In order to formulate and condense my concept for Beatspinner I was influenced by several existing sound toy projects.

Before I made Beatspinner v1.0 for submission as my sound applet I needed to gather inspiration in order to settle on an idea. I knew I wanted to make a sound toy aimed at children so I started by googling ‘sound toys’. The first site I stumbled across was soundtoys.net [1]. Here I found a lot of interactive flash/shockwave movies that provided varied interaction with sound but many without any notable structure. This then led me to seeing the obvious pitfalls of going down a similar route and producing something too random and unintuitive in flash.

I then stumbled across another flash sound toy online called Jam Tarts [2]. This movie used a tray of jam tarts as the basis to form a sequencer. The jam tarts themselves were a metaphorical representation of a 3x4 matrix where sound triggers could be placed, while a horizontally looping sprite activated them when it became adjacent vertically. Although the concept may seem puerile and the functionality flawed, it was the first online sound toy I had encountered that had an intuitive interface and a logical structure. This then made me take note of my interaction with the piece which was one where I wanted to keep playing with it until I could make something that sounded pleasing. I thought it was simplistic and flawed but found this concept very engaging and therefore, influential.

I then found online a second sound toy I cite as an influence, dubbed: Snythesite 1 [3]. This movie was again, simplistic, but with obvious structure. The segregation in sound was defined by the colour and size of the sound triggers. But also what was important was the overall cyclic shape that many of these triggers together formed. I found myself spinning the mouse franticly in an effort to activate as many sounds as I could as fast as possible. It was then I thought that if I combined this action of sound triggering via rotation with the logical process of sequencing I could have a sound toy that could offer many different sound structures but was also refined and intuitive.

Once I had loosely formed the concept that became Beatspinner I had too choose whether I implemented it using hardware or software. I was aware that both were valid forms of sound interaction but had to decide which was applicable to my idea. I chose to further investigate online and looked into tactile sound interfaces. I came across Soundgarten [4]. This tactile interface offered a way of interacting with sound by plugging ‘mushroom’ shapes in a layered plastic dome. These ‘mushrooms’ trigger samples that can then be subsequently modified by plugging in other attributes. This tactile form of interaction that was designed for the younger brain with intuition in mind really appealed to me. What I found most compelling in relation to my project was how Soundgarten made use of physical space and used it to make a relationship with sound. No knowledge of computers of GUI’s was required to play with the sound toy.

“Compared to common graphical sound software interfaces, the principle of Soundgarten has the following advantages: It involves haptic-tactile senses and spatial reasoning, taking advantage of basic human skills and therefore enables children, Illiterates or handicaped to make use of new electronic and digital technology. Contrary to WIMP (Windows Icons Menu Pointer) Interfaces Soundgarten allows multi user activity.” – Soundgarten website [4].

Although I still liked the idea of sequencing from Jam Tarts [2]. Reading the Soundgarten [4] website made me realize the metaphor was based on a graphical user interface anyway. I also didn’t like how with the interactive movies on screen the temptation was to simply click everywhere and see what happened. I like the idea that with hardware, users would have to reason their interaction more logically by thinking about physical space. It was then I modified my concept to fit a more intuitive model of hardware design.

What became my project, Beatspinner, was really an amalgamation of ideas from Jam Tarts [2]. Snythesite 1 [3] and Soundgarten [4].


My Production process for Beatspinner v2.0 was one driven by hardware development in order to deal with some of the design flaws associated with Beatspinner v1.0 (discussed in Sound Practice Evolution). Here I will give a technical insight into the development of this prototype.

When I started this project, I wanted to develop my interface out of an old Marantz record deck I bought in a second hand shop. I had not modified it at all when I made Beatspinner v1.0 and so my primary aim was to adapt this piece of commercial hardware to suit my needs. However I felt that this was more than a simple hacking exercise. Because the project uses a turntable for the method of rotation it draws on people’s cultural awareness of what a turntable does. Just like the needle detects prerecorded sound as a record rotates underneath it, Beatspinner’s micro switches detect generative rhythm as pegs pass underneath them. This concept of rotation and sound activation can also be applied to people’s working knowledge of compact discs & laser heads as well as other media formats. The transference of this knowledge onto the Beatspinner concept gives an insight into peoples preconceptions about how the hardware operates. Hacking a record deck would then seem to add a level of intuition to the interface that I had not considered before by combining the concept I had evolved with the cultural reference of sound control via ‘DJ-ing’.

Once I had settled on the idea of taking the usage of the record deck further I then developed the disc shaped peg board that would keep the interface tactile. I made it by firstly cutting a circular piece of 10mm thick board. I then bought two different thicknesses of dowelling, one larger than the other. This allowed me to make larger ‘trigger pieces’ that would slot over the pegs (made from the smaller dowel) protruding from the disc. I endeavored to make the disc, the pegs and the larger trigger pieces as accurately as possible with the tools available to me. This is so there were no discrepancies with spacing that could have an adverse affect on interaction.


Fig1: A Photo showing when the peg board had just been made.

Once I had built the hardware that spatialized the sounds (fig1) it was time to hack the record deck into something that could loop them. I opened it up and removed the needle and all the audio playback mechanisms that were not needed. I then rewired the power that operates the turntable to a switch that I installed in the top right hand corner. After this I made my own mechanism to support the micro switches that become activated by the rotating disc I have already made (see last post). Making this armature to support the micro switches was quite an endeavor. Firstly I had to make some small thread bar that would go through all 8 switches and hold them all together by using a tap and die set and a steel rod. Once I did this and joined all the switches together I used some larger (M10) thread bar to hold them above the turntable in place. This M10 thread bar was then bolted to two steel arms either side of the record deck. The result effect is that the system is very strong and sturdy but also completely adjustable. To retrieve data from the micro switches and send it to MAX/MSP via USB I decided to hack a USB game pad. I needed to use nine out of the ten digital inputs and it took several hours and a very steady hand to solder wires to the minute copper strips that are normally bridged by the game pads buttons.


Fig2: My hacked USB gamepad circuit board

Once I had sprayed the whole deck silver I hooked up the hacked game pad to the micro switches and put the whole piece together. I found that with once I had learned to use the ‘Hi’ object the switches seemed to pump data into MAX/MSP quite happily.


Fig3: Beatspinner v2.0

I then made my MAX/MSP patch to process the data. This time however I originally started using the ‘groove’ object in correspondence with different zones on the disc rather than with specific sounds. This meant that any sounds could be used in conjunction with these groove objects. However I found it did not take full advantage of the interaction with the hardware. I modified my patch so it would send MIDI data to Ableton live. This way I can use lives synthesizers to generate a multitude of sounds that can be applied to loops and record them digitally. I had to use a driver called MIDI Yoke in order to effectively pass MIDI note data to Live. Once I implemented this with Max’s ‘mtr’ object I was able to record specific samples from one revolution for mixing in Live. I installed a ninth micro switch on the side of the disc so that it can detect every revolution of the disc. I did this deliberately with further expansion in mind. If every revolution can be detected then the data that is being inputted into MAX within a revolution can be isolated by the mtr object and used to store individual samples. Not just that but with further development the data could also be potentially used to create visualizations that give instant user feedback. This would shift the emphasis of Beatspinner more towards sampling and live experimentation which would mean users could create their own Beatspinner mixes much more easily. I feel that this potentially adds an extra dimension to the projects Usage as a sound toy.


For my first version of the project I was awarded 71. It appeared that the concept was sound but the implementation was flawed. To quote some of my feedback:

“The accuracy, robustness and latency of the triggers are the shortcomings of this project, limiting musical interaction somewhat. As an excellent basis for further development…..”

I felt that aside from my positive feedback this was a fair criticism of v1.0. This is why that when it came round to my negotiated sound project I wanted to continue my work and address these issues.

Version one of Beatspinner was implemented using the unmodified Marantz turntable. It used red, blue and green coins to trigger sounds through a use of a live video feed.


Fig4: Beatspinner v1.0

While I was happy that I learnt how to use some aspects of the jitter library in quite significant detail, in retrospect this was not the best way of staging my concept.

In version one I only had three zones because this is all the camera could focus on at short range; however on each of these zones I could place a blue, red and green coin. This gave me effectively 9 trigger points (3x3). Without this understanding of operation though the hardware was not intuitive because each colour made a different sound depending on where it is placed. This made the segregation of sound triggering confusing because it was dependant on both ‘what’ and ‘where’. The most irritating thing about this method of implementation was the calibration required to set it up. I didn’t realize until I developed the patch how sensitive the video feed would be to ambient light conditions. It was very sensitive and it had a crippling impact on the project as a robust and tactile interface.

I firstly sought to revise the triggering mechanism by discarding the use of the video feed and exploring the use of an electronic solution. I thought the use of micro switches to generate a 1/0 input through some sort of physical suppression would be ideal. This is what led me to initially build my peg board idea using doweling to see if it would work. I needed to see if the larger pegs (that are 10mm taller) would maintain support when placed on top of the smaller ones. I also needed to see how tactile the interface was and whether the pieces could be easily manipulated (unlike with v1.0). When I built the peg board I also dealt with issue of the interaction space and sound triggers. With v2.0 as I have previously described I gave each sound its individual zone. There are a total of 8 zones in total.

The main logic behind ‘8’ being the magic number determining the number of zones was routed in basic musical theory. With 8 individual sounds that could potentially be triggered at any one time Beatspinner could effectively play 8 notes within one octave of music. This could allow for basic melody to be generated in one key. I thought that this would give the project even more grunt this time around because as well as using the 8 different sounds from a 5 piece drum kit, 8 different notes could be used, or maybe 8 different guitar chords. This would give this new robust interface and revised spatial interaction more flexibility and as a wider range of samples could be created.


I conclude that I have achieved what I set out to do and have followed the advice for further development from v1.0 of my project. I feel the interface is both tactile and robust which makes for a usable, interesting, sound toy.

Although the system is not one that gives users haptic feedback, it does make basic composition easy and intuitive. Users are required to be imaginative and logical in order to generate desirable output. And because the system is not haptic and the sounds are actuated mechanically the user is completely free to explore the relationship between timing and space.


[1] Soundtoys.net website: http://www.soundtoys.net/
[2] ‘Jam Tarts’ Shockwave sound toy: http://www.theworldisoursoundtoy.com/index.php?page=jamtarts
[3] ‘Snythesite 1’ Flash sound toy: http://www.sas.mdx.ac.uk/first.html
[4] ‘Soundgarten’ tactile sound toy for children: http://www.soundgarten.com/project.htm