3D Modeling for Profit
3D Modeling for ProfitBy Nathan Segal. August 8, 2003.
In an earlier column, I wrote about the Effective
Export/Import of 3D Models. This month, we’ll look at the idea of
creating 3D models for profit, with the intention of selling licenses to use
models for commercial applications. This is a variation of stock photography,
where images are licensed on a limited basis for a variety of advertising purposes.
This can create a decent, ongoing royalty stream.
To find out how to build 3D models for licensing, I spoke with Tom Avgikos, Director of the Digital Assets Division of Digimation, who offered some valuable pointers to people who wants to create models for resale. He said: "The first thing to consider is the issue of Polygonal vs. NURBS models. Our collection has a lot more polygonal models than NURBS models. And we find that in general, people are more interested in polygonal models than NURBS. But if the option presented itself, it’s much easier to convert NURBS models into polygons, but not from polygons to NURBS. It’s much more difficult and I wish it weren't because we do get requests from time to time for somebody who wants a model in a NURBS format and it’s not easy to do."
N: Are there translation formats that handle that, or do you have to start from scratch?
Tom: "One program that we’ve heard of is made by Baren-Boym. They market a tool called ShapeWorks, an add-in for SolidWorks. ShapeWorks is able to take a polygonal model and strip away the faces, leaving a point cloud. From there, they rebuild the model in NURBS using the point cloud as a kind of a template, but it’s pretty labor intensive."
N: Is it mostly a stylistic choice to use NURBS as opposed to polygonal modeling?
Tom: "Well, if I had to choose a model format, I’d choose polygonal, because so much of our business is in that space, but because of the translation problems, I would choose NURBS so we could offer models in both formats."
N: I’ve heard that some people prefer to create their spline paths in applications such as Illustrator instead of a 3D program. What are your comments about that?
Tom: "What’s important to us is the final product, the construction of the model and also the format that it’s delivered in. We need a format that’s translatable. With polygonal models, we want a basic format like .OBJ or .3DS. Now there are an awful lot of tools out there, plug-ins for Max which create hair, cloth, etc. On our end, it’s important to create models that don’t use any of those plug-ins. In other words, what we’re looking for is a bare-bones model because it’s important to deliver a model that’s available to the widest audience. And if a user needs to have a hair package in order to use your model, you’re really limiting your audience or potential user base."
"As an example, if you have a bee that was created before a hair package was available, then that bee hair has actually been made out of vertices. And that is a huge amount of work. It would be easier to deliver a bee without the hair and leave it to the end-user to put the hair on. You might also create a version with hair as well, with instructions on needing to use a particular plug-in to achieve this result. But I believe that basic models are the best way to go. They translate better. Also, there’s no one dominant package out there, so you want to make your model available to as many users using as many packages as possible."
N: What’s your policy when you do a translation and you find a model with holes or breaks?
Tom: "First off, translations have gotten better. If there is a problem such as holes or interpenetrating geometry, etc. and if we catch it before we publish it, we will ask the developer to correct. If not, and the customer finds it, we’re have a dilemma. We have to either let the customer return it if they’re unhappy or give them a discount or make some other arrangement. But if we catch that up-front, we don’t put those kinds of models out there."
"Very detailed models can take awhile to evaluate and what we do here is to look at a wireframe view, a shaded view and a lot of those flaws show up in a shaded view. If you can spend five minutes rotating a model around in a shaded view and it looks good, chances are that it’s a good model. And that will take care of the issues of seams, holes, gaps, uneven surfaces and interpenetrating geometry."
"In a wireframe view, what we’re looking at is grouping information, a topic that’s covered in our publishing packet. It’s really important to make the grouping information logical. Using a car as an example, the tires and the rims are separate groups, which makes it easier for the end-user to apply textures to those groups. And it’s also helpful if the groups are named logically with a descriptive name as opposed to calling them group one, group two, etc. If you look at the model as a product that you’re delivering to the user as their essential raw material, the easier it is for them to use it, the more money it’s going to be worth."
Created: June 5, 2003
Revised: August 8, 2003