Question

I’m currently trying to create a game for a tablet device for Android using a SurfaceView. I have sprites and what not, but I’m struggling if I should use a whole spritesheet (With every animation in it), or just seperate the sprites on image files.

For example, my spritesheet is named bear.png, which contains 16 images inside. I can animate them beautifully using Rect (Thanks to mybringback for this):

canvas.drawBitmap(bitmap, rectSrc, rectDest, null);

But, I fear doing so would grow to larger than the heap size eventually.

However, there is another method. By seperating each image (bear1.png, bear2.png until bear16.png), I will also be able to create an animation with these sprites via:

Bitmap bitmap = Bitmap.decodeResource(context.getResources(), context.getResources().getIdentifier("bear" + index, "drawable", "com.example.game"));

then calling that to the canvas:

canvas.drawBitmap(bitmap, bitmapX, bitmapY, null);

Which also works fine for now. And of course, it greatly reduces my heap size since I’m only using 1 image at a time. But, because well.. Because, I’d be recalling Bitmap.decodeResource() at almost all times, I fear that when putting many animations together at one time, it may cause a significant lag.

So in short:
– Spritesheets: Increased heap size (which may eventually lead to an OutOfMemory error), but better performance
– Sprites on different image files: Decreased heap size, but may hinder performance

I’m not entirely sure which approach to use. Would it actually be okay if I go with the second method (Sprites on different image files)? Or would it be better with just spritesheets?

Answer 1

Never mind, I just answered my own question. It appears that rendering Bitmap.decodeResource() on every iteration lags the whole canvas, rather than rendering actual spritesheets and using Rect. Performance-wise, if experiencing too much lag on the canvas, there’s no other way other than to shrink images using inSampleSize or using GLSurfaceView.