>>Do I understand it correctly that the attractor is a surface (or could
>>be represented as one)? It looks like it, so you could try to build a
>>(smooth) mesh from it. Then you could also do fancy stuff like nice
>>finishes etc. Of course you'd need POV again to do so.
> 
> 
> No, it's not a surface. It's a bunch of points on a 2D-coordinate system.
> They look 3D, but they aren't.

Which is not to say that you can't have 3D attractors. I wrote a quick 
and dirty hack to search for sets of coefficients for a 3D quadratic 
that produce nice attractors a while ago, I've been meaning to work on 
it, perhaps I'll clean it up and post it here if anyone's interested.

The trouble with rendering it as a solid object is that you have no more 
information than the location of each point - other than simple 
proximity there is no way of knowing whether two points should be 
connected forming part of a surface, or whether there is a gap between 
them that should not be filled. In the end you either just plot so many 
points that it looks like a single surface, or use a guessing algorithm 
designed to form a surface from a collection of points.

The sample attaced is rendered by plotting each of the 65536 points as a 
sphere. It looks a little flat, but you can see where a bit of it sticks 
in front of the rest at around 8 o'clock, and a bit of twisting from 
around 11 o'clock to 2 o'clock.

Post a reply to this message

Attachments:
Download 'attractor3d.jpg' (25 KB)

Preview of image 'attractor3d.jpg'