Arkanoid, first game attempt

[steeg]

Hi, I’m programming an arkanoid/breakout clone (very basic since this is my first game) and for the collision detection between the ball and the bricks, I wrapped each ot these objects with a rectangle (Rectangle class) and used its intersects() method.

I am not using any angles or anything, only position and speed variables. The thing is that if the ball hits a brick on its top or bottom, then I multiply its ySpeed by -1, and if it hits the brick on its side, the same has to be done with the ball’s xSpeed.

Is this the best approach for this type of game? Also, how can I detect on which side the ball has hit the bricks? If there is a more efficient way of doing this, please let me know.

Thanks.

[oNyx]

Is this the best approach for this type of game?

No. The best approach is priori-ish. You predict if a collision will happen and act accordingly. Like you have the normalized time interval of 1 from one frame to the next, you check if a collision will happen (with some sweeping algorithm), the you subtract the (normalized) timespan from your remaining timespan (started with 1) and continue with that until you run out of (normalized) time. Thats it… basically.

Its horrible complicated… mind you. But the advantage is that the simulation is very accurate. It does look and feel a tad more solid/real than breakout games usually look like. (You also dont lose speed when collisions happen.)

Well, I suggest to keep it like that. I mean the simple approch, you’re currently using If your timestep is small enough it works pretty well. If the displacement gets too big the ball might warp through bricks and/or the paddle.

copy&paste

http://kaioa.com/k/max_speed.png

Eg the brick is 16px heigh and the ball radius is 6. 16+6+6=28. So don’t move more than 27px otherwise the ball can warp through bricks. (It’s really “fun” if the ball warps through the paddle ::)).

[steeg]

Let’s say I check to see if a collision will happen within the ball’s path (imaginary line starting at the center of the ball), how could I detect from which side the ball will be hitting the bricks? Thanks.

[swpalmer]

The ball follows the path… it hits a brick on the side that intersects the path and is closest to the balls current position.

[steeg]

Thanks for the opinions, I’ll take them into consideration for when I’m a bit more advanced in game programming. For the time being, I used outcome(Rectangle e) in the java.awt.Rectangle class and checked its return value (eg. Rectangle.OUT_BOTTOM) to determine where the collision took place.

Thanks.

[Nekromant]

ahh thats my topic! i can only underline to use a very basic collision detection. my first java project
was a (never finished) arkanoid. as more as i advanced in the game and used more complex
techniques the problems really grew exponentially!
there are questions like how to measure if a ball hits exactly a corner of a ball. or the problem
of multiple reflections, e.g. the ball hits a brick, so that its velocity vector gets reflected. after each
collision you would need to check another time for a collision (if the ball gets “reflected” into another
brick) etc.
another issue is how you generally deal with movement. if you add a dx/dy to your with bigger values
for fast movement the ball can jump over small objects without collision or even get stuck in them.
most methods to avoid this result in another compromise: precalculating a collision, to move a stuck
ball out of its position etc.

i really noticed that an atomic clear arkanoid becomes a physical snooker simulator. to make things
short, take the easiest way to have a collision, test, if any of your little inaccuracynesses are detected
by your eye or by the game feeling. and then consider more complex improvements.

[steeg]

@Serethos

ahh thats my topic! i can only underline to use a very basic collision detection. my first java project
was a (never finished) arkanoid. as more as i advanced in the game and used more complex
techniques the problems really grew exponentially!
there are questions like how to measure if a ball hits exactly a corner of a ball. or the problem
of multiple reflections, e.g. the ball hits a brick, so that its velocity vector gets reflected. after each
collision you would need to check another time for a collision (if the ball gets “reflected” into another
brick) etc.
another issue is how you generally deal with movement. if you add a dx/dy to your with bigger values
for fast movement the ball can jump over small objects without collision or even get stuck in them.
most methods to avoid this result in another compromise: precalculating a collision, to move a stuck
ball out of its position etc.

i really noticed that an atomic clear arkanoid becomes a physical snooker simulator. to make things
short, take the easiest way to have a collision, test, if any of your little inaccuracynesses are detected
by your eye or by the game feeling. and then consider more complex improvements.

lol, I pretty much had all of those problems myself. Since it was (is) my first game project, I didn’t expect certain things to happen, that now seem obvious…like the ball skipping over objects where it should collide, as a result of it travelling too fast and not implementing a collision detection system where it would check throughout the ball’s path, etc…but I guess the best way to truly learn is finding these things out for yourself…