• Create Account

## Spritesheet Algorithms

Old topic!

Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.

6 replies to this topic

### #1EricsonWillians  Members

288
Like
0Likes
Like

Posted 10 March 2014 - 06:51 PM

Suppose that you don't use an engine like "Unity3D" that has some built-in ways to deal with spritesheets, how would you deal with the "spritesheet problem"? As it is known, spritesheets are better than loading separate .png files for animation purposes (Considering that a character has movement, atack, defense, death, etc, animations). Most of the people, I guess, would take the 0x, 0y pixel colorkey and make it transparent for the whole image AND cut manually all animations and store them in a collection of some sort.

The key point here is automation. If a spritesheet has irregular sprites (For example, the first one is a rectangle of 30 by 25 pixels, and the second one is irregularly far from the first sprite image), one cannot implement a function to cut all next sprites based on the rectangle of the first, because all sprites would have parts missing, etc.

Manually storing every rectangle position in the sprite sheet seems to be a great option for a general game, but the same does not apply to an engine. I'm developing an engine on Pygame/Python, and, therefore, I want a clever way to separate/cut the inner sprite rectangles and return them as a list.

The solution? Looping pixel per pixel and applying some logic based on the colorkey. How would you do that? Would you bother to implement such a function? What do you think about it? For the sake of the topic, here's my method for cutting based on the first rectangle's position and size (It does not work if the spritesheet is irregular):

def getInnerSprites(self, xOffset, yOffset, innerRectWidth, innerRectHeight, innerRectQuantity):

"""
If the grict is a sprite sheet, it returns a list of sprites based on the first offsets and the width and the height of the sprite rect inside the sprite sheet.
"""

animation = []
if self.isSpriteSheet == True:
for i in range(xOffset, innerRectWidth*innerRectQuantity, innerRectWidth):
print i
animation.append(self.getSubSurface((i,yOffset,innerRectWidth,innerRectHeight)))
else:
print "The Grict must be a sprite sheet in order to be animated."

return animation


I'll try to implement the "getInnerSpritesByPixel()" method. Spritesheets are a key thing in a complex game like a MMORPG, where almost every item has its own animation. Such method is more than necessary.

Creator and only composer at Poisone Wein and Übelkraft dark musical projects:

### #2Danicco  Members

449
Like
0Likes
Like

Posted 10 March 2014 - 08:25 PM

Would you bother to implement such a function?

No, I think your problem isn't really a problem for your engine to do... you can first implement a simpler option to cut the spritesheet in equal rectangles, and that might work for simpler games, so you'll only have a single rectangle for that object and you just change the UVs of the texture.

As for irregular rectangle sizes, instead of trying to automatize (ie. figuring what the user wants) you can let the user TELL you what he wants. If you have an GUI you can put a list that he can add the sprites' areas manually via typing, and maybe then even evolve to a graphical tool that the user selects an area and you insert a new sprite area to the list.

I'm coding an engine and this is what I do to deal with spritesheets:

//This will load "MyImage.png" a 500x500 image, and cut it in pieces of 50
Sprite* myCharacter = Resources.Load("MyImage.png", 50, 50);
myCharacter->currentFrame = 10;

//This will load "MyImage.png" a 500x500 image, and cut it in pieces of 50
Sprite* anotherCharacter = Resources.Load("MyImage.png", 50, 50);

//And now I'm changing frame 0's width and height to something else
anotherCharacter->Frame[0]->SetSize(75, 75);

//But I want to be lazy to type in code so I got a .txt file like this:
//frame=0 x=0 y=0 w=75 h=50
//frame=1 x=50 y=0 w=50 h=50
//frame=2 x100 y=0 w=25 h=50
//And a method to auto load it:


This will let the user (game coder) decides how he deals with it, either by using pre-defined rects for everything, or manually typing, or loading from another file (maybe even generated from a map editor of their game).

### #3EricsonWillians  Members

288
Like
0Likes
Like

Posted 10 March 2014 - 08:49 PM

Would you bother to implement such a function?

No, I think your problem isn't really a problem for your engine to do... you can first implement a simpler option to cut the spritesheet in equal rectangles, and that might work for simpler games, so you'll only have a single rectangle for that object and you just change the UVs of the texture.

I did that already, but as I've said, it does not work with irregular spritesheets (And I want a more definite function to work with in any kind of project).

As for irregular rectangle sizes, instead of trying to automatize (ie. figuring what the user wants) you can let the user TELL you what he wants. If you have an GUI you can put a list that he can add the sprites' areas manually via typing, and maybe then even evolve to a graphical tool that the user selects an area and you insert a new sprite area to the list.

I don't have a GUI for now, and in a certain way, the "automatization" would still let the user tell what he wants. I want to cut perfectly all sprites in a row in the spritesheet. Generally, each row represents a certain kind of animation, so, it would just be necessary to call the "pixel-perfect-cutting function" for each row, and fire the animation according to the necessity (Separating each row in a list of sprites, and using another function to loop/animate them). (And at the same time, providing a "manual way" to deal with the matter).

Edited by EricsonWillians, 10 March 2014 - 08:52 PM.

Creator and only composer at Poisone Wein and Übelkraft dark musical projects:

### #4Stainless  Members

1875
Like
0Likes
Like

Posted 11 March 2014 - 02:21 AM

The way I have done this in the past is to define a single RGB value as a control.

The artist fills the image with this control value and then cuts out areas for each sprite.

The area has to be at least one pixel bigger than the sprite in all directions to handle blending artefacts in the texture lookup.

A more modern approach would probably to use the alpha channel as well. Use a control value of 0x00FFFFFF (fully transparent white).

### #5TheComet  Members

2678
Like
1Likes
Like

Posted 11 March 2014 - 05:42 AM

If you cut them irregularly using some kind of recognition algorithm, you're just asking for the animation to screw up. How are you going to determine the correct offset for every frame so the animation frames are correctly aligned over one another?

Don't over-engineer things. Sprite sheets have always been designed so every frame has equal size. What purpose is there in changing this? The few bytes of storage space it could save is insignificant, and frankly, I don't see any other motivator to do what you propose.

Edited by TheComet, 11 March 2014 - 05:43 AM.

"I would try to find halo source code by bungie best fps engine ever created, u see why call of duty loses speed due to its detail." -- GettingNifty

### #6haegarr  Members

7189
Like
1Likes
Like

Posted 11 March 2014 - 06:06 AM

As said earlier, you should try to move the responsibility for such a thing into the toolchain. So a sprite sheet (generally a texture atlas with a single LOD) consists of both the texture and a list of texture clips. Each clip is named (for retrieval) and provides the sequence of texture co-ordinates.

When the sprites are delivered e.g. in a PSD file with several layers, then the toolchain can use e.g. a shrinking frame based on alpha (or color key) to determine the smallest axis aligned bounding-box (or an octagon box) on each layer, and pack them into an atlas using one of the bin packing algorithms. The name of the layer can be used as name for the clip.

When the sprites are delivered e.g. in a single layer of a PSD file, and the sprites can be separated by axis aligned non-overlapping boxes, then PSD's slice mechanism can be used. Each slice is translated into a clip. Here the artist is responsible for the correct packing.

If all of the above don't work for you: When the sprites are delivered in a layer and no slicing is available (or could not be done because the slices would not be rectangular), then the toolchain can generate a background layer, flood fill all pixels in the background layer that are not covered by the transparency alpha / the color key in the foreground layer, apply a thinning algorithm (a kind of morphological operation), and vectorize the pixel traces by using a neighborhood search. In dependence on how you do the vectorization, you may need to flatten the sequence of line segments by averaging or a more sophisticated algorithm, just to reduce the number of segments.

### #7EricsonWillians  Members

288
Like
0Likes
Like

Posted 11 March 2014 - 09:33 AM

Don't over-engineer things. Sprite sheets have always been designed so every frame has equal size. What purpose is there in changing this? The few bytes of storage space it could save is insignificant, and frankly, I don't see any other motivator to do what you propose.

I did not pretend to make frames of different rectangle sizes, for I know that that would not be a good thing. I did iterate through all pixels in the spritesheet image, but soon I realized that that approach was not allowing me to see the whole forest, for being too much distracted by the trees (Iterating each pixel proved that I can't see the sprite, even if I detect where the colorkey is different (Nevertheless, I still think it's possible.)).

If all of the above don't work for you: When the sprites are delivered in a layer and no slicing is available (or could not be done because the slices would not be rectangular), then the toolchain can generate a background layer, flood fill all pixels in the background layer that are not covered by the transparency alpha / the color key in the foreground layer, apply a thinning algorithm (a kind of morphological operation), and vectorize the pixel traces by using a neighborhood search. In dependence on how you do the vectorization, you may need to flatten the sequence of line segments by averaging or a more sophisticated algorithm, just to reduce the number of segments.

But then, we got this good answer. That's what I was trying to do, but had no idea of how it worked (Nevertheless, I still think that there's a way of detecting and separating the sprites by iterating through all pixels). Unfortunatelly, it is a bit more complex than what I expected. I've implemented a general solution to it (An enhanced function compared to the first one that I posted):

def getInnerSprites(self, innerRect, times, step):

"""
If the grict is a sprite sheet, it returns a list of sprites based on the first offsets and the width and the height of the sprite rect inside the sprite sheet.

# [0] = x, [1] = y, [2] = w, [3] = h.
"""

innerSprites = []
if self.isSpriteSheet == True:
for i in range(innerRect[0], innerRect[2]*times, step):
innerSprites.append(self.getSubSurface((i,innerRect[1],innerRect[2],innerRect[3])))
else:
print "The Grict must be a sprite sheet in order to be animated."

return innerSprites


Before, I had problems with irregular sprite sheets, when I supposed that the next slices would be exact as the first one. But, I realized that instead of placing the next rect right on front of the old one, I could change the "step" of the iteration and get precision with both regular and irregular spritesheets.

Edited by EricsonWillians, 11 March 2014 - 09:39 AM.

Creator and only composer at Poisone Wein and Übelkraft dark musical projects:

Old topic!

Guest, the last post of this topic is over 60 days old and at this point you may not reply in this topic. If you wish to continue this conversation start a new topic.