This text is an account of the first workshop led by Tarek Atoui and Inigo Wilkins for Infinite Ear in 2013. This article is composed of a description of the different dimensions of the concept of noise, and of their different applications during the workshop.

The concept of noise crosses many fields - biology, chemistry, physics, computer science, finance, music, crowd movement, etc. Though the common usage of the term generally refers to a subjective decision about unwanted sound, there are many ways of objectively defining noise in terms of statistical distributions of randomness. Noise is a generalized condition irrespective of hearing capabilities.

For musicology and psychoacoustics noise may be defined in terms of non-linearity or non-periodicity of the waveform. The characteristics of such waveforms are disambiguated by the auditory system, however they may also present vibro-tactile information to the whole body especially at lower frequencies.

Information theory, developed at Bell labs following hearing research, quantifies noise according to probability.

Harvey Fletcher, one of the employees whose work would influence Shannon’s formalization, and the ‘father of stereo’, defined noise as unwanted disturbance. However Shannon’s formulation of information entropy allows for a generalized systematic application that is desutured from intentionality and so need not be described in subjective terms as ‘unwanted’. Norbert Weiner’s development of cybernetics has a related but different conceptualization of noise as extraneous disturbance. The continued research at Bell labs and elsewhere into the auditory system as an information processing mechanism moved further into the cognitive domain, enabling rigorous explanations of phenomena such as auditory masking and non-linear compression, and for the development of perceptual coding.

Conceptualizing Noise

Noise can be understood as incomputable or non-cognizable randomness. A complexity beyond the present means of calculation.

In thinking this we should be careful to make a distinction between digital computation and cognition. As Alan Turing recognized, the computer (as we know it) is a logical machine, or discrete state machine. The brain is an ‘unorganized machine’ allowing for exponential drift. It is sensitive to initial conditions; the processes are non-linear and involve stochastic randomness.1 So we should think about cognition in terms of the deterministic chaos that Poincaré discovered in complex orbital systems.

The brain can be understood as a pattern completing device (association, manipulation, completion). That the brain’s unconscious capacity for the completing patterns is of a higher complexity and range than conscious though is shown in Mismatch Negativity Response tests. (MMN) Effectively the brain is primed for the detection of noise: anomalies, interruptions in patterns, or the disturbance of regularities, since these may signal danger or may reveal new knowledge.

Biological feedback

Modern biosemiotic conceptions of noise trace back to Jakob von Uexküll’s seminal work on ethology, where he developed the notion of the organism as an information processing systems equipped with sensors and effectors and operating through processes of feedback control.

He calls the world revealed to the perceptual array of an organism its umwelt, or milieu, while what exists outside of its perceptual capacities constitutes the umgebung, or environment. This conceptualization is implicitly non-hierarchal in the sense that it does not favour any perceptual system as more developed. However complex the sensory apparatus there will always be an umgebung that exceeds its processing power. Uexkull talks of nature in musical terms; the cellular level is rhythmic, organs are melodic, the milieu is harmonic or contrapuntal, and nature is a symphonic composition. Though the umwelt constitutes a horizon for sensory knowledge it is not fixed, and crucially technology allows for a prosthetic extension of the umwelt, for example the microscope, telescope, or computer etc. For example, using sound synthesis technology we may speed up and slow down a waveform, or zoom in and out in order to hear its microsonic constituents. At certain thresholds in the microsecond time range there are perceptual boundaries for timbre, pitch and tone, beyond which the continuous waveform disintegrates into discrete beats. The audio technology set-up that Tarek Atoui and I devised can be thought as a prosthetic extension of cognition through a distributed external system designed to explore through the horizon of the discrete and continuous or the way they are dialectically nested in the multi-scale complexity of the waveform.

Day 1

We created a situation of creative improvisation where several performers share the same speaker, so the collective production of sound was the starting point. There were 4 players connected to one sub-woofer, each had control of an infrared sensor which produced a digitized pulse: Each pulse had a different frequency so members could identify their own sound. We introduced the concept of this music space first using pictures and then with audio examples. Then we encouraged the students to explore this space through exercises that progressed from simple to complex rhythms. We also built up from two players to four players. We started by trying to conduct four players together but found it very difficult to maintain coherence. We then tried the same tactic with 2 players, this was also difficult. To overcome these problems we demonstrated an exercise where one player copied the rhythm of another -- building from simple to complex. At the end of that session we played a guessing game using a microphone and subwoofer -- students had to guess what word or sound was being relayed. Another problem we encountered on the first day was the different sensitivities to loud sounds -- some students found it overwhelming while others wanted more volume.

Day 2

For the second we devised a more complex set-up with more parameters for control -- this included a foot-controlled pressure sensor for repeating a single sample, a dial to control speed, an infra-red sensor to change samples and a volume control slider. We introduced the concept of discrete and continuous sound and the way these form a continuum. We did this first though a tactile example : students closed their eyes and felt a series of bombs (door knobs) and then over a more complex carved wooden pattern. The former were countable while the latter were a continuous texture. We introduced the new control parameters step by step starting with the repeat function, then the speed dial and volume slider, and lastly the sample selector.

On the first day the set-up was very simple and open and this led to difficulties in coherent collaboration. This is why on the second day we added further constraints to enable directed exploration of sound space. Students needed to have notions of music and rhythm, which they were not used to, so we needed to build up these abilities through structuring the space of exploration.