Before we start, I have to tell you that the title of this post makes no sense.
Why? Because there is no such thing as a true primary in paint. At least, not in the sense that the word is often used – that you can mix any colour from the three primaries of red, yellow and blue.
So why are we still taught using a colour wheel that shows these primaries, and the secondaries and tertiaries that can (supposedly) be mixed from them?
If you don’t already have a lot of experience with colour, I think that colour wheel can be quite misleading. Because you can’t mix greens, blue greens or low value browns (which are basically oranges) with it, at a reasonably high level of chroma.
It has quite a limited gamut – i.e. the range of colours that you can mix with it is restricted. It gives the impression that, given three primary colours, you can mix all the others.
Conventional wisdom says that using a limited palette can make learning colour easier, because instead of taking on too much at once, you can learn what you can do with a small number of colours first, and then move on.
But I think that’s a mistake. A lot depends on the palette you use, of course, and even more on whether you’re in the company of an experienced teacher who understands colour, or you’re going it alone.
From what I’ve found trying to mix and then paint with this palette, there are two main difficulties:
- There are areas of the colour space you simply can’t reach, so you could spend a lot of time trying to mix unobtainable colours – that’s certainly happened to me in the past when I tried to learn colour using limited palettes.
- Many colours are significantly harder to mix, especially at either end of the chroma spectrum. High chroma colours are problematic in some hues, downright impossible in others. But what surprised me most when I worked with this palette was how tricky it is to mix low chroma colours.
I started an investigation into this “primary” palette with the last post by seeing what range of colour you can mix with them.
In this post, I’m trying to actually paint with them. I want to find out what it’s really like to try to paint realism with this palette.
I’m doing it for two reasons: Firstly because I hope it will be informative, and secondly because I want to challenge my own assumptions.
I use and teach a different method from the one we were taught at school, one that is much more effective and more practical – at least I find it so. It comes originally from Graydon Parrish and his (I think) visionary work with Munsell.
But I wanted to see exactly how far I could get with using just these colours.
And actually, I was surprised. I could get a lot more out of them that I expected I would be able to. You can mix a pretty useful range of colours with this palette.
It also meant me using complement mixing when I needed to reduce chroma, something I never normally do. And now I know why – because it’s hard! The method I usually use for colour mixing treats each aspect of colour – hue, value and chroma – separately as far as possible. It makes it much easier to hit target colours exactly.
I do think there was a lot of value just for me personally in doing this experiment though, because using complement mixing really stretched me, and being stretched is a good thing.
As long as your’e not being stretched too far, though.
The sweet spot for learning is to be stretched just beyond what you’re currently capable of. If you’re just learning to mix and especially if you’re learning on your own, in my opinion trying to hit colours exactly with complement mixing is going to be too hit and miss to teach you much.
Most likely you’ll spend a lot of time mixing mud, and wondering why. Well, if my email in-box is anything to go by, you’re not alone.
One last point before I get on with this post: if colour for you is about what you feel and not what you see, you’re reading the wrong blog. I write here for realist painters who want to learn to paint better. I have no argument with your approach if it works for you. I’m talking about something different here.
All clear on that? Good. Now we can start.
Can you paint realistically with red, blue, yellow and white?
I decided to start by trying to do a simple painting twice, once using my usual tube paints and once using this primary palette, and at the same time. Just so see how close I could come.
The subject I chose to start with was an orange, because I wanted to test out the orange end of the colour space. In this case, I also had an orange with a leaf on it, so I got to test green at low chroma a bit too.
Here is the set up I used:
And here are the resulting paintings:
And here they are next to the subject. First the full-palette painting:
Then the primary palette version:
Kinda wish I’d done a better job of both of them now, they were both done pretty quickly just to test the colours. But you get the idea. Have a look at how close you think the colours are.
Here’s the palette with some of the colours pre-mixed.
I’ve got the range of colours I would usually use for a subject like this, including the background down the left:
- titanium white
- cadmium yellow
- permanent orange
- cadmium red
- transparent red oxide
- raw umber
- burnt umber
- sap green (for the leaf)
- black
That gives me a full range of values and lets me easily mix a range of low and high chroma oranges right through that value range.
The “primary” palette is on the right – cadmium yellow, quinacridone rose, ultramarine blue, titanium white.
- titanium white
- cadmium yellow
- quinacridone rose
- ultramarine blue
The colours I have premixed there are the main modelling factors of the orange. If I can judge and mix those colours well, I can paint a pretty convincing orange.
As I was saying before, though, what really surprised me was trying to mix the low chroma colour for the background. Here’s a quick vid of me attempting to do it. It’s downloaded from facebook and was broadcast live so the quality isn’t great, but hopefully it’s informative nonetheless.
Now let me just go through what struck me about working with the limited palette:
The table top
Firstly, and most obviously, I couldn’t get near the colour of the table top the orange is sitting on. This is because I matched the value as closely as I could, and you can’t get a brown (really a low value orange) at that value with this palette. To get the value low enough you have to use ultramarine, which takes the mix in the direction of magenta.
That hurts the realism because the hue of the table top is very different to the hue of the front edge of the table. They don’t look like they belong to each other.
The occlusion shadow
There’s a shadow underneath the orange of very low value called the occlusion shadow where there’s no light at all. For the painting I did with my normal palette, I just used black and some burnt umber for that.
With the primary palette painting, I couldn’t get the value low enough, and it went blue, because that’s the lowest value I could get. It just looks odd. Flat.
The orange
Here the primary palette did best, because it can mix fairly high chroma oranges. Where it fell over, though, is the darkest part of the shadow. I couldn’t get the chroma high enough because to get the value right, I had to add blue. So again, the shadow went decidedly magenta. It hurt the feeling of three dimensional form a little I think.
Everything took much, much longer
Behind all this was the amount of time it took me to mix each colour. That had multiple effects on the final painting, not least that I didn’t paint it as well because I had less time actually painting. But it also meant that the last stage – final adjustments to the hues, chromas and values of various parts by painting over or scumbling – didn’t happen at all. The sheer effort involved in having to mix each colour, and the subtle differences I need to create made it just too difficult to do.
That might seem a minor thing but I think that had as much effect on the final painting as anything else.
It was hard. Too hard!
The biggest thing that struck me as I was painting though, and which will stay with me, is that attempting this painting felt forcibly like what it felt like to paint before I’d learned much about colour.
I struggled to hit the colours I was seeing.
Thinking that I must be seeing the colours wrong because they were so hard to match, I kept remixing even though I knew intellectually that I couldn’t hit the target colour with the paints I had. Think about that for a minute. Even though I knew it wasn’t possible to hit the colours I was seeing with this palette, some insidious part of my brain was telling me that I was the problem, I wasn’t good enough. That is not a helpful thing to be happening at the easel!
The sheer frustration of mixing, mixing and not getting the colour I wanted made painting a deeply unsatisfying process.
Knowing as I was painting that the painting just was not working was tough! That one hurt. At least now I know enough about colour to have a good idea why, whereas before I was simply left with the frustration.
It was like going back in time to before I knew how to judge and mix colour and the experience was not a pleasant one.
So can you really paint with just primaries and white?
Yes, you absolutely can. And I did better with it than I thought I would. But it has some serious limitations, that you’d be well advised to be aware of before you try it. Especially if you’re doing it to help yourself learn colour.
Frankly, using a palette like this is going to make colour mixing, and painting as a whole, much, much harder. You may even begin to think that you are the problem, that you just don’t have what it takes, when all the time, the problem is actually our palette.
So my advice would be not to use this palette thinking that it’s going to help you learn colour. It’s not. It’s going to get in your way and hold you back.
I’m not done yet with this, I’ve got more experiments still to do and a lot still to learn about the differences between different palettes – and how useful (or not) they are. To an extent, I was just trying to get together a way to test them out with this and had a lot to work out as I went.
Next I’ll be looking at using this palette to mix skin colours, then I’ll try to paint a fairly high chroma green apple and see how it does. I kinda know already, but it’ll be interesting to see. I might be surprised.
I’m pretty sure it’ll teach me something and I hope it will do the same for you too.
As a last word, I’d like to encourage you to try some experiments like this. See for sure what range of colours you can reach with a limited palette. It teaches you so much about both colour in the real world and what happens on the palette.
Best wishes and thanks for reading,
Paul
The Keys to Colour - Free 6 step email course
Learn how to:
- mix any colour accurately
- see the value of colours
- lighten or darken a colour without messing it up
- paint with subtle, natural colour
Paul, I really want to buy your painting course, but I have no way to get the Munsell Book (it is so expensive, for me is about 2 months of hard work). My question is: does it wor with the Munsell Students book, or another edition less expensive?
Thank you for your answer and for sharing your knowledge in such an interesting way.
Best wishes.
Balta
Hi Balta,
Yes the course is designed to be done with the Student Book.
I know the big book is outrageously expensive, which is why I built it that way. I’m giving serious consideration to producing my own chips in future too, which will obviate the need for the student book too. Not there yet though!
By the way if you do get the Student Book DO NOT GET THE 4TH EDITION! There were some problems with the printing of the chips and a lot of them come with the chips stuck together.
I think it’s up to the 5th edition now and I’ve heard from at least one student that that edition is ok. I’m waiting to hear feedback from more people though.
Ok, thanks for your answer. I will search for the Students book in Amazon.
The feedback on the 5th edition is terrible. See Amazon.
thanks Gaye, checked. A lot of the reviews talk about the chips being stuck together but many of those reviews are also from before the 5th edition was published – Amazon lumps all the reviews of the different editions together, and doesn’t say which edition the review is for.
It would be really disappointing to hear that the 5th edition still has the same problem.
do you have a 5th edition yourself?
If Iām thinking about a primary palette, I think about 4 color process printing. I use pthalo blue, quinac rose, and cad yellow light of bismuth yellow, plus titanium white and black. Black is a necessary addition. It should allow you to get to brown. I think it makes for useful color mixing charts and I have learned a lot using this palette. It is also a good travel palette. In reality, I also use a red oxide color, ultramarine, and a mid gray. Sometimes I even have a full complement of grays on the palette (red gray, green gray, blue gray, yellow gray, gray), and an orange (permanent orange, cad orange, or transparent orange), and sometimes a couple of greens. It simply saves time in the field, and with plein air painting, time is meaningful.
CMYK palette yes? I plan to look at this one, I have a lot of questions about it that I want to answer with a palette knife in my hand š
Interesting that you add those low chroma colours – I think that will likely help you with one of the main problems I would see with that palette – that, similarly to this one, mixing low chroma colours will be complicated.
With your range of low chroma colours, you’ll be able to quickly reduce the chroma of colours in a wide range of hues. That’s a really sensible addition I think, and having tried to mix low chroma colours with high chroma tube paints, I can see how it would save you huge amounts of time!
Ok, thanks for your answer. I will search for the Students book in Amazon.
Hi Paul, great experiment. I use primaries to mix most of my colours (with the exceptions of also adding Burnt Sienna and Pthalo green to my palette). The trick is to use 6 primaries instead of just 3. I use 2 yellows, 2 blues and 2 reds. For example ultramarine and cad yellow will create a rather dull green because neither of those pigments will reflect much of the green spectrum. The primaries that I use are Cadmium yellow (reflects orange), Arylamide (lemon) yellow (reflects green), Cerulean blue (reflects green), ultramarine (reflects violet), quinacrodine red-violet (reflects violet) and cadmium red (reflects orange). With 6 primaries the possibilities are endless.
Yes, the “split primary” palette from Michael Wilcox. I’m going to be doing similar tests with this one too. Basically it just means more tube paints = a wider range of available colour š
I have a feeling that one of the main drawbacks I found with this palette – it being more difficult to mix low chroma colours – will be the same with that one, since all the tube colours are still high chroma. The value range will still be limited too, with no black. But the range will certainly be greater than with this palette.
I’m planning a series of experiments with different palettes, so it made sense to me to start with the most restrictive and go from there.
I think your burnt umber and phthalo green are good additions, too, since the umber will give you higher chroma browns at low value, that you probably couldn’t get otherwise, and the phthalo will let you get right down the value range to near black.
Out of interest, how do you go about mixing low chroma colours with that palette? Complement mixing?
yes, mostly compliment mixing to reduce the chroma and ‘grey’ the colour, although sometimes just mixing alternate primaries gives fairly low chroma results also, eg cad red and cerulean gives a very brownish violet. I use Burnt Sienna and Ultramarine for an almost black value, or sometimes pthalo gr with cad red but that’s not quite as dark.
I’m looking forward to seeing the results of your experiments Paul š
That sounds like the warm/cool palette where you choose the warm and cool version of each primary color.
It is, yes – at least that’s what a lot of people call it.
Personally I try to stay away from the terms warm and cool because I find them too vague, relative and not very descriptive. Hue is a better way to describe it I think.
Those tube paint choices will give you a wider range of colours, yes. But you’ll still have mostly high chroma tube paints on your palette, so mixing low chroma colours will be more difficult. It’s likely that you’ll end up painting with too high chroma – which funnily enough is the most common mistake I see in colour, and this is a popular palette choice.
Please try substituting manganese blue for ultramarine blue and see how it goes. It is impossible to get a good result when using ultramarine blue. The red that is present in the ultramarine pushes the green and orange towards gray/brown. Good for seascapes and landscapes, not good for still life with an orange.
I will. I suspect that whilst it will allow for higher chroma greens, it will mean loss of chroma in the magenta and purple area of the colour gamut.
The manganese blue will not help with higher chroma low value browns (really oranges, in terms of hue) or the higher chroma mid value oranges on the orange itself, however. Mixing any blue into those colours will lower the chroma.
Try these colors… Titanium white,
Mars black
PY3–lemon yellow
PY65–opaque orange yellow
PY83–golden transparent yellow
PR254–DPP red
PR122–rose magenta
PR177–anthraquinone red
PB15.3–cyan phthalocyanine blue
PB29–ultramarine blue
You can also use cadmiums, if you wish, and both opaque and transparent iron oxides.
I can almost guarantee one can get any basic hue they want, from high chroma to unsaturated, muted colors. All the colors listed have very good to excellent lightfastness.
Others I find useful: some pre-mixed sap greens, umbers and siennas of your preferences, carbon black, ivory black, indanthrene blue–PB60–and nickel azo yellow…a warm, kind of lemony transparent permanent yellow of vast capabilities. Also, dioxazine violet–PV23–a very intense permanent violet.
For painting nature, I’ve become convinced that it takes split-primaries plus one or two greens and two Earth colors ; especially in watercolor. You might not need then depending on the topic, but you need them available. I try to use the ochres, siennas and umbers were possible and avoid the phthalos, the phthalos are too overpowering. Fortunately there are alternatives, but they’re a bit pricey…but a little paint goes a long way once you have it. And I can economize in those paint-tubes I’ll rarely-use.
Without a low-chroma, low-value near-Neutral you just can’t get there. This is related to why the Carder Palette includes Burnt Umber to create Darks and a Chromatic-Black made of blue and orange (yellow+red)…and there are many limitations to the Carder Palette in terms of mixing “costs” to hue-value-chroma and pigment chemistry.
Many underestimate the subleties of near-Neutrals and their undertones;
Warm gray – red, orange, yellow undertones. (Biege)
Cool gray – green, blue, purple undertones.
Neutral gray
French gray – yellow/green undertones.
Greige – gray with warm black (Warm Gray)