It would be cool if this could also do dithering based on error-diffusion methods. But that would probably be tough to add if the code is entirely based on fixed matrix methods like Bayer.
Still a really nice tool, thanks!
