Here's my newest test-animation.

The metal tube is a CSG-Spline made
of spheres and cylinders. The Spline was
calculated using a Bezier-Spline-Algorithm
I implemented myself.

The Particles are emitted from the designated
Spline and are forced outwards due to centrifugal
forces.

Actually, I implemented an algorithm to calculate
a centrifugal-force on a given Bezier-Curve. This
was not that easy, since I had to adjust the algorithm
to not try to calculate a position after or before the
defined spline (when thinking in values of 0 to 1,
-.1 and 1.1 aren't allowed, simple as that).
Also, when the calculations made in detail seem to
result in no centrifuge on a bezier-spline, the
algorithm expects that the intervals had been taken
too small and increases them.

Another new addition running in the background
of the system is to leave particles out of calculation
process. I hadn't implemented it beforehand, simply
because it escaped my mind that I could have a
new random-stream for every particle.

So, these are 250 frames, with about 10sec parsing
and 25 secs tracing per frame, with up to 400
particles, 250 spheres and 250 cylinders, along with
a media-sphere for the glow-effect, a lightsource
and the camera. Due to tracing interruptions (never
trace and modify the code when using random-streams)
I can't tell the overall time of parsing and tracing.
(Oh: 1.4 GHZ, 512 MB Ram, Win98)

Comments? (Aside that I should have used more particles
to make it look more like a cloud than individuals)

--
Tim Nikias
Homepage: http://www.digitaltwilight.de/no_lights/index.html

Post a reply to this message

Attachments:
Download 'centri.mpg' (331 KB)