Dan Margulis Applied Color Theory

Primary and Secondary Colors

   Date: Tue, 22 Mar 2005 18:56:46 -0000
   From: "seismo1899"
Subject: Very Basic Color Theory Question...

We all know that in "light" the primary colors are red, green and blue. Work in Photoshop demonstrates this completely.

In the subtractive world of "pigments" as used by painters, the primaries are often called out as red, yellow and blue.  However in the printing world, cyan, yellow and magenta seem to be the primaries.  Coming from the "light" side of the world, the printing primaries seem correct in my mind.  Why are pigment primaries commonly called out as red, yellow and blue instead?  Is it just a matter of terminology (i.e. when an painter talks about red, they really mean something closer to magenta, and likewise a painter's blue is really closer to cyan)?

Many thanks.
Gordon Ross
___________________________________________________________________________

   Date: Tue, 22 Mar 2005 11:28:26 -0800 (PST)
   From: Mike Bevans
Subject: Re: Very Basic Color Theory Question...

Gordon asks: "We all know that in "light" the primary colors are red, green and blue.  Work in Photoshop demonstrates this completely.

Why are pigment primaries commonly called out as red, yellow and blue instead?  Is it just a matter of terminology (i.e. when an painter talks about red, they really mean something closer to magenta, and likewise a painter's blue is really closer to cyan)?"

A primary color is any color that can not be made through the combination of any two other colors within a system.

The "within a system" part is important.

For example, when combining lights, as in the case of RGB (because in order to have color you must add illumination) green is a primary color. ding red and green lights produces a yellow light.

I the additive world of mixing pigments, RYB, green is a secondary color. It can be made by combining blue and yellow pigments. Yellow is a primary color in the RYB model.

So, the color model you use really depends on the way you are making colors, whether lights, pigments, or something else (what, I can not think of).

Now I introduce a new topic. Interestingly, you could argue that CMY is not a subtractive color model at all. In fact, you could argue that CMY is another name for RYB. In this case C=B, M=R, Y=Y. It is clear to see that CMYK is not a subtractive model, becuase you are adding pigments together to form new pigments, such as C+Y=G (the arguement of subtracting from the white of the page is false because you are subtracting white from a page when you add paint to a page, too). If CMY were truely a subtractive model, then the combination of C,M, and Y would produce darkness, or black. But, because CMY is really an additive process, it forms a brownish color (just like R,Y,B paints do). This is why we cheat and add black pigment.

Of course, CMY was once a subtractive process in the days of color seperations. A photographer would provide a prepress house with a transparency. The first step of making this into a plate was to perform a color seperation, or a series of four negatives, cyan, magenta, yellow, and black sheets of film. This process of making "internegatives"(forgive me for not knowing the term) is indeed a subtractive process; however, in the world of digital imaging and direct to plate printing, there is no subtractive process.

-Mike Bevans
________________________________________________________________________

   Date: Tue, 22 Mar 2005 11:24:03 -0800
   From: Paul D. DeRocco
Subject: RE: Very Basic Color Theory Question...

They call them blue and red because first graders don't know what cyan and magenta are. ;-)

I think that when oil painters mix paints, they don't necessarily start with true CMY primary colors, they start with similar colors that they have, and play with the mix until they get what they want. In other words, they're practicing art, not science.

Ciao,               Paul D. DeRocco
________________________________________________________________________
 
   Date: Tue, 22 Mar 2005 20:00:51 +0000
   From:Richard Rosenberg
Subject: Re: Very Basic Color Theory Question...

CMY is a subtractive process whether digital or analog.  It is subtractive because each color subtracts 1/3 of the visible spectrum. The paper is just the reflector of light. The ink acts like filters.  E.G. Cyan subtracts Red leaving Green and Blue coming back to your eye. Green and Blue give the impression of the color Cyan. The reason black is used is because ink are impure.  They do not reflect or absorb all the colors 100%.  This is called Hue Error. Black is used to help counter the imperfection of process ink and add contrast to images.

Richard Rosenberg
________________________________________________________________________
 
   Date: Tue, 22 Mar 2005 15:20:35 -0600
   From: David Dyer-Bennet
Subject: Re: Very Basic Color Theory Question...

Mike Bevans writes:

Now I introduce a new topic. Interestingly, you could argue that CMY
is not a subtractive color model at all. In fact, you could argue
that CMY is another name for RYB. In this case C=B, M= R, Y=Y. It is
clear to see that CMYK is not a subtractive model, becuase you are
adding pigments together to form new pigments, such as C+Y=G (the
argument of subtracting from the white of the page is false because
you are subtracting white from a page when you add paint to a page,
too). If CMY were truely a subtractive model, then the combination
of C,M, and Y would produce darkness, or black. But, because CMY is
really an additive process, it forms a brownish color (just like
R,Y,B paints do). This is why we cheat and add black pigment.

Nope, the dirty brownish color is just because the pigments are impure/imperfect.  To demonstrate this, go admire the rich, deep, blacks of a dye-transfer photo print, made with only C, M, and Y dyes (no black even).
--
David Dyer-Bennet
________________________________________________________________________

   Date: Tue, 22 Mar 2005 15:17:52 -0600
   From: David Dyer-Bennet
Subject: Re: Very Basic Color Theory Question...

Painters formed their habits before carefully-formulated pure pigments were available.  Red, yellow, and blue work pretty well as an informal starting point.  Painters don't limit themselves to using just those!

Printers, who were trying to reproduce all the colors everybody else was using, and who mostly couldn't use more than 3 colors plus black, needed a more precise basis for their color space, so they refined things a bit.  (Yes, I know that high-end offset printers sometimes use 6 colors, and that some inkjet printers now do too, though not the same six).
--
David Dyer-Bennet
________________________________________________________________________

   Date: Tue, 22 Mar 2005 12:57:36 -0700
   From: Andrew Rodney
Subject: Re: Very Basic Color Theory Question...

On 3/22/05 11:56 AM, "seismo1899"  wrote:

In the subtractive world of "pigments" as used by painters, the
primaries are often called out as red, yellow and blue.

These are additive primaries. This describes the use of light to create color.

However in  the printing world, cyan, yellow and magenta seem to be the
primaries.  Coming from the "light" side of the world, the printing
primaries seem correct in my mind.

For the creation of color using light, Red, Green and Blue are correct. These are the primaries used.

 Why are pigment primaries
commonly called out as red, yellow and blue instead?

They shouldn1t be. We are not creating color with light, we1re producing color by subtracting. Here there is a subtraction of red, green and blue from the white light striking the paper. When white light strikes cyan pigment, green and blue is reflected and red is absorbed. When white light strikes magenta pigment, red and blue is reflected and green is absorbed. When white light strikes yellow pigment, red and green is reflected and blue is absorbed.

 Is it just a
matter of terminology (i.e. when an painter talks about red, they
really mean something closer to magenta, and likewise a painter's
blue is really closer to cyan)?

Not all colors systems necessarily use red, green and blue as primary colors. Green is not a primary color when mixing paint but instead yellow, red and blue. Further, look at synthetic color spaces like CIEXYZ. Like the RGB color model with three additive primaries, CIE XYZ uses three spectrally defined imaginary primaries; X, Y and Z.

Andrew Rodney
http://digitaldog.net/
________________________________________________________________________

   Date: Tue, 22 Mar 2005 13:55:55 -0800
   From: David Cardinal
Subject: RE: Very Basic Color Theory Question...

This may be too well understood by everyone to mention, but just in case: Remember that R,G,B work because the human (and many animal) vision systems happen to use similar response curves roughly corresponding to those colors. There isn't anything magic about them, per se, although there has been some interesting research that the choice of the HVS primaries helps optimize the apparent contrast of the natural world and possibly explains why we evolved as we have.

--David Cardinal
________________________________________________________________________

   Date: Tue, 22 Mar 2005 17:02:08 -0800 (PST)
   From: Mike Bevans
Subject: Re: Very Basic Color Theory Question...

I should have prefaced the entire string with IMHO, and an academic exercise. Keep in mind, that color, like evolution, is just a theory. (I throw that in to honor Darwin's 196 birthday back in February)

I geuss it comes down to how you define a color model. To me, a color model describes the interaction of hues in a given state, like lights or pigments. I think of a color model like a set of instructions how to combine colors.

So, if I take blue paint and mix it with yellow paint, we know I will get green. I add one color to another to form a third.

The way I intuit making green with CMY pigments is to add cyan to yellow. The combining of two colors forms a third. So, to me, I consider this an additive color mixing model.

"The ink acts like filters," writes Richard. If CMY were a subtractive color model, I would subtract the color I wanted removed from whiteness(the presence of all colors, or simply: light). Following this reasoning, I would subtract magenta from white to make green.

"Cyan subtracts Red leaving Green and Blue coming back to your eye," This describes the absorbtion and reflection of light, but not a system of combining colors, so by my limited definition above, I would not consider this a color model useful for combining colors.

"Green and Blue give the impression of the color Cyan." An "impression" is a description of cognition. Cognition is often classified as an opponent color system, RYGB, that's a whole 'nother ball of wax. Whether or not an observer need be present to be impressed upon is a discussion I would not care to venture into.

"The reason black is used is because ink are impure," Repent! This raises issues of formalism. Is there pure ink? If so, why don't we use it? If not, then how do we measure our ink as impure?

-Respectfully,

Mike Bevans
________________________________________________________________________

   Date: Tue, 22 Mar 2005 19:03:12 -0700
   From: Andrew Rodney
Subject: Re: Very Basic Color Theory Question...
 
The phenomenon we call 3Color2 only exists inside our brain. This is a sensation created by various frequencies of light falling upon the retina of our eye. A 3red3 apple doesn't emit red light. The apple absorbs all the shorter wavelengths of light shining on it and reflects the longer wavelengths. A receptor in the retina or our eye is sensitive to longer wavelengths and becomes stimulated and sends a signal to the visual cortex of our brain. The shorter wavelength receptors do not send a signal. The visual cortex processes these patterns and associates them with a sensation of color. Another part of the brain associates that sensation with the word 3red2 which is a word used to describe this sensation using the English language.

Or we could all just be fuel batteries for machines and our perception of color is simply the stimulus in our brains produced by the Matrix...

Andrew Rodney
http://digitaldog.net/
________________________________________________________________________

   Date: Wed, 23 Mar 2005 07:34:52 -0600
   From: "Maris V. Lidaka Sr."
Subject: Re: Very Basic Color Theory Question...

Bruce Fraser has written a very good article called "Why is Color"

http://www.creativepro.com/story/feature/13036.html

Maris
________________________________________________________________________

   Date: Wed, 23 Mar 2005 10:01:39 -0800 (PST)
   From: Mike Bevans
Subject: Re: Very Basic Color Theory Question...

Maris sent in a very good article about color theory; however, it does not answer the question of why RGB, CMY, and RYB are all considered primary colors. A primary color is any color that can not be made by combining any two other colors within a given system, or should I say, within a given color model.

A convenient way to think of this is to draw the color wheel for each of the three systems, and to start to fill in the secondary colors. You will see that in CMY blue is the compliment to yellow, while in RYB blue is the compliment to orange, and of course, in RGB blue is also the compliment to yellow but yellow is a secondary color in RGB, not a primary.

It can be confusing. To add to it, I make the statement, one can consider CMY, at least digital CMY, an additive color model. I do not do this to offend, simply as an academic exercise while in between prints.

It is interesting to note that CMYK color seperations, when made on film (back in the day) were negatives, not positives. The color layers were meant to subtract illumination. In this state, yes, CMY was a subtractive process.

Filtering yellow from light(interestingly by adding yellow filtration to the light), as in the case of priting from a negative, is a subtractive process that will produce a blue positive(the values being inverted in printing).

Since we don't invert our images to work in CMYK as negatives, the combinations are additive in nature.

In the positive to positive world of digital printing, I find it easier to conceptualize the interactions of hues(the blending of colors) an additive process. C+M=B, for example an additive process.

This is why Epson printers that use CMY inks are often refered to as really being RGB printers. Although, I think it more the case to say the Epson printers that use CMY inks are additive printers(most people sub in the term RGB to mean additive, when really they mean RYB). I consider the Epson printer to be an RYB printer, just a bluish red, yellow and a pale blue.

-Mike Bevans
________________________________________________________________________

   Date: Wed, 23 Mar 2005 13:25:36 -0600
   From: David Dyer-Bennet
Subject: Re: Very Basic Color Theory Question...

Mike Bevans writes:

Maris sent in a very good article about color theory; however, it
does not answer the question of why RGB, CMY, and RYB are all
considered primary colors. A primary color is any color that can not
be made by combining any two other colors within a given system, or
should I say, within a given color model.

That's not a workable definition, at least within the usual usages of people doing color digital printing.  You've just defined any out-of-gamut color as "primary"!

A convenient way to think of this is to draw the color wheel for
each of the three systems, and to start to fill in the secondary
colors. You will see that in CMY blue is the compliment to yellow,
while in RYB blue is the compliment to orange, and of course, in RGB
blue is also the compliment to yellow but yellow is a secondary
color in RGB, not a primary.

It can be confusing. To add to it, I make the statement, one can
consider CMY, at least digital CMY, an additive color model. I do
not do this to offend, simply as an academic exercise while in
between prints.

I don't think parading academic exercises, especially those that severely challenge the conventional models and usages, in front of people trying to learn to make practical applications of a theory is a productive, or even a friendly, thing to do.  People get confused enough trying to understand color in the conventional models!

This is why Epson printers that use CMY inks are often refered to as
really being RGB printers. Although, I think it more the case to say
the Epson printers that use CMY inks are additive printers(most
people sub in the term RGB to mean additive, when really they mean
RYB). I consider the Epson printer to be an RYB printer, just a
bluish red, yellow and a pale blue.

No, they're referred to as RGB printers because you need to send RGB data to the Windows or Mac printer drivers.  
--
David Dyer-Bennet
________________________________________________________________________

   Date: Wed, 23 Mar 2005 18:09:17 EST
   From: Dan Margulis
Subject: Re: Very Basic Color Theory Question...

Gordon Ross writes,

Why are pigment primaries commonly called out as red, yellow and blue
instead?  Is it just a matter of terminology (i.e. when an painter talks about red,
they really mean something closer to magenta, and likewise a painter's blue
is really closer to cyan)?

I've never heard painters talk that way. When they say red they mean red. *Pressmen*, on the other hand, use "red" to mean magenta and "blue" to mean cyan. It's traditional, and the likely reason is that in the pressroom the pressmen are very close to deaf, because they are wearing ear protection to ward off the extremely noisy environment. Under those circumstances, you have to *scream* the instructions, and "MORE RED" is a lot easier to yell (and to lip-read) than "MORE MAGENTA".

From there on things have deterioriated. Mike Bevans writes,

In the positive to positive world of digital printing, I find it easier to
conceptualize the interactions of hues(the blending of colors) an additive
process. C+M=B, for example an additive process.

These terms "additive" and "subtractive" are so confusing that I refuse to use them either in books or in classes--particularly since a good two-thirds of what "experts", including many academics, have written indicate that they don't understand the concepts.

All systems are aimed at controlling the amount of red, green, and blue light that hits the eye. The system can produce its own light (e.g. a monitor), or it can use pigments to block the reflection of light (e.g. inks). In one case increasing the intensity of the system's tools makes for lighter colors (adds lightness) in the other for darker colors (subtracts lightness). So, C+M=B is "subtractive", but it's a totally semantic distinction that is of highly limited utility.

To make a blue, a monitor adds blue light. In print, we use cyan ink to block the reflection of (=subtract) red light, and magenta ink to block the reflection of green light, allowing the substrate to reflect--in theory--only the blue.

I would like to comment on three other misconceptions.

1) "RGB is additive and CMY is subtractive." This is true only to the extent that RGB is theoretically the optimal way to design a light-transmitting system involving only three channels and CMY is theoretically best for a pigment-based system, although in practice neither model is perfect. Nothing stops us from making an additive system with, say, RBY, or a subtractive system with CGM, other than the fact that these aren't particularly efficient ways of doing things--in the first case, we couldn't produce a dark green and in the second we couldn't get a bright red.

2) "A desktop printer is an RGB device." A desktop printer outputs with CMY (plus possibly some other) colorants just like any other print-making device does. The fact that it wants to have RGB input (which it then converts on the fly into CMY+) is neither desirable nor relevant.

3) "Black ink is needed because the CMY inks are impure". Black ink was needed for 400 years before anybody ever thought of printing with CMY. We use black ink because jobs contain type, not because there's anything special about black as a complement to CMY. If we were choosing an ink to compensate for the impurities of CMY inks, it would be blue, not black.

Dan Margulis
________________________________________________________________________

   Date: Wed, 23 Mar 2005 13:20:51 -0700
   From: jim donovan
Subject: Re: Very Basic Color Theory Question...

Very Very, incorrect as far as The epson 10600 goes. The epson 10600 can be set up any way you like. We send only cmyk data to our 10600. It bounces anything rgb,wont  process data that is rgb. It's a matter of setting up  correctly. The amazing thing about these new epsons is they don't drift at all like old matchprint color key systems. After well over 10,000 prints made,our test patch matches number for number, % for % from print one to print 10,000. I didn't believe it myself until it was proven in our shop with our files. Jim Donovan
________________________________________________________________________

   Date: Wed, 23 Mar 2005 14:26:43 -0600
   From: David Dyer-Bennet
Subject: Re: Very Basic Color Theory Question...

Andrew Rodney writes on 23 March 2005 at 13:07:12 -0700
 
 The primary colors of a color space (which define a gamut) are at
 the edge of color gamut. R255 is the redest red in both sRGB and
 Adobe RGB (1998) but that primary has a different scale (gamut)
 each space. So the primary is in gamut and at the edge of the
 gamut.

You snipped the bogus definition of "primary" that I was responding too, though.  

They should not be referred this way. The inks (colorants) are not
 RGB (there are only a tiny number of output devices that are,
 usually those using lasers to light valves). That same 3RGB2 Epson
 can be hooked up to a RIP that will happily send it CMYK data. An
 Epson isn1t an RGB output device.

I'd agree that they shouldn't be referred to as RGB devices.  However, calling them CMYK devices causes confusion, when people try to send CMYK data from Photoshop to their Windows and Mac printer drivers.
--
David Dyer-Bennet
________________________________________________________________________

   Date: Wed, 23 Mar 2005 22:57:28 +0100
   From: "Francisco Bernal Rosso"
Subject: Re: Very Basic Color Theory Question...

Primary color are only the colors you have and you use to mix and get other colors. The rule is that the mix of all "primary" colors must render a neutral gray. Primaries is only a selection.

See at the CIE color tongue. Select two points of that grahic. You have selected two colors with the mix of those two colors you can get all the color over the line beetween the two points. the far the points are from the "gray" the more saturated colors you get. Now take 3 points, then you can mix that 3 colors and get all the colors of the triangle. Those are the "primary" colors of your color polygon. If you take the care to select the three colors leaving the gray 8the neutral tones) into the polygon, then you can almost render "all" the other colors. Actually you can render all hues. The saturation up to the distance beetween the primary color and the neutral tone. The far away it is, the more saturated you can render the colors. You can use more than 3 point, but 3 is the minimun, so it is logic the mother nature only use 3 elements for color selection.

Francisco Bernal Rosso

Luz-color-fotografia
Redacción y traducción
________________________________________________________________________

   Date: Wed, 23 Mar 2005 13:07:12 -0700
   From: Andrew Rodney
Subject: Re: Very Basic Color Theory Question...

On 3/23/05 12:25 PM, "David Dyer-Bennet"  wrote:

That's not a workable definition, at least within the usual usages of
people doing color digital printing.  You've just defined any
out-of-gamut color as "primary"!

The primary colors of a color space (which define a gamut) are at the edge of color gamut. R255 is the redest red in both sRGB and Adobe RGB (1998) but that primary has a different scale (gamut) each space. So the primary is in gamut and at the edge of the gamut.

No, they're referred to as RGB printers because you need to send RGB
data to the Windows or Mac printer drivers.

They should not be referred this way. The inks (colorants) are not RGB (there are only a tiny number of output devices that are, usually those using lasers to light valves). That same 3RGB2 Epson can be hooked up to a RIP that will happily send it CMYK data. An Epson isn1t an RGB output device.

Andrew Rodney
http://digitaldog.net/
________________________________________________________________________

   Date: Wed, 23 Mar 2005 18:43:56 -0500
   From: Rafe Bustin
Subject: Re: Very Basic Color Theory Question...

At 02:26 PM 3/23/2005 -0600, DD-B wrote:

I'd agree that they shouldn't be referred to as RGB devices.  However,
calling them CMYK devices causes confusion, when people try to send
CMYK data from Photoshop to their Windows and Mac printer drivers.

Exactly.  I made that mistake early on, and it took a while to realize my mistake.

If you send a CMYK file to an Epson, it does a lousy conversion back to RGB, and then a second separation out to whatever inks that model happens to use.

AIUI, next-generation Windoze OS's will support driving printers with CMYK -- through the OS.

rafe b.
http://www.terrapinphoto.com
________________________________________________________________________

   Date: Wed, 23 Mar 2005 18:40:21 -0500
   From: Rafe Bustin
Subject: Re: Very Basic Color Theory Question...

At 01:07 PM 3/23/2005 -0700, Andrew Rodney wrote:

They should not be referred this way. The inks (colorants) are not RGB
(there are only a tiny number of output devices that are, usually those
using lasers to light valves). That same 3RGB2 Epson can be hooked up to a
RIP that will happily send it CMYK data. An Epson isn1t an RGB output
device.

What makes it an "RGB output device" is the Epson driver.

I think we all understand that most of Epson printers use CMYcmK inks... although R and G inks are now used in the very newest models (R800, R1800.)
 
rafe b.
http://www.terrapinphoto.com
________________________________________________________________________

   Date: Wed, 23 Mar 2005 16:24:29 -0800 (PST)
   From: Mike Bevans
Subject: Re: Very Basic Color Theory Question...

Perhaps we should recall Gordon's original question, "We all know that in "light" the primary colors are red, green and blue...Why are pigment primaries commonly called out as red, yellow and blue instead?"

Seacrhing on the internet, you can find a lot of definitions for primary colors. Most commonly, primary colors are refered to as the minimum number of stimuli needed to create all colors, and a primary color is a color that can not be made from the combination of any two other colors.

Of course, we all know that colors occur in different states. Colors can be projected illuminants, such as the RGB triplets of a computer monitor, or reflected pigments, such as paint on canvas.

We also know that using RGB triplets, we can make white light, and therefore we can mathmatically derive all colors. We also know that you can get yellow by combining red light and green light; therefore yellow is not a primary in the RGB color model. We know that you can not make blue by combining any two other colors in the RGB color model, so blue is a primary color.

We also know that when mixing paints, we can make green paint by mixing blue paint and yellow paint. Therefore, in this system of reflected hues, we know that green is not a primary color, since it can be made from two other colors. We also know that you can not make yellow by combining two different colors, such as green and red... not with paint you can't.

So you can see that primary colors depend on what state color is in. How the colors combine and interact is what I have defined as a color model. A color model allows us to predict the interaction of colors in a given state.

I will leave aside discussion of CMY for the moment.

You can agree or disagree, but we ought to keep the tone civil. Although the question may be academic, judging by the number of responses and the tenor of the conversation, we can see that color theory, the name of this group, is an evolving field. I think it would be hard to point to an absolute truth on this subject, which is why the word theory exists. An excellent webpage to learn about color models is www.colorsystem.com

-Mike Bevans
________________________________________________________________________

   Date: Wed, 23 Mar 2005 18:00:07 -0700
   From: Andrew Rodney
Subject: Re: Very Basic Color Theory Question...

On 3/23/05 4:43 PM, "Rafe Bustin"  wrote:

If you send a CMYK file to an Epson, it does
a lousy conversion back to RGB, and then a
second separation out to whatever inks that
model happens to use.

ONLY using the Quickdraw or GDI driver. This has nothing to do with the inherent hardware we are referring to. Place a RIP in the print stream and the Epson will gladly accept and print CMYK.

Andrew Rodney
http://digitaldog.net/
________________________________________________________________________

   Date: Wed, 23 Mar 2005 17:41:37 -0800 (PST)
   From: Mike Bevans
Subject: Re: Very Basic Color Theory Question...

So, I think by now we all agree that Epson printers are sometimes refered to as RGB printers because they combine colors additively, not because they only accept RGB data. (I stated earlier that refering to an Epson printer as an RGB printer, as some are wont, is a minomer)

By additively, I mean C+M=B.

This may bring us back to my theory that CMY is an additive method of combining reflected pigments.

CMYK film separations were negatives. Negatives block, or subtract, illumination. So, using a negatives implies subtractive color mixing. Magenta filtration is added to filter out certain wavelengths of light, so that when the image is inverted(made positive), you get green. This truely was a subtractive process.

The oft repeated argument that CMY is a subtractive color model because it subtracts whiteness from the page is misleading. Red, yellow and blue paint, commonly agreed as an additive color model also reduces the amount of whiteness from the page. (The page is not the source of illumination. The source of illumination is external) The colors combine in an additive manner, just like the CMY pigments.

RYB: B+Y=G

CMY: C+Y=G

This is a theory, and obviously one that is not widely held. But it is easy enough to observe.

A good way to practice subtractive color theory is to make prints from negatives.

-Mike Bevans
________________________________________________________________________

   Date: Thu, 24 Mar 2005 13:39:33 -0000
   From: "lwb_guitar"
Subject: Re: Very Basic Color Theory Question...

Mike Bevans wrote:

So, I think by now we all agree that Epson printers are sometimes
refered to as RGB printers because they combine colors additively, not
because they only accept RGB data. (I stated earlier that refering to
an Epson printer as an RGB printer, as some are wont, is a minomer)

Er, no, I don't agree (and I'm pretty sure few others will). Referring to these devices as 'RGB printers' might have some very loose merit in that they 'expect' to be sent RGB data, and give better results if so, because their drivers are set up to perform custom conversions from RGB to CMYK on-the-fly. Under no circumstances do these devices 'combine colours additively' in the technical sense of the term 'additive'.

By additively, I mean C+M=B.

And there's the problem. As I understand it (and I doff my hat to Dan's refusal to use the terms 'additive' and 'subtractive'), the technical term 'additive' when applied to colour models demands that adding two or more primaries in any quantity increases luminosity. It doesn't just mean that 'adding two colours together makes another colour'.

see: http://en.wikipedia.org/wiki/Additive_color
 
This may bring us back to my theory that CMY is an additive method
of combining reflected pigments.

See above. Until we have pigments that can emit their own light, any colour model describing pigments will be subtractive.

CMYK film separations were negatives. Negatives block, or subtract,
illumination. So, using a negatives implies subtractive color mixing.
Magenta filtration is added to filter out certain wavelengths of
light, so that when the image is inverted(made positive), you get
green. This truely was a subtractive process.

It's *a part of* a subtractive process, in the same way that making a plate is a part of the offset print process. It's no more 'truly' subtractive than platemaking is.

The oft repeated argument that CMY is a subtractive color model
because it subtracts whiteness from the page is misleading.

It's not an argument, it's just a conventional way of describing something. You are choosing to use an unconventional way, which is your right, but you will have hard time getting anybody to understand you.

Red, yellow and blue paint, commonly agreed as an additive color

No, it most certainly is not commonly agreed. The opposite is true.

model also reduces the amount of whiteness from the page.

And is therefore based on a subtractive model.

(The page is not the source of illumination. The source of
illumination is external) The colors combine in an additive manner,
just like the CMY pigments.

Well, if by 'additive' you mean 'adding colours together results in another colour'.

There is something interesting (to me at least) being discussed here. I come from a fine art background, having painted a lot when I was younger, studied at Art School and ended up with a degree in Fine Art Photography. I'd heard vaguely of the CMY colour model when I started to be taught how to use a colour enlarger, and wondered why everybody (not just painters) I had spoken to previously had told me by rote that 'the primary colours are red, yellow and blue'. They are primaries only by convention, with no real applied basis (i.e. our physicality defines the additive primaries as red, green and blue and therefore the subtractive primaries as their complementaries). The interesting thing -- to me -- is: where does this convention come from? I'd suggest that red simply has so much resonance in human terms (blood etc.) that it is just there by default, and once it's there magenta is simply no use; yellow is yellow, and cyan, egad, is really just a fancy name for 'blue'. Thus the 'RYB model', effectively an early, intuitive guess at the more efficient CMY model.

I guess it persists in painting because in general painters are not trying to reproduce photographic images. I'll bet you, though, that Seurat, e.g., or Chuck Close are (were) perfectly well apprised that the colour model they are using is effectively CMY.

Robert Johnston
________________________________________________________________________

   Date: Thu, 24 Mar 2005 09:09:11 -0500
   From: Raphael Bustin
Subject: Re: Re: Very Basic Color Theory Question...

FWIW...

Red, green, blue describe or characterize the long, medium, and short wavelengths of visible light.  They appear to play fundamental roles in the physiology of human color perception.

Yellow/blue and green/red also have human physiological counterparts in the opposition model of color perception (modeled by Lab color space.)

Ie., (to humans, at least) there's no such thing as yellowish blue or greenish red.

As for using any of these as pigment or ink dye primaries, I think there's as much technology at play as there are "first principles" or theoretical physics.

/rafe
________________________________________________________________________

   Date: Thu, 24 Mar 2005 07:50:26 -0700
   From: jim donovan
Subject: Re: Very Basic Color Theory Question...

Again Very Very incorrect as far as the 10600 goes. The epson 10600 can be set up any way you like. Our 10600 WILL NOT ACCEPT RGB data,unless we tell it to. No conversion what so ever takes place when we send cmyk data and that's all we get. Again it's a matter of setting it up correctly the set up possibilities are endless with the 10600. Jim Donovan
________________________________________________________________________

   Date: Thu, 24 Mar 2005 08:24:13 -0800 (PST)
   From: Mike Bevans
Subject: Re: Re: Very Basic Color Theory Question...

--- "lwb_guitar" wrote:

the technical term 'additive' when applied to colour models demands that adding two or more primaries in any quantity increases luminosity. It doesn't just mean that 'adding two colours together makes another colour'.

That is an excellent definition. In light of this definition, you are most certainly correct that both RYB and CMY are subtractive in nature.

I think what it comes down to is how I had defined color model, which is a means of predicting color interactions in a given state. By this I meant a set of recipes, for example, for combining colors.

The definition of color model that you have cited is less a model, or recipe, for combining colors, and more a model of how light and color interact. Two similar but different definitions.

What this exercise has proven to me is that in order to discuss color you have to be very precise and state your definitions at the outset.

I have found this discussion to be most illuminating, and therefore additive.

-Mike Bevans
________________________________________________________________________

   Date: Thu, 24 Mar 2005 12:26:36 -0600
   From: David Dyer-Bennet
Subject: Re: Very Basic Color Theory Question...

Mike Bevans writes:

So, I think by now we all agree that Epson printers are sometimes
refered to as RGB printers because they combine colors additively,
not because they only accept RGB data. (I stated earlier that
refering to an Epson printer as an RGB printer, as some are wont, is
a minomer)

No, absolutely not. I do not agree with that.  (The part you're claiming we all agree to; I *do* agree that calling my Epson R800 an RGB printer is incorrect, it's a CMYKRB printer being driven, in my case since I'm not using a RIP, via drivers that accept RGB data.)
--
David Dyer-Bennet
________________________________________________________________________

   Date: Thu, 24 Mar 2005 10:51:01 -0800
   From: "Mike Russell"
Subject: Re: Re: Very Basic Color Theory Question...

lwb_guitar wrote:

[re color primaries as red, yellow, and blue]  The
interesting thing -- to me -- is: where does this convention come
from?

These colors were probably a convention for centuries , but they were codified in Michael Eugene Chevreul's book on color - The Law of Simultaneous Contrast among Colors - was published in several languages in the 1850's.  He used red, yellow, and blue as primaries in a color system that described many physical and perceptual characteristics of color.

Chevreul describes almost every imaginable color phenomenon including, for example, the mixing of colored threads, which mimics CRT behavior.

Mike Russell
www.curvemeister.com
________________________________________________________________________

   Date: Thu, 24 Mar 2005 14:52:50 -0700
   From: jim donovan
Subject: Re: Very Basic Color Theory Question...

Andrew 100% is correct on this one we do it daily,and only in cmyk,no conversion to rgb then back to cmyk can take place because that is how we have set it up. Our 10600 is a cmykcm printer not a cmykrb printer. Again with the 10600 it's up to whoever owns it to set it up however they like. Jim Donovan
________________________________________________________________________

   Date: Thu, 24 Mar 2005 17:40:02 -0500
   From: Rafe Bustin
Subject: Re: Very Basic Color Theory Question...

Is this a function of the printer, the driver, or some particular RIP you're using?

I've worked with lots of inkjet printers of several brands -- but only through their standard drivers.  And I've never seen an Epson driver that didn't mangle CMYK files.

So there must be something very unique about the 10600 driver, even compared to (say) the Epson 7000 driver.

I understand that the Epson ESC/P command set allows detailed control of every ink nozzle -- but I assume we're not talking about that.  The data flow must be either
 
Photoshop -OS -driver -Esc/P -Printer
-or-
Photoshop -RIP -Esc/P -Printer
 
My understanding was that, without a RIP, it was the OS that forced the extra/hidden/ undesired CMYK->RGB conversion.

rafe b.
http://www.terrapinphoto.com
________________________________________________________________________

   Date: Thu, 24 Mar 2005 16:17:38 -0700
   From: Andrew Rodney
Subject: Re: Very Basic Color Theory Question...

On 3/24/05 3:40 PM, "Rafe Bustin"  wrote:

Is this a function of the printer, the
driver, or some particular RIP you're using?

The driver. GDI and Quickdraw drivers (which ship with the Epson) don1t know how to handle CMYK data. The 3black box2 in the driver from Epson expects RGB to do (in the case of say a 2200) an RGB to CcMmYKk conversion. If you send it CMYK, it creates a heinous CMYK to RGB conversion, you end up with an ugly print. Now place a RIP in the print stream (I use ImagePrint). The RIP will send, and the Epson will gladly take either RGB or CMYK data. This black box is simply different.

I've worked with lots of inkjet printers
of several brands -- but only through their
standard drivers.  And I've never seen an
Epson driver that didn't mangle CMYK files.

The standard drivers from Epson don1t handle CMYK, period.

Andrew Rodney
http://digitaldog.net/ ________________________________________________________________________
 
   Date: Thu, 24 Mar 2005 16:39:41 -0700
   From: jim donovan
Subject: Re: Very Basic Color Theory Question...

  Thanx for the great reply Andrew. It is the driver, we use full press and a harlequin rip to make digi pages. It's all in how you set it up and what you drive it with. Jim Donovan
________________________________________________________________________

   Date: Fri, 25 Mar 2005 16:53:10 +1100
   From: Johnno Subject: Re: Very Basic Color Theory Question...

As someone working in customer support, what you tell people about colour depends on their level of knowledge and/or usage. If a home user lifts the lid on their Epson printer and sees Cyan, Magenta, Yellow and Black etc ink cartridges and then prints a CMYK file through their non-PS driver they get crap.

Therefore you tell them its an RGB device and they happily print RGB files from PhotoShop.

If your talking to someone with a RIP system that cost them five times the cost of the printer - its a CMYK device.

Printers have always called the Cyan plate 'Blue' and the Magenta plate 'Red' - as Dan says, its less syllables and easier to yell! It doesnt mean that Blue, Red, Yellow is another set of primaries.  There is enough confusion out there without adding to it...

Just my AU0.02c..and therefore practically worthless ;)

____________

John Dee
Australia
________________________________________________________________________

   Date: Fri, 25 Mar 2005 16:43:39 +0100
   From: "Francisco Bernal Rosso"
Subject: Re: Re: Very Basic Color Theory Question...

I always teach so: Primaries are, for light ONE blue, ONE red and ONE green. For pigments: ONE blue, ONE red and ONE yellow. In press the blue is what is called "cyan" blue, the red is what is called "Magenta" (or purple) red...

And about printers, I tell my alumns to tray to get a yellow from red, blue and green inks... when they cannot, the convince that primaries are red, blue, yellow for pigments.
/*--------------------------------------*/
"If quality is important, sRGB is not an option"
(From the European Color Initiative web page www.eci.org)

Francisco Bernal Rosso

Luz-color-fotografia
Redacción y traducción
________________________________________________________________________

Adobe Photoshop training classes are taught in the US by Sterling Ledet & Associates, Inc.