[comp.graphics] To dither or not to dither

mike@relgyro.stanford.edu (Mike Macgirvin) (02/03/89)

	Since we don't have a color machine available to us, I have been
converting a lot of Sun color images to our monochrome systems, using
a variety of programs. I have found that most of the conversion programs
available use several different dithering techniques, but almost always
the resulting image has a very 'grainy' look to it.
	I have found that for most 8 bit deep images, dithering is 
unacceptable.
	Instead, I start with an 8 bit greyscale image (converting from 
color if necessary), get 8 bytes from input, and for each, if > 127, 
then I set the corresponding  bit of an output byte.
	This results in a very sharp monochrome image. Admittedly, a lot
of shades are lost, but isn't that what's supposed to happen when you 
compress color to mono?
	I just bring this up for folks who are writing conversion programs.
Dithering is almost taken for granted as 'the ONE TRUE CONVERSION'.
Don't overlook simpler methods that may achieve better results.

/*
************************************************************************
	
 m     m     k              Mike Macgirvin
 m m m m  o  k  k   eee     Stanford Relativity Gyro Experiment
 m  m  m     k k   e   e    Stanford University
 m     m  i  kk    eeeee        (415) 725-4117
 m     m  i  k k   e        ARPA: mike@relgyro.stanford.edu (36.64.0.50)
 m     m  i  k  k   eeee    UUCP: /dev/null

	This room has been punched out.
		Now where are all the gazingas?
*************************************************************************
*/