copyright © 2006 Andrew G. Morgan
OK, whilst this is the 'advanced' guide to passes, you've already tackled the tricky stuff (well, assuming
you read the other guide). What we're doing here is taking all the stuff you already know about passes
and extending it in order to make it even more flexible and to enable you to 're-light' rendered shots.
You've already seen how it's possible to reconstruct a shot using the various passes supplied by
Buffer Saver, the next step is to break down the lighting in your scene so you can render separate
lighting passes for reflections, diffuse lighting, shadows and specular shading. By combining these
with your other passes (raw RGB, depth, ambient occlusion etc.) a compositor can easily tweak
individual light intensities and shadows in your shot without you having to re-render it - neat!
When working on feature film visual effects, this level of control over a shot is essential - typically,
film fx shots are broken down into dozens of passes - what we're going to demonstrate here isn't that
advanced but it should give you a feel for how it works. Take a look at the Photoshop project below:
(You can download this PSD here)
What I've basically done here is render my shot twice - once with just the key light turned on (as
seen in the 'Key_Light_Render.png' above) and once with just the fill light. After putting all the passes
together in Photoshop I brightened the shadows and the ambient occlusion to get a nice balance
in the lighting - not bad when you consider the shadows in the renders are completely black...
The number of renders you do is entirely based on your lighting setup and your compositing
requirements - i.e. if you have 5 lights in your scene you may want to render it 5 times. Typically,
this is unnecessary though and separate passes for key, fill and a single pass for any remaining
lights should still give you plenty of flexibility in comping your shot.
Obviously, this level of complexity in your comps won't always be required - particularly as it requires
re-rendering of your scene but if you've got the time (or render nodes) then it can really help in getting
the exact look you need.
And now the bad news...
OK, so you've got a pretty solid grounding in passes now - you know how to get them out of
LightWave, you know which ones are useful and which ones aren't so useful. You also have a good
idea of how they're put back together into a comp (be it in Photoshop/Shake/Fusion etc.)
The bad news is that, unlike Maya or XSI you can't get all the info you might need out of LightWave.
Refraction in particular is a problem as there's no buffer for this. If you're rendering something with
refraction (like say the crystal ball in the image below) you or your comper will have to isolate the
refractive elements from the final render (possibly using an alpha channel).
Buffer Saver can't output a refraction pass
In addition, all the info in these guides specifically relates to object/scenes without any 3rd party
plugins (well, apart from Buffer Saver) - this is strictly out-of-the-box LightWave stuff. If you use shader,
image filter or volumetric plugins (HD_Instance for example) you may get unexpected results -
try and do a quick low-res test to get an idea of what Buffer Saver will output before sending
your shot off to the render farm.
Is that it? Can I go now?
Nope - whilst rendering in passes is incredibly useful, it's only half of the equation - if you want
Real Ultimate Power as a CG artist you need to know about Rendering in Layers too - see you there!
Home