This idea was the fortunate outcome of a problem that arose during development of the mystery lights prototype.
What you see
The exhibit consists of a box divided into two chambers. The chambers are painted white inside, and have lights inside them. Two holes in the front of the box allow you to see the illuminated interior of each chamber. A control box allows you to adjust the colour of the light inside each chamber.
With your eyes close up to the box, you adjust the knobs until the two chambers appear to be lit by the same colour of light. It is a pale green. Then you walk away from the box and look at it from 5 metres away. The chambers no longer appear the same colour. The left-hand chamber appears pink!
Because cameras and eyes work in different ways, this effect cannot be photographed. Therefore I obtained the photograph on the right by foul means. It gives an impression of the appearance of the lights when seen from a distance. Seen from close up, both lights look roughly like the right-hand light. However the green in this picture is more vivid than the real thing.
Why?
The first thing to know about the lights is that when you adjust the lights to look identical, they are not physically identical even though they may be visually identical. The left-hand light is made by mixing a blue-green light and a red light, and the right-hand light is made by mixing a deep blue light with a yellow-green light. The visual identity of the two lights depends upon the exact nature of your colour vision. Someone with different colour vision (for example many so-called colour-blind people*) would not see the lights as identical. However, they could adjust the controls until the lights were identical to them - but now the the lights would not look the same to you.
The next thing to know is that in the centre of the retina of our eyes is a region of yellow pigment - the macular pigment - that filters out some of the incoming blue-green light. This means that the colour vision of our central vision is different to the colour vision of our peripheral vision - we are less sensitive to blue-green light in the centre. And just like the colour-normal and colour-blind person, the central and peripheral parts of our vision won't always agree that two lights are identical.
When you are close up to the box, the two chambers occupy most of your field of vision, and therefore you are using mainly your peripheral vision to see them. Your settings of the controls satisfy the periphery's requirements for the lights to appear identical.
When you stand back, the image of the box on your retina is much smaller, and it falls entirely within the area of the macular pigment. Because your colour vision not the same here as it is in the periphery, the two lights no longer appear identical.
*I'm referring to the common kind colour deficiency known as colour-anomaly. I'm writing on the assumption that you aren't colour-blind. This is on statistical grounds.
But...
...when you are close up to the box, you may be seeing most of the light in peripheral vision, but you must still be seeing some of it in central vision. So why doesn't part of the left-hand light appear pink?
The answer is that it does. You see an indistinct pinkish spot in whatever direction you are looking in the left-hand chamber. This was the problem that I encountered during development of the mystery lights prototype. I solved that problem by not using any blue light in my mixtures - which is why the lights in that exhibit are deep yellow.
And what happens if you...
...adjust the lights to look identical when you are standing at a distance, and then walk up close to the box? Do they look different at close quarters?
Yes, this works, but not as dramatically as doing it the other way round. I don't know why this should be.
Turning it into an exhibit
Making a hands-on exhibit from this effect presents a few challenges. These stem from inter-person variability in colour vision. The density of the macular pigment varies many-fold between different people. This means that a setting of the controls that makes the two lights look identical to one person will not necessarily make the lights look identical to their companion. An extra source of variability is the lens of the eye, which gets yellower as you get older. Old and young people will disagree about what is the correct setting of the controls.
This might not be so bad except that, with four knobs to adjust, setting the lights to look identical is very difficult indeed. Even with quite a lot of practice and an understanding of how colour-mixing works, I find it a long and slow business. Expecting someone to walk up cold to the exhibit and make the adjustments is unrealistic.
The key to making this a practical exhibit is therefore to reduce the number of controls. As most of the variability between people is in their sensitivity to blue-green light, it's possible that we could set three of the knobs permanently, leaving the user to control only the blue-green light. I haven't tried this yet, so I don't know whether it will work.
Inside the box
Each chamber contains LEDs of two colours, placed under a shelf so that no light from them can shine directly out of the hole or illuminate its edges. The two colours of LED in each chamber are alternated along the circuit board, and complete mixing of the light occurs by reflection from the matt white floor of the box. The LED brightnesses are controlled using the same pulse-width modulation system that I used for the mystery lights prototype. The control system is in the small compartment at the left end of the box.