Sunflow Procedural Fractal

[doncasteel8587]

I've been trying to learn how SunFlow works, so I'm doing a lot of playing in the code.
Here are two of my latest images



And just for fun:

[Nahee_Enterprises]

I have always had a fondness for L-System type images, especially when they come out looking so well as yours.  Nicely done!!

And that "tea pot" as the object on your second image is hilarious!!     
You got me thinking about the use of several different kinds of objects for doing similar images.   

[lycium]

hehe, teapots are a staple of computer graphics see the great article at http://www.sjbaker.org/teapot/ (don't miss the teapotahedron - http://accad.osu.edu/~waynec/history/lesson20.html)

btw, my own (very poor) attempt from last year at making a fractal teapot: http://www.fractographer.com/wip/badfractalteapot.png

[doncasteel8587]

btw, my own (very poor) attempt from last year at making a fractal teapot: http://www.fractographer.com/wip/badfractalteapot.png

Huh?    I think it's excellent!

For a long time I've been trying to figure out how to "wrap" a polygon object with instances of itself (or some other shape) then wrap the instances and so on...  I envision the instance mapping following the objects topography, for each Quad of the surface, a unit cube is mapped aligned with the quad's normals the height along the normals would be proportional to the surface area of the quad.

If the shape to be mapped is tileable, with openings where the tiles would join, the tiles could be sewn together resulting in a contiguous surface.

I can't figure out how a 4x4 matrix transform can do it because the mapping would have to flare open or closed depending on if the quad lies on a convex or concave surface.

I haven't researched what kind of transform would work but would love to figure it out someday.

Anyway....  Nice Work!

[doncasteel8587]

I have always had a fondness for L-System type images, especially when they come out looking so well as yours.  Nicely done!!

And that "tea pot" as the object on your second image is hilarious!!     
You got me thinking about the use of several different kinds of objects for doing similar images.   

Thanks you very much!   

The teapot has a certain appeal to us old timers.

I needed to test geometry importing, so the teapot seemed appropriate   

[lycium]

it's definitely not possible to both place and orient objects on the surface AND have it taper off or expand; if you're willing to ignore the expansion / contraction then it's pretty easy to produce such a matrix, provided of course that you can make a consistent vector field on the surface -- itself a very difficult problem (cf. http://en.wikipedia.org/wiki/Hairy_ball_theorem and http://graphics.cs.uiuc.edu/svn/kcrane/web/images/proofI.jpg). btw, that wikipedia article also mentions why my camera code from that funky.exe app fails sometimes (it's in the computer graphics section)

however, there is a bunch of great literature about generating "fins" for use in modelling hair on surfaces. no particularly good references come to mind, but googling around ought to produce some good results.

[alister]

Wow, that is very nice.
I've done a little work with 3D fractals, but none as good as that!

[doncasteel8587]

Thanks alister