Working out peak of a curve

[valhallen]

ok am making a flash game and need some help in writing a formula for it

basically user will select power and angle
character will then jump depending upon the values user entered

i need however to figure out depending upon power, angle and a fixed value for gravity where the peak of a curve will be so that character then starts to fall again

so that the characters x/y postion increases until they reach peak of curce and then start to decrease again

what equation could i use to calculate this?

v_Ln

[valhallen]

to give you some idea what i mean look at attached image

ok red line :
user enters high power setting & low angle setting

blue line :
mid power setting mid angle setting

green line :
high power setting high angle setting

you can see that depending upon the values entered by user shape of curce changes - now i will know the values for Power (P), Angle (A) and Gravity (G) but what equation could I use to figure out the peak at which the upward acceleration stops and the user then begins to fall down again?

v_Ln

[Striek]

Wow. Long time since I took calculus. I'll try to answer, though.

At the peak of a curve, an object is neither accelerating or decellerating (vertically). To find the rate of acceleration, you need to find the first derivative of the function which calculates the object's movement. This derivative, rather than plotting height as a function of time, will plot acceleration as a function of time.

The peak of that curve will be where the first derivative crosses the x-axis, or where acceleration is equal to zero. The value of x at this point indicates the time at which this occurs.

Now, as for how to incorporate that into a flash animation, I have no idea you may need some higher math libraries. I don't even know if this will help, but wanted to see how good my calculus skills still are.

I think I'm right. Somebody correct me if I'm wrong.

<edit>

to give you some idea what i mean look at attached image

What attached image?
</edit>

[valhallen]

heh image is now attached - lol soz for that &gt;.&lt;


hmm wonder if I could use something like :

Power (P) = 50
Angle (A) = 45
Gravity (G) = 5

B = P-G
C = A-G

then make B & C equal to P & A
and then add B & C onto characters current x + y postions

they wil start out as positive values but as am constantly taking G away from them will eventually become negative values and character will start a downward arc......would just be a matter of getting a seemingly correct value for G so that the arc falls at aprox the right time.

v_Ln

[Striek]

Hmmm...

Unless you have wind or something similar (I'm thinking Scorched Earth-type game here?), the horizontal velocity will remain constant. So, I think you will need to calculate a horizontal velocity and a vertical velocity from the initial power and angle variables, like

[vH = P cos A]; should give the horizontal velocity. This will remain constant.

[vV = P sin A]; should give the initial vertical velocity, which will be reduced by G for every loop iteration.

And the loop would look something like this:
do {
X = X + vH;
Y = Y + vV;
vV = vV - G
} while Y &gt;= 0

The top of the curve should then be where Y equals 0 (or shortly before it becomes less than zero). Maybe calculus isn't needed after all, but I don't think you'll get away without trigonometery.

[valhallen]

ok code i came up with

//set inital variables
onClipEvent(load) {
	alt = 30;
	speed = 30;
	gravity = 2;
}

/next part runs every time frame loads (12fps)
onClipEvent(enterFrame) {
	Ymov = alt-gravity;
	Xmov = speed-gravity;
	
	this._y -= Ymov;
	
         //added next part in to stop pause looking as bad when equals 0
	if (Xmov == 0) {
		this._x += 3;
	}
	
	if (Xmov > 0) {
		this._x += Xmov;
	}
	if (Xmov < 0) {
		this._x -= Xmov;
	}
	alt = Ymov;
	speed = Xmov;
}

works kinda but gives a very unatural looking curve - i knew i should have listened more in maths.....to think i took bleedin advanced maths as well, seems like 100 yrs ago now

v_Ln

[valhallen]

btw thought would add as am doing this in flash it adds its own lil quirk to it

the top left hand corner of screen is 0,0
so as you move -&gt; x increases
and as you move ^ y decreases - just to make things extra simple &gt;.&lt;

v_Ln

[Striek]

I think causing X movement to be reduced by gravity would cause some problems. I imagine the downward side is steeper than the upward side, perhaps doubling back on itself?

Discounting wind and air resistance (which I doubt you're adding in yet at this point), the horizontal velocity should remain constant. Something like Xmov = Xmov + speed might work better.

//set inital variables
onClipEvent(load) {
	alt = (whatever);
        Xpos = (whatever);
	speed.horizontal = (power cos angle);
        speed.vertical = (power sin angle);
	gravity = 2;
        TopReached == Flase;
}

/next part runs every time frame loads (12fps)
onClipEvent(enterFrame) {
        /Set the initial position
        this._y == alt;
        this._x == Xpos;

        /determine how far to move
	Ymov = speed.vertical;
	Xmov = speed.horizontal;

        /move the object
	this._y -= Ymov;
	this._x += Xmov;

        /adjust vertical speed as a function of gravity
        speed.vertical -= gravity;
	
        /has the top been reached?
        if (speed.vertical <= 0 && !TopReached)
        {
          TopReached == True;
        }
}

Then checking the TopReached flag will tell you when the top of the curve has been reached.

Does that make any sense?

[sec_ware]

Hi


If I understood your coordinate system correctly, it is

(000,000) ------ (320,000)
| .............................. |
| .............................. |
| .............................. |
| .....x....................... |
(000,200) ------ (320,200)


Let's say, there is a guy waiting at (50,180),
which is somewhere in the left bottom region.

I would suggest as the three start-parameters:
- velocity (v0) (instead of "power")
- angle (alpha)
- gravity (g)


Then:
-----
I call startposition x0=50
I call startposition y0=180

Take a small "stepsize" (timestep) dt = 0.5 (or something like that. Adjust this for
the correct "feeling").

repeat
  t=t+dt;
	1. x coordinate: x(t) = x0 + v0*cos(alpha)*t
	2. y coordinate: y(t) = y0  - [ x(t) - x0 ]*tan(alpha) + g/(2*v0*v0*cos(alpha)*cos(alpha))*[ x(t) - x0 ]*[ x(t) - x0 ]
 
until (floorreached(y(t))) || (bumbed_in_a_wall(x(t))).

You can test it as follows: If the floor is straight (no steps, ...), then the x(tmax) you have
reached should be:

x(tmax) = x0+v0*v0*sin(2*alpha)*sin(2*alpha)/g.

The maximal y-height that can be reached is
y_maximal = y0-v0*v0*sin(alpha)*sin(alpha)/ (2g).


This should give you the correct parabolic motion. I hope, this was not too technical.
I tried to adjust the coordinates to your y-reversed coordinate system

Have fun!

/edit: In case you want to incorporate the airflow, I could elaborate ( ) about it.
In short: the additional force (acceleration) is proportional to the area of the
object (person, ie. feetsize ) times the velocity squared.

/edit2: a little error in the main formula corrected.
Reasonable initial conditions (example)
v0=10
alpha=Pi/4
g=1
t=0..14

Cheers

[gothic_type]

Just so you know, the coordinate system with 0,0 top left isn't really weird. It's the coordinate system used in all (AFAIK) computer graphics. Any graphics coding I've done has used that coordinate system anyhow.

ac