Evolution of my watercolor palette


Published on October 14, 2021

In this post I’m going to focus on my watercolor palette. I will describe the choice of colors first and then I will explain the principle by which my setup has been evolving.

The colors

I currently use a 12-color palette:

Hansa yellow mediumI2NoPY 97
Hansa yellow deepI2NoPY 65
Raw siennaI2YesPBr 7
Pyrrol scarletI3NoPR 255
Indian redI3YesPR 101
Burnt siennaI1YesPBr 7
Burnt umberI2YesPBr 7
Cerulean blueI1YesPB 35
Cobalt blueI2YesPB 28
Ultramarine blueI3YesPB 29
Phthalo blue (GS)I4NoPB 15:3
Neutral tintI2NoPBk 6, PV 19, PB 15

All colors are Daniel Smith extra fine watercolors. You can find more information about the lightfastness rating (L/f in the table above) by following the link.

I use a limited palette for the following reasons:

  • It is easier to achieve color harmony.
  • It is rather hard to master even a 12-color palette. By “master” I mean having a clear idea how every pair of colors mixes. I don’t want to hunt for colors, I need to know precise shades that I can obtain. I also don’t want to stick with a few colors that I already know well and neglect others. For me, 12 colors produce an abundant range of possibilities.

My current palette Daniel Smith

This color chart shows on the left hand side the 12 colors of my palette in their pure form. We can also see how a given color mixes with others. For example, the color at the left bottom is neutral tint and the bottom row shows its combinations with the remaining 11 colors. Furthermore, the topmost color on the left hand side is hansa yellow medium and the second column from the left shows how it mixes with the remaining 11 colors. The leftmost 4th color from the top is pyrrol scarlet, in its row we can see how it combines with 3 different colors, the rest being in the 5th column from the left, and so on.

Perhaps the most obvious thing about this particular palette is that it does not allow us to mix arbitrary colors. Indeed, it is not possible to mix an “acid” green because there is no cool yellow; pure purple is also out of reach. This is because I consider my palette an essential tool in defining my style. The colors that I often want to use should be easily to obtain, and I don’t care about colors that I never use. It is even better that I simply cannot mix them. Color is very personal, and so is the palette.

My rendition of Edward Wesson’s Martham Church, Norfolk

The palette has a strong bias towards warm colors. There is no cool yellow, only a medium yellow and a warm yellow. Of these, I almost never use hansa yellow medium as it seems to be still too acid and lacking subtlety and mystery. I would like to explore it more, though. I tend to use hansa yellow deep as a mixer for greens which results in dirty olive greens and, combined with neutral tint, dirty yellowish colors for painting autumn scenes. If I need a light color, I usually reach for raw sienna which is one of my most used colors—indispensable for softening intense colors, painting nature, and skin tones. Pyrrol scarlet is fairly warm and will not mix pure purple, instead it mixes dirty purples that I find rather mysterious and perfect for shadows and dirt. I feel very strongly about what kind of red I want to use. Pyrrol scarlet is perfect because it matches the idea of the quintessential red that I have in my head. I have also tried a medium red—pyrrol red and I simply could not stand it in pure form. I added indian red to have a darker version of red that is different from two other darker red colors—burnt sienna and burnt umber. These 3 colors are my “dark red triad” and I use them a lot. Indian red in particular I use more and more because it is not as obvious as pyrrol scarlet, so it can be mixed in in a lot of situations. I use it for cloudy heavy skies and shadows. Lately, after studying Edward Wesson, I have acquired a certain bias towards dirty purple and my vehicle of predilection for achieving shades of dirty purple is indian red.

My rendition of Edward Wesson’s Blythburgh Church

Burnt sienna is a palette staple. I try to mix grays in various ways, but burnt sienna is the most obvious. It is also great for achieving even dirtier greens. With phthalo blue it is not strong enough to suppress the blue completely, so it produces a lovely grayish green. Burnt umber also mixes grays and my use of it is very similar to that of burnt sienna. I often use burnt umber as a way to add subtle chromatism to neutral tint. On to the blues, I don’t use cerulean blue often, but I do want to keep it around. It mixes a lovely greenish gray with raw sienna and one of the weirdest purple shades with pyrrol scarlet. Cobalt blue is probably the blue that I use most often. To me it is the standard blue that I reach for e.g. when I need to mix a gray color or blue for the sky. Cobalt blue mixes very light greens, which I rarely want. With hansa yellow deep it mixes a gray. I also use ultramarine blue fairly often for the same purposes as cobalt blue, and I find that it mixes more interesting greens. Especially with hansa yellow deep it doesn’t even mix properly, but forms a strange combination that is highly granulating and interesting. Phthalo blue (green shade) is the blue I use most for mixing greens. It mixes a convincing simple green with hansa yellow medium and a more interesting olive green with hansa yellow deep. I do not use phthalo blue often otherwise, because it is very strong and tends to overpower other colors. It should be noted that it mixes a very dark color with pyrrol scarlet. Speaking of darks, I know that neutral tint is sometimes frowned upon by watercolor snobs because there are other ways to make a certain color less intense or to mix a black. True, but with that approach we would be painting with 3 primary colors. Having a palette is about mixing efficiency and reproducibility of colors. Nothing beats having a dark color already available on the palette. It saves so much hassle. It is a quick and reliable way to decrease color intensity (saturation) and to make a color darker. It is also straightforward to foresee the results of adding it. Neutral tint also mixes interesting greens with yellows. I consider myself rather a tonal painter and neutral tint is my single most used color. It is also the only color in my palette that is made of more than one pigment. I tried lamp black and it is also okay, but I prefer neutral tint because it doesn’t alter the original shade so much.

I’m not sure why I tend to prefer these colors and their combinations. One hypothesis is that it seems that these colors match those found in nature, but it is just a guess.

Evolution of the palette

In this section I document how my palette evolved and what made me change it. I hope that the principles I developed might be of interest to others.

Andrew Pitt’s palette

The palette organization that resonated with me early on is described in this video by Andrew Pitt. Since most painters consider 3 primary colors (red, yellow, and blue), and it is universally accepted that there should be at least two (cool and warm) versions of each, why not go further and have 4 versions of each primary color? This gives us a 12 color palette. To this Andrew adds 4 colors that he is “trying out”. However, if you watch his videos you will see that he uses the 4 “auxiliary” colors just as much as 12 “main” colors, so it is fair to say that his palette has 16 colors.

Cadmium lemonIYesNoPY 35
AureolinIIYesNoPY 40
Raw siennaINoYesPR 101, PY42
Raw umberINoNoPBr 7
Cadmium redINoNoPR 108
Light redINoNoPR 102
Burnt siennaINoNoPR 101
Burnt umberINoNoPBr 7, PR 101, PY 42
Cerulean blueINoYesPB 35
Cobalt blueINoYesPB 28
French ultramarineINoYesPB 29
Winsor blue (GS)IIYesNoPB 15
Alizarin crimsonIVYesNoPR 83
ViridianINoYesPG 18
Cobalt violetINoYesPV 14
Payne’s grayIIYesNoPB 15, PBk 6, PV 19

All colors are Winsor & Newton professional quality watercolors.

A few nitpicks: I don’t think Andrew uses cadmium lemon much preferring aureolin. I also doubt anyone would call raw umber a yellow. I think Andrew just likes to see certain symmetry in the organization of his palette and thus he describes it as 4 by 4 by 4 base plus some extras.

Because my palette only had 13 wells for colors, and because I felt that I should try to use as few colors as possible, I decided to use only 12 primary colors from Andrew’s palette. However, since I was also trying to do this with Sennelier instead of Winsor & Newton, I could not really find a replacement for light red. After some thinking, I realized that I didn’t need light red because it is very close to burnt sienna and mostly has the same function. I don’t know what Andrew thinks about that. Thus, I arrived at this 11 color palette:

Lemon yellowIIPY 3
Sennelier Yellow LightIPY 153
Raw siennaIPBr 7
Raw umberIPBr 7
Sennelier red?PR 254
Burnt siennaIPBr 7
Burnt umberIPBr 7
Cerulean blueIPB 28
Cobalt blueIPB 28
Ultramarine deepIPB 29
Phthalo blueIIPB 15:3

All colors are Sennelier professional quality watercolors. I could not find any information on staining and granulation for them. I tried to choose only single pigment paints, but I am not sure this is as important as some people think.

Here is the color chart I made back then:

My old Sennelier palette

Neutral tint

Almost immediately I realized that I had trouble achieving high contrast and very dark darks. If you compare the color chart of my old Sennelier palette with the one of my current Daniel Smith palette, you probably will notice that Sennelier is less intense. I took photos of these charts in identical conditions, so this should give you an idea of how bleak my paintings made with Sennelier were1.

Magical tree

I sought to paint with darker colors. Around that time I discovered Daniel Smith and bought their neutral tint. It was a game changer. I have been using it regularly ever since. I thought several times of getting rid of it but always decided against that. The reason to get rid of neutral tint is that unless you pay special attention you end up using the same black everywhere, which is less interesting than mixing your own blacks that always vary. However, I have solved that problem by always mixing a bit of other colors into my neutral tint so that there is still a lot of micro-variation of color.


For a while I was inspired by Javid Tabatabaei and had white paint in my palette. Many people say that white watercolor is a big no-no. I don’t think white is useless. White is a way to achieve light values but with high concentration of paint. Sometimes it comes in handy, for example when you need to paint a very dark area and then join it with a very light area but without the dark paint migrating completely to the more watery light area. Oftentimes, it is helpful to have fairly thick light paint or to lighten a mixture without changing its concentration. Finally, white can be a way to have a light mixture with higher opaqueness.

Giverny flowers

I have since stopped using white on a regular basis, but I think it is a valid addition to a palette.

The switch to Daniel Smith

After a few months of using Sennelier I succumbed to the idea of switching to Daniel Smith completely. As much as I wanted to support a French company, I found that Sennelier watercolors were simply not my kind of paints. It was also a good occasion to reconsider my choices as I gained more experience and more understanding of what I wanted:

  • I realized that I never ever use lemon yellow preferring Sennelier yellow light. The rule said that I needed one warm and one cool yellow, but I had no use for a cool yellow, not as a pure color and not as a mixer. So I got rid of it.
  • With my old Sennelier palette I always had to add something (most often burnt sienna) to my greens to make them less acid and obvious. I wanted to be able to mix nice usable greens from just 2 components. To manage that I decided that I needed an even warmer yellow, this is how I ended up including hansa yellow deep.
  • I added quinicridone gold as my 4th yellow still trying to follow the 4 by 4 by 4 principle.
  • For the red I knew I wanted something warmer. When I finally tried Pyrrol scarlet it was an obvious choice. I did not like any other reds and the colors they mixed.
  • Indian red is also a new addition. The principle was that I wanted 4 reds in my palette, and indian red had the shade that no other red in my palette had.
  • I removed raw umber because I found that I didn’t use it often and a similar shade can be achieved by “cooling” down burnt umber with a blue.

All other colors stayed the same. I just found their analogues in the range provided by Daniel Smith. I have to say, Daniel Smith is more intense and much more granulating paints than Sennelier. I have never looked back.

Exclusion of quinaridone gold

My palette has 13 paint wells and so for a while I used a 13-color palette. However, I realized that I used both hansa yellow deep and quinacridone gold for the same function—mixing nice olive greens and dirty yellows. Once I started favoring quinacridone gold I would abandon hansa yellow deep completely and vice versa. To me, this was a sign that I had to choose between the two. The function of every color must be clear and well-defined. There should be no hesitation during mixing. I chose to keep hansa yellow deep because it is brighter and a single-pigment paint. It seems to be more versatile than quinacridone gold.

Quinacridone gold removed

16 colors?

I will eventually have a palette that will have 16 paint wells. For a while I thought about the 4 extra colors I could add. I considered adding something similar to Andrew Pitt’s choices: viridian, cobalt violet, and some kind of purple (perhaps quinacridone liliac). I could also add back raw umber. After giving it some good thought though, I decided that I will stick to 12 colors as there is really no benefit in adding extra colors to my current palette. They will only introduce confusion and possibly destroy the color harmony that my current palette benefits from. If anything, I would rather remove colors than add more.

Final thoughts

I think the main principle for building a palette is the same as for developing your own style—try to notice what you like and what you naturally tend to do. Adjust to make that kind of stuff easier. This is also a road to self-discovery.

  1. Well, to be absolutely honest, the paper on which the charts are painted plays a huge role in color intensity. The Daniel Smith chart is painted on Bockingford, while the Sennelier chart is painted on Arches. Arches eats up a lot of color.