🚀 go-pugleaf

      Everything you always wanted to know about color (almost) and were'nt
afraid to ask about.
      
   First Rule:  Color is a royal pain in the ass.
   Second Rule:   A human beings eyes are the worlds best camera.
   Third Rule:    Analog is natural, Digital is a mutant from alpha centuri.

   Way back in the mists of antiquity, someone at apple must have liked color
hence they included a little blurb in Inside Mac about drawing in color. What
this enabled you to do was draw on one of 32 color planes at a time. i.e. you
have thirty two colors to draw in or effectivly thirty-two different screens.
   This is the way the IBM-PC (boo hiss) works with an EGA. You have four planes
Red, Green, Blue, and Intensity. Giving you 16 colors. In addition what exactly
a "red" pixel means etc. is changable giving you 64 possible colors 16 at a time
This is referred to as planar graphics. The advantage of this is the hardware 
for the video card is easy to build, each gun simply scans through the 
appropriate (rg or b) screen and turns on or off as appropriate. The software is
a bitch to write (esp on non-68xxx family uP) because drawing a line consists of
turning on and off the appropriate bit in a section of eight (rg or b) pixels.
This is also slower.
    The new machines handle color differently. A sepecific color is difined as
16 bit intensity levels of RG and B.  This color is then converted by a color
lookup table into the nearest 8-bit color. (this table is changeable so that
the user can define what colors he needs most, incendentally). These colors
are then stored in memory in 8-bit chunks (or 4,2 or 1 depending on the memory
on the video card). Apple refers to this as chunky graphics. This makes the
hardware a little more difficult to design (but not greatly) and the software
much faster and simpler (except in xor drawing etc. all the uP needs to do
is store to the appropriate location vs. load, set bit, store). 
   Thus the new machines take a 48-bit color spec, convert this to an eight
bit color-pixel. And save the pixel/colors together.
   The old machines effectivly drew in 32 monocrome screens at one at a time.

   picture:   old pixel:   R-----screen size in bytes-------G----screensize---B
              new pixel:   RGB
   where R is the red bit/pixel G is the Green bit/pixel blue is the Blue
   bit/pixel.

   Thus the old-machines had at most 32 different colors since pixels were 
  either on or off. ( assuming each of the different color planes was actually
 a different color not intensity level or anything similar. i.e. using 6-bits
 you could have Red(on/off) Red Intensity(dim/bright),g,gi,b,gi giving 4^2
colors. IM seems to imply, though that the layout assumed by QD was similar
to the EGA.
   The new machines have (2^16)^3 different colors 256 of which can be used
at once. 

  Where the three rules come in: You can see, (and tell the difference between)
all (2^16)^3. (if you don't believe me assume there are a million different 
shade of color in your room. Now turn off the lights. There are now a million
new shades of color in your room.)  This is because your eyes work on an analog
level. 2.3 is 2.3 etc. The computer has to divide all the colors your
eye can see in some fashion. Presently, into 256 slots. To go to sixteen bits
of color would require 600k of video memory. to go to a full RGB would require
1.8 megs of memory to store the screen. (ouch)
  Thus the practical limit on display depth is eight bits. or 300k. This means
that no computer screen will ever be flawless. For an artist, the computer
will always be too limited a medium for serious work. (unless memory becomes
REALLY cheap.)

   I hope that clears up some of the arguments about what the new machines, do
and do not, do.

   Pierce Wetter



C, n.:
	A programming language that is sort of like Pascal except more
like assembly except that it isn't very much like either one, or
anything else.  It is either the best language available to the art
today, or it isn't.
		-- Ray Simard

--------------------------------------------

wetter@tybalt.caltech.edu

--------------------------------------------