Re: Bit of a Con Really - Follow-up ...

[Leonard Caillouet]

Probably. Not as many stations, either.

Leonard

[meow2222]

What that has to do with it I dont know. If you find an RGB display with violet output, I'm all ears.

NT

[meow2222]

With each colour channel you've got everything available from backlight output x LCD max down to backlight output x LCD minimum. AFAIK that covers every flesh tone on this planet, unless one goes down to 2000K backlight or some other very extreme value.

NT

[Leonard Caillouet]

This is simply not true. Every display has a color gamut that is limited by the maximum saturation of its primaries. You can produce any color within that gamut but not any outside. Even if every flesh tone is in that gamut, that does not mean that you will get the right flesh tones for a given combination of RGB. In order to do so, you must have the same spectrum in the primaries that you have in the camera filters, the correct colorimetry for the white point, and the correct application of the decoding matrix. If you depart from any of these, you can adjust a display for ONE color to be correct, but everything else will be off.

Leonard

[Leonard Caillouet]

It has everything to do with accurate reproduction of color in video.

What you seem to miss is that the underlying assumption in color reproduction in video is that the display and the camera both approximate the CIE standard observer curves for red, green, and blue spectral response. If this is the case, and you encode properly, you can use a standard decoding matrix on the display end and get a reasonable reproduction of what was recorded. If you have a very narrow spectrum on either end, some colors will be reproduced with less energy than with the proper spectrum. This can be compensated for using a customized matrix or LUTs.

Again, while it is true that you can make any color within a given gamut with some combination of R,G, & B, it is NOT true that you will get the CORRECT color for ALL colors if the decoding matrix is not correct (very common in many consumer sets over the years, if the gamut is wrong, if the gray scale is wrong, or if the spectrum is wrong. To get the right mix of colors for all colors in a given system, you have to play by the rules for that system. If you change them, such as is the case when you deviate in spectral response from the CIE curves, you have to make it up somewhere else. This gets very complicated and is precisely why some people who are sensitive to color reproduction have noticed that LED based displays have had trouble with some colors.

Leonard

[Geoffrey S. Mendelson]

I think that the important point is that the CIE standards are a subset of colors, not the millions of colors people think they are getting from a computer display.

Geoff.

[Bob Larter]

You can't tweak the colour response of film, you can with CRTs or LCDs.

[Bob Larter]

Go read some textbooks on the subject of colour management for screen & print. The topic is far too complex to sum up from first principles in a newsgroup post.

[Bob Larter]

[Bob Larter]

And LCDs use a matrix of RGB filters over a backlight, which makes them behave similarly to a transparency in a projector, with the exception that you can tweak the RGB values to change the colour balance - which you *can't* do with film transparencies.

[Dave Plowman (News)]

You can during processing.

[Bob Larter]

Which process is an entire art & science of its own. It is *not* easy to do a good job of converting an RGB light image into CMYK pigments.

[Bob Larter]

Um, this sounds a little confused. The important distinction is between

*additive* colour (eg; RGB) vs *subtractive* colour (eg; CMYK). With RGB, you're *filtering* a full spectrum illuminant, with CMYK, you're *absorbing* part of a full spectrum illuminant.

[Bob Larter]

Ayup. When I was a kid, our first colour set was a 27" Baird, which had a huge, hinged PCB with a kazillion pots on it, & an acetate overlay covering the whole thing, with diagrams printed on it to show what each pot adjusted. *Lots* of fun!

[Bob Larter]

Of course.

Those of us who take colour reproduction seriously use colourimeters to calibrate the screen & display card LUTs.

[Geoffrey S. Mendelson]

People have been known to do it with Ektachrome type film (Ektachrome, Fijuchrome, AgfaChrome, Ilfochrome, etc) with varying success by tweaking the exposure and development.

At one time you could get Kodachrome developed to spec, which was intended to correct for mistakes in exposure, but it certainly would change color response.

In the printing process, either to film (used in movies) or paper, you could do all sorts of things.

Standard C-41 color film has an exposure latitude of less than one stop underexposure, but 4-5 stops of overexposure. As the exposure increases, color response, contrast and graniness change. If you like fine grained over saturated colors, try shooting a roll of ISO 100 color negative film and have it processed normally.

Technicolor which is actually a black and white negative process (producing seperate red, green and blue negatives on black and white film) could very easily be manipulated and often was. Look at the recent prints (or the DVD from them) of The Wizard of Oz.

The original intention was to produce a movie that was almost cartoon like in its color, later prints were much more subdued, almost "normal" in color as well as the video tapes made from them.

Geoff.

[meow2222]

I agree with what you're saying, it just wasnt the point I was addressing.

NT

[The Natural Philosopher]

You can actually.

[The Natural Philosopher]

Why do color enlargers come equipped with color filters then?

[William Sommerwerck]

Correct.

This bothers me. It might be true in a practical sense, but it's always struck me as being theoretically wrong (mostly because of the extreme overlap of the eye's blue and green receptors). I won't start an argument, though, because, even if my intuition is correct, I don't have the "science" to back it up.