Collision Works

Have you ever wondered how Have you ever been in the air next to a monster and landed a critical Today we are talking about how collisions work. Okay, to recognize every kind of collision, no matter if it is about a character that is on the ground, against a wall or a ball First, we have to define the limits on w And secondly, we have to find out whether different limits are cut and what should happen if t First, the If I now put a ball on a character fire, to check whether a ball is in the model of a character, that is usually a fairly complex curved surface, we may have to carry out many calculations and calculation intensive divisions that all appear in the scene for each frame on each object. Therefore, we have to take some approaches. It would be more efficient to only notice a few values instead of a huge amount of points and lines. We could put the But as you can see, you could attack the entire empty space around the character and evaluate it as a So let's take the next best compromise, a box instead of a ball, to put the character in. In most objects and characters in modern games, there is still enough additional empty space that would make this pretty confusing and decisive in game, but it works well for games In most games, especially in game-oriented games, we have to compromise with the computer performance to match the accuracy. Instead of having a large box, we put the character in many parts and give it many different Some people study these very carefully in game-oriented games, because it is not that it is necessary to aim for the character model, that you can see on the screen, but that you can distinguish itself from several pixels on its invisible To put the In these games, for example, you have a separate And sometimes these opponents have very specific and not intuitive Therefore, our characters, monsters and bosses are all divided into many different But where do we know if they actually have contact? Well, there are two ways to go about it. We can try to calculate the exact time point by colliding two objects in advance, or we can check for every frame whether the two objects have already cut off. So we take a character that fires a ball on a wall. With the first method, we know that the ball flies 2 meters per second and the wall is 2 meters away. So we point out the game, to let the ball burst off in exactly one second from the wall. The second method would be to check all object pairs in every frame to determine whether they have already collided. So it would be the ball in the wall, the ball in the person, the wall in the person and so on. The first method is much more efficient, because it only requires a single calculation, but rarely is used in games. When the environment between the fire of the ball and the burst on the wall changes, a single calculation cannot be calculated. The second method is also checked in every frame whether collisions occur. So it can calculate all these changes, but it is incredibly inefficient. If a level has N objects, there are N-selectable two combinations of pairs that have to be checked. With 100 objects, that would be 5000 pairs, w There are a few options, how we can improve the performance here, w Basically, we can easily express it. If to become a bit more technical, we could say that t If we now know that an object is located on top left in a level, we do not have to see any of these deviations or e's and have saved a lot of calculations. That's what cutting trees informatics means. So there is much more to be done, w I hope that you now understand a bit more about collisions and why It is a compromise between accuracy and calculation performance. I wish you a nice day, wherever you are, and until next time. Bye.