Raymarching; a potental solution to limitless growth.

You may have noticed a game I posted to the Unclaimed List. In this game, you match various fruits to increase the size of a female character. If you’ve actually played the game, you’d noticed that such growth on a polygonal model doesn’t work so well. Breast scaling works okay as they are attached to the body at constant points. That game’s belly, hips, and butt mods fail to exhibit any meaningful growth in the same way.

This problem occurs because connection points (where each bodypart joins the main form) need to change in response to the growth in a way that can’t easily be programmed as a mesh morph. A potential solution to this problem is to forgo using polygonal models all together, and to use a shape-based raymarching technique. With shape-based raymarching, a model is drawn pixel-by-pixel and are defined geometrically.

Basically we can define a model to be a bunch of spheres, increase the size of those spheres to simulate fatness, and the raymarchign technique handles the rest. If we want to be able to simulate limitess (or even simply massive) growth, we’d need either something like this or a polygonal mesh with hundreds of thousands of vertices.

What do y’all think? Is this technique worth perusing or is the more traditional method of polygonal mesh and shapekeys/mesh morphs still worth the trouble?

[quote=“PeachClamNine, post:1, topic:1132”]You may have noticed a game I posted to the Unclaimed List. In this game, you match various fruits to increase the size of a female character. If you’ve actually played the game, you’d noticed that such growth on a polygonal model doesn’t work so well. Breast scaling works okay as they are attached to the body at constant points. That game’s belly, hips, and butt mods fail to exhibit any meaningful growth in the same way.

This problem occurs because connection points (where each bodypart joins the main form) need to change in response to the growth in a way that can’t easily be programmed as a mesh morph. A potential solution to this problem is to forgo using polygonal models all together, and to use a shape-based raymarching technique. With shape-based raymarching, a model is drawn pixel-by-pixel and are defined geometrically.

Basically we can define a model to be a bunch of spheres, increase the size of those spheres to simulate fatness, and the raymarchign technique handles the rest. If we want to be able to simulate limitess (or even simply massive) growth, we’d need either something like this or a polygonal mesh with hundreds of thousands of vertices.

What do y’all think? Is this technique worth perusing or is the more traditional method of polygonal mesh and shapekeys/mesh morphs still worth the trouble?[/quote]

That’s actually what I was attempting in the past. Couple it with spring physics, and you have the potential for limitless growth that should more or less match how things actually get fat, that you can interact with. The challenge is in the details though, the biggest being doing it all efficiently enough for it to run on most people’s computers.