View Full Version : Flat Mesh for UV

02-19-2005, 03:54 PM
Hey guys, trying to get into the nitty gritty of UV mapping these days.

I've seen a great deal of talk about Smart Maps, for example on newtek's tutorial site:

The thing I think I'm just having trouble pulling out from these tutorials is how they flatten their meshes like that. Are you going through and just pulling the points around 1 by 1 in a painstaking long process or is their a more streamlined way to do it?

In that link it says at one point: "The dinosaur is flattened from the side, -100%, after pulling the middle row points out evenly." Flattened -100%? Is there a 'flatten' tool I'm not seeing? haha.

I'm sure these questions are pretty lame and I'm sure I'm just looking for the 'do this exactly how I want it' button that doesn't exist, but if someone could kinda help me out in explaining this process a bit more, I'd really appreciate it.

In addition, I've been thinking about it and I think I have an idea about how to automate this in a lw plugin:

First off, there are a lot of things about this I'm not sure of, like how it starts, ends, puts it all together, etc. Also I am worried that it might take a long time to run on a high poly mesh. But regardless, here is the core part of the idea.

What I understand to be the problem with making good UV maps on complex meshes:

For explanations sake I'll use a super simple object. (View figure 1 below. sorry for the quality, made it in paint haha also, the 'inches are not accurate, just for examples sake.). Okay so say you have a 2 poly object, and the polys are bent so that they make a triangle. Lets say the polys are both 12 inch squares. Now if you made a UV map, you'd use the Y Axis to do it. And the map would look like the one in the bottom right corner. The only problem is now on the UV map, the polys arent squares, so there would be distortion on the mesh. Well on a simple model thats an easy fix of just dragging the points in the UV map around. But what if there was a poly that was facing the X axis, then the poly is just a mere sliver on the UV map, etc etc etc.

So what I propose the plugin would do. It would start on one polygon (yet again, not exactly sure what the best place to start would be). And by using the point info from each of the polys points, it would determine the exact size of the poly, as if you were looking at each poly directly (your line of sight being perpendicular to the polygon surface. It would then layout a poly for it in the UVmap and continue going on to adjacent polys doing the same thing, laying each next to their adjacent polygon. So essentially in a sense, you are going polygon to polygon and taking a front projection UV of it, from an angle directly perpendicular to its surface. As you can see in figure 2, if thats what actually happened, then the polygon would be the correct size. But rather then doing that, its using the X,Y,Z values from the points to determine what the true shape of the poly is.

I have trouble explaining this completely but I really hope that makes sense.

I've never programmed Lscript before, though I have done quite a bit of programming in other languages, just never have tried it. So I probably wouldn't be the guy to make this.

But I would really like to hear what some people who have experience with Lscript think of doing something like this.


02-19-2005, 05:30 PM
For the first part of my question, I found this, havent watched it yet, but it seems like it'll go through it.