virtual jellyfish
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish4.jpg)
I was thinking about Iterated Function Systems (IFS) after reading about them in the book, and I thought back on some weird stuff that I made using tetrahedrons and OpenGl a couple of years ago. I am not sure if this stuff counts as an IFS but I think it might.
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish.jpg)
The shape you see is a single freezefram of a constantly morphing object which is rendered by drawing the same tetrahedron over and over, each time applying an affine transformation which both rotates and translates the shape by a small increment. When applied over and over again, frame after frame, the position of each tetrahedron becomes a function of each of its previous states, all of which originate in the same spot at the original tetrahedron.
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish4-5.jpg)
Here are a couple of shots of some other ones that I made. Its something else entirely to see them in motion, but I doubt that I can post an .exe file on this blog.
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish4-2.jpg)
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish2.jpg)
![](http://photos1.blogger.com/blogger/6686/2189/400/virtual_lellyfish3.jpg)
4 Comments:
Nice colors! You may not be able to post an executable on the blog (and if you did, nobody would run it). But I'm sure a video would be okay. :-)
I wouldn't call those iterated function systems, mostly because there is no probabilistic aspect to their generation. They're really much more like 3-D maps akin to the 1-D maps of your last post.
IFS's are annoying because 2-D IFS's have the tendency to look like profoundly cool 3-D figures, but you can't rotate them or put them in a 3-D world.
Anyway, as far as I recall, an IFS is a set of functions together with probabilities:
f(x) = x+1 (25%)
= 2x-1 (50%)
= x/3 (25%)
Then you just fold that function onto itself and plot the results (where applying the function picks one of the implementations with the given probability). I just pulled that out of my ass, so it's probably not a very interesting function.
I wonder if you could generate 3-D figures using a point (rather than polygon) plotting mechanism and slowly varying the parameters.
Thanks luke!
You are bringing an air of correctness to this otherwise full-of-horse-poo blog!
I like your idea about rendering an IFS in 3d... do you remember that one that I made that was like a shooting star last summer? that was definitely an IFS and it was 3d...
maybe I can dig that up.
Anyway, how would I go about making a video? if anyone knows how to capture a video from your computer's display, please let me know. My next post will be a cool video if so!
-Max
Another note for those who don't get Luke's inside joke about the colors of the "virtual jellyfish" screenshots, here is the explanation: While I programmed the construction and geometry of the figures, I was early in my investigations of computer graphics, so I asked Luke (who was not) to write me an algorithm to create a nice color scheme.
Thus luke himself is responsible for the colors!
ho ho ho!
-Max
Send me the exe, Max! Do you still have my email???
I wish I had something to say about the video. I've seen people doing it before, though, so it can be done! Hehe.
Post a Comment
<< Home