David Tudor’s Untitled: Feeding Forward

Words by Matt Rogalsky

Tudor devices. Photography by Matt Rogalsky

I spent the month of July 1966 in David Tudor’s basement, in Tomkins Cove, New York. He was upstairs in a fair amount of distress, having suffered a number of strokes which had left him blind, partially paralysed, almost unable to communicate verbally. He had a 24 hour nurse. It was hard to tell how much he was really in the world; occasionally he had moments of peace in which I felt I could connect with him, by reading a letter from a friend or talking to him about what I was doing, and my interest in his work. I had been introduced to him only two years before, and met with him occasionally for interviews or other research.

Most of the day and night I was downstairs, sorting, cataloguing and (where possible) testing hundreds of Tudor devices. I was pretty much overwhelmed by the sheer number and the obscurity of homemade devices, simultaneously incredibly charismatic and enigmatic. Most boxes had numerous connection points and knobs, switches, etc, but if any of these was labeled it was usually with some cryptic mnemonic code that it seemed only Tudor could interpret. It was very frustrating knowing that he was just upstairs but reduced to such a state that he couldn’t assist in the work I was attempting.

What became apparent to me was the likelihood that any of the connection points on any of the boxes was both an input and an output, depending on the situation. It seems that Tudor’s use of electronics did not depend on rules of ‘proper’ use but rather on intuition informed by deep background study in the subject; and there are many accounts of how he employed circuits ‘backwards’ or made use of the chaotic behaviour of failing components. Study of his devices (they are all accessible for use as part of the World Instrument Collection of Wesleyan University’s Music Department, in Middletown, Connecticut) needs to proceed according to Tudor’s own experimental working methods. It is rarely the case that a single device forms the identity of a composition; it is rather a synergetic combination of devices which is required.

Possibly as early as 1969 Tudor was employing principles of electronic feedback to generate sound ‘spontaneously,’ i.e. without any external input. Photographs of his setup for Rainforest (the version for the Merce Cunningham dance of the same name) in 1971 show a number of boxes which upon examination turn out to be little amplifiers; Tudor made use of them as sound generation devices by connecting outputs back into inputs, allowing the noise in the system to blossom into controllable chaos: whoops, chirps and irregular rhythms. John D. S. Adams has written about the importance of Tudor’s experiences at Expo 70 in Osaka, working with an elaborate sound modifier console designed by Gordon Mumma, for the development of these ideas.

The feedback soundmaking approach grew into a major composition, Untitled, which Tudor first performed in the summer of 1972 on a European tour with John Cage (the duo’s last major collaborative outing, and the first for which Tudor received composer billing). Untitled employed virtual chains of dozens of devices. For practical purposes, Tudor divided his system into ‘Source Generation’ and ‘Performance Processing.’ ‘Source Generation’ involved double chains of amplification, equalisation, fixed and variable phase shift circuits and a couple of Gordon Mumma-designed modulators, which produced feedback-generated sounds that Tudor recorded to tape. ‘Performance Processing’ took as input the prerecorded material from the ‘Source Generation’ phase. The practice of effectively multiplying the complexity of a system through reprocessing source material in this way (keeping the overall size of a system manageable, particularly for touring) is a feature of many Tudor pieces from around 1970 onwards.

Between May and July 1972, Tudor performed Untitled across Europe, with Cage declaiming his text Mesostics Re: Merce Cunninghan (hear Track 3 on the accompanying CD), it was premiered on 8 May 1972 at Radio Bremen, Germany, and the simultaneous performance of separate works mirrored the Cage/Cunningham method of collaboration (any synchronicities acceptable). Each performer had an independent multi-channel sound system and Cage used an array of microphones to move his vocalisations around the space, from loudspeaker to loudspeaker.


Part of David Tudor’s score for “Untitled”

Tudor’s published note on Untitled (in the liner notes for Three Works for Live Electronics, Lovely Music LCD 1601), begins: “Untitled is a part of a series of works composed in the 1970s that were developed through experiments in generating electronic sound without the use of oscillators, tone generators, or recorded natural sound materials.” Another version of this statement, written out longhand and found in the Tudor papers at the Getty Research Library in Los Angeles, uses significantly different wording: “Untitled is part of a never-ending series of discovered works in which the electronic components are found to be natural objects”

This suggests to me that Tudor thought of this series of feedback pieces as having always existed in potentia; somewhere there was energy waiting to be released and he was searching for means of letting it express itself. The electronic devices, though human-made, are following their ‘natural’ tendencies: the performer’s role in creating the piece is to herd the electrons in one direction or another. The sounds are only semi-controllable and the instrument’s barely manageable complexity is strategically designed to thicken the plot.

Tudor’s interest in ‘natural’ processes was an integral part of his career as a composer/performer. In 1979 he stated: “It seems to me that the way I use the technical medium, it’s just more of what’s already there. I don’t see, for instance, what’s unnatural about a parabolic reflector. They exist in nature already, perhaps not in perfect forms, but neither am I after a perfect form” (1979 interview with Billy Кlüver). It would perhaps be correct to say that it was Tudor, rather than Cage, who most sincerely practiced composing as ‘imitating nature in her manner of operation’ (one of Cage’s professed goals).


Matt Rogalsky in performance

In 1975, Tudor developed Untiled into a commission for Merce Cunningham entitled Toneburst (the dance is entitled Sounddance). The live performance of the 1975 Toneburst differed from Untitled in that it used no prerecorded input; most or all sound was created by feedback loops. Interestingly, the piece went through a process of simplification when Cunningham brought the dance back into repertory in the early 1990s. Due to Tudors deteriorating physical condition, this late version was essentially an exercise in mixing CDs of material recorded in the 1970s; it was still an astonishing sonic tour de force, however. After hearing that version in a Cunningham performance for the first time, I approached Tudor and told him what a stunning experience it had been for me; he laughed and said, ‘Yes, it’s shameless, isn’t it?’

In 1996 I became interested in exploring some of Tudor’s feedback loop principles, and finding new ways to implement them if possible. Searching for an affordable, programmable digital signal processor, I found one which Digitech had discontinued that seemed to offer the possibility of configuring internal feedback circuits, and also had unusually good MIDI control: up to 30 parameters could be altered remotely using a bank of MIDI faders. I found one second-hand in a New York City music shop, and bought it without knowing if I could in fact use it to model Tudor loops. Soon I had it up and running, figured out how to chain together virtual effects boxes — phase shifters and filters and gain stages, à la Untitled — and sure enough, when I routed the signal output back to the input, the box began to speak. It had its own ‘digital’ character but the range of sounds I was able to produce was impressive, and after a lot of experimentation I settled on four simple configurations which I programmed into its memory. Suddenly I had an extremely portable, unique (as far as I knew) new digital instrument deriving from a rich analog heritage. It also pleased me that I had been able to subvert the intended application of the device, turning it from a sound processor (great for ‘ear candy’) into a sound producer: a non-obvious use of the technology which I thought owed much to what I had been able to learn from Tudor’s work.

The feedback patch which proved the most interesting combined two phase shifters with a ten-band graphic equaliser, as shown in the diagram below. Sometimes small changes can cause enormous sonic shifts, and sometimes in surprising ways. For instance boosting high frequencies with the graphic equaliser can have the effect of bringing out low frequencies in the feedback loop.


Applying the instrument in improvisational performance (calling the piece Tudor Loops as a study of and hommage to Tudor’s practice), I took various approaches on different occasions. In 1996 I made a multichannel version that overlaid live performance with multiple prerecorded performances, which I could mix in and spatialise. In 1997, at STEIM in Amsterdam, I made a more complex version (excerpted on Track 4 of the Resonance CD) that employed a system of light-sensitive triggers which allowed me to access sampled fragments of previous performances. Additional sound processors were available (primarily pitch shifters), and the multiple layers of sound were dynamically spatialised via a digital mixer controlled by my software, which moved sounds among four channels.

My feedback loops are extremely stripped-down version of Tudor’s ideas, partly due to limitations of the Digitech processor’s memory, and partly due to my interest in focusing on the core principles of Tudor’s work circa Untitled. My loops are very different from his, not only in the technology used to realise them; Tudor’s phase shifters typically provided a fixed amount of shift, which might be manually adjustable but otherwise did not vary. They caused subtle changes in the signal, which became amplified in the feedback loop. My Digitech box has phase shifters which sweep continuously, and it’s the interference between two phase shifter speeds that makes things interesting.

Although since 1997 I’ve more or less completely migrated to a notebook computer as the heart of my performance setup, the Digitech box with its Tudor Loops patches still sounds unique and I recently used it as a sound source within a large installation of Tudor’s 1973 group piece Rainforest IV in California. That to me was true feedback in action: my sounds, inspired by a 1972 Tudor composition, reintroduced as input for a realisation in 2001 of a Tudor piece from 1973.



  1. David Tudor, Three Works for Live Electronics, Lovely Music CD 1601 (liner notes on Untrt/ed by Tudor).
  2. John D.S. Adams, ‘Giant Oscillations: The Birth of Toneburst’ Mus/cworks 69, 1997, Toronto Canada.
  3. David Tudor Instrument Collection (part of the World Instrument Collection), Music Department, Wesleyan University, Middletown CT 06459 USA.
  4. David Tudor papers, Getty Research Institute, Los Angeles CA USA (finding aid accessible via www.getty.edu).
  5. David Tudor, 1979 interview with Billy Kluver, Experiments in Art and Technology papers, Getty Research Institute, Los Angeles CA USA (finding aid accessible via www.getty.edu).