Zdepth jaggies

[kriscabrera15]

Hey guys am trying to figure out how to get ride of the zdepth jaggies on the edges when rendering a depth pass, i though i remember seeing a filter option in a previous build... i could be wrong

[kimgar]

there's an outline checkbox in adaptive AA sampling that should fix this

[kriscabrera15]

i am not using adaptive subdivision am using Adaptive DMC

[dlparisi]

The filter option is still there - part of the "Selected Element Parameters" groupbox right under render elements list (i.e., it's not part of the Vray ZDepth Parameters" rollout.

[Macker]

AFAIK zdepth isn't antialiased, and with good reason - the greyscale value represents a distance to an object, thus if the edges were antialised it would essentially "move" the edges either closer, or further away than they are, thus skewing the zdepth.

[dlparisi]

AFAIK zdepth isn't antialiased, and with good reason - the greyscale value represents a distance to an object, thus if the edges were antialised it would essentially "move" the edges either closer, or further away than they are, thus skewing the zdepth.

This isn't what the OP was asking about, but in general you're correct. If you're doing z depth compositing or a post DOF you'll want it to be aliased for the reason you mention. There are times though when you may want an antialiased z-depth element where an aliased one would lead to problems around the edges. For example, I routinely render out an antialiased zdepth pass to be used as a mask in photoshop to add some distance fade or fog to a shot.

[kimgar]

i think the jaggies in question is not AA related. i've had this before, but now i can't remember how i solved it except with the edaptive outline option. i think it was related to rendering without light or shadows or something...

[rivoli]

i think the jaggies in question is not AA related. i've had this before, but now i can't remember how i solved it except with the edaptive outline option. i think it was related to rendering without light or shadows or something...

in that case, if it was a problem of bad sampling, raising the min subdivisions for the dmc sampler should solve the problem.