I've noticed the bands as well, and they get worse as the simulation
continues. I have a random element already in place, which randomises the
initial velocities of the particles. The fact that the bands become more
prominant with time suggests that it is the code exectuting after
collisions that is to blame.

I am using a very basic system, with Euler integration to determine future
velocities, and using a constant for gravitational accleration. At present
the particles simply fall and bounce, losing a preset amount of velocity
each bounce, to simulate very crudely a loss of energy.

I plan to alter the collision detection next, as it works pretty badly at
the moment.

Post a reply to this message