z-depth pass on smoke sim

Hello all, I’m trying to figure out how to run a z-depth pass on a smoke simulation. My problem now, if I render the smoke and mask out the layer with the object, the entire layer with the object is rendered alpha and no smoke is visible within the alpha parts of the render. What I am looking to do is have the smoke render with the object layer masked out in alpha, but still have the smoke that should be in front of the object render visibly.

Any ideas?

how do you define a zdepth based on a volumetric material? the reason why we see something which is volumetric is because it is dense… light penetrates into it a bit but then disperses… if it was a straight zdpeth pass, which value to you assign to the pixel? do you assign the value of the first point in space that it hits with the volumetric material? or do you set a threshold? then it wont actually align properlly to your scene…

this is why deep compositing is awesome… its not in blender atm, but what it does is it stores multiple samples per pixel… so you can hold true volumetric data in it, aswell as geo zdepth values without any aliasing effects.

first step for blender to support this sort of thing is to support openexr2 first… then to be able to write / read this data… then for the render engines to save out… its months / years away, but i am sure it will happen.

Then save up aaaaaaalllll of your money for storage, cause deep comp sucks the life out of memory/hdd allocation!

Heh, i have been dealing with 100+meg multilayer exr image sequences already… putting multisample data in there aswell will turn it into gbs i think.

Read the fxguide story on Rise of the Plane…blah blah and it was stunning how much data the deep field soaked up!

So, thanks for the replies. I ended up finding out what was wrong. I have a few main objects, that are cycles materials, which by default, render out in the blender internal as alpha, which is awesome. And of course have the smoke layer, which is only a blender internal function. My problem, which could or may be a bug, which I will report and probably get trashed for, was my camera’s location. through trial and error, I moved my camera outside of the smoke domain box and got perfect results with the depth of the smoke layer. (meaning smoke partially covering some objects based on their location in 3d space). If I move the camera anywhere inside the box, I will only get the smoke rendering and no objects.

you may need to change your transparency type to raytrace instead of z-transparency