0

was8bit 2019-04-20 15:12 (Edited)

.. study of the math...

.. more to come..

Mrlegoboy 2019-06-07 06:30

If you really want an exact collision detection, you must treat time as a dimension in your math

was8bit 2019-06-07 11:54

I've been working on the math for years, with help from different online resources even... I have never got it perfected yet.. :/

Mrlegoboy 2019-06-09 16:38

I might draw a diagram for you, i used to be obsessed about this exact problem.

was8bit 2019-06-09 17:14

Oh, I’ve researched this up and down and left and right... I get the math on paper OK... it is implementing them into the programming that is difficult.... partly because of the limitations of the math, but mostly because of the method of handling the multiple collisions...

To save processing time, you first don’t use the real distance formula, you first use approximate distance to determine if two balls are close enough for a collision, and flag them for further processing... then you go thru again and process only the collision pairs...

.... multiple balls colliding causes the biggest issue, when 3 or more are colliding.. I get the “sticky balls” syndrome where the balls stick together rather than bounce apart....

was8bit 2019-06-09 17:16

I’ve researched for actual programs on this, but most all of them only use VECTOR MATH not real normal algebra with angles and points in space, and of course lowres doesn’t have vector processing...

Log in to reply.