pSY and The Copenhagen Interpretation: Software or Composition?

Abstract

To what degree can the software used in creating music be considered responsible for the final result? Clearly the answer depends on the type of software, the type of music, the composer and other factors, but should software used in the creation of music be credited with the piece? If I use an audio editor to manipulate a sound in a particular way, is not the programmer of the material at least partially responsible for it too? Can the software 'tool' be entirely without 'colour'? And what about software that contains 'actual' compositional procedures?

...a nauseating musical experience, but one not without interest...Is the computer composing? The question is best unasked, but it cannot be completely ignored. An answer is difficult to provide. The algorithms are deterministic, simple and understandable. No complicated or hard-to-understand computations are involved; no "learning" programmes are used...the machine functions in a perfectly mechanical and straightforward manner...

M Matthews and L Rosler, "A Graphical Language for Computer Sounds" in H von Foerster and J W Beauchamps, eds., Music by Computers, p96, New York: John Wiley, 1969

pSY, The Copenhagen Interpretation and the Perception of Music

Abstract

A rough version of a paper given at the APU Music Research Conference 1999.

Programme Note to the first performance, 20th June 1999, West Road Concert Hall, Cambridge, UK

Describing who or what writes and/or performs The Copenhagen Interpretation is a surprisingly tricky business. I sometimes feel that this is deliberate on my part - a deliberately disingenuous attempt at protecting myself from true responsibility, perhaps. I at least know that whatever the case, pSY, the programme which will do the creating, took far more time and trouble to write than any other note-based composition for which I have been openly responsible. I should say that I have been trained and taught as a 'standard', dot based composer - at least one of the reasons for undertaking work on pSY was that I could, in the process, learn more about computer programming. At least in this regard, the project has so far been successful.

The Copenhagen Interpretation is one manifestation or instance of a group of settings controlling two related computer programmes. The collection of programmes is, collectively, called pSY and this was written by me. pSY is MIDI based. Essentially, at present, one programme (the Score) controls which settings are applied to the two mentioned above, and of those two, one controls the settings of a Yamaha SY77/99 synthesiser via system exclusive messages and the other controls general MIDI messages, that is, note on, note off, velocity, pitch, etc.

pSY itself is therefore the organiser and interpreter of these settings. Currently, while the experiment progresses, pSY is a tamed, inhuman beast, chained and shackled by my programming. The next stage is to plan the gradual release of pSY into the environment increasingly under its own control. Ultimately, the idea is to produce a programme which produces, as independently as possible, pieces of music which satisfy and speak to us in the same way that any live human performer/composer does. What does this mean? That the music should behave according to certain levels of predictability while never becoming dull. What would it tell us? To some, nothing. To others, I feel that - given full liberty to be poetic and anthropomorphic - these are how we might imagine communications from an alien life-form to be like if we pointed an antenna into space and turned up the sound. It is alien, almost mechanoid, but above all, devoid of human interference. (This is not true, of course - it's just bad poetry - I have censored the programme shamelessly in order to keep its utterances within the bounds of human acceptability!)

The basic idea behind pSY is quite straightforward and is itself based on another programme of mine, Arpeggiator, which, taking an arbitrary starting point, creates streams of general MIDI information through mainly probabilistic algorithms. pSY does the same, but utilises system exclusive information to control the sounds that the SY makes itself.

The surprising thing about the original versions of pSY, originally written for Macintosh computers and auditioned as an audio installation at the 1998 CDAF, was the amount of the output that 'communicated'. In general, although not exclusively, the programme works best when between three and five computers are operating the same number of synthesisers and where the settings involved are similar, at least at coordinating moments. In this alignment, under certain cirumstances and to certain people, the output seems to 'speak' or 'perform', even given the fact that most 'decisions' made by the programme are based on complex webs of probabilities.

I would like to think that pSY and to some extent the 'piece' itself is in a constant state of development. In this way, I would like to compare it to a human performer or composer - it could be argued that the process of learning and interpreting any given score is a cumulative process, where there is no 'definitive' performance, rather a continuous process of refinement of experience and understanding. Only after the performer has died or retired is it really possible to talk of 'definitive' performances. Currently, the programme's form of 'learning' is through me - I decide whether the programme's reaction to certain settings is more or less appropriate and I change the programme (or not) in accordance with my own criteria. Ultimately, I would like to see what happens if the programme were enabled to do this itself. Whether the results of such an experiment would be 'acceptable' to human ears is another matter.

For me, the development of pSY has raised a number of questions about the nature of musical performance, the nature of composition, the nature of the human-machine interface, the nature of artificial intelligence and ultimately the nature of perception. I am currently investigating these.

R Hoadley June 1999

No Javascript • Last Modified: 18/01/02 • All original material on this website copyright Richard Hoadley 1999-2002
Make a comment or report a fault • Site questionnaire