How to Render Passes in Cheetah3D (OWIP tips/tutorial)

How to Render Passes in Cheetah3D (OWIP tips/tutorial)

Hi folks!

I was recently in the need of faking some render passes in Cheetah and found some quick n' dirty tricks to create some of them. I know most of you already know these ones (and maybe some others) but I didn't find any "explicit" tutorial about that in this section so I thought they might be useful for someone...

I'm starting this as an Open WIP "quick tips in a post" thinghie, if anyone has some more/better info to share... please do!
also, if the whole thing will grown up enough at some point perhaps it would be worth to make a real .pdf tutorial... who knows?

so, here we go:

1) build up a scene as usual, place objects, assign materials, tweak camera and lighting settings and so on.

2) when you're fine, group and make a copy of the entire scene; if the scene is quite complex simplify it as long as you can (i.e. apply modifiers, import children etc.); delete all materials; make a copy of your main camera too, keeping all the tags you used for your standard render (hdri, dof, etc.)

3) the basic concept is to use a copy of the main camera and the scene for each pass, playing with camera settings/tags, materials and render tags. (you can actually keep one copy for each pass OR use multiple camera with a single copy of the scene and change materials/settings for each pass... it's up to you and probably depends on scene complexity/polycount)

4) start rendering one pass at once, using the appropriate camera and "visible in rendering" toggle in Mode tag to set what will be shown in the render. to keep it simple (and short), camera, tag and material settings are shown in each sample image (any other value is set to default or is the same as the original item)

5) save each render pass, open it in your favorite image processing tool (I use PhotoShop or Pixelmator) and have fun. (note that some of the passes may need adjustment like levels, power/opacity, blending mode etc...)

note: these sample images are from my personal project; an example file (with very basic objects and standard C3D 5.5 materials/hdri) included.

more to come...

cheers,
Alessandro

P.S.: I'm quite sure I messed up something either in the example file or image "caption" infos... please report any oddity you'll notice

alternative example of Alpha pass (objects only)

and a quick composition example (Zdepth used for DoF, Alpha for masking, AO to enhance shadows/contrast)

thanks for sharing the depth fog trick !

archie said:

thanks for sharing the depth fog trick !

Click to expand...

you're welcome...

of course the first thought was a StateNode driven ZdepthMAT: it "almost" worked but I wasn't able to set up the scene as easily (and univocally) as per the fog tag.
look at the 4x4 cubes and 10x20 planes in the shot:
1) is X=Y=Z=4 and scale set to 1;
2) is X=Y=Z=2 and scale set to 2;
3) is the same as 1, after "make editable" and burn scale (obviously nothing changes); pivot point moved to foreground face (does nothing)
4) is the same as 2, after "make editable" and burn scale;
A) is a 2x2 plane, scaled at 5x10;
B) is a 10x20 plane, scaled at 1x1;
so my (simple) mat only works right with "post-scaled" [-1;1] objects due to the state node output... so a correct Zdepth pass (2nd shot, left) needs a 2x2(x2) plane/cube scene reference and all other object have to be dropped onto it and "imported" as children... then came the for trick :smile:

cheers,
A

yeah, i too made a depth map mat ago (pig thread) and found the necessary tweaking annoying, the advantage of the fog tag is the grid help in the editor.
there also is the cache epix option in the render prefs though i haven't tested in 5.x yet.

- archie