Fooling the eye is a tricky thing and duplicating reality is no easy task.This turorial shows one way to get a relatively fast photo-real result for a project using Autodesk Maya.

While Maya does have what we need to create just about anything, we’re always looking to see where we can cheat to get the same result faster and how we can optimize in order to save render time without compromising on quality. Here’s one way to get a relatively fast photo-real result for a project that integrates a CG car into a high-dynamic range environment. In this image, we have a sphere attached to a surface shader with an HDR image.

For the ground, a pre-rendered image of the world sphere is projected through the same camera on a surface shader. Ironically, this approach to lighting starts by removing lights from the equation. What’s most important in this case, and in all photo-real projects, is observation- no shader or tool will achieve that for you. If you look closely in the world surrounding you, most lighting environments (indoor or outdoor) come from indirect sources, even in direct sunlight. After all, the sun is just a big point of light. The main issue here is to get the best ambiance/diffuse values as a solid starting point and later fine-tune our shading and lighting on top of that. This way, you’ll optimize your workflow to get the best integration.

First, you need to render a final gathering (FG) pass to get color sampling from the HDR sphere. Maya 8.5 offers some new and improved features that generally render FG much faster than previous versions, although the default values might work on simple scenes. I still prefer using the Radius controls to fine-tune the result. The entire car is assigned to a basic white lambert with a diffuse value set to 1.

Tip on using FG: You should scale all your scenes to fit into a 100 x 100 Maya unit. This way, you won’t have to start all over again for the next project you’ll be working on.

You should also render an ambient occlusion pass that includes the ground. This one is fairly easier; just increase the samples to 64 or higher to get good results.

By assuming that your camera and animation won’t change, those two elements will always stay the same and you’ll save a lot of time by re-using them. They’ll also render faster because you won’t need motion blur at this point. If something changes you’ll have to re-render those two sequences.

Obviously, this can all happen in comp, but remember that we’re trying to optimize the process and get faster feedback while fine-tuning. You can save a huge amount of time by using camera projections.

We’ll build a shading network that will act as a comp node by combining our two passes through a layered texture.

Note: You could add many more layers to it like pre-rendered shadow passes.

Here’s a simple comp-shading network to integrate the ambient values. The ambient occlusion pass is multiplied over the final gathering into a layered texture node. You also should add a blender to have more control on the black levels, which sometimes need to match the values from the background plate. Use a camera projection to connect in the ambient of the shader.

Let’s make sure the camera projection is set to Proj Type -> Perspective and in the camera Projection Attributes. It should link to the same camera you’re rendering your shot with. You might have to try a few settings for the projection to match the camera resolution (Maya does not provide much visual feedback for this) but once it’s done, you’re all set to build your shading and lighting.

The car now feels right in there when combining the occlusion and the final gather, even when assigned to the default shader. We can now concentrate on the shading. The final result will clearly depend on the quality of your shaders (rubber, chrome, plastic, etc.) but by connecting this ambient network to all your materials- even the Mental Ray car paint phenomenon- they’ll acquire all the subtleties that you need to achieve the photo-real look. This image was rendered in one pass only in less than three minutes at HD 720p resolution.

Although I’m not suggesting using Maya as compositing software, it doesn’t hurt to use it to pre-comp. Once you’re happy with the result, you can still re-render your source images with motion blur.

From an artist’s point of view, even if you have access to high-end compositing resources to put the final touch to your work, it’s extremely valuable to be able to see the big picture as you’re working. This approach will allow you to spend the time where it matters the most- in the details.

Patrick Gray
Digital Supervisor
Reel FX Creative Studios
Dallas, TX

Patrick Gray is a 3D generalist with more than 12 years experience in all aspects of CG production. Patrick comes from an architecture and design background, which led him to his specialization in photo-real lighting and rendering. He’s worked as the lead of the lighting team at Meteor Studios in Montreal where he took on many feature film projects, such as Fantastic Four and Final Destination 3, and television specials for Discovery Channel. Currently at Reel FX in Dallas, he’s devoting himself to commercials.

Patrick Says Keep in Mind…
Efficiency is the key to any successful project. Planning right from the start is what will save you time (and money), allowing you to focus on the most important things, therefore finishing the project on time and with a higher quality level. It’s particularly true when it comes to photo-real projects. Fooling the eye is a tricky thing and duplicating reality is no easy task.

Pay attention to details. Observe methodically; it’s often those small features that will make all the difference. Be your own judge and be honest about it. If you can’t believe it’s real, nobody else will.

Reel FX Creative Studios
www.reelfx.com
301 N. Crowdus St.
Dallas, TX 75226
ph. 214.979.0961