Max/MSP patch for headphone installation

The initial patch was built around 15 individual stereo sound files. These were arranged in three libraries so that, at most, three sources could trigger at any one time. I reasoned this was a good level of complexity to start with, one capable of creating an interesting layering effect, while avoiding cacophony. In the earliest version of the patch, playback was initiated using a simple metronome to drive three random number generators, which, in turn, sent triggers to the individual sound file players. Not surprisingly, this delivered a rather mechanical sounding output, but this was not an issue when material was being selected and edited for composition – the selection criteria simply being interesting and unexpected collisions between audio events. However, building a ‘soundscape engine’

capable of creating a mediated, real-time output with a grounded environmental sensibility, required much greater flexibility.

While searching for alternative approaches for triggering files I observed that the installation’s wireless technology facilitated mobile listening and hence participants would be able to explore the physical site and the superimposed soundscape

simultaneously. An analogy was noted here between the installation and soundwalking practice, especially with reference to Schafer’s (1994, p. 213) original brief for, “an exploration of the soundscape of a given area using a score as a guide.” The use of a text score to orchestrate playback was subsequently considered, trialled and implemented.

The role of the score was primarily envisaged as a form of poetic grammar (see Street, 2012, p. 14), something designed to initiate a cyclic, and if necessary, rhythmic feel. One reason for its inclusion followed the observation that while John Cage used many

permutations of random and chance procedures in his work there was always an

underlying set of authorial directions. For example, the score for Williams Mix, is a tape cutting and splicing guide derived using I Ching operations, which enables the work to be produced using different source recordings (Gilbert Chase, 1966, p. 592).

Playback timing was controlled by predetermined triggers sent by a ‘reading’ of the text score incorporated into the patch’s main page. This reading began as an actual spoken word presentation that I recorded into Pro Tools. I then noted the start time of each word in the score and transferred this to a timing grid to create a temporal map of the score. A triggered pulse was then sent thought the grid so that at the point a word appeared in the spoken delivery, the ASCII codes for the word were output. This provided me with a set of numerical codes that could be used to initiate the playback of sound files or activate sound modifiers. Most importantly, this meant that words with specific meanings could be used as triggers for sound files with specific content. I also added a second mechanism that would allow me to send all the triggers embedded in a sentence at once. This was a tool I designed for use in the field so that if, while invigilating, I needed to generate a denser soundscape for demonstration purposes, I could.

The importance of the score in relation to the architectonics of the work varied considerably over the course of the installation programme. For the first three

installations I produced entirely new scores based on memories I, or in the case of the Bournemouth installation my mother, had of the location. Subsequently, I produced a generalised score, based on the recording instructions I devised for myself during site visits. As a result, this score was appropriate for a range of different locations. At the time of its introduction I felt that by removing a variable, as a result of standardising the score, I might find it easier to assess the primary differences in the generated soundscape output for each new site. I had also been refining the output for a while and had moved away from the denser, more rhythmic and arguably composed approach to triggering in favour of a sparser aesthetic. This, I felt, offered an output that was still clearly mediated, but one that was more like a coherent environment. This was achieved by thinning out the numbers of triggers embedded in the grid. Thus the score still served a useful function as it facilitated the overall control of timing, sound density, file access and sound treatment, but the importance of its ability to impart rhythm diminished.

In order to function as intended the sound files needed to be archived in themed libraries.

For example, if the word ‘weather’ appeared in the score the trigger could be sent to a library containing a range of meteorological phenomena. Rather than specify an outcome a file would be chosen at random, which allowed the work to unfold in an unpredictable way. The first sound bank classification I devised was to differentiate between ‘people sounds’ (sounds of human actions, voices, motions) and ‘place sounds’ (environmental atmospheres, machine sounds, natural phenomena). These two basic categories proved useful and were eventually developed (prior to the SPR Colloquium at Goldsmiths in July 2011) to differentiate files in relation to five criteria:

• Stereo acoustic spaces and keynote sounds

• Binaural acoustic spaces and keynote sounds

• Stereo and binaural sound events – emphasis on people

• Stereo and binaural sound events – emphasis on places and materials

• Monophonic sound events – objects

The Max/MSP patch used for the first installation at De Montfort University had four-voice polyphony and accessed twenty stereo sound files. These ranged in duration from approximately three seconds to three minutes. During the course of the installation programme the number of sound file players was increased to five, and the file storage capacity to thirty. This was found to reduce the number of sound repetitions to a level that more accurately mirrored the behaviour of sounds in everyday environments. The library function was also developed to incorporate monophonic sound recordings and a sub-routine to randomize their placement in the stereo field. Generally speaking,

monophonic events were found to add a sense of realism and directionality to the overall mix. Binaural sound recordings helped add a realistic three-dimensional feel to the soundscape (especially for motion events such as aircraft or bicycles). However, careful management was required; too much binaural activity, especially when layered, proved to be psychoacoustically disorienting. It created problems with the localisation of sound and generated unpredictable masking effects.

Various forms of sound file manipulation were explored over the course of the

installation programme including automated looping, filtering, speed, direction and repeat functions. Once again, the score was used to trigger these functions. While the effected

sounds added a novel, compositional aesthetic I soon found that participants became fixated on the source of the sounds they did not recognise. This led many discussions back towards Max/MSP and its sound processing capabilities. As the software was not the focus of my enquiry I began to move away from obvious treatments.

As the installation programme progressed the desire to treat sounds also lessened as the novelty wore off. At the same time the focus on recording, preproduction, sound quality and curating sounds increased, and the work became more knowingly phonographic.

Less complex mixes relying on the interaction of ‘open’ materials controlled by cuts, juxtapositions, layering and the native characteristics of different microphones and recording techniques proved to be more appealing and engaging in the long term.

By programming a zero option within the playback field of each file player it was possible to create an audio output which included periods of silence, into which the prevailing soundscape could easily bleed; so as to enhance the ambiguity between recorded and ‘live’ sound. This subtle device became one of the most powerful in

establishing a link between the work and the quotidian sound environment and enhancing the sense of site-specificity. Details and discussion concerning the individual headphone installations can be found in chapters 5 and 6.

Fig 3: Max/MSP patch: example main page

Fig. 4: Max/MSP patch: example file player bank (one of two per patch)