Wind Chimes




“During retreats Thay (Thich Nhat Hanh) encouraged participants to give calm, bright-eyed attention to each daily activity, whether eating a meal, drawing a Buddha, or just walking quietly, aware of the contact between our foot and the earth which supports it. In order to encourage this kind of mindfulness, a ‘bell master’ sounded a large bell regularly, and everyone stopped their activity, breathed three times, and recited silently, ‘Listen, listen, this wonderful sound brings me back to my true self.’ ‘A bell is a bodhisattva,’ Thay said, ‘It helps us to wake up.'” (Peter Levitt, 1988)

Windchimes are in their natural state a musical instrument, a percussion instrument, but when augmented with sensors, windchimes become a minimal interface to controlling sound and creating electronic music. A tangible interface to sound performance. Interfaces that are transparent and invisible to the user and that illicit a sense of surprise when used.

Windchimes were the last piece that we developed for the exhibition and in many ways one of the more interesting pieces with a vaired influences which led to the concept being developed. The need to develop this installation came from the Quietplease exhibition itself. The idea was that the exhibition needed an instillation, a piece of work, that would bring people into the entrance of the gallery. The gallery space itself does not have a very well defined entrance and the exhibition space itself, being all black, does not as well have a very well defined entrance. So we wanted something to lure people in from the gallery entrance into the entrance to the exhibition. We wanted a piece specifically for that purpose. So in coming up with the concepts which would fit this use, one of the first ideas I had came from my experience traveling in Thailand and in Nepal; and that is as you enter a temple in Thailand first of all when you are inside the temple itself you walk around in a clockwise motion, touching bells and touching cylindrical shapes as you walk around. But in Chiang Mai in Thailand I believe, at the Doi Sut, that temple, there is at the entrance to the temple a large collection of bells which lead you into the temple.So my first concept was that we could create bells themselves from the entrance to the gallery until the entrance of the exhibition that people would touch and ring as they went through and of course these bells would be augmented with sensors to relay even more digital information. Of course in Buddhist practice ‘Buddha’ means to awaken so the sound of bells and bells themselves are very central to Buddhist practices and these bells that you find generally in these kinds of temples tend to be at least the large ones usually heard in isolation and their sound is actually very very complex and this sound had a very strong influence on the kind of sound I was going to create in this instillation. So generally the smaller bells are less tuned or untuned and are arranged in large sets around temples in Thailand. Devotees strike each bell for forgivness of a sin as they ascend into the temple. I cannot understate the complexity of sound that this creates when they are all rung at the same time. But as we came closer to developing the prototype we realized that the construction of a bell is fairly complex. What we eventually settled on was something which is very common to the surroundings of Taiwan and that is bamboo. Bamboo has a sound when struck all of its own which also, though in a minor sense, fits in well into the concept that we were trying to create. And the idea of windchimes themselves, the action is generally, you gradually from one end to the other will put your hand across the chimes. A very similar action as you would see in the Thai bells. So we settled on using bamboo windchimes in replace of cast iron bells for primarily practical reasons. But I stuck with creating the sound that I remember hearing in these bells in Thailand and Nepal. So that when the people enter the gallery they would touch the windchimes in a forward motion and not only would the sound of the wood hitting one another create sound but the sensors would activate a very complex sounding iron bell influenced sound as well. And when all of these are touched at the same time the sound is very much like what I heard when in Thailand and Nepal.

The sounds themselves were created digitally based on various bell sound samples and various effects and reverbs in LogicPro 7. The complexity of the sound that I created was naturally quite a bit different from what you hear in a temple but it reflects a little bit of my experience in that I created not only an individual bell sound but then for each sound triggered by the windchime interface I would play a specific chord. This would be what we call a dominant structure chord which would contain a root, a third, a fifth, and a dominant seventh note. The windchimes themselves were tuned but not in a scale fashion. I choose to use the cycle of fifths throughout all the windchimes so that when you brush past them you would go through the cycle, the cycle of fifths, is considered the most powerful root movement in music and I used that on purpose to try and lend an even more power to the sound coming out of the interface itself. In the end the result is something which is very, very tense and very, very strong and really does tell people to wake up, open your ears, and come into the exhibition.

Technical Designsketch
The windchimes consisted of 15 tuned bamboo poles hung with wire and connected to an IBM laptop that played the sounds through common computer speakers. The bamboo poles had embedded which communicated through the serial port to the PC. A simple flash client was activated and played pre-created sound.

In Windchimes the main affordance is sound through motion. The user interaction with windchimes was designed to be one directional but in practice there are more complex interactions that can take place. In selecting a sensor I was looking for something that could sense movement and be acquired cheaply and quickly. I found that simple mercury switches suit my needs. The mercury switch is a very small glass container with mercury inside which when balanced properly will detect motion and complete a circuit. I got lucky in acquiring these switches as quickly as I did as I was told initially that because they are apparently also well suited to the creation of explosive devices permission must be granted for their purchase. Luckily I had no problems.

The serial interface was again a disassembled PC keyboard but a different approach was taken from the first attempt to use this method in Hullabaloo. Instead of working with the plastic circuits the breadboard of the keyboard was dealt with exclusively. This did involve some minor customization.

After taking apart a number of keyboards a model made by LEO was chosen due to an easier to understand breadboard. The spec. of the keyboard used was a LEO Model no. KWD-203.

One the main reasons for using this serial interface as it doesn’t require any low level programming to communicate with a PC – the drivers and libraries needed are already in place. The client environment I chose was flash which allows for rapid development using Actionscript . This approach has worked quite well with the exception that two keys cannot be pressed at any one time. Perhaps a different approach can be attempted in a later revision. When we use a this particular serial interface we are replacing with our witch the action of a key press. I substituted the keys with the mercury switches. The keyboard has 2 layers of plastic printed circuit and a blank slide with holes with respect to each keys between them. The top layer and bottom layer are separated in the default state. When a key is pressed, 2 layers will meet at certain hole and form a circuit. There is a small breadboard dealing with the signals form the layers and sending to PC. The top printed circuit has 18 pings and the bottom one has 9 connect to the breadboard. Pair them to get all the signals of a keyboard. The technique I use to map wires to the breadboard I call the matrix. Next I will attempt to explain the reasoning behind this and show how I created this very durable interface.

After I opened the selected keyboard and took apart the breadboard from the plastic print circuit, I used an avometer board2to locate which ping is used by which keys and created a table. I can use a 9 by 18 matrix to represent all the keys of a keyboard (I usually just fill in what I need so some cells are blank). So you can find that:

the entry in the (3,15)-th position of the matrix means “a”
the entry in the (4,18)-th position of the matrix means “b”
the entry in the (4,10)-th position of the matrix means “c”, etc.

You can see that every ping is shared by many keys, since we only need those keys of letters, it is enough to prepare 5 connections for each ping and so I used 1 to 5 sockets. With the help of this matrix, I melded all the wires to the breadboard and collected all the wires’ other side to make sockets for switch’s wires to plug in. And in the last step, I fixed it on a wooden plate. The following is the pictures of the finished result.

The rest of the procedure is combining the keyboard and switches. The diagram shows how a switch connects to the matrix. Every switch has 2 wires(+,-) and the wires will be pluged into x-sockets(top layer) and y-sockets(bottom layer) respectively. It is exact the same as that a key is pressed/released when the switch is turned to on/off. Using this procedure we can connect any key of a keyboard to any kind of on/off switches. And we can also program an application that will be executed when certain key is pressed or released.