doyle@b11.ingr.com (Doyle Davidson) (08/05/89)
I have a question about how color allocation is done on a StaticGray
color display. I have a client (twm for example) that has its color
parameters specified by names like "red", "blue", "thistle" etc...
Upon running this on a StaticGray display, I discovered that the
following occurs:
Color R G B Gray level
-------- --- --- --- -------------
"red" 255 0 0 33%
"blue" 0 0 255 33%
"green" 0 255 0 33%
"black" 0 0 0 0%
"white" 255 255 255 100%
Now, I have read several interpretations and none really match this
except for the code itself!
1) O'Reilly & Assoc:
p.207 intensity = (0.30 * red) + (0.59 * green) + (0.11 * blue)
really happens? WRONG! (no flame intended at O'Reilly & Assoc)
2) O'Reilly & Assoc:
p.207 next sentence:
"If only a GrayScale visual is available, the red value ONLY will be
used for gray scale intensity"
really happens? WRONG! I thought that #1 just said... owell.
If #2 is right then "green" and "blue" should map to black.
3) I read the protocal on the distribution tape: nothing that I found
said anything specific and is not part of the DDX layer :-(
I remember reading in something that R G and B should all be set to
the same value by the client because it cannot be sure which one of these
three that the server will use as an intensity. I cannot find it
now to quote from it unfortunately.
4) I looked at the code: it searches through the RGB values for the Grayscale
color table (which are all the same) and finds the color that minimizes
the following (in pseudocode):
in dix/colormap.c: FindBestPixel()
for (all RGB pixel values in table)
value = (red_asked - red_table)^2 + (greed_asked - green_table)^2 +
(blue_asked - blue_table)^2
/* like a 3D distance formula but no square root */
if (value < best_so_far)
{
pixel = current_pixel_being_testing_from_table;
best_so_far = value;
}
really happens? Yes!! Hey this is it but not documented ANYWHERE
Hey you guys at MIT, got the answer?
Thanks in advance,
Doyle (I RTFMs and am still confused :-O ) Davidson
------------------------------------------------------------------
Doyle C. Davidson |
Intergraph Corp. | These comments are...
Workstation Graphics Standards |
1 Madison Industrial Park |
Huntsville, AL 35806 |
(205) 772-2000 | X-clusively my own.
|
..!uunet!ingr!b11!doyled!doyle |
------------------------------------------------------------------rws@EXPO.LCS.MIT.EDU (08/07/89)
I have a question about how color allocation is done on a StaticGray
color display.
You get the "closest RGB values provided by the hardware". That's all the
protocol says, for better or for worse.
I remember reading in something that R G and B should all be set to
the same value by the client because it cannot be sure which one of these
three that the server will use as an intensity.
This is only when storing (StoreColors) into a GrayScale map.
finds the color that minimizes the following ...
really happens? Yes!! Hey this is it but not documented ANYWHERE
Because it's not part of the protocol specification. (Also, you missed an
important part of the implementation, which is that ResolveColor gets called,
and it can transform the RGB values. For example, in the server I'm typing at,
it transforms to a gray value according to the 30/59/11 percentages. Since the
GrayScale map happens to have been initialized similarly, the FindBestPixel
code pretty much ends up actually looking for equality. Since ResolveColor can
vary with each ddx, it's kinda hard to provide a single description of what
the sample server chooses to do for this implementation-dependent detail.)