I'm doing some research in trying to replace the current bilinear resize with something more intelligent. Since I'm hopelessly bad at math I don't think there is going to be any really cool solutions like vector transformation. I have started by implementing scale2x and it could be the basis of the final routine. Or not. The good thing about resizing is that it's quite logical and easy to understand. You need to figure out how to translate the information to bigger or smaller size without losing the form of the original.
With tile/sprite editors there is a special need to maintain the original pixel's color which means interpolation routines are giving somewhat poor results. Developing a really good resizing routine is difficult, but an interesting subject.
Subscribe to:
Post Comments (Atom)
0 comments:
Post a Comment