View Full Version : Who sez you can't mix primaries!

John H
04-22-2002, 09:45 PM
The other day I mixed dioxazine purple and thalo green and got a fairly blue blue. Greyed, of course. It looked similar to prussian blue.
Does anyone have a "special mix" you'd like to share?

04-25-2002, 06:13 PM
I was surprised at the result of that mix myself when I tried it, not an easy mixing result to foretell and a good example of what subtractive colour mixing can be like in practice. It makes a good "iron blue" but a bit too grey, at least with the colours I used.

You can mix a huge range of transparent reds, oranges and orange-yellows with Quinacridone Rose and one of the good high-chroma arylide yellows. More opaque results can be had using Cadmium Yellow Medium. Same goes for green-blues, greens and green-yellows using Phthalocyanine Blue GS with the same yellows.

One of the things I like to do is match convenience colour mixes and other pigment colours that I don't have on my palette. Some of those I have tested and/or use myself:

Buff Titanium
Titanium White, Raw Umber and a touch of Raw Sienna

Sepia (Asphaltum is very similar)
Transparent: Bone Black, Burnt Sienna and Raw Sienna
Opaque: Mars Black, Red Oxide and Yellow Ochre

Van Dyke Brown
Burnt Umber, Bone Black and perhaps a touch of Burnt Sienna

Naples Yellow
Titanium White, Red Oxide and Yellow Ochre

Payne's Grey
Ultramarine and Burnt Sienna (more interesting that Ultramarine with black)

Davy's Grey
Titanium White, Mars Black and a touch of Yellow Ochre

Sap Green
Phthalocyanine Blue GS and Cadmium Yellow Medium

Hooker's Green
Phthalocyanine Green BS, Burnt Sienna and a touch of Raw Sienna

Emerald Green
Phthalocyanine Green BS, Azo Yellow Light and Titanium White

Cadmium Orange
Cadmium Yellow Medium and Cadmium Red Light (near-perfect match)

Ultramarine Violet
Cobalt Blue and Quinacridone Violet (more opaque than the real thing but close in hue)

Mars Violet
Red Oxide and Dioxazine Purple

Benzimidazolone Maroon
Quinacridone Violet, Mars Black and a touch of Burnt Umber (good match in hue, not quite as opaque)

And FWIW my favourite neutrals are:
Cobalt Blue and Burnt Sienna
Ultramarine and Burnt Umber
Phthalocyanine Blue GS and Raw Umber
Phthalocyanine Green BS and Quinacridone Rose


John H
04-25-2002, 11:16 PM
I use ultramarine & burnt umber for neutrals too.

I tried mixing cad. orange and sap green to see how close to a yellow I could get. I got a pretty good "raw sienna".

04-29-2002, 12:58 PM
One of my favorite mixtures, though it has nothing to do with mixing primaries, is yellow ochre plus a bit of diox. purple, plus white to make the highlights of rocks. It makes a complex-looking
yellowish or brownish grey.

04-29-2002, 04:47 PM
Hi Patrick, I like mixes like this a lot too - colours that are hard to 'pin down' visually - and I'm making an effort to use them more. I hate seeing obvious pigment colours used straight in paintings, "spot the Burnt Umber shadows"!


04-29-2002, 05:04 PM
Am I correct to think that this mixture I suggested looks neat because the mixture has complex peaks and valleys? If so, then trying to replicate it with a brown plus a white, althought it might get the overall hue, value and chroma right, wouldn't have that same spectral complexity, thus would lack the 'neat' look of the mixture I suggested?

If so, this is fascinating, because I find an analogy between color and music. Hue is like the note/frequency, and specrtral complexity is like tone. You might have two colors that are the same color (hue, value and chroma) just like you might have two different instruments playing the same average note, but they are clearly different to look at/listen to because of the spectral complexity; overtones.

05-04-2002, 05:28 AM
Hi Patrick, sorry for the delay in replying but I was thinking about the issue and forgot to come back. I think you are almost certainly correct (good thinking!) that two different mixes that have the same hue -essentially an average of the reflectance spectra - and one looks more interesting because of what parts make up that average. But if so this is a very subtle point in colour mixing and it would be very hard to take advantage of in practice, you might have to approach a given mix using starting colours that were as different as possible for it to be noticeable at all. What would be more practicable and interesting perhaps would be where you aim for a given colour in the painting with only partly mixed colour using scumbles or half-mixed paint maybe, where the final colour is much more obviously an optical effect.

I don't know enough about music but I think you're right there too: the hue being the note, with differences in tone equivalent to spectral variations. I wonder if there are any colour analogies with chords or if that's stretching the analogy a bit? Larry might have a thought on that one.


Keith Russell
05-04-2002, 01:33 PM

One can mix primary colours--printers do it every day. (Cyan is not 'blue', magenta is not 'red', but...)

The colour system that Maxfield Parrish used was comprised of variants of secondary colours, whereby magenta and orange make red, violet and cyan make a sort of navy-esque blue, and so on.

(No, as far as I know, even in this system, 'yellow' is 'yellow', and cannot be made by combining other colours.)