A Low-Level Modelling Tool (Viability Test)

Eketek

Hello. I am an Independent Game Developer. I have an in-house 3d modelling tool which I developed to suit a personal need, and am now considering developing further, with the intent to sell.

The need was simplicity - I wanted a bottom-up process that would allow me to operate and think of models as vertices and triangles, rather than as a series of operations performed on a set of primitives. I found the existing tools quite difficult to use, but I knew what I wanted, and so I built my own. For now, I call it the "Eketech 3d Editor" (ek3d).

At its core, it is an efficient and precise 3d point selector. Vertices are defined through a simple set of operations, then objects are defined as sets of vertices (and manipulated by manipulating the generating vertices).

So far as I know, ek3d should be great for building low-poly models (or as an alternative for what is called "blocking" here, should more arbitrary shapes be desired).

Ek3d in action, with only a few minutes of semi-random usage:


I would like to know whether or not others might find such a tool useful. I occasionally see discussions wherein others want the sort of functions I put in my editor. To that end, I have posted a demo of it on my website (requires the Unity3d browser plugin). Please try it and tell me what you think, since I am unsure right now whether I should proceed with development or let it remain just an in-house modelling tool.

Ek3d demo:
http://www.eketech.com/demos/ek3d.html

WarrenM

Lots of primitives and super simple snapping and scaling/rotation controls would give this it's best chance of success. If it's fiddly, they might as well use Blender or another 3D app...

Eketek

There's a lot of features I am planning for the editor, even if it remains for private use. At present, I only use vertex groups to define triangles, but I plan to support a variety of primitives and other objects. Cylinders would be defined as two end points and some point on the surface. Spheres by a center and edge point or by top and bottom points or by a set of 4 non-coplanar points. I'm planning to add curved surfaces (also controlled as expectable). I could add a convex hull builder which uses an arbitrarily sized set of points.

Another major feature I'm thinking about is the ability to use the various objects indirectly as picking-surfaces, rather than directly as mesh data - this would allow you to build a primitive shape or a curved surface with a customized topology. Points defined this way would [optionally] remain bound to the underlying object, so that changes to the generating points of that object would result in cascading adjustments to dependant objects.