Depth & Overlay Cameras

[cobalt oyster]

Heya!

I have a setup where a camera renders the world to a 640x360 Render Texture. That's passed to a shader graph material, that's then passed to a URP renderer that just renders that 640x360 texture to the screen's full size via a display camera. Pretty simple, just to get a pixelated look.

My problem is that I'd like to use an overlay camera for things like item sprites and other sprites in general, because if they're just rendered by the base camera in-world, their resolution is 1/3rd of what it should be, and is therefore jittery/wobbly when upscaled. I fixed this using an overlay camera on the display camera, but because I'm passing a render texture to the display camera through a shader graph material, it doesn't have access to the render texture's depth.

Therefore, the sprites are just on top of everything. (See attached videos for the comparison. The choppier ones look terrible but sort correctly.)

Of course, using the overlay camera on the base camera instead inherits depth just fine, but then it runs into the issue of looking choppy again.

Is it possible to just write a fullscreen shader that does nothing but sample and pass a render texture's color buffer + depth buffer to the fullscreen pass renderer feature?

[cobalt oyster]

Just bumping this, still looking for a solution.

[proven meadow]

Yeah, sounds like you need some custom pipeline scripting and grabbing the depth map from the first rendering then compare against it with the overlay objects.

The alternative is do this pixel effect local to the shader and pass on the blitting of the world render

Oh, but then you'd also have to deal with how the shadow is projected with this effect, so your idea does make things easier.