Room for Jovito and Lasse

8:49 PM

I don't understand. Why a 90 degree angle?

you can use pythagoras theorem if it's 90 degrees

I'm typing, a sec

origin = sqrt(ballPositionAfterHit^2 + hitPointOnPaddleInWorldCoordinates^2)

something like that

that might or might not work, I'm not sure

lol thats just silly, a sec I'll think it thru better :D

haha, ok

okay

let's say point A is where the ball hit the paddle, and point B is the new position of the ball after the hit according how it reflected (like it normally would without rotation)

wait!

I gotta go take a shower, can you wait?

yeah, I'll just write it here meanwhile

it's not hard

8:59 PM

k, brb

subtract A from B, then you get C, and if you flip the components (x = y, y = x) and normalize you will get the normal for that vector from A to B. Now you can use that normal to set where you want the origin of the circle to be by multiplying it with some scalar value that determines how much curve you want for the ball

that can be got for example like 1 / paddleMoveDelta

ahh the Pythagorean theorem

a^2 + b^2 = c^2

or

sqrt(a^2 + b^2) = c

and in case you don't know, paddleMoveDelta = paddleCurrentPos - paddleLastPos

yes that is pythagorean

it could be used here too, but this is just simpler

lol

the result is the same

also, when the ball hits something else you might want to reset the motion to be linear, or just put the origin of the circle really far away

like, really really far away

derp "linear"

I have no idea what's the english term for something travelling straight line instead of something else

"straight path" might be better

oh, wait, the normal is x => -y, y => x

you need to add the sign there

9:26 PM

I'm back

so C would be a vector of length distance_travelled right?

yeah, it's just vector from point A to point B

I want to go over this step by step until I understand the logic

but I'll try not to bore you.

So, why flip the components?

that makes it a normal for that vector

yeah, I visualized it now

it exactly flips the vector

yep

9:30 PM

by normalize you mean to set its length to one?

yes

k

The paddle last pos would be the one before the collision or the one on the collision?

no you need to know the ball position where it hit the paddle

in world coordinates

31 mins ago, by Lasse

and in case you don't know, paddleMoveDelta = paddleCurrentPos - paddleLastPos

that's what I'm asking about

oh, that

it's the last frame's position

so I guess it's the position right before the hit

9:35 PM

ok

I don't need to keep track of it, right? I can just calculate it back

it's easier to keep track of it, but you can calculate it too if you want

ok

so let me see if I got this

the bigger is paddleMoveDelta, the larger will be the curve

I mean

the radius

no, the opposite, because you do 1/delta

yes

yeah, that was exactly what I was going to ask you

uh, wait, no

9:38 PM

I wanted the reverse

:D

let's say delta is 10

1/10 = 0.1, the radius will be then 0.1 (for example, in reality you multiply that with the value you want)

if the delta is 20, and the paddle moves faster

it's 1/20 = 0.05

so it's closer

so yeah, the bigger delta, the smaller the radius, and more curve will be apparant

okay, so to sum this up

I get the vector between A (lastBallPos) and B(nextBallPos), flip that, normalize it then multiply by something like 1/paddleMoveDelta * scalar

and that's the origin

yep

that was beautiful

k, could you tell me again those circular movement equations

I couldn't find them in the transcript for some reason

they got deleted by raging SE employee

newx = sin(travelAmount) * radius;
newy = cos(travelAmount) * radius;

the travelAmount starts from zero, and should cumulate

9:45 PM

travelAmount being just the linear distance?

yeah

the distance from the starting point

what do you mean 'starts from zero'? why?

you mean the distance since it started spinning?

you need to tweak that a bit though, larger radius will move the ball more with the same values

yes

also you need to find out the starting point on the circle

why should it cumulate?

I meant, if the ball travels like 1 unit per frame, you need to cumulate those, so it'll be 3 units after 3 frames

9:48 PM

okay, its getting weird now

it's like the deltatime needs to be added to the totalTime

shoudn't the travelledAmount be the distance from lastpos to currentpos?

the original example was this:

totalTime += deltaTime;
newx = sin(totalTime);
newy = cos(totalTime);

no, it should be from where the ball got hit by the paddle

Damn, everytime I try to understand this I get confused

alright

I can explain this part

9:52 PM

please try

@Jovito travelledAmount should start over every time it collides

with the paddle

so, make a lastCollidePos

and the travelledAmount will be the distance between the currentPos and the lastCollidePos

done

the thing is

if you tell me "just do this and it will work" I can go back, code it over and everything will be cabbagewonderful

what I don't understand is why

to understand you need to know how sin and cos works

Sine

In mathematics, the sine function is a trigonometric function of an angle. The sine of an angle is defined in the context of a right triangle: for the specified angle, it is the ratio of the length of the side that is opposite that angle to (divided by) the length of the longest side of the triangle (i.e. the hypotenuse). Trigonometric functions are commonly defined as ratios of two sides of a right triangle containing the angle, and can equivalently be defined as the lengths of various line segments from a unit circle. More modern definitions express them as infinite series or as sol...

"Relation to the unit circle"

sin() and cos() will always return a value from -1 to 1

okay, you think I should study some trig before attempting this?

If you want to fully understand it, yes