Newbie Qn: RPF, occluded objects, and layers.

[Anonymous]

Hi All,

I have a simple questions regarding if the RPF format is able to achieve
this:

In my 3dsmax scene, I have 3 objects, A, B and C.
From the camera viewpoint, A is nearest the camera and fully visible.
B is behind and partially blocked by A, and C is behind B and partially blocked
by B. Thats the full 3d scene.

Now I want to render A and C in 1 pass and B in another pass.
Can I render an RPF with just A and C (B hidden)
and then render another RPF with just B (A and C hidden),

And the compositor package (like combustion) would know how to
slot in B between A and C using whatever needed G Buffer info?

Thanks!

[azfar mohd azhar]

You’d might use z-depth. I’d render all the layers seperately in single passes. That way you won’t have to muck about using extra channels for a simple comp. When you render your passes don’t hide the objects that aren’t there, just make them invisible to camera(right click on the object and choose object properties). That way you’ll get reflections and shadows, and have more control over your individual passes. z-depth in rpf’s is not anti-aliased so you might get a weird halo around your objects when you composite using that method. If you really want to do two passes give your objects unique object id’s and use an object id render element(or seperate matte elements for each object) That way it’s anti-aliased and you don’t run into zbuffer issues at all.

[azfar mohd azhar]

You might use z-depth but don’t. I’d render all the layers seperately in single passes. That way you won’t have to muck about using extra channels for a simple comp. When you render your passes don’t hide the objects that aren’t there, just make them invisible to camera(right click on the object and choose object properties). That way you’ll get reflections and shadows, and have more control over your individual passes. z-depth in rpf’s is not anti-aliased so you might get a weird halo around your objects when you composite using that method. If you really want to do two passes give your objects unique object id’s and use an object id render element(or seperate matte elements for each object) That way it’s anti-aliased and you don’t run into zbuffer issues at all.

[Author]

Posts