Wednesday, September 30, 2015

"Secondaries" as Primaries

In yesterday's post about charting limited palettes, I mentioned that the colors you choose for your palette don’t have to be blue, red, and yellow.

Autochrome by Louis Lumière"Madeleine, Suzanne et Andrée à travers les vignes"
You can use what we think of as "secondaries,"—orange, green, and violet—as primaries and come up with very interesting color schemes. The Autochrome process, an early form of color photography, did just that.
Jean-Baptiste Tournassoud ( French, 1866 - 1951); Le Phonographe;
Autochrome, circa 1912 courtesy the 
Photography Museum
Autochromes used grains of potato starch dyed red-orange, green, and blue-violet.

Through a magnifier (below), the individual colored grains are visible (left courtesy, right courtesy Univ Delaware).

I'm not sure how accurate the color in these examples are. The first one looks Photoshopped to pump the colors, and the second one looks yellowed. But you get the idea.

Yellows are mixed from orange and green, similar to the way they're mixed from red and green in computer screens and theater lights. Yellows are the hard color to achieve this way, because they come out weak and low value, so they have to be tinted up with white, and at best they'll be sort of beige.

But the experience of building a color scheme where the colors we think of as "primaries" have to be mixed from "secondaries" is a strange exercise that will rewire your color brain.


Robert F said...

If you only use green, orange, and purple on your palette, you wouldn't be able to mix high chroma reds, yellows, and blues though would you?

James Gurney said...

Robert, true, in the same way you can't mix high chroma oranges, greens, and violets from red yellow and blue.

Robert F said...

I use liquitex acrylic paint and they sell what are called Primary red, yellow, and blue and they seem to mix pretty high chroma secondaries. Also cyan, magenta, and yellow seem to mix pretty high chroma secondaries. Do the premixed greens,oranges, and purples that paint companies sell have even higher chromas then the secondaries mixed from paints labeled primary red, blue, yellow and a palette of cyan, magenta, and yellow?

Stephen Henderson-Grady said...

They aren't pre-mixed. They are discrete chemical compounds, such as dioxizane purple, pthalo green, and pyrolle orange.

James Gurney said...

Robert, the basic principle is that any mixture made from any two hues will result in a secondary hue of lesser chroma than the original hues, just as a straight line between any two points on the outer edge of the color wheel. The closer the two starting hues, the higher chroma the mixture. As Stephen said, a given paint color will often be either a single pigment or a composite of multiple pigments to achieve a certain hue.

Chris Beaven said...

Very interesting, I wonder if using this palette will make a painting look like an old photograph... I may just have to re-write my color brain as some point in the future, maybe for that graveyard I still need to find :)

Jenna Berry said...
This comment has been removed by the author.
David Briggs said...

Some clarification might be helpful here. Red-orange, green, and blue-violet work as primaries in autochromes because autochromes are viewed by transmitted light, and so the colours mix additively, and these additive primaries give a near optimally large gamut. The same colours will give a near optimally SMALL gamut in largely subtractive processes like paint mixing, where mixing paths between pairs of colours vary from convex or straight (between subtractive primaries) to deeply concave (between additive primaries), for example as in Fig. 5.3.4 here:

James Gurney said...

David, thanks for that additional information. Of course we're not speaking of such experiments with the objective of achieving the largest gamut, but rather achieving gamuts that have unconventional harmonies. If the only goal were to achieve the largest possible gamut from three initial subtractive primaries, then people would probably use something resembling cyan, magenta, and yellow. But from a painter's point of view, large gamuts are not necessarily desirable.

David Briggs said...

Undoubtedly! My comment was just to clarify the answer to Robert's question at the top.

Jenna Berry said...
This comment has been removed by the author.