Recently I used the official XNA 2D per pixel collision checker to replace my rectangle collision checker in my game. Now I find I have an issue with the fact that the XNA collision checker uses single sprites while I use a sprite sheet. Looking all over the internet I have yet to find anyone who has figured out/attempted to figure out a solution to this problem. So here it is:

I have the code done, it works, but it only works on large images with a lot of pixels strung together. (Solid objects)
I have a couple of objects that have only 1x1 pixel lines that are the top/bottom of the images. Such as a player controlled lander with thin landing struts (4x1 pixel) and the landing pad (50x1 pixels).
As I understand it the Color[] ImageData and Image.GetData works by taking all the pixels in a texture image and stringing them out along an X x 1 pixel line and checks each pixel along that line. Unfortunately my sprite sheets are large, about 200X300 pixels or more while my viewable sprites are ~32x32 pixels. So I have this collision checker checking 200X300 pixels with each pass to find out if a single pixel of a 32x32 frame of the sprite sheet has collided with another image. The problem is with only a 1X1 pixel collision point the chance of "missing" a collision check and passing through an object are extremely high and even if I were to "smudge" the numbers my player controlled object would most likely land inside the pad rather than on it.

For those who need it here's my code.
Note: I use an OOD that goes (Physics - Sprites- Object - Compiler- Game1) so what you'll see here is the flow of the code, not its actual location in my design scheme:



Now I've tried to change how the color data is collected to shorten the collection zone to that of the displayed image from the sprite sheet. Now, although I've gotten the code to work and it doesn't crash the game the texture I "build" is merely a frame that's the same size as the display area and has no color data to collect so no collision is ever detected. Like so:


If I attempt to smudge the above code by collecting frameSize.X * frameSize.Y and then giving it to "spriteImage" I get an error that says "The size of the data passed in is too large or too small for this resource." which in this case I assume is too small.
I've thought about lying to the program to make it think that the spriteImage is the same size as the frameSize.X * frameSize.Y, but I have no idea how to convert a Texture2D image into and INT that can then be altered.

If anyone has any ideas I'd love to hear them.
I've also posted this on Ziggyware.com and someone mentioned passing the Rectangle argument through the GetData code, but to do that I'd have to turn the Color array into a Texture array and I don't know what that would do to the rest of my code or how it'd interprete the data so that's out unlesss someone can explain to me how to do that.

SOLVED!

Thanks go to ThinkTank from Ziggyware.com for leading me down the right track.

Here's the new code that only checks the viewable sprite area of the sprite sheet and allows 1x1 pixel collision!



As you can see it's a little chunky, but this returns the mipmap level (0 since my sprites are not scaled or rotated), the rectangle size which is the same as my drawing rectangle size, where to send the color data, the current start point of the data collection and the total number of pixels in the data collection area.

All of the area and pixel calculations are the same as the draw area calculations so the code repeats itself a bit. But this is how you can use sprite sheets and per pixel collision detection together!

If you need to know more on how this all interacts with my current OOD scheme feel free to ask.

Otherwise, Enjoy!

Hi there,
First of all, i want to apologize for my bad english.
I'll do all I can to be clear.

I try to integrate your GetData Line, but i've some problem to understood all of this.

My problem is that all times i use this line :
Instead of :
My program just can't compile with a problem of index or somethng like this.

My code is the next :
(Problem is between //////////)




I use the same Intersect fonction that in the 2D pixel intersect tutorial.

And to control my collision :


Thanks in advance.

Best Regards.
Maldus

Sorry to re-post on this thread, but i'm still searching a better solution (or a working solution ^^) to do collision with spritesheet !

So if someone can explain me what can i do to solve my problem, because it's the only thing that i don't know how to do...

Thanks.

Best Regards
Maldus.

I finally manage to make it work but in another way.

If someone need it, ask me ;)

Maldus.
Best Regards.

Yeah, please post it. There must be a way of getting the array for the current frame only, and iterating that, rather than iterating that in the entire sheet.

I use another fonction :


I draw the playerRectangle by enter the size of sprite and his position.
After, if this two rectangle collides, i see if there is some pixel that are over  20 in alpha that collide
.

Maldus.

Color[] myData = new Color[20 * 20]; texture.GetData( 0, //Mip map level, for standard sprites this will always be 0 new Rectangle(0, 0, 20, 20), // The section of the sprite that you want to retrieve myData, // The array to put it in 0, // Index offset of the array to start copying into myData.Length // The number of elements you want to copy ); Just so you know.