I’m afraid there are many incorrect statements and conclusions in your friend’s commentary and I’d like to clarify them.
We do not see white as white because we know it is. Our eyes have a chemical "color balancing" system which corrects for color temperature up to a point. Our cones, which are responsible for our color vision come in three versions, interestingly, sensitive to red, green and blue, which are the primary colors of light. We have the most green cones, followed by red and then blue. As we age we lose more blue cones due to damage by UV exposure, and for most of us our lens yellows to a fairly deep orange by the time we are in our 60s. This actually helps filter out UV to protect the few blue cones we have, but between the loss of blue cones and the orange lenses, our blue perception is quite poor as we age. This in part explains why old people wear purple clothing, think it is brown, and older ladies tend to wear blue hair dye, thinking it is just toning down the intensity of the white hair.
Anyway, each cone produces chemical set points, and the receptors saturate to the predominant color and become less sensitive to it. If you want to see this in action, to prove it isn’t just what we "expect" try these there experiments.
1) Go into a room which is lighted with a strong color lamp, such as a red lamp. Wait until your cones adjust so that the red begins to "fade" and the colors in the room begin to become more "normalized" (they never will fully because simply put a red light source doesn’t contain the full spectrum, or anything close to it, and since color of objects is perceived by reflection of certain spectral wavelengths, if they aren’t there, the color just isn’t visible (an example might be a dark green piece of paper)). Red light emits almost no green, and therefore the paper will look nearly black. Anyway, once your vision has acclimated to the red bias go into something approaching white light and notice how green it is. Basically, in an attempt to compensate for the overly red bias, the green cones produced excess chemistry to try to readjust for that bias.
2) Test two. Go out on a nice sunset night and look at the horizon or sunset sky (do not look directly into the sun, even at sunset!). Once your eyes have accommodated to that color balance, which is quite red, turn around so your back is to the sunset. Now, when no one is looking at you (if you are easily embarrassed) bend over and look at the sunset sky, this time through your legs. Suddenly the colors will be much brighter and red will be stronger. No, it is not because the chemicals spilled, but close… You have now reversed the area of the retina exposed to the sunset sky colors. The area which was probably dark and silhouetted ground has not become the area where the sunset sky is being perceived, and that area hasn’t been saturated with reds.
3) Find a graphic image in a book, one with many primary and secondary color areas boldly showing. Things like stripes of circles of color are good. Stare at it for 30-45 sections and try not to move where your eyes are looking during that period. Now, rapidly take a look at some well lighted white paper with nothing else on it, and you should notice something referred to as an "afterimage" and that image will have "negative color" relative to the colors you were staring at. Dark portions will be brighter, and vice versa, Greens will look red and vice versa, etc.
Although optical dyes can create color casts on the white wedding dress, since they are UV sensitive and our eyes generally are not, but film is, that is why it is so much more obvious when film is involved.
As for color being subjective, each of our eyes sees and perceives color differently. I have one eye which is considerably warming than the other in how it sees light and color.
Now, as to dithering and the ALPS printers. I am assuming he is using the dye sub dye material and not the hot wax semi opaque "inks".
Dye sub technology cannot be compared spec-wise to inkjet or laserjet, because the technologies mean very different things.
Dye sub prints are referred to as continuous tone. This means that for all intents and purposes the number of variations in color depth and density are such that the human eye won’t normally see bands or breaks in the graduated color or density. The way a dye sub product works is the head heats a small area of dye which is literally vaporized and made gaseous. It then reforms as a solid within a specialized usually plastic coated receiving surface on the print. Most dye sub printers use 64, 128 or 256 levels of heat to determine how much of the dye gets vaporized off the ribbon and redeposited onto the print. Most dye subs used three ribbons to create all their colors, Cyan, magenta and Yellow (plus a protective clear UV coating) some also provide a black ribbon or panel. In the case of a dye sub head that provides 256 levels of heat per color, that’s 256 density levels for each of C, M and Y, or a total of 16.8 million colors. Since the dyes are transparent they are laid right on top of each other. Each dot on a dye sub when looked at in detail usually has a soft fading peripheral edge to help them to blend with each other well. Since each "dot" corresponds directly to one pixel of data, these printers only need 200 to about 400 dpi.
Inkjet printers and to some extend color laser printers use a very different approach. Unlike the dye sub with 256 levels of each "ink" color. inkjet printers basically only can produce one color density of each ink color. Now, this has been somewhat altered by inkjet printers which use light and dark cyan, for instance, and they may also have or more dot sizes they use, but the principal is still similar. Unlike dye sub where each dot is approximately equivalent to one pixel, inkjets need to make up a matrix of dots of each ink color which together dither to create the illusion of a color. The smaller these dots and the more of them, the less obvious the lack of color variation becomes. So the reason inkjet printers, for instance, require 1000 up to 5000 dpi is because each color is represented by 9 or 12 or more dots plus white paper showing through the spaces between them.
So, in conclusion, dye sub printers do not need to have as high a dpi as inkjet and laser for equivalent color smoothness (and potentially accuracy).
The disadvantages of dye subs are:
They may be fast fading
They cost more to run (consumable)
They tend to be noisier and slower
If you are using the cartridge type versus the ribbon type a lot of the dye is wasted because all three of 4 color panels are used per image. You must used their receiver paper.
The printers can be expensive
Today’s inkjet printers with their forever decreasing nozzle sizes and variable dot sizes plus using more than one ink color density, helps to provide a near contone image.
CCDs used in scanners and even in digital cameras are actually colorblind and only detect black and white and those shades in between. They use color filters to color separate out the colors using some complex algorythms. However, CCDs are corrected in scanners for a black and white point each time they are used.
Lastly, while color perception is subjective, colors are not. They are measuarable by photo spectrometers and densitometers, and are reproducible, and indeed that’s how professional printers keep the things they print in spec to client’s expectations.
gecko wrote:
On Fri, 07 Mar 2008 20:03:36 -0500, tacit wrote:
Yes, you do need dithering for 256 colors. 256 colors is an extraordinarily limited amount; that’s very poor color fidelity indeed. Remember that each shade of the same color counts as a different color.
My friend responds:
The thing about color — particularly about fine differences between hues of color — is that it is completely subjective. Film — and I suppose CCDs as well — is not subjective. Color depends entirely on the color temperature of the light. Try taking a color photo with fluorescent light to see the difference. The curious thing is that we have color memory; if we know what color a thing should be, we will see it that way whatever the color temperature of the light. Whites look white to us because we know they are white. But white is made to appear more white by tinting it most generally with blue, but also with yellow and even black. When all the other colors were perfectly balanced, Keeping bride’s dresses from appearing blue in photos was a challenge for me because of the optical whitening added to the fabric.
So, for purposes where actual color is not known or where critical rendition of actual color is relatively unimportant, 256 colors will suffice. For 4×6 snapshots, 256 is good enough. The toughest challenge for film and CCDs is good skin tones. For a large print of a head shot, 256 colors are inadequate. Dithering is a poor substitute for the infinite modeling provided by film. But I shall need to learn to apply the finest dithering the Alps is capable of to be happy with its 600 dpi resolution. I thought that my MD-4000 was capable of 1200 dpi, but that may be for scanning only.
-Gecko
