Same animation as before. I coded it from scratch, and... it came out 
differently. I suspect what I've done is set the magnet coordinates 
differently or given then a slightly different strength or something, 
resulting in a different image.

The animation I got the first time round was *very* noisy. I did wonder 
if it was to do with the precition of the numbers in the save file. But 
I recorded to not use one, and it still looked the same.

...then I changed the time step, and all the artifacts got smaller.

The movie you're seeing has 25 fps, but computes 8 steps per frame for 
the physics simulation. Anything less than that and you see "waves" 
round the sides - and sparkling noisy areas exactly between the magnets. 
(IOW, where the system is most unstable!) The whole thing uses Euler's 
method; I really must learn about RK4 sometime...

Rendered at 160x160. Took about 5 hours. AMD Athlon64 (in 32-bit mode!) 
3500+. Has 1GB RAM - not that it used any of it! Windows XP Pro.

Post a reply to this message

Attachments:
Download 'chaos-2b-8x-cut-low.m1v.mpg' (582 KB)