Halftone processing
The grayscale image and color image are subjected to halftone processing.
The following two methods are available:
Error diffusion method, which has high quality but requires many calculations.
Blue noise method, which has low calculation and the same quality as the error diffusion method.
Original image (256-level color)
RGB 4-levels using error diffusion method
RGB 4-levels using blue noise method
Original image (256-level grayscale)
4-levels using error diffusion method
4-levels using blue noise method
The error diffusion method produces sharp images, but the gradient parts become very apparent.
The blue noise method adds a slight amount of noise compared to the error diffusion method. This results in smoother gradient parts, as noise is also added to them.
In order to improve the apparent parts in the error diffusion method images, a process that combines random noise with the error distribution is performed, resulting in the same images as the blue noise method.
The processing time is significantly faster for the simple blue noise method algorithm.
A program for quickly creating threshold matrices for the blue noise method (15 seconds for a 256×256 matrix) is also provided, allowing for customization.
