Uniform bouncing

[Rolig Loon]

I'm just playing, and I can't seem to figure out the right geometry to get a silly toy to work right.  It's basically a trampoline, for which I can easily write

default
{
    collision(integer total_number)
    {
        vector AvPos = llDetectedPos(0);
        vector MyPos = llDetectedTouchPos(0);
        vector Go = llVecNorm(<0.0,0.0,AvPos.z - MyPos.z>);
        llPushObject(llDetectedKey(0), Go * 30,ZERO_VECTOR,TRUE);
    }
}

 That works fine, except that the push is always directed from the center of the trampoline prim. That means I get a better bounce at the middle than at the edges -- OK for a tiny trampoline, but not for a big one.  I thought maybe I could scale the vector Go by my angular distance from the center of the trampoline, something like this...

vector Go = llVecNorm(<0.0,0.0,AvPos.z - MyPos.z>)*llSin(llRot2Angle(llRotBetween(llVecNorm(AvPos),<0.0,0.0,1.0>)));

 but it makes very little difference.  I'm clearly not visualizing the geometry correctly.  Any ideas?

[Kayaker Magic]

I think you wanted llRotBetween(llVecNorm(AvPos-MyPos),<0,0,1>)

AvPos is in region co-ordinates, and you want the little vector from the center.

HOWEVER, doesn't llPushObject have a wierd-strange fall-off with distance from the center of the prim and the mass of the pushee, or the pusher, or which ever one doesn't make sence? I recall trying to compensate for this once and just running out of "energy" instead.

 

[Qie Niangao]

Gee, I don't think it's that complicated, if the object is always supposed to push upwards with the same result regardless of where on the surface the collision is made.

The only thing is that the force falls off as a function of the distance cubed, so... something like this, maybe:

default{    collision(integer total_number)    {        key av = llDetectedKey(0);        vector AvPos = llDetectedPos(0);        vector myPos = llGetPos();        vector up = <0.0, 0.0, 1.0>;        float distance = llVecMag(AvPos - myPos);        llPushObject(av, up*30.0*llPow(distance, 3.0), ZERO_VECTOR, TRUE);    }}

I think you may also want some kind of initial de-bouncing because I think you're likely to get two collision events right at the beginning.

(Or maybe I'm not understanding the project.)

[Rolig Loon]

You're right.  I spotted that after I posted too. :smileytongue:   It still doesn't help, though.  The real problem is exactly as you said: the intensity of llPushObject does fall off from the center of the prim.   That's the effect that I'm trying to counterbalance, without any luck.  I thought I could do it by applying some sort of gain to my Go vector that increases with distance from the center. Either I just haven't hit on a decent function (a very likely possibility) or it can't be done (a disappointing one). 

This isn't a major issue, BTW.  I only started playing with it because I got curious about how llPushObject works.  I still don't know.

[Rolig Loon]

Hehehe... It shouldn't be too complicated, you're right.  As I just replied to Kayaker, I think I just haven't hit on the right function to describe the falloff yet.  I should have thought of your idea, though .... a simple inverse cube law.  Now I can't get in world to try it out until this evening.  :=p

[Rolig Loon]

Yup.  Works like a dream.  Thanks, Qie.  There's one fewer mysteries in my world now.

[Randall Ahren]

For uniform bouncing, you need a more well endowed avatar. Momentum is proportional to mass and velocity, so increase the mass. That should make the uniform appear to bounce more.

[Void Singer]

::here's me hoping that was dry humor::

[Rolig Loon]

I was ignoring it gracefully.  :smileyindifferent: