* > I will push the positioning fix in a few hours and the moire pattern > is an artifact of our 16-level graphics blender. Looks really good > indeed, we only need to improve the speed. *
Thanks for working on this. Since the output is only 8-bits now I might be able to use more bits for precision and thus reduce artifacts. As for performance, I see that I'm doing way too many small allocations which do not serve any purpose in the current implementation. I'll try to work on these things. Anuj
