Interactive Music Performance

Rainer Linz

The term interactive when applied to music performance can be problematic. This is because in a broad sense, music has always been an interactive art. We could describe any musical performance as a process of real time control over a complex system (instrument), through gesture and based on feedback between performer and machine. Even with a more recent perspective, and with hindsight to advances in computer technology, the distinctions between interactive and traditional music practice can sometimes be difficult to define.

Music is the oldest of the electronic arts, with many instruments predating even the development of electronics in the early part of the 20th century. As far back as the 19th Century it was possible to speak of 'electric music', performed on instruments which generated tones by means of simple electromechanical circuits [1] . American entrepreneur Thaddeus Cahill's organ of 1897 for instance, was the size of a train and intended to transmit music over wires directly into peoples' homes as a kind of proto muzak .

The invention of the Theremin in 1920 demonstrated for the first time an instrument which could be played without any direct physical contact. The Theremin is played by moving the hands near antennae, controlling pitch and volume by gesture alone. Just as importantly, the Theremin stands apart from other electronic instruments of its day by not being based on the traditional piano keyboard, which also brought about new musical possibilities. (Although scored for by a number of composers from Edgard Varese to Percy Grainger, the theremin is perhaps most popularly known through the ethereal gliding tones of the Beach Boys' Good Vibrations, or from film scores such as Bernard Herrmann's The Day the Earth Stood Still .)

Modular voltage-controlled (analogue) synthesisers of the 1960s provided another alternative to keyboard-based instruments. Live performance on these synthesisers called for new performance techniques, such as adjusting panel controls and using more unusual devices such as joysticks or ribbon controllers. Through these we can see another pointer to interactive performance on today's computer-based instruments.

More recently the use of computers in the performer/instrument relationship has made a different type of instrument possible , one which is capable of more complex real time responses to gesture. Yet even here, the distinction between interactive and traditional performance techniqes can be unclear, and there are many instances of traditional instruments with interactive extensions built into them. The Midi Bow of violinist Jon Rose, for example, has a pressure sensor which detects varying pressures on the hairs. In performance the sensor sends information to a computer which in turn creates a corresponding, real time counterpoint to the violin solo .

A growing number of instruments are not based on traditional models, and it can be instructive to look at some of these in more detail. The performance group Sensorband provides a good illustration of some new approaches to interactive music performance. Sensorband members are Atau Tanaka, Edwin van der Heide and Zbigniew Karkowsky. Each performs on a virtual sensor-based instrument, using various types of sensor technologies to communicate with the computers which form the basis of the instruments.

Tanaka's instrument is called the Biomuse [2] , and is a bioelectric controller designed to map the muscle and brainwave signals of a performer directly to synthesiser commands. An electrode headband is used to detect brainwave (EEG) and eye movement (EOG) signals from the performer, while bands on the wrists and forearms detect voluntary muscle signals (EMG). Special analysis software inside the Biomuse can detect features of these biological signals and isloate voltage peaks, particular brainwave frequencies or other patterns from the raw input. This analysis of the performer's actions is then used to create and control the sounds.

Two basic decisions affect the performance on this instrument. Firstly, the analysis software must be programmed to register certain kinds of gesture, and provide a meaningful value for the quality of any related gestures. Here, the performer effectively decides in advance what types of actions the instrument will respond to. Secondly, the value of the movement must be translated to a meaningful sound process.

Edwin van der Heide performs on the Midi Conductor, an instrument developed at STEIM studios by Michel Waisvisz in 1994 [3] . It consists of two independent handgrips studded with a variety of switches, a pressure pad and a two channel ultrasound system, linked to a Sensorlab computer.The left hand grip contains 12 switches which are used to sound the 12 notes in one octave, while the right grip contains a number of special function switches to control musical parameters, such as transposition, synthesiser program changes and so on. The instrument is played directly by pressing on the switches, but also by moving the hands closer to and away from each other to activate the ultrasound system.

The Midi Conductor looks like no previous instrument, but its performance technique is more clearly a musical one. The switches on each handgrip can resemble the keys of a clarinet or a keyboard for instance, and so draw on a performer's previous musical skills.

Zbigniew Karkowsky has developed a unique instrument, consisting of a two-metre high frame, or cage arranged with velocity-sensitive infra red beams. The performer stands inside the frame and uses hand gestures to break the beams. In performance, the velocity-sensitive arrangement of the beams calls for different speeds of movement, and enables a strongly visual and gestural control of sound. According to Karkowsky, "we want to make a pure, primitive, direct connection to our audience by playing music with our bodies".

The impression of watching Sensorband is a strange one, since the performers' gestures are unlike any that are used to play a traditional instrument, yet they clearly create the music that one hears.

An artist not usually associated with music performance is Stelarc, although he has worked with similar interactive sound technologies since the 1970s. Initially, using biofeedback and medical equipment to monitor his biological signals (EEG, EMG, ECG), he was able to create and control sounds by these phenomena. More recently [4] his sound interface - there is some doubt about whether it should be called a musical instrument - consists of motion, pressure, flexion and light sensors attached to his torso and limbs. These are activated by movements of the body, which in turn are caused by electrical stimulation of the muscles.

The term Interactive in music performance can be interpreted widely, and this is reflected in the great variety of approaches to it. Historically, electronic and interactive music have occured almost simultaneously and there are important precedents to today's instrument designs.This confirms that interactive music performance is part of an ongoing musical tradition, and is helping to redefine what we will consider to be music in the future.


  1. P. Lertes. Elektrische Musik Theodore Steinkopf Verlag Leipzig 1933 return
  2. Knapp, R.B. and H.S. Lusted, A Bioelectric Controller for Computer Music Applications. Computer Music Journal Vol 14 No 1, Spring 1990 pp42-47. return
  3. Spider Manual - Sensorlab programming language reference and software guide. STEIM Studios Amsterdam Dec 1994. return
  4. Linz, R. Towards the Design of a Real-Time Interactive Performance Sound System. Leonardo Music Journal Vo 6 pp 99-107.MIT Press 1996 return

This article first appeared in The Interactive Art , catalogue of the 3rd international festival of computer arts in Maribor, Slovenia 1997.

© 2003 NMA Publications and Rainer Linz

back to index page.