Since they are no longer actually taking fixed point parameters the 'x' suffix is
no longer appropriate. To maintain support for sub-pixel precision the
corresponding interfaces that were taking integer parameters now get patched
to take float parameters instead.
