This is my learning log for TAOP exercise 3-2.
Before going into the exercise itself, it is useful to make some definitions and reflect on the colour theory presented in the course material.Before Isaac Newton in 1666-72 proved that all the colours of the rainbow exists in white light, people thought that colour was created as a mixture of light and darkness. Newton used a prism to split the light into the colours of the rainbow, and, crucially, combined them again using another prism to make white light. That proved that the colours did not come from the prism, so the only place it could come from was the white light. Using this knowledge, Newton arranged the colours in a conceptual colour wheel, which showed the relationship between the colours and their “opposite” or complementary colours.
Several different ways of organising colours have been constructed since Newton’s days. Each of these “ways” forms what is called a colour space, that is, a theoretical model of how colours relate to each other. There are three colour spaces that are relevant for this exercise:
- RYB: This is often referred to as the “painter’s colour space”, and is what we are taught as children in kindergarten and during some art training, including the colour wheel in the TAOP training material. It is based on the primary colours red (R), yellow (Y) and blue (B), with secondary colours orange, violet and green. It is based on the (partially flawed) idea that the three primary colours cannot be created by mixing any other colour, and that any other colour can be created by mixing these three primaries.
- RGB: This is the colour space that underpins all modern electronic colour representation: TVs, computers, digital cameras, etc. As opposed to RYB, which is a subtractive colour space, RGB is additive and describes what happens when you mix light with different wavelengths (colours). This is an important distinction because pigment and light does not behave the same way. The primary colours in RGB is red, green and blue.
- CMYK: This can be thought of as the opposite of RGB, a modern set of subtractive colours. The three primary colours here are cyan (C), magenta (M) and yellow (Y). K is an abbreviation for key (black) and is used because mixing the three primaries in practise gives a muddled brown/grey colour rather than pure black.
Converting between RYB and RGB is far from trivial. After an afternoon’s research, the best explanations I found is in this paper by Gossett and Chen. Gossett and Chen suggests using trilinear interpolation (nope, I am also not 100% sure what that means) in order to convert between the two colour spaces. Helpfully, they also define RGB values for the three RYB primaries and for the three RYB secondaries.
Going back to the exercise, I was asked to find scenes, or parts of scenes, which are each dominated by a single one of the primary (in the RYB colour space) red, yellow and blue, and their complementary (or secondary) colours green, violet and orange. It’s cold and miserable outside and it rains. So even though it feels a little bit like cheating I did most of my scene scouting in the supermarket’s greengrocery department.
The exercise suggests taking three pictures of each scene, with one as metered, one slightly over exposed and one underexposed, and pick the one most closely matching the RYB colour wheel in the course material. As I was working in the studio (kitchen counter covered by two pieces of white foamboard) I set my camera to manual, f/20, 1/250, ISO 1600 (which gave a black frame, thus ensuring the ambient light didn’t contribute) and dialled in a flash until the scene was properly exposed. That happened to be at 1/16th power. This was my “metered” exposure, and to get the over and under exposure I changed the ISO 2/3 stop up and down to 1000 and 2500. I set the white balance to “flash” in Lightroom (colour temperature 5500, no tint).
In order to check how closely my scenes / objects match the intended colour, I took the pictures into Photoshop. Here I selected the subject, inverted the selection (to get a selection of the background), deleted the background, and finally filled the background with Photoshop’s version of the pure RYB primaries and secondaries.
Here are the results, starting at the top of the colour wheel as shown in the course material and progressing clockwise around the wheel:
It was an interesting exercise, and while the actual picture-taking didn’t take long, I ended up spending several hours in front of the computer researching colour theory and trying to come up with a way to assess the accuracy of my chosen scenes. I had initially expected to compare the RGB values of the scenes against the values given by Gossett and Chen’s article, but after having done the calculations I realised that a triplet of numbers measured in a different colourspace is a very unintuitive way to compare things (even if it mathematically may be accurate). I feel that the visual approach above gives a far better way to compare the colours.