Monday, January 24, 2011

On 3D Modeling

It has been a long time since my last post, so allow me a quick, unpolished post this time.

I have been learning how to create 3D models the last 10 months. Here are what I learned from my modeling lessons.
  • Modeling is basically putting vertices on 3D space in such a manner (for example, 2 vertices are connected to form an edge, 3 or more edges form a polygon) that the computer can create an image that we can recognize as objects.
  • Polygon modeling using 3ds Max, to me, means defining edge loops at points important to define a shape; then let the computer interpolate the area in-between (using the TurboSmooth modifier). As a computer science graduate, I see this style of modeling as a way to selectively add details.
  • Modeling using Mudbox begins with defining a basic shape (called the base model). Vertices on the base model should be spread as uniformly as possible; equivalently, quads on the base model should be as close as possible to a square. Next, we sub-divide the base model and add details by pulling out or pushing in vertices. As a computer science graduate, I feel this style of modeling reminiscent to brute-force algorithms. We rely on the computer hardware to handle the large amount of data generated as we keep sub-dividing the base model in order to add finer and finer details.
I found myself agreeing a friend when he said that modeling using Mudbox is just using a fancy FFD (Free Form Deformation). However, further use convinces me that the tools provided by Mudbox are much more powerful than FFD. The difference among the tools are subtle enough to fool me into thinking that sticking to one tool is enough. They, however, serve different purposes and are useful on different scenarios/problems.

All in all, I find modeling using Mudbox much closer to drawing. I can concentrate on getting proportions right, enhancing shapes, and other high-level thinking instead of worrying about the topology of the model.

No comments:

Post a Comment