Well, let's supose i have a spritesheet like this:

If you want the block for drawing, you can use the SourceRect param on the SpriteBatch.Draw overloads and don't need to copy each part out into a separate texture.

Intresting! That can be usefull! Thanks for the help!

You could make an extension method like so:



Note: I just typed this, it's not tested.

Edit: Added the where constraint.

I already solved with the first method. Thanks!

Craig Martin:

You could make an extension method like so:

public static Texture2D Transfer<T>(this Texture2D srcTexture, Rectangle srcRect) where T : struct

{

T[] buffer = new T[srcRect.Width * srcRect.Height];

srcTexture.GetData<T>(0, srcRect, buffer, 0, buffer.Length);

Texture2D result = new Texture2D(srcTexture.GraphicsDevice, srcRect.Width, srcRect.Height);

result.SetData<T>(buffer);

return result;

}

Note: I just typed this, it's not tested.

Edit: Added the where constraint.

I don't mean to be rude but I don't think this is a good idea to use at all. You'll end up with a bunch of textures that need to be swamped if you are using it on a animation sprite.
The whole point of using sprite sheets and rendering just portions it to avoid the GPU to be switching textures.
Imagine that you are always rendering you main character, it has an animation, every few milliseconds you switch it's texture, and if the GPU has is memory full it was to erase the previous texture and load the new one. If you have everything on the same texture swamping is not necessary. It's already there. Usually I use a Texture2d for each animation.

I agree with you David, it is not the right way. I already showed the right way, but they asked for an alternative, so I gave one :p