View Full Version : Twin primaries

07-14-2001, 05:20 PM
I was wondering how many of you use twin primaries as the basis of your palette? For anyone unfamiliar with this idea it basically throws out the old idea of a single red, blue and yellow to mix your violets, greens and oranges (basically 'cause it don't work!) in favour of at least two of each 'primary' hue in order to more easily mix clean secondaries.

Also, curious as to whether most of you generally mix your greens, violets and oranges, or do you prefer tube secondaries (either single-pigment or convenience mixes)?


07-14-2001, 06:05 PM
Yes. I use a double primary pallette as my basic selection. A warm and cool version of each; which ones vary. Plus white and burnt sienna. I mix all my secondaries and don't use black. Black does make for some interesting mixtures though; so I'm not one to advise against it. I just don't feel the need for it most of the time with my plein air landscapes.

07-18-2001, 02:08 PM
Yes, I use 2 of each primary. One transparent and one that is opaque. Like AC (alizarin) and CRL (cadmium red light). Indian yellow and Cad. yellow light etc.

I also prefer to mix my greens and violets Alizarin Crimson and Ultramarine Blue make a very lovely purple. and Orange I rarely use. I may have variations of it, ie. in the browns or reds, but never straight orange.

07-19-2001, 01:06 AM
Einion, to an extent, I also use a split-primary palette, although I don't limit myself to only these colors.

-hansa yellow light or bismuth vanadate as a cool yellow
-hansa yellow medium as an orangy yellow

-quinacridone magenta as a cool 'red'
-a middle red like naphthol red medium or pyrrole

-ultra. blue as a purplish blue
-phthalo blue GS as a cyan/greenish blue

But I found a very informative website that says that the split-primary palette is inferior to just using a 6-color secondary palette (three primaries plus three secondaries):


It says, among other things, that the primaries that people often choose to "split" around are incorrect (they should be split around magenta and cyan, which are truer primaries, not red and blue). This is why many people find it hard to mix clean purples and greens.

When I have time, I will do an experiment (in acrylics) to compare the mixing abilities of different palettes side-by-side. I will paint color wheels using the following palettes:

-3 color primary based or red ,blue, and yellow
-3 color primary based on magenta, cyan,and yellow
-6 color split primary split around R B Y
-6 color split primary split around M C Y
-6 color primary + secondary
-12 color primary +secondary + tertiary

I will take a picture and post it on this forum to show the differences in mixing power of these common palettes.

TPS, I also agree that black is unnecessary in landscapes. In fact, over a year ago, whn I learned why black is unnecessary in landscapes, this was a quantum leap forward in my learning to paint landscapes...learning that shadows in landscapes are rarely black (and often surprisingly light in value and colorful) was a huge revelation dawned upon me. (even though I'm still not much more than a beginner and still learning).

07-19-2001, 01:43 AM
Yellow-Cyan-Magenta-Black are the primaries used by commercial printers. These '4 process color' inks allow them to print most any variation. However, those printing to replicate an artist's painting may use 16 or more.

My experience using these with paints is that you lose the character of the traditional artists' pigments, resulting in rather sterile and bland looking images. The complexity of the colors with their various undertones, overtones, transparency, tinting skews, etc. is one of the things that makes painting so interesting and challenging for me. Using the "true" primaries seems to negate those wonderful qualities, and makes all paintings seem to have a sameness about them.

This is my personal preference. Each should try their own and decide the "look' that fits their expression.

07-19-2001, 03:00 AM
TPS, although I will always defend cyan, magenta and yellow as technically the truest primaries (regardless of whether it's in printing inks or in paints...both are subtractive color mixing), you are absolutely right that C M Y is not perfect, and not always the best choice for landscapes.

First, although C M Y mixes colors quite evenly, these mixed secondaries will almost always still fall short of the same secondary color from a single pigment. This is why, as you said, much more than 3 colors must be used if you wish to print a painting with good color accuracy. And also, C M Y don't make a black as dark as a single-pigment black.

Secondly, for landscapes, I agree that red yellow and blue often give better results. It's because the R Y & B palette has a bias in that it mixes cleaner, more intense reds, oranges, and middle-deep blues than C M Y. It turns out that this apprximates the way it is in nature: in nature you see very intense oranges/ reds around sunset (not just the sun & sky but the light cast on objects) and to a lesser extent, blues in the sky. And where R Y B cannot mix very clean purples and greens, nature is also this way. In nature, you rarely see pure purples and greens. If you've ever painted foliage or grass using a green that's too intense, you'll get a garish, child-like effect (I've learned this through experience!).

Thirdly, it's easier to mix earth tones with R Y B than C M Y, because of the orange-bias of R Y B (brown earth tones are based on orange).
Since brown is a rich, warm color, it provides warmth. Lack of it could easily give a sterile feel.

But a skilled painter can make wonderful, vibrant paintings using CMY as anchoring primaries. On example is Don Jusko (although he doesn't just limit himself to these three colors):


Although I would probably not use intense greens this liberally in my landscapes, he is obviously skilled enough to make it work well.
His paintings have a distinct vitality and freshness that, in my opinion, can be attriuted at least in part due to the fact that he has mastered the use of CMY as primaries.

There are some areas where CMY is clearly superior. Try mixing a clean purple with a middle red and a blue like cobalt or ultramarine and compare it to a singl-pigment purple like dioxazine purple. It will be somewhat brownish or greyish. Also try to mix a clean, bright cyan with cobalt or ultra. blue plus yellow. You can't. It will be dark and dirty. You can add white to it to make it brighter, but it won't make it cleaner. It's no match for phthalo blue GS (which approximates cyan).

Here's another link that explains the RYB vs. CMY palettes. It's pretty basic, but it's good info nonetheless. Its only major weakness is that it fails to say that there are some instances where RYB is the better choice.



So to sum up, CMY is superior to RYB in some ways (especially when your main concern is mixing the widest range of other colors from only 3 primaries), but is still far from perfect in reproducing all the colors that humans can see, and sometimes, like in landscapes, RYB can be better and prbably easier for beginners. As you said, it's up to the artist to choose the palette that will work best for their work.

07-19-2001, 11:01 AM
Hi Patrick, a million thanks for these links to pallette theory..wow, what a wealth of information...Bingo! a real payoff.

07-20-2001, 06:25 PM
The Handprint site is one of my favourite stops on the 'Net and while I have the utmost respect for the author and his staggering dedication one must not loose sight of one thing: it's for watercolours. Watercolour palettes are not necessarily transferable to acrylics and even more so to oils because they are much more concerned with undercolour, not masstone. Obviously if you glaze a lot in oils the same information can be relied upon, more or less. Anyone interested in this should take a look at the <A HREF=http://www.handprint.com/HP/WCL/IMG/cwheel.gif>Handprint pigment colour wheel</A> and compare the colour positions to their oils and acrylics. I can see any number of inconsistencies, particularly with oils because of the wetting effect of the binder. It probably goes without saying but just in case: ignore the colours you see on your minitor as they are extremely inaccurate and just go by the relative positions.

Now we could obviously debate the pros and cons of various palettes until the cows come home but one thing is pretty universally accepted and that is colour choice is very much a matter of personal preference and need, and not so much a right/wrong issue.

Personally I favour the split-primary system and I wouldn't even consider a secondary unless it is a single-pigment colour; even then I prefer to mix these hue positions as much as possible. As far as mixing clean violets, greens or oranges goes, this is really a snap when you learn to recognise colour bias and buy accordingly: it also forces you to learn the character of your paints and remember them. One thing I particularly like about the system is that the underlying principle is so simple you can learn it in a single sitting (literally minutes) and then it is only a matter of observation and work to put it into practice. Another is that is helps to form the smallest practical palette for your needs; limited palettes are a good idea both for economy and because it is next to impossible to be on intimate terms with all your colours if you own 70 of them! :)

As regards cyan, magenta and yellow this is, theoretically at least, a perfect way to mix observable colour and I would love it if it did. However the operative term here is theoretically, as there are a number of shortfalls. The first is the pigments have to have precisely the right reflectance spectra (impossible): in real terms the cyan pigment (based on phthalocyanine) is pretty close, the magenta is not so good and the yellow is abysmal which is why reds and oranges are the poorest-represented colours in 4-colour process printing and don't forget the very obvious necessity of using black to increase depth and adjust certain colours, a sure sign the system isn't working. Second, if we wanted to use them in painting, they would have to be completely lightfast: well, we've all seen printed items that have bleached in the sun and they loose their yellow very fast, eventually ending up with just the cyan showing typically. The final nail in the coffin that will forever doom this principle as far as painting goes, is that again there is a difference in practical terms between masstone and undercolour so opaque and transparent painting techniques can't both work.

With regard to those mixing palattes you are going to try, that's a great idea and will teach you a great deal about your paints. One thing I think you might be surprised at is how close you can get, in masstone at least, to any number of single-pigment secondaries. For example one can mix almost identical (and if you colour vision isn't perfect, completely indistinguishable!) hues to Cadmium Orange, Chromium Oxide Green and even Dioxazine Violet from the right primaries.


07-23-2001, 06:33 AM
Sure, CMY isn't perfect. As I stated before, no 3-color palette is. But if it is so awful for color mixing, Golden Paints wouldn't have used CMY as the primaries in their color mixing guide.

07-23-2001, 01:26 PM
Originally posted by Einion
I was wondering how many of you use twin primaries as the basis of your palette?

Also, curious as to whether most of you generally mix your greens, violets and oranges, or do you prefer tube secondaries (either single-pigment or convenience mixes)?


I do not use twin primaries. I use one red, one blue, one yellow, one brown, and one white for each painting. I have enough colors to choose from, (at least 60 or 70), that I can use a reddish orange or reddish violet for red, a bluish violet or bluish green for blue, a yellowish green or yellowish orange for yellow, what ever works best for the current picture, but for several years now I have been sticking to 5 tube colors per painting. I choose my tube colors very carefully for each painting trying different combos on a white sheet of paper, making sure that I can get a good mix of cool and warm with the choosen colors, and, that I can get the right tube colors to mix the secondarys that I need and the greys, blacks, and browns that I want. It does take a while to get the right combination, and there are a few colors that I am partial to, so I have to be careful to use the right colors for the painting and not just choose my favorite colors. My favorite colors may not be the right ones for the color tone/color mood of the painting. When I first started painting I was using as many of my favorite colors as would fit into each painting! hehehe My paintings were not nearly so cohesive and it was much easier to get muddy colors. I guess you could call me a color minimalist now. Over time I have learned what works best for me :D

I have nothing against twin primaries, in fact, it does make sense to do it that way, but it would take me longer to find the right mixes and it seems to me that it would be a lot harder to work with. I try to keep the drawing stage and color picking to a minimum so I can get right to the painting stage as that is the part I like best ;) I end up with very clean colors and I like that. It's almost impossible to get muddy pictures with such a limited pallet. I keep only the colors I need for the current painting in my art box so I don't have to worry about picking up the wrong tube, and I write down the tube color name beside the final color mix smudges and keep this in a folder with any reference material the goes with the current painting. This way, if it takes me a while before I can get back to a painting, I am not lost as to what I was doing before the break. Sometimes I cannot paint for a few months.

It just seems like such a lot of work to use more colors and I know it would give me the same problems that I had before with color balance and harmony, so I keep it as simple as possible for now. Perhaps one day I can master juggling more colors at once, but for now this is what I do.

07-23-2001, 02:54 PM
Patrick, believe me, I work in the graphics industry, CMYK really is that IS that awful, just take a look at any printed colour guide and compare it to your paints :D

Originally posted by Domer
But if it is so awful for color mixing, Golden Paints wouldn't have used CMY as the primaries in their color mixing guide.
Not being a smart-ass, you would have to ask them that. If you check the link that StarGate posted in the acrylic forum for their <A HREF=http://www.goldenpaints.com/jp06art5.htm >Historical Color Matches</A> for example, there are some really bizarre suggestions in it that I can see wouldn't work without even trying them.

Seriously, one of the things that continues to surprise me is how different manufacturers offer mixes of different pigments to reproduce unstable or unavailable single-pigment colours - one would think that if they all know the same detailed information they would all come to the same conclusions but it probably comes down more to interpretation than one would think.


Midwest Painter
07-23-2001, 10:48 PM
I too use a multiple primary palette. I am a fan of Michael Wilcox and his Color Bias Wheel theory. My primaries, and their respective bias, are as follows:

Cadmium Red - biased toward orange
Alizarin Crimson - biased toward violet
Cadmium Yellow Light - biased towards orange
Lemon Yellow - biased towards green
Cerulean Blue - biased towards green
Ultramarine Blue - biaed towards violet

Michael Wilcox has a good book about color mixing. It's somewhat geared for boneheads like me. But his theories are sound and they make sense. His book is titled "Blue and Yellow Don't Make Green". He also has a web site for those interested.


07-23-2001, 11:08 PM
Einion, I agree about printed paint color charts. I got several printed color charts for Grumbacher oils and acrylics (which I like) and many of the colors are so inaccurate that it's almost useless. Grumbacher says that hand-painted charts are too expensive. So I said I'd be more than willing to pay for one but they still say they have no plans to make hand-painted color charts. I'm glad that some paint manufacturers do make hand-painted charts. There really is no substitute. Sure it might be expensive, but since I can't properly see what their colors are like beforehand, I am very hesitant to buy their colors that I haven't tried before. In fact this happened just this weekend. I was ready to buy several Grumbacher paints over the internet this weekend, but decided not to at the last minute for this very reason. This cost saving by Grumbacher actually cost them some sales. It's too bad, because I like their paints.

07-23-2001, 11:14 PM
Midwest Painter, that's a book I want to get (Blue And Yellow Don't Make Green). I hope it's at my local art store, if not I'll try Amazon.com.

07-24-2001, 03:07 PM
Blue And Yellow Donít Make Green is a seminal work on colour mixing and was the beginning of a deeper understanding of the subject for me when I first saw it about eight years ago. Although it's not perfect I can't recommend having a look at it enough (your local library might have a copy) as even a few minutes will show you the benefit of the system.

You will find very polar opinions on the book online (just check out the Amazon reviews to see this) but the basic theory is sound and is borne out time and again by practical experience. One thing about the system that makes it valuable in my opinion is it gives you the ability to think about colour intellectually, instead of in a fuzzy, intuitive way, so you can predict mixing results very accurately. This is a great boon to help prevent waste when looking to mix a specific hue.


07-24-2001, 09:52 PM
On my watercolor palette, I use 3 of each, a warm, a middle and a cool of each....makes mixing more interesting and I can get the contrasts I need.

Quin gold
New Gamboge

Scarlet Lake
Winsor Red
Aliz crimson

French Ultra

07-25-2001, 02:00 PM
HI PAMPE, Would you expand a little on How the mixing is made "more interesting" with the triple primaries? thanks

07-25-2001, 03:43 PM
Mario, my greens are certainly more varied....but more than that, I can add depth with temperature and not just value.