Commercial usage of APE
onedayitwillmake
the APE engine.
I stripped out a lot of the drawing stuff, and in the end APE was very
easy to work with.
I wish the ball went through the line less often, but hey what can you
do (unless anyone has any suggestions).
One thing i did try was turning on multisample for the ball once it
passes X velocity, but that introduced stuttering.
As i said, still was very enjoyable to use APE to make this game.
Wes
onedayitwillmake
http://www.candystand.com/play.do?id=18253
if anyone has any suggestions on how to limit the ball going through
the line problem, i'd love to hear them.
thanks,
mario!
onedayitwillmake
The sample course is pretty good, as are some of the ones on the
highest voted column.
If anyone has any suggestions on relieving the ball through line
problem, i'd love to hear'em.
>
Chris Green
> If anyone has any suggestions on relieving the ball through line
> problem, i'd love to hear'em.
I only played with the API a little bit but isn't that 'fixable' by
playing with the simulation step? I think you might also run > 1 APE
step per frame to increase your chances of not missing the collision?
--
Chris Green <gre...@gmail.com>
Troy Gilbert
> > problem, i'd love to hear'em.
Yeah, increasing the sampling is the only way to do it (through either
multisampling or shortening the time step). Or, if you can determine
the "inside" and "outside" of lines you could test the ball against
the lines and "pop" it out if its on the inside of lines. But that's
probably hard/impossible to do given the freeform drawing tools.
Pretty cool... are the lines just RectangleParticle segments, or are
they CircleParticles connected with Springs or what? Any optimizations
that you made to APE?
Troy.
onedayitwillmake
of stripping out what i don't need.
Not so much in the form of crazy math optimizations.
Like for example, all the lines are drawn into 1 clip.
The lines are are springcontraints, and i re-use spring constraint
particles (the end of the last line, is the same particle used for the
start of the next one).
I tried multisampling, and that helped somewhat but not as much as you
would think and especially not for the cost.
I even had it toggle on and off per the speed of the ball, but that
just introduced stuttering and slowdown since I had to use say
multisample = 100 before anything changed.
The time step is set to,
APEngine.init(1/4);
should that be different?
onedayitwillmake
rid of the ball thru line.
What i did was change the init delta time parameter to 1/8, thus
cutting it in half.
This caused it to run at half the speed, so i then called
APEngine.step() twice per frame.
Oh yeah, one thing i forgot to mention was that i modified the
collision checking, to only check if (using a rough distance equation)
public function distRough(p:Vector):int
{
var ox :int = p.x - x;
var oy :int = p.y - y;
var r :int = ox + oy * 0.5;
if(r < 0) r = -r;
return r;
}
is less than 100 pixels.
That definitely made a big difference.
On Jan 17, 6:35 pm, onedayitwillmake <onedayitwillm...@gmail.com>
wrote: