directly to an Unlit shader graph's color output, you get a predictable result.


Example using a simple test scene with some random objects (ignore the outline effect):


However, when doing the exact same thing in a Fullscreen shader graph, I'm not sure what it is that I'm getting as a result:


Could someone tell me what it is I'm getting exactly when using the Position node in a Fullscreen shader graph?

And of course, the data I'd like to have is the one from the first instance. Is there any way to get the same data in a Fullscreen shader graph?

EDIT: Seems like this is really only an issue in orthographic mode (when combined with fullscreen shader graph material) ... ?

 

After a few hours of fiddling around, I was able to recreate the world space positions by sampling the camera's depth texture


Based off of Cyanilux's post on depth (https://www.cyanilux.com/tutorials/depth).

Still, in case anyone could enlighten me on what exactly we're getting from the Position node (object space) in a Fullscreen graph (orthographic), that'd be appreciated.

 

I am also confused about the Position Node in FullScreen Shader. In World mode, the Position Node seems to obtain the world coordinates of each pixel in the current frame buffer. Instead of the world coordinates of the object where this shader is located, as usual. This point is poorly explained in the manual.

 

@NymThePseudo hello can you share the GetCameraInvProjection Custom Function please ?