Nah, though that would make sense, too. I'm just using a simple exponential curve that samples higher in the curve the closer a pixel is to the middle of the image.
