Ok, this time my problem is, I have a sprite sheet with my player's animations frames in it. Follows example:

Using the get() method to iterate what part of the sprite sheet i want to display doesn't seem like an option, since 1 - the image I'm cropping must be inside the screen area (better saying, I need it to be outside the screen but if doing so the get() method will return me a black rect) and 2 - i loose my alpha channel.
Of course I can create arrays and load each frame as a separated image file, but it fells like a huge mess.
Is there a way to render my sprite sheet outside the screen area and not only use it, but not loose the alpha channel as well?
I tried pixels[] but I just haven't fully understand how to do it – and I'm not sure it can do what I want.
Thanks in advance, guys and ladies!

Hello, guys.

I'm trying to work with this collision check for my Zelda-like game (controls and view like “A Link to the Past”).
Speaking only for “wall collisions”, simple rect collision check will not work because I need to know which side has collected with the wall and, once collided, I need the player to be able to “slide” - X axis if colliding with a lateral wall, Y axis if colliding with a top/bottom wall. This is to happen when the player is pushing two directions at the same time.
I've tried a few options (for everyone of 'em I keep track of the previous player's x/y coordinates so, once colliding, I can move her back to position):

. Method 01, checking the corner pixels of the player's collision box – works fine as long as the walls are never smaller than the player – if so, the edge pixels won't detect what they should.

Method 02 – Checking only the collision box and deal with possible mixed inputs the player may be holding – totally fail.

Method 03 (the one that should work) – creating four rects, one for each edge of the player.
rect 01 (upper one)
rect 02 (left one)
rect 03 (lower one)
rect 04 (right )

The four rects have intersections so, let's say the player collide with an upper wall. Collision check with rects 01, 02 and 04 will return true – rect 03 will return false. And so on.

That said, if I get collisions with the upper or lower parts of my player o recall the last Y position but not the x position. Same goes for lateral collisions. All 4 rects returning true means I hit a corner so I recall X and Y.
I know that some combinations can mean two different types of collisions (like when testing true for rect 1 and rect 2 – it can mean that half the height of the player is colliding with a left wall or half of the width of the player is colliding with an upper wall). I suppose this can be dealt using my arrow keys variables. But before worrying with it I have to solve the first issue.
The following code works with lateral walls (90% of the times, don't ask me why) but you get stuck with horizontal walls:
  

1. background(0);
          q1=false;
          q2=false;
          q3=false;
          q4=false;
          oMundo.checkCollisions();
          println("Q1: " + q1);
          println("Q2: " + q2);
          println("Q3: " + q3);
          println("Q4: " + q4);
         
          if (q1 && q2 && q3 && q4)
          {
            oPlayer.restoreStateX();
            oPlayer.restoreStateY();
          }
         
          if ((q1 && q2 && q4) || (q3 && q1 && q4))
            {
              oPlayer.restoreStateY();
            }
           
          if ((q1 && q2) || (q1 && q4) || (q3 && q2) || (q3 && q4))
          {
            oPlayer.restoreStateY();
          }
         
          if (q1 || q4)
          {
           oPlayer.restoreStateY();
          }
           
          if ((q2 && q1 && q3) || (q4 && q1 && q3))
            {
              oPlayer.restoreStateX();
            }
           
          if ((q2 && q1) || (q2 && q3) || (q4 && q1) || (q4 && q3))
            {
              oPlayer.restoreStateX();
            }
   
          if (q2 || q4)
           {
              oPlayer.restoreStateX();
           }      
         
          oPlayer.storeStateX();
          oPlayer.storeStateY();    
   

Printing the variables to the screen shows that they are in fact working as they should. The problem is that even with proper care in the code, I still stop moving once hitting a wall and pressing two arrow keys at the same time.
Any ideas?

I know it's a stupid question, but, here we go.
When rendering my game i use a matrix stack to translate the scene coordinates while keeping the player at the center of the screen. So far so good.
When rendering the map elements I also render a rect with boundaries representing the limits I'm trying to use to detect collisions. The rects have alpha set to 0% unless I want to set it to a nice semi-transparent lime green during the game using a key that changes their alpha variable.  Bellow are some demonstration screenshots (note that my bizarre z_index layering method is working \o/):


After implementing a collision detect function based in this method ( https://forum.processing.org/topic/simple-collision-detection-21-9-2012 ) I can't seem to detect the collisions, unless I stop tracking the camera and don't use the matrix stack anymore, but if doing so, there goes my screen scrolling.

My map elements are rendered in a function that reads a series of coordinates from an Array (previously fed by a text file) and draws their images. Bellow is the code:

1.       for (fores=0; fores<layer1.size(); fores++)
           {
               latlong = oTile.setSprite(layer1.get(fores).getType());         
               image (latlong, pix2unit(layer1.get(fores).getX()), pix2unit(layer1.get(fores).getY()));
                g.removeCache(latlong);
                if (layer1.get(fores).isColideable())
                 {
                   fill (55, 255, 55, boundsAlpha);
                   rect(pix2unit(layer1.get(fores).getX())+layer1.get(fores).pushX(), pix2unit(layer1.get(fores).getY())+layer1.get(fores).pushY(), layer1.get(fores).get007X(), layer1.get(fores).get007Y());
                 }
           }
   

Let me explain.
First I probe what image should be drawn (latlong = yadda yadda...)
Draw the image
clear cache
check if the element (not the image) is set as having collisions
draw a rectangle that i'm trying to use as a collision map
*pix2unit multiplies my coordinates per 32 so my map file can be a lot cleaner.
*² – get007 is just a joke, it's my boundaries getter function. (get bounds, get bond, get 007).

Ok, in my draw function I first check the keyboard's input, push a matrix, call the render function, translate coordinates according to the input, pop the matrix, finally render the player. I'm using an int to track any collisions but not a single one is registered, however, if I manually feed the collision check function with overlapping coordinates (first rect coordinates manually set, second rect gotten from my array), if registers as it should.

So, is the matrix stack the reason I'm having such problem? And, if so, any suggestions about how to fix it. Finally, is this the best way to scroll the screen so the player characters will be followed?
Ok, thank you all in advance :)

5yhMe again :)

My game is a top viewed dungeon crawler. To build each room I use inteces os tiles, repeating 'em based in the coordinates i get from a text file.
I separate the sprites based in their "depth". There are 6 levels of depths so I render them in series - first one first and so on. This is a strange way to act like image layers.
I would like to know what could I do to improve my performance, since it gets unusable once i load a full featured room (about 500 sprites based in ~30 objects).
I would also like to know how much is Processing suitable for such use (sprite based full featured games(not simple demos)). Also, I'm usin minim to deal with music. Is it a good choice?
Thanks for your help!
Here's my draw code

       fill(0);
       rect(0, 0, 480, 320); // i need this rect to clear the previews frame screen
       oPlayer.controls(); //Use to check players input. Is there a better way?
       pushMatrix(); //using push/pop matrix to simulate a camera
       translate(cam_x, cam_y);
       oMundo.render0(); // render first layer (it reads the coordinate array everytime that is called)
       oMundo.render1(); // render seccond layer (it reads the coordinate array everytime that is called)      
       popMatrix();
       oPlayer.render(); // render the player (3rd layer);
       pushMatrix();
       translate(cam_x, cam_y);
       oMundo.render2(); // render 4th layer (it reads the coordinate array everytime that is called)
       oMundo.render3(); // render 5th layer (it reads the coordinate array everytime that is called)
       popMatrix();
       inter.drawUI(); //// render 6th layer (gui showing selected itens and player energy)
      

Ok, I'm kinda lost (again).
Here's the problem. For each layer of my game that I want to render I use a function that, based on an Arraylist of elements (called 'mundo').
The problem is that when running this specific part of the code my memory usage bumps like 50mb/sec. In 30 secs the game is obviously unplayable.
If it was C++ I would think about pointers. In java/processing I don't know how the memory is handled but I guess I'm filling it with n elements each frame - they are not using the same memory address. I'm I right?
If so, what do you guys suggest?
Trash collecting or something like that (does processing collects garbage)?
Bellow follows the code:
Thank you in advance


  void render0()
    {
      for (i=0; i<mundo.size(); i++)
        {
          if  (mundo.get(i).getZ_index() == 0) //z_index defines the number of the "layer"
            {
           
              latlong = oTile.getSprite(mundo.get(i).getType()); //type defines what sprite I should use 
              image (latlong, pix2unit(mundo.get(i).getX()), pix2unit(mundo.get(i).getY()));
          //I assume that for every new frame the image will be erased. Does it leave the memory?
            }
        }
    }


--------------------------------------------------------------------------------------------

//a sample from the switch/case that returns the image:

         case 0: //here's where the 'getType' from the render0 will land.
           tile = loadImage("imgs/sets/green_dungeon/simple_floor.png");
           return tile;

Inside this game I'm working on I maneged to make my player locked at the center of the screen, while scrolling the scene by doing like the following post:

https://forum.processing.org/topic/2d-moving-objects-chasing-by-camera

but here's the problem. Inside my draw() method i use the following "render" order:

render bg
render tileset layer0
render tileset layer1

render player

render tileset layer2
render fg
render GUI

The problem is, once I pop the matrix, the other "layers" (tileser layer2 and fg) will naturally lock'emself with the camera.
Does anyone have any idea how to avoid it?
Thanks in advance!

I'm creating this project that is a classical top down RPG and I would like to create a dungeon with a locked Z perspective but the camera should follow the x and y axis. My walls and some objects wold be polygonal and textured with oldschool textures.
The thing is, is there no other way to apply textures besides the beginShpae(), endShape() and vertex() method?
I'm I not allowed to import some 3D object (blender made maybe) and throw a texture on it?

Thanx ins advace :)

Hello, guys. I'm new here at this forum 

Here's the thing. I'm creating a simple game that consists in a small dungeon, viewed in a 2D aerial perspective – much like Zelda a Link to the Past.

Until now everything is going good, but I was wondering if I could use some lightning system to simulate the player's line of sight.

Thanks in advance!