Page  00000001 RE-WIRED: REWORKING 20TH CENTURY LIVE ELECTRONICS FOR TODAY Dr Richard Polfreman Music Department University of Southampton David Sheppard Sound Intermedia London, UK Ian Dearden Sound Intermedia London, UK ABSTRACT Music has a long tradition of using electronic technology in performance - from early electronic instruments such as the Theremin, through to today's plethora of digital devices. With rapid technological change, there is always the risk that important works using live electronics can be become un-performable over time, due to the lack of working equipment available, lack of the associated expertise necessary to operate the equipment, or damaged/obsolete storage media containing the performance data. This article reports on some of our experience of reworking pieces for newer technology and preliminary work on a research project examining the problem in general. 1. INTRODUCTION Preservation of electroacoustic music is rising in profile with centers independently finding solutions, and moving toward collaborative work [1]. The main focus has been electronic instruments and audio recordings (disc, tape) together with their playback devices. Issues of preservation and digitization of analogue sources, formats to use and audio repair of degraded material have been discussed elsewhere, including [2] and [3]. Battier [1] outlines the information to be recorded for works involving live electronics and computers, although his main concern is the analysis of works. Our concern is performance: the current archived material that publishers have available, and the practical realization based on that archive. The 'Digital Age' appeared to offer a panacea in terms of providing long-term storage of all types of material such a scores, texts, patch settings, audio and video. However, it may actually accelerate the potential loss of works by reducing the perceived need to record information as text on hard copy in addition to saving settings onto magnetic or optical media. The media is often fragile and easily subject to damage, e.g. floppy disks, while media readers are also at risk of obsolescence, either the device itself (e.g. a Yamaha synthesizer for its data cartridges), or an external reader. Formats have come and gone (5.25" floppy disk, Syquest cartridges, 2.8" Quick Disk), and 3.5" floppy disks are fading fast. Software and computer platforms can also cease production (e.g. Atari systems). Publishers and composers rarely automatically update their archives and transfer to new media as these become outdated. The approach is typically reactive, with demand for a performance leading to updates. Works can be left for a number of years are therefore are in danger of becoming impossible to perform. In this context we discuss work being carried out in collaboration with Sound Intermedia examining the scale and nature of the problem and seeking solutions within the scope of readily available technology, and for archiving works in a manner providing maximum longevity. We hope to stimulate sharing of practice, since these issues must confront many in the field. 2. PERFORMANCES 2.1.Quando Stanno Morendo, Diario Polacco No.2, Luigi Nono (London Sinfonietta, London 1995) This first UK performance was also the first reworking the electronics for different hardware, although some of the equipment had been modelled on the MARS workstation in the early 1990's by Alvise Vidolin [4]. While this performance was realised some time ago, it is good illustration of the problem and a challenging example of live electronics, given the custom nature of the original equipment and the sophisticated routing involved. Further, it illustrates the problem of reworked versions themselves becoming outdated. This piece is for two sopranos, mezzo soprano, contralto, bass flute, cello and live electronics, realised in 1982 at the Experimentalstudio der Heinrich StrobelStiftung des Siidwestfunks, Freiburg, Germany. The live electronics are defined by a series of "patches", signal routings from stage microphones through various effects and into various patterns of ten speakers (e.g. figure 1). To reproduce the patches authentically, a sophisticated audio re-routing system was devised combining analogue mixing desk, 32x32 MIDIcontrollable audio patch bay and digital mixing desk. ti51- i-"i- I.......... 6f~~~ , G I'`E H D I~ ^M'a =E^ Figure 1. Patch 14 from Diario Polacco no 2. Effects were replaced with off-the-shelf units and Opcode's Max software controlling the digital mixer via MIDI to achieve halaphon rotations (sounds moving smoothly between speakers) and gates, with input from breath-controllers and a sound card (figure 2).

Page  00000002 T:T TM I)T M T M2 M3 M4 M5 M*, M: N Cx: BO"C2 N1 Contrler (-PC 1)001 S: 3 7 77 -- - - - -,- 7.... --^- 1:MIDI n. i, Hunn a-iDrao, r -- Auldi ~ flr-nIect Il. n. I -I Multiw u Audti cmnecftnl O1rator Figure 2. Summary schematic of reworked electronics for Diario Polacco no 2. The specification detail varies. The reverberation effect is only given a duration (is it a plate or spring reverb? are there any filter or early reflection settings?). We elected to use a high quality Lexicon reverb as we believed that this worked best musically. In other areas detailed information is given, such as precise frequencies/bandwidths for the two filter banks used. changes causing audible artefacts, and while these were minimised by the mixing desk operator, during rapid sequences of patch changes some problems remained. 2.2. Omaggio a Gyorgy Kurtaig, Luigi Nono (London Sinfonietta, London, April 2002) In this first UK performance, we built upon our previous experience since similar hardware is specified. MSP [5] now available for Max, the electronics could be carried out in software, although for mainly practical reasons, some effects were applied externally (figure 4). Ml M2 M3 M4 `7 7JJT AIMIContr:mie r C--1600)', ++ OK_ I norp t r OPM co A llu ký o ý~ ace (8 2 8Bý, M\ i TP 1| "IS^HH^ & cV ^*oag-a^ Figure 3. Extract from Diario showing rapid patch changes (circled along top of score). The patches are changed throughout the piece, sometimes in rapid succession (figure 3). These were driven by commands from the Peavey MIDI controller, recalling settings on the patch bay, digital mixer, effects units and Max patches. The patch bay did not mute - - MIDI O -- >- Audio) connectona - unarway mAudn scmnn n ^ Fe'reHun cpnn:r... aam.. l i!. ' '.,M.i~:n:3 w.i;jn A!.cld:,':,,qrirl:,:ctiaolla < 9 ~~p Flr~9<; w [ It,2:..:1/n:1]-':,ic,:; Figure 4. Summary schematic for reworked electronics for Ommagio A Gy6rgy Kurtag.

Page  00000003 This is a piece for contralto, flute, B-flat clarinet, bass tuba and live electronics, composed in 1983. Each performer is amplified and processed. There are four patches for the electronics, not particularly complex, although parameter values change, simultaneous effects are applied and sounds move across six loudspeakers. -,d 21 2 3 4 5 6 7 8 9 10 ad1c 1 21341 1 2 3 4 5 6 7 8 9 0 IfI Z Current E Reset C FI Cl Tbn input levels In rehearsal set filter i/p levels. *4=46 wi ch on halaophoos....l before beginning! Delay output level adjusted by MIDI controllers 5,6,7&8 J, |o J ^. |.Luigi Nono a I T Iphons fit Sound Intermedia & Richard Polfreman a m nize l pdacN 12 345678 ~ whereby a number of signal inputs can be routed to specified outputs). A series of routing messages were triggered by program buttons (for rehearsal) and by the space bar stepping to the next program (for performance). The filter-bank was this time implemented by cascading bandpass filter objects in MSP (reson-) to achieve the steep roll-off specified in the score (>60dB per octave). The halaphon now used Adrien Lefevre's vtboule Max objects. These allow smooth interpolation of parameter sets by moving a cursor around a 2D space containing circles for each parameter set. Our sets represented send levels to speakers, and a calculation engine was created to provide adaptable circular motions for sound rotation (figure 6). 3. ASSESSMENT AND REWORKING TODAY Today a 'Diario' system could be a mixing desk, computer with 8 input and 10 output channels and a set of Max/MSP patches. In fact, hardware elements from our reworking, such as the Akai patch bay, might be difficult to obtain today. With 'Ommagio', much of the work was carried out within a single software package, as have more recent Nono performances (April 2005). For 'Con Luigi Dallapiccola' a simple bank of ring modulators with programmable modulation frequencies was constructed, while for 'A Pierre. Dell'azzurro Figure 5. Max main control patch for On Gy6rgy Kurtig. While in the first design a GRM Tools used for the harmonizer effects, in the actual an external unit was used since this elsewhere in the concert and reduced the dem computer system. Also it seemed that thi might give results closer to the original (figui Halaphon 2002 Us thisasaspchwith ru io thr<.gh Fwr MSP, or,Js, MIDI control sets up the circles in the VTboule set channel levels for each position does the interpolation Clubb-elick,> makes circular orbits 'off direction,-tire3 ) radius x pos y p | I X] [^] FT I ^550~| ^50 -240o I atoer c irtlis I |Tpl,magio a silenzio, inquietum', steep-sided filter banks were constructed more accurately by cascading biquad- filters (both high and low cut) to precisely match the original specifications, delays were constructed as before and plugin was tap.shift- used for harmonizer effects. performance While occasionally hardware devices are revived and was needed return to manufacture (e.g. the Theremin), software lands on the emulators have recently emerged, driven by widespread s processor use of analogue synthesizers in electronic dance music. re 5). These systems may provide an immediate aid, with products such as FM7 (Yamaha DX7 emulator) used to replace increasingly rare originals. However, these may have dependences limiting their future viability such as plugin format, or host program. For publishers, the non-trivial task of transferring materials from original format to the emulator remains, and these replace one proprietary format for another, with no guarantee of longevity. More desirable in our view, is the production of a complete system-on-disk for distribution, where one program provides all the necessary processing in an integrated manner. Among the requirements for such programs are that they be: &LUECT a Flexible: programs should function with a variety of audio hardware and operating systems. This is perhaps WTO: the only way to ensure any longevity in a rapidly.........,,,o,"i,, changing technological environment. Control Mode - Audio Input E HO motitE Audio Outputs Q Q r Q r h Q,, o1 MSP n in interpolator window. rbit controls %t slrcify n. Edit multi-s liders for position - here,,e ha'e g.hanne Is contro" able. ( ttntrdl nn<, - using or MII control Figure 6. Halaphon control patch. Delays were implemented with tapin- and tapoutobjects, with feedback controlled via MIDI. Routing into the delays used a matrix- object (a mixing patch bay, Open: enabling other hardware (or software) devices to be inserted in the audio chain to replace the software elements provided. This gives operators as many options as possible, including using original devices or contemporary external hardware if deemed appropriate. Authentic: the program should aim to reproduce the sound of the original system as far as is reasonable.

Page  00000004 However, a composer may have been dissatisfied with results at the time, and would be keen to improve the work. Then again, the effect may have had a particular timbral quality that the composer favoured. With some effects (e.g. Ring Modulation) the process does not fundamentally change, while an effect such as reverb is more difficult, since it can be used either to apply a specific transformation (the original is likely to be appropriate), or simply to provide a sense of space (a current processor may be an improvement). Consultation with the composer is preferable, in order to discover their intentions. In other cases it may be necessary to research into the origins of the work and writings by the composer, contacting technical assistants/sound engineers present at the time, or finally left as a matter of musical judgment. Complete: the program should encompass all the required processing for a work so that no other software (other than device drivers) or hardware (other than i/o devices) are needed. This is the counterpart of the Open requirement: that the electronics can be carried out completely on computer if none of the original systems can be obtained for the performance. Documented: thorough instructions for use, including how it relates to the original specification and external devices needed (microphones, loudspeakers, etc). The original specification should be kept in all cases. Where digital data has been used, this should ideally be printed to hard copy in a human readable format, if not already available in this way. Locatable: in rehearsals the systems need switching to the correct state for any position in the score. This includes sending program changes and other messages to remote hardware if necessary. The mapping of settings to score locations must be clear. Robust: the program must be reliable enough for effective use in rehearsal and performance. Our preferred platform has been Cycling '74's Max/ MSP, which is mature, cross-platform, can build standalone applications and provides a wide range of (extensible) MIDI and DSP functionality. By supporting VST technology [6], commercial emulators/processors can be integrated. Open Source relatives of Max, such as Pure Data [7], provide further choice of platforms, e.g. Linux. Open Source solutions (Pd, CsoundAV, OSW, jMax, etc) may in fact be preferable in terms of distribution and long-term maintenance of programs and we aim to assess these further. 4. CONCLUSIONS The problems of performing works for live electronics escalate as the time from first performance increases. If steps are not taken to update works for today's technology, pieces will eventually be lost since there will be no way to extract the original data and translate it onto contemporary technology. While much effort has been concentrated on the restoration of audio archives and tape compositions (e.g. [8], [9]), the issue of live electronics has often been left to individuals updating works for performances. Preservation of the instruments themselves [10] is useful, particularly as a source for tools aiding the re-specification of a work, but is unlikely to provide availability for performances and does not address the issues of associated expertise. We arrive at similar conclusions as Battier [1], in that there is a need for large quantities of precise text information to be archived alongside works for live electronics to allow reworking subsequently with completely different technology. Exactly who will undertake the production of these materials is not clear, although it is in the interests of composers, publishers and commissioning bodies to ensure that it is done. Our reworking of compositions has moved from offthe-shelf hardware to bespoke software solutions. Placing live electronics within a single software program has a number of attractions, including: simplification of the setup; reduced space required for equipment in the venue; ease of delivery and storage by publishers; lowered (post-development) costs aiding further performances of a work. We are involved in a 12 -month funded project investigating the nature and scale of the problems discussed here and developing an effective strategy for rescuing works at risk. We are grateful to the UK Arts and Humanities Research Council for part funding this research. 5. REFERENCES [1] Battier, M. Electroacoustic music studies and the danger of loss. Organised Sound 9(1): 47-53, Cambridge University Press, 2004. [2] Novati, M. The Archive of the "Studio di Fonologia di Molano della Rai" Journal of New Music Research, 30:4, pp 395-402. Swets & Zeitlinger 2001. [3] Teruggi, D. Preserving and Diffusing Journal of New Music Research, 30:4, pp 403-405. Swets & Zeitlinger, 2001. [4] Chadabe, J. Preserving Performances of Electronic Music Journal of New Music Research, 30:4, pp 303-305. Swets & Zeitlinger, 2001. [5] Zicarelli, D. "An Extensible Real-Time Signal Processing Environment for MAX". Proceedings of the 1998 International Computer Music Conference San Francisco, USA, 1998. [6 ] Steinberg. VST 2.3 Software Developer Kit,, 2003. [7 ] Puckette, M. "Pure Data", Proceedings of the International Computer Music Conference, San Francisco, USA, 1996. [8] Teruggi, D. Electroacoustic preservation projects: how to move forward Organised Sound 9(1): 55-62 Cambridge University Press. 2004. [9] Canazza, S., Vidolin, A. Preserving Electroacoustic Music Journal of New Music Research, 3 0:4, pp 289-293. Swets & Zeitlinger, 2001. [10]Davies, H. The Preservation of Electronic Musical Instruments Journal of New Music Research, 30:4, pp 295-302. Swets & Zeitlinger 2001.