Agenda for the 4/22/98 Grex Board of Directors Meeting 0.00000077 Initial Gavel Pounding - janc < 1 minute 0.00000063 Treasurer's Report - aruba 10 minutes 0.00000060 Fund Drive - aruba 15 minutes 0.00000054 Publicity Committee - mta 5 minutes 0.00000049 Technical Committee - staff 10 minutes 0.00000044 New Business - all ??? 0.00000041 Final Gavel Pounding - janc < 1 minute The meeting will be at 7:30pm, probably at the ITI cafeteria (though I still have to confirm that). If there are other items that should be on the agenda, please let me know.44 responses total.
i may not be able to make it, as i am working the Harry Connick, Jr. show at Hill Auditorium that day. On the other hand, I might be able to make it, though i would likely have to leave by 9:30...
Probably we should have another topic on procedures for selling/loaning Grex equipment. I'm not sure that we wouldn't prefer to let that be discussed more here on-line before we try to establish procedures for this, but I think we could reasonably approve selling Scott the monitor.
What? Nobody guessed the agenda item numbers? Hint: It's another physics/chemistry thing.
hrm.. i can get you folks in, but i need to get out by 9:30 at the latest from iti. There's an important revelation at 10: that i need to see.
Thanks Jared, that will be fine.
I haven't the foggiest about the agenda item numbering scheme this time. It's probably the sizes of atoms in miles across, though, or something like that. (-:
This response has been erased.
Another hint: The units are meters.
Given the small amounts (in meters) and the fact that there are seven of them, I'll guess: the wavelengths of the seven colors in the visible spectrum.
I would make a snide remark about the foolishness of this item numbering business, but then I'd have to scribble it... ;-)
How do you pick the 7 colors in the visible spectrum? There are an infinite number, or referring to the seven simply named colors of the "verbal spectrum" - has someone actually defined a fixed wavelength for the center, or mean (arithmetic or geometric or harmonic?) of each? See, item numbering isn't so foolish as it raises interesting issues (which some wet blanket will tell us to take elsewhere).
These are physicists we're talking about, Rane, of course they
think they can define reality with nothing but numbers... >8)
I have vague recollections of seven "official" wavelengths
for the colors on one of those annoying posters in my HS
physics class. Unlike janc, they didn't make us memorize them.
Here's something from a Web page I just snagged:
Approximate wavelengths for various types are:
Wavelength
Type in Centimeters In Å
y-rays 10^-12 to 10^-10 0.01
X-rays 10^-9 to 10^-7 0.1 to 10
Ultraviolet light 10^-6 to 4 X 10^-5 100 to 4,000
Visible light 4 X 10^-5 to 8 X 10^-5 4,000 to 8,000
Infrared light 8 X 10^-5 to 0.1 8,000
Radio waves 1 to 10^5
Note that the visible spectrum is defined as being between
.00004 and .00008 cm - or, between .0000004 and .0000008 meters.
And the range in janc's list is .00000041 to .00000077 meters.
I think I may be on to something.
People have studied how many colors are recognized in different cultures. As it happens, there is actually a great deal of similarity, which suggests that there is actually a great deal of commonality in the way we all see colors. There are quite a few interesting exceptions, of course. Various forms of color blindness are common. There are also several interesting mutations in color pigments that really can make colors look "different" to different people. Many of these mutations are sex linked, which means men are somewhat more likely to see colors "differently" than women.
We, of course, only see *three* colors - that is, we have three types of color receptor cells, each of which respond to a narrowed range of colors (which over lap somewhat). But the brain only receives three types of stimuli from these cells, and creates all the colors we "see" from those. This, of course is why only three color guns are needed in our monitors for "millions of colors". You are definitely on to something, Rob, but I doubt very much that real physicists set those numbers, or if they did it was because some inquiry arose from a layperson (or a secondary school teacher). So the question is, who chose those arbitrary wavelength values for the arbitrary seven-color subdivision of the visible spectrum?
janc said something about an ANSI standard....
How is it implemented?
no clue. i suppose to some extent it would have to be completely arbitrary.
There are only 7 colors that we can see. ROY G BIV or Red, Orange, Yellow, Green, Blue, Indigo, Violet. These collors cannot be broken down into sub-colors. That is, a prism will not break these colors down any farther. If a prism is placed in front of sunlight, then another prism placed in front of just one of these colors, it will remain that color. All other colors are seen only in our minds.
We can "see" only three (3) colors. All others are combinations of those,
according to how the three (3) types of color sensing cells in the eye
are stimulated. We can *distinguish* thousands of colors, but that's
because the brain can distinguish different admixtures of just those three
(3) colors. Haven't you ever looked at a spectrum, bru? There is a
continuous variation of color across it - not seven stripes, each of
a solid color.
I was uncertain what "256 colors", "thousands of colors ('16 bit')"
and "millions of colors ('32 bit')" meant, until I started working with
a video board. It is simple enough. 16 bits can encode 1028 combinations
of the three (3) colors reported to our brain. However I have wondered
what admixtures (relative proportions of the three (3) colors are encoded.
forget
The item numbers are the boundaries between the wavelengths of lights associated with the different colors. They are, of course, not real physical constants, but rather somebody's guess at where typical human perceptual boundaries (which are probably at least partly cultural) lie. I've seen the same table with the same values in several intro physics books, but this one is from Halliday & Resnick. It is a useful thing for a physicist to know roughly how frequencies of light map into spectral colors, just as it is useful to know if she is talking about visible light, x-rays, or radio (also arbitrary divisions from a purely physical viewpoint). It's been a while since I taught about color in a computer graphics class, but no computer can display the full range of colors the human eye can perceive. The spectrum contains only pure, single frequency colors. Many things we perceive as different colors are not pure frequences (look for brown or beige in the spectrum sometime), but mixtures. You can not make all possible spectra by mixing 3 standard spectra. You can do well enough for most purposes though.
If you pick the *right* 3 colors, the the only colors you can't reproduce by mixing are way out on the fringes. This problem isn't restricted to computers - it applies to any color video monitor, and it also applies in painting, theatrical lighting, textile dyes, or anywhere. There are variety of ways computer video systems can represent colors. However, most common systems today use "color maps", because it is cheap and extreme versatile. The classical example of this is most "256 color" systems. What most of these systems do is to store bytes in memory, each byte representing one pixel on the screen. Rather than directly mapping a numeric byte quantity into an RGB triple, however, the byte is used to look up a 24 bit quantity in a separate color map -- a separate segment of memory in the video adapter. The 24-bit quantity is then used to separate distinguish between 256 possible intensities for each of the electron guns. The quality of the resulting picture can be *almost* as good as a real 24-bit deep picture, but in addition, it is also possible to play very interesting animation tricks - by changing the color map values used, things can be changed very quickly on the screen without the necessity to redraw things. Newer displays tend to have more than 8 bits for a pixel, of course, but most still have a color map that can be changed on the fly. Even though we know a lot about the biochemistry of the pigments in the human eye, as it happens, we *still* don't know how color is represented in the neural signals sent. There are, as a result, a variety of effects we can't explain scientifically. For instance, certain rapidly spinning black & white shapes can create the illusion of color to the human eye.
So do I win anything for coming the closest to getting it?
You win our eternal (well, temporary) gratitude for stimulating this fascinating discussion (which, of course, deserves its own item). Re #21: when I was working with filters, etc, I knew a central wavelength for the seven "colors", in order to talk about filter choices more easily. The boundary values are, in the context of seven "colors", not even colors (since they are between two colors). Re #22: I don't understand yet...is the "256 colors" coding an 8 bit word? If so, it can only code for 256 colors, not 256 of each of the three guns, which would require either 3 8 bit words, or one 10 bit word. I guess I don't understand the lookup scheme for the 24 bit color map. In addition, a lot of the combinations from the 24 bit map are the same color - just different intensities. Is intensity coded separately?
So far as the computer is concerned, a "different intensity" is a "different color". You might like to think of colors differently, but from the computer graphics standpoint, it's simplest to consider white, light pink, red, and black as 4 different colors, even if some of the RGB values happen to fall more or less on a line. Of course, one of the nice things about computers is you can hide the actual implementation details from the user, and present a completely different user interface that give people (for instance) a color wheel to select colors from, and then maps it internally into the actual RGB values for the screen.
resp:24 I think marcus was saying that there is an 8-bit lookup into a table of 24 bit values, which are the 3 8-bit quantities you mentioned. Thus, you can only have 256 colors at one time, but which 256 colors you use can change. That's the cause, if I'm right, of the funky "Screen goes funny colors" you get in unix when you switch to our netscape window. Netscape want's a *different* 256 colors than the rest of X-windows. That's alwo why you don't get it on 32-bit displays, because they can indes all the colors at once.
Which 256 of the 256^3 24 bit colors are allowed?
It works like this, Rane. Somewhere in the computer are two blocks of memory, which I'll call the "screen array" and the "palette array". Lets assume that the video mode is 1024x768, with 256 colors. Then the screen array will be a block of 1024*768 bytes - one per pixel on the screen. The upper left corner might look like this: /----+----+----+-- | 00 | FE | 00 | +----+----+----+-- | 00 | 02 | 01 | +----+----+----+-- | FF | 03 | FF | +----+----+----+-- | | | | The pallete array is an array of 256 3-byte blocks. Each byte represents a value for one of the 3 colors red, green, and blue. The beginning of the palette array might look like this: /----------\ | 00 00 00 | (color 0 has all colors turned off - i.e., it's black) +----------+ | FF 00 00 | (color 1 has lots of red and no green or blue, so it *is* red) +----------+ | 00 FF 00 | (color 2 is green) +----------+ | 00 FF FF | (color 3 has some green and some blue in it.) +----------+ | | So when the computer decides what color to make a particular pixel on the screen, it's a two-step process. First it looks up the value in the screen array. Then it uses this value as an index into the palette array, from which it retrieves the actual color. So, for instance, the pixel in row 0, column 0 has color value 0, which, according to the palette, means it will be black. The pixel in row 1, column 1 has color value 2, which means it will be green. The advantage to this method is that it uses only about 1/3 as much memory as it would if each pixel had a full color value associated with it. The disadvantage is that you can only have 256 colors at a time.
This response has been erased.
Thanks^2!
(And you can change what's in the pallette table, so you can have a *different* 256 colors at different time, but only 256 colors at one time.)
Is there a standard (or algorithm) for the 256 colors?
I imagine there's a standard for which 256 colors are used by default, but I don't know what it is.
<wonders why we're discussing colors in the agenda item for the board meeting>
(because the agenda identifiers are colors - why are we wispering?)
(Because it's way too early to shout.)
(..good point..it was way too late to shout when I wrote #35 - and now..)
IS IT OKAY TO SHOUT NOW?
WHY ARE YOU SHOUTING?
(Because they said it was too early to shout, so I was wondering if 11 AM was late enough to shout.)
I think so.
Congratulations! This item has now established a new record for drift in a Board Agenda item.
The tether got cut after 4/22, so drift is expected, even without some help from physics.
Maybe intense solar wind?
You have several choices: