1
votes

Situation: I am programming a 2D isometric styled Java game with libGDX. Right now I have a moveable player that properly collides with tiles of solid objects.

Problem: Now it comes to the renderorder of tiles. By default the library code renders from the bottom maplayer to the top layer. (Ground is layer 0, object layer is layer 1) Which makes sense. I draw the player on top of that. This means that the player is always on top of everything which doesn't make sense in some situations.

Goal: Since an isometric look means to have a kind of a 3D perspective, the player can be behind or in front of objects. So I have to come up with some code that decides whether the player is rendered behind or in front of it. I have this fridge as an example:

Example

I hope it is comprehensible what I mean with "logical collision". I have some glimpses of ideas how to achieve that but that would be a mess in the code. So I wanted to ask if anyone has experiences with that or can hand me some nice sources that can help me.

Thanks for reading!

1
Surely the obvious solution is to maintain the player in the object layer either behind or in front of other objects.OldCurmudgeon
Yes, and my question is what the best way to do this would be. You simply repeated my problem/question.Pixell
Painters Algorithm could be a good start. Without knowing which rendering library you are using it is difficult to advise. The negative attitude doesn't help either.OldCurmudgeon

1 Answers

0
votes

It depends HOW you render. One thing that comes to mind, is to render objects/tiles from up to down and left to right, that way in most cases the objects should always be in correct layer. Now depending on how you make the graphics, there could be flaws in this. If so, you could also have few layers of of priority, to draw different parts of objects/tiles, or draw something before other. If you also have tiles and objects as separate type, you could have objects ability to draw before and after the tiles, to fulfill different needs. You could also implement some tiles to have objects, for cases like this, however that could also be waste of processor times comparing to other methods I mentioned.

How you would go about comparing object positions to tile positions is not actually very difficult. Probably a sufficient method would be comparing tile positions to object positions. Lets take a hypothetical situation, where you have tiles that are 32 x 32 in size, and there is object at 25 x 18. The object would be in front of tiles at offset 0 x 0 and 1 x 0, but behind 0 x 1 and 1 x 1 (If we imagine the tiles start from upper corner). Therefore, we first draw tiles at 0 x 0, and 1 x 0, then object, then tiles at 0 x 1 and 1 x 1. It should naturally fall in it's correct place, with fairly simple code logic.

Hope you got some ideas about how to implement it :)