The “Cafe Wall Illusion” Project

Things that Think – Spring 2001

Leo Burd & Jennifer Peterson

1. Abstract

The goal of our project was to create an exhibition-size version of the Café Wall Illusion. The Café Wall Illusion is a popular optical illusion that has been puzzling people’s minds for a long time. The illusion was first reported by Richard L. Gregory and Priscilla Heard in 1979. Its name is derived from a curious effect found in a pattern of bricks on a cafe wall in Bristol, England.

Similar to other optical illusions, the Café Wall results from our brain trying to make sense of conflicting information received from the environment. More specifically, it is based primarily on a concept called border locking. According to Gregory and Heard, the way the human visual system registers shape and color is not very precise, especially when it has to manage contrasting colors such as the ones found in the bricks and mortar of this illusion.

The basis of the illusion is a wall that contains bricks of two alternating and contrasting colors. The bricks are divided by thin strips of mortar like those you would see in a regular brick wall. When the bricks are shifted so that the colors do not line up and do not form a checkerboard pattern, the brain tries to make sense of the inputs it receives and generates a plausible image. In the case of this illusion the horizontal mortar lines do not appear parallel. Instead, the horizontal mortar lines appear or converge on alternating sides of the board. For more information about this effect, please check http://www.cs.ubc.ca/nest/imager/contributions/flinn/Illusions/CW/cw.html .

Today there are many implementations of the Café Wall Illusion available as dynamic pages on the web (for instance see, http://www.exploratorium.edu/exhibits/cafe_wall/cafe.html , http://www.essex.ac.uk/psychology/visual/cafewall.html , http://www.sandlotscience.com/Distortions/Cafe_wall.htm), but we couldn’t find any physical implementation of it that could be used in a science museum or in a class demonstration.

Our idea was to build something large and robust enough so people would be able to interact with. Eventually, individuals would be able to measure the distance the bricks moved before they could see the illusion, and with the help of a computer, compare it with the measurements taken by other people.

2. Design

We decided we were going to try to implement the Café Wall Illusion after visiting a couple of Science Exhibit sites available on the web including http://www.exploritorium.com, and reading the book The New Way Things Work by David Macaulay. We thought that the Café Wall Illusion was a strong enough illusion that it could be implemented. We also thought it contained the right scope complexity for the limited time we had to complete the project, and that it would also provide an interesting automaton. We were particularly interested in the fact that we could make it large, and something that we a science museum could potentially use.

2.1 Materials, Construction and Mechanism

We built the automaton using a large piece of ¾” ply wood as the base. We made eleven slats of bricks that were divided by small strips of mortar. The titles alternated red and white. Six of the slats also contained a top and bottom piece of mortar. These six slats were then glued to the plywood base with enough room in between them to slide the other five slats. The last five slats contained only the bricks and the thin strips of mortar dividing them. These slats had to be able to move in as one unit. In order to accomplish this, we connected the five slats with a vertical board on each side of the wall. A wooden three-armed pinwheel controlled the movement of the rows. The pinwheel was driven by a small 1 rpm motor that was mounted to the foot. This foot held the wall in a vertical position. The user has a control box with two buttons, one button to switch the motor on and off, and another button to control the direction of the movement.

2.2 Revising the Design

During the design phase we started by drawing several diagrams. From these diagrams we began by building the brick slats, attaching the stationary slats, and adding the moving rows. Even though we did not have the motor with us – it had to be ordered –we thought that we would be able to add it later on. We also thought that we would be able to add a frame and the foot, without major difficulties to the project. It turned out that mounting the motor to the rest of the structure was more difficult then we expected. As a result we ended up having to cut slits for the rows that slid in the slides of the plywood, and we also had to devises a way to mount the motor to the foot, while still allowing for a single drive shaft that had to be in the middle of the plywood backing.

The mount engine drive shaft went through several revisions. We started by using piece from erector sets. This allowed the motor shaft to slip. We tried adding rubber, but still we had not luck. Finally we devised a collaring system that allowed us to mount a small collar on the motor shaft. We then super glued the small collar inside a larger collar. This collar contained a pin we drove through the pinwheel arm and into the back plywood board.

2.3 Obstacles

During the development of the Café Wall Illusion automaton we faced a couple of challenges:

• The size of the whole project.
• The Café Wall Illusion automaton was not the kind of project that we could carry around and use our spare time to work on. We basically needed the kind of tools available in the lab, and due to the short time available and the lack of flexibility in the materials themselves, there was not much space for playing with a number of design alternatives. We had to plan for something we considered simple and try to build it the best way we could.
• Beyond Lego-Logo.
• The size, friction, and weight of the materials used required components and connections were beyond the capability of the ones commonly offered in Lego kits. In our opinion, Legos are easily assembled and disassembled, which brings a lot of flexibility to the implementation of automata-like projects. In our case, we had to order a more powerful motor. We then had to connect this motor to the power line, build a strong wooden structure to hold everything in place and allow the user to control the whole thing in a nice way. Although this functionality is probably required in a great variety of projects, it could not be implemented with Legos and crickets. In fact, crickets would have to be connected to the motors via a combination of relays – something beyond our knowledge of electronics – Additionally there was no way to have their IR-receiver pointing to one direction – the remote control – and the IR-transmitter pointing to another one – a cricket measuring distance among bricks.
• Mounting mechanical parts together.
• One of the hardest challenges we had related to the connection between the motor shaft and the wooden arm that would drive the bricks from one side to the other. The shaft was very narrow and it did not provide enough surface for mounting the large arm. After trying with a different combination of Erector parts, we ended up going to McGuckins Hardware to talk to more experienced people. The final solution was a combination of two metal collars.
• The illusion parameters.
• Reportedly, the Café Illusion effect is affected by a variety of parameters such as the color contrast among dark bricks, light bricks and mortar, the relative thickness of the mortar, the size of the bricks, the distance to the audience, and room illumination. We carried out a number of tests and noticed that with the colors we used, the illusion works better in dim or black light.

2.4 Original Idea

Originally, our idea was to have a cricket operating as an infrared remote control for the motor. This would eliminate part of the wires. We even attempted to implement this. One cricket would be the transmitter while the other would work as a receiver controlling the motor. However, due to the lack of time, lack of experience and the lack of adequate materials available with us – we only had two single pole, single throw, solid state relays – we had to go back to a more traditional and simpler wire-based implementation.

We were also planning to provide users with a way to measure the distance between bricks and use the computer to help them compare results among different people. In order to do so, the idea was to use a linear potentiometer similar to the volume controls that slide from top to down in hi-fi systems. Eventually, that would address our needs.

However, we would have to test and check whether the values returned by the potentiometer could be read by the cricket and be precise enough for the analysis we intended to do. According to our colleagues from other projects, using a cricket to manage potentiometer is not a trivial task. Perhaps the construction of a cricket-friendly potentiometer interface is something that should be tried. It would definitely enrich the variety of projects that could be built.

3. Evaluation and Education

Our automaton was designed to help people think about optical illusions. People would be able to explore it from different positions and analyze how the distance between bricks would vary from person to person. It is very hard to face an optical illusion and avoid start thinking about how the brain and our vision system work.

3.2 Figuring It All Out

The basic mechanism is very easy to figure out. It is the optical illusion per se that is hard to understand. To understand the mechanism all you have to do is turn the wall around and look at it from the back

3.3 Comparison With Other Objects

There are no other objects to compare the illusion with that we know of .

3.4 Intended Response

There is an intended response for the automation. People usually get excited when they experience a nice optical illusion. The goal is then to get them to stop, look at it, ponder, and try to make sense of the illusion

4. Things We Learned

• In projects like this, where we are experimenting with a lot of new materials and mechanisms, it is very important to focus in the central parts first and develop prototypes to prove the basic concepts. Only when we have an overall picture of how all the parts will fit together we should go to the real implementation. This does not mean the we need to understand everything in detail, nor that the picture will guarantee that the project is going to succeed. However, without it the chances of success are very low.
• If we go beyond the protected frontier of Lego-Logo projects, an understanding of basic electronics is required. It is virtually impossible to implement projects like the ones we developed in class without knowing at least how relays, transformers, energy sources and electric motors can work together.