🚀 go-pugleaf

>In article <3csf57$l...@mojo.eng.umd.edu> John Sabean writes: 
>
>In article 
<Pine.SUN.3.91.941204190400....@aloha.cc.columbia.edu>, 
Tom Hong <th...@columbia.edu> writes:
>>On Sun, 4 Dec 1994, Steven Mark Hartland wrote:
>
>>>Some systems do use upto 64 bit colour representation,
>>>but you wont find it on a PC. I believe that 32 bit colour
>>>representation is the highest that is found on PC graphics cards.
>>>Steve. 
>>
>>	I have never seen *any* computers display 64-bit color, let alone
>>PCs that display 32-bit color.  The highest is 24-bit-per-pixel palettes
>>(a PC can't display all the colors simultaneously anyway), at 16.7 million
>>colors, e.g. 2 to the 24th power is 16,777,216 colors.  I don't think
>>there is such a thing as 64-bit color, let alone 32-bit color. 
>>64-bit-color per pixel is simply impractical because 2 to the 64th power
>>is 1.84 x 10 to the 19th power, or about 18 *quintillion* colors.  32-bit
>>color is flat-out impossible on a PC because 2 to the 32nd power is 4.3
>>*trillion* colors.  I don't know if even graphics workstations can do
>>this.  As I said before, when people say 24, 32, 64 or 128-bit video
>>cards, they don't mean the amount of colors that can be displayed nor do
>>they mean how wide the bus width is.  The designation simply means the
>>highest resolution at which the card can display 24-bit color palettes. 
>>
>
>Actually, some of the SGI's can do more than 24 bit color. I forget if it
>is 32 or 64 bit, but I know its more than 24 bit. I had the same discussion
>with a few friends, "Could you feasibly ever need it or use it?"  We couldn't
>really think of any applications for it. We got this information from a catalog
>from Silicon Graphics. It does indeed exist..... (Those guys are crazy! :)
>

There seems to be some confusion here.

The use of palettes in graphics modes for boards stops at around 16 bits or 2 
bytes
(e.g., 64K entries).
When a palette is used in a graphics mode, it is an intermediary between the
screen's stored values and the actual color generation.  The values stored in
the palettes are those 24-bit, 32-bit or 64-bit color values.  The screen's 
stored
values just index the palette.

When you need to allow for more than 64K possible colors, boards go right to
having the screen's stored values be the actual color values (i.e., no palette),
thus being limited by the board's color representation scheme (3, 4, or 8 
bytes).
Someday, there may be boards supporting multiple representation schemes.

Earlier graphics modes used fewer bits for color representation.  For instance,
18-bit color, with 6 bits each for red, green, and blue, has been used.  This
was adequate when the pixels were very coarse (remember EGA, etc.).  But
as the resolution increased (i.e., much smaller pixels), shading was found
to be too coarse to be acceptable.  A standard example that was used was
shading on a shiney pool ball reflecting an off-axis light source.  The shading
of the light's reflection on the ball came out in blatantly visible layers.
If you go to 8 bits per color (24-bit color representation), you get a more
tolerable gradation in shading, particularly on the current mass-market level
of higher resolution.  This (24-bit color) may be visibly pushing its limits
as 1600x1280 becomes more common.

Graphics systems that want to support high resolution photographic quality
work can go beyond 24-bit color.  Some of those color representations are
not based on red-green-blue color generation, either.

Now, some of the talk of 32-bit or 64-bit boards is actually refering to the
width of the bus to access video (i.e., the screen's) memory rather than the
number of bytes in the color representation.

I hope this helps some.




Good Day
Dick
Dick.Me...@DaytonOH.NCR.COM