Lighting Vs. Renderer

After I just finished around 100 shots for two video game promos, I became truly aware of what made those shots pretty to look at versus how technically realistic a lot of vfx shots look these days. Granted, the shows were supposed to look cartoony, and I had only myself to fall back on, and I had about 1/24th the time allowed a Pixar film, but the images looked decent and satisfied the client nonetheless. And this time around, I found myself finally going back to artistic decisions and fundamentals instead of technology to get my final images.
This brings up my point. It is not really necessary, and often unappreciated, to bring a crazy amount of technical realism to the final images. What is really important, however, is that you imply realism where needed and focus it where it counts. For example, I found myself needing to produce a jungle type forest at night, with character lighting and set that would have a feel similar to Pixar's The Incredibles. So off I went, looking into all types of forest and vegetation simulators, terrain generators, and rendering technology. After about ten days of R & D, I realized that the animatic suggested lots of shallow depth of field shots. I realized that I was trying to mimic reality, so I could then blur it beyond recognition in post. In any given shot, I would only see maybe 15 trees and bushes in all their high quality glory, the rest could be lower resolution textures and geo, even cards.
Ever since I perused the matte painters' web sites - (http://www.goodbrush.com/ ... http://www.mattepainting.org ...etc) , I started to try and incorprate more art into my technical shots. Color hue/ value are so important to understand how we perceive sharpness, foreground/background, and what is the focus of a scene. Then there are the limitations and strengths of the human vision and perceptual systems that play a big role in getting a good shot. For example, lighting reds, oranges, yellows up front, and having greens, blues, purples in the background automatically creates a sense of depth to the viewer due to the speed at which our brains process different wavelengths of light.
What surprised me was that I was able to achieve the result I wanted with very little overhead (ie. rendering time) just by going back to art fundamentals of light, hue, values, and composition. In fact, I used 3dsmax's scanline renderer for everything. As long as I let my textures and surface properties do most the work, I could get great looking elements very quickly - about 30 sec to 1 minute per frame. If something was in the shadows or was getting blurred, all I cared about was its value/hue, not its detail. And if it was off camera, it did not exist. It was very liberating not having to worry about details in the renderer that were going to get lost in comp. In the end, lighting was what made the images, not the renderer.
Some stills are here: http://mmoses00.cgsociety.org/gallery/