"Craft is what enables you to be successful
when you're not inspired." - Brian Eno

The Wire MAY 2001 - by David Toop


Contemporary music's biggest challenge is to liberate sound from the tyranny of the composer. David Toop tracks the escape routes being opened up via digital media, natural processes and evolutionary systems, encountering Oval's Markus Popp, Brian Eno, Autechre, Evan Parker, field recordist Steven Feld, multimedia artist Yoshio Machida and more.

In 1986, Jae-eun Choi, a Korean artist and film maker, initiated a series of experiments that she calls the World Underground Project. She buried sheets of Japanese paper in the soil of eleven locations around the world. The first pieces were excavated from the site in Kyong-Ju, Korea, after four years. Others, including those buried sites in Kenya, France and Italy, were still underground in 1998. Japanese paper begins with a strong character, before a single mark is made on, or into its surface. The absorbency and texture encourages accident and generates unpredictability. Those sheets that were excavated had been transformed by the years of their interment into gorgeous maps of organic growth.

According to Chinese myth, wrote Akira Asada in a commentary on the World Underground Project, in an attempt literally to give the featureless face of Chaos eyes and a nose, a new hole was dug into its face each day, until finally, when its features were completed, Chaos died. In order to capture Chaos, human beings cannot inscribe order into it. Not to represent it ourselves, but to let it represent itself naturally. In this way, Jae-eun Choi is attempting to capture living chaos.

Now think about Japanese paper in relation to compact discs. Unless the silver disc malfunctions or aborts, the promise of this carrier is to remain true to an original state throughout its so-called life. False optimism, no doubt, but aside from the occasions when they go drastically wrong, CDs don't exhibit the slight variations in playback sound and gradual deteriorations and fluctuations that characterise vinyl and tape. A CD is more or less a dead thing, or seems that way until it really dies.

At the polar opposite of that inertia is Christian Marclay's Record Without A Cover. Marclay's instruction in how to initiate the process of Record Without A Cover was embossed on the surface of the vinyl: Do not store in a protective package. I've had mine since the mid-'80s. Two years ago I used to lay it on a pile of twelve-inch singles by the window. Heat absorbent black vinyl, it made an attractively warm bulls-eye on which our cat would sit and gaze out of the window at birds in the cherry tree. A lot of unmentionable stuff got embedded into the grooves through that particular example of functionality, and when she was out trying to catch those birds, sunshine warped the disc into a picturesque wave. And then there's the dust, collecting on the record, as a record of my ambivalent attitude to order.

Just from a simple instruction, a supposedly 'final' artefact is transformed into an ongoing musical piece that the initiator cannot control. Like an awful lot of music enthusiasts, in my own house I'm vanishing into a vast housing estate of miniature tower blocks built from CDs. The more oppressive this static, one-sided arrangement seems to become, the more I'm interested in the idea of a music that can generate itself over time, giving itself up to the user in the way that Jae-eun Choi's Japanese paper surrenders to a colony of micro-organisms under the earth.

The notion of music that can be generated by an instruction or set of rules is not particular to our time. Within their description of Vox Populi, an interactive evolutionary system for algorithmic music composition, four Brazilian researchers give the example of Guido d'Arezzo, an Italian monk who, in 1026, resorted to using a number of simpler rules to map liturgical texts in Gregorian chants due to the overwhelming number of orders he received for his compositions. Guido d'Arezzo believed that his method of finding what he called an unknown melody was given to him by God, so forestalling any impulse to own, copyright and police the system. Instead, he offered his eleventh-century shareware in a spirit of generosity, offering apologies for his unavoidable absence but adding the caveat that his mono-chord based method could help beginners but was no substitute for proper music study.

What I wanted to provide was to suggest some additional criteria so music has to be discussed in terms of user interface design, interactivity or on-screen editing, and not so much discussed in terms of frequency or music as we know it. It's not interesting at all for me to discuss music in terms of frequency. I see it as my obligation to provide mildly innovative multimedia authoring.
- Markus Popp / Oval

An email from Richard Ross, programmer for Markus Popp's Oval Process, asks me a question: I was wondering what constituted generative music, and were computers necessary? I came to the conclusion, he writes from California, that if you dispensed with computers as a component of it, then things like wind-chimes and Aeolian harps might arguably fall into that camp. Other possibilities might be Cage's Imaginary Landscape No. 4 as a live performance. If generative music is music created on-the-fly, by some king of rule-based system, then these things follow very loose sort of rules but rules nonetheless.

In issue five of Musics magazine, published in 1976, sound sculptor Max Eastley wrote a short history of Aeolian harps, including the story of St Dunstan, who narrowly avoided incineration at the stake in the Middle Ages for the suspiciously demonic crime of making a harp that played by itself. Eastley also related the interesting case of Ichabod Angus Mackenzie, a sculptor and musician who produced fifty-three wind sound sculptures in 1934. During an interview he was asked if it disturbed him to leave his instruments performing alone without a human audience, Eastley wrote. He replied, 'That's up to humans. They're never without an audience.'

This raises some of the core issues challenged by twentieth-century music, and twentieth-century thought in general: the relationship of the composer to the audience, for example, or the use of chance and accident in the creation of music; the construction of feedback systems or self-generating and adaptive mechanisms that shape sound; the exertion or abdication of control of a musical result; the modelling of music based on ecosystems and similar complex environments and the setting in motion of events that question the definition of music as a cultural production distinguished from noise or unorganised sound by human agency and intentionality.

In the twenty-first century, such ideas have been expanded dramatically by the evolution of the Internet, in itself a self-propagating Web lacking any central control. Sound Drifting was a large scale generative sound installation by Collin Fallows and Heidi Grundmann for the Ars Electronica 99 festival. A web of sub-projects, sourced from six different countries, could be heard simultaneously and continuously, either onsite in Linz, Austria, online as a virtual installation and on air via Austrian National Radio. The sub-projects included Seppo Gründler's Dune And Redundancy, a room in which sand was blown by ventilators and vibrated by loudspeakers; Tim Cole's Intermorphic Koan^Oasis composed in real time by his generative Sseyo Koan system; and the orchestrated sounds of the air supply, the extraction system and the chilled water circulation pumps of the plant room of the Liverpool Institute for Performing Arts, recorded on microphones set up by Colin Fallows in an arrangement that replicated the recording method used by the sound engineer with The Liverpool Philharmonic Orchestra.

More recently, the introduction to Sound Drifting explained, there has been a growing interest in generative systems by artists working with the Internet, especially using sound, but increasingly with the appropriation of games software, search engines and so forth. Some of this work is highly critical of the ubiquity and unseeming power of generative systems in modern decision-making. But the most conspicuous cultural use of generative systems has been in the field of music - which means that the word 'generative', when used in relation to sound, usually causes people to think of music. However, although some music drifted in, Sound Drifting was not about 'music' - nor was it conceived as a concert hall, showcase or gallery space for the works of individual artists. Sound Drifting was about networking, communication and collaboration; about control-sharing between artists, users and machines; about letting go of one's own art and making ecological use of existing things; about listening to the world without adding to it; about the different concepts of duration and evolving processes at work in the material and immaterial realities of which we are part; about the aesthetics of different but connectable sounds, images, texts appearing on line - on air - on site as fugitive interfaces to a complex, invisible and not yet properly understood system of data processing.

In March this year, Brian Eno gave a lecture at the ICA in London, linking his ideas on generative music with the model of David Conway's Game Of Life. Conway, a Cambridge mathematician, invented Life as a cellular automaton, a game regulated by three logical rules: (1) Every counter with two or three neighbouring counters survives to the next generation (i.e. the next move). (2) Every counter with zero to one neighbours 'dies' (of loneliness), and every counter with four or more neighbours dies (of overcrowding). (3) Every empty cell with exactly three neighbouring occupied cells gives birth to a new counter. With these simple rules of birth, survival and death, Paul Davies wrote in God And The New Physics, Conway and his colleagues have discovered the most astonishing richness and variety in the evolution of certain counter configurations. In other words, out of a set of very basic conditions, or limitations, surprising events will emerge.

A week after this lecture, sitting in a patch of sunlight outside his studio, speaking on his mobile, Brian Eno talks about connections between that proposition, developed from ideas investigated by mathematicians such as John Von Neumann and Stanislas Ulam, and the compositions that first sparked his own interest in generative music. I think the Steve Reich pieces and Terry Riley's In C, he says. I would call those the predecessors of this. I would say anything where the composer doesn't specify a thing from the top down. What I think is different about generative music is that instead of giving a set of detailed instructions about how to make something, what you do instead is give a set of conditions by which something will come into existence.

The Steve Reich pieces he refers to are the early voice works for tape - It's Gonna Rain and Come Out - both of which explore the strange accretion of phenomena that occurs when two identical tape loops play in sync but then run progressively out of phase due to slight variations in motor speed in the tape machines. I thought the economy of them was so stunning, says Eno. There's so little there. The complexity of the piece appears from nowhere. You think, my God, it's so elegant to make something like that. Of course, I was hearing this at the time when twenty-four-track recording had appeared and people were making huge, vast, heavy, soggy pieces of music with no economy whatsoever, Suddenly to hear this Steve Reich piece, which I thought was the most beautiful listening experience, and to realise that it was made from just a few molecules of sound. That really impressed me.

As a member of The Scratch Orchestra, he was also impressed by Paragraph 7 of Cornelius Cardew's composition, The Great Learning. Written in 1969, this piece was described by Michael Nyman in his 1974 book Experimental Music as an example, alongside specific compositions by Christian Wolff (Burdocks), Frederic Rzewski (Spacecraft) and Alvin Lucier (Vespers), of Contextual processes [which] are concerned with actions dependent on unpredictable conditions and on variables which arise from within the musical continuity. Using a Confucian text, Cardew gave a few rules for an ensemble of any number of untrained voices. Each singer chooses his or her own note, then sings the phrases or individual words from the text for the length of a breath. For each subsequent line choose a note that you can hear being sung by a colleague, Cardew instructed. It may be necessary to move to within earshot of certain notes. There are a few other rules, but that, in essence, is the piece.

In Generating And Organising Variety In The Arts, an article written for Studio International in 1976 (an 'Art & Experimental Music' issue, guest-edited by Michael Nyman), Brian Eno wrote an analysis of the way in which the 'variety reducing' clauses in the score suggest that the piece might thin out to a small stock of random notes, yet in practice, variety is encouraged by a set of 'accidents'. These include the 'unreliability' of a group of singers of mixed ability, transposition of notes, the occurrence of beat frequencies, and the resonant frequency of the room in which the performance takes place. The composer, instead of ignoring or subduing the variety generated in performance, he wrote, has constructed the piece so that this variety is really the substance of the music. One example of how Cardew's contextual process could result in quite breathtakingly beautiful music can be heard on Organ Of Corti's recent reissue of The Great Learning.

A day after our first chat about this subject (though in retrospect, all our conversations over the years seem to have been about this subject), Eno comes back to me with an aphorism: Generative music is like trying to create a seed, as opposed to classical composition which is like trying to engineer a tree. Gardening and engineering are key metaphors. I think one of the changes of our consciousness of how things come into being, of how things are made and how they work, he says, is the change from an engineering paradigm, which is to say a design paradigm, to a biological paradigm, which is an evolutionary one. In lots and lots of areas now, people say, How do you create the conditions at the bottom to allow the growth of the things you want to happen? So a lot of the generative music thing is much more like gardening. When you make a garden, of course you choose some of the things you put in, and of course you have some degree of control over what the thing will be like, but you never know precisely. That's the wonderful thing about gardening. It responds to conditions during its growth and it changes and it's different every year.

Even professional gardeners are on to generative music. A recent article in The Guardian exhorted amateurs to throw out their wind-chimes and water features and plant sound-rich ornamental grasses such as arundo donax, evening primrose which opens its blooms with an audible pop, or common broom, with its exploding seed pods. Just as Cornelius Cardew's method for The Great Learning produced an implicit social model, this approach to constructing a sound environment connects music to ecology and, perhaps unwittingly, connects gardening to the chance methods and theories of John Cage (who was in turn strongly impressed by the Zen garden created by the painter and gardener Soami at Ryoanji Temple, Kyoto).

In the late 1960s and early 1970s, collaborative group music was one of the most powerful available tools for experimenting with new models of society, forms through which individual expression might thrive within collective organisations of inclinations ranging from hippy anarchism to Maoism, along with new ideas about the importance of ecology in the relationship between humans and the environment (in the early 200s, the equivalent seems to be events like Sounds Drifting, the New York-based Unity Gain, described by organiser David Linton as a systematic JAM of 'systems', or the California-based network computer group known as The Hub). At roughly the same time that Brian Eno was experiencing The Scratch Orchestra first hand, I was taking part in performances of Eddie Prévost's Spirals, a composition for large ensemble that attempted to apply some of the psychological characteristics of small group improvisation to very large groupings, as well as attending weekly workshops held by drummer and improvising pioneer John Stevens.

During the period in the late 1960s when he shifted his group, Spontaneous Music Ensemble, away from the compositional models of Ornette Coleman, Eric Dolphy and George Russell into uncharted territory, Stevens began to formulate pieces that could help musicians who were new to this way of playing (and that included just about everybody back then). Click Piece, for example, was a simple instruction to play the shortest sound possible on your instrument. The difficulty of this varied from instrument to instrument, player to player, and quite a considerable amount of concentration was needed to pare each sound down to its smallest event and keep it there. As a player, you became aware of the way in which a group sound was emerging only after sometime had elapsed. The paradox lay in the way that a complex group interaction, quite ravishing to listen to on occasions, could emerge from individual self-absorption. The piece seemed to develop with a mind of its own and almost as a by-product, the basic lessons of improvisation - how to listen and how to respond - could be learned through a careful enactment of the instructions.

Evan Parker remembers the way in which his duo with John Stevens pushed this atomistic way of playing to a limit. The moments of interaction got shorter and shorter, he says. You couldn't go any further than that. So a method that stimulated considerable variety in a large group comprising players of mixed ability and experience, quickly became an unproductive limitation for a duo of two well-matched, skillful and confident musicians.

I put it to Parker that Brian Eno's gardening analogy might be applied to his solo playing for soprano saxophone, along with many of the theories of webs, swarms and emergent phenomena found in books such as Kevin Kelly's Out Of Control: The New Biology Of Machines. We all are delighted if we can find some way of talking about something that is very difficult to talk about, he admits. Fractal maths and chaos theory are very useful for talking about the solo playing, though of course the number of calculations involved to arrive at a fractal diagram or drawing is probably a magnitude of millions different from the number of calculations involved in me playing a solo. But in the the sense that the whole design is built up from one calculation, the output of which becomes the input for the next calculation, there is in some way a connection with the way I work in the solo thing. I set up loops of stuff and then observe the loop and listen closely to the loop and say, ah, now I'll emphasise that note, or now I'll bring out that difference tone, or I'll try and put something underneath it in relation to that or on top. Gradually the centre of attention in the loop shifts somewhere else. The loop is suddenly a different loop. It's something that's still bearing fruit for me. I'm not saying that's exclusively the method I'm using in solo playing but it's the core method.

This sets up a complex feedback system between the saxophone and independently functioning regions of his own distributed consciousness, enabled by Parker's circular breathing and his knowledge of the overtones available through advanced fingering techniques. Absolutely, he agrees. It's the key notion of the twentieth-century. I'm not an expert on cybernetics but bringing an ability to generalise about feedback is a twentieth-century phenomenon. Before that there were specific applications but I don't think there was a general awareness of how many control systems can be analysed in terms of the feedback between inputs and outputs. It's certainly high on my list of analytical tools.

In 1966 and 1967, Pauline Oliveros produced two tape pieces - Alien Bog and Beautiful Soop - using Don Buchla's 'Buchla Box' 100 Series synthesizer and her own tape delay system. Working at the Tape Music Center at Mills College in Oakland, she had been influenced by the sounds of frogs living in the pond outside her window at Mills. Tape delay systems were means of creating unpredictable variety in music. Terry Riley's system, the time lag accumulator, was a technological equivalent of the feedback system later developed by Evan Parker and one of the inspirations behind Brian Eno's use of tape loops.

For Eno, the system that allowed him to create Discreet Music was fine, except it was limited to the length of a vinyl LP. All of those phase systems, he says, they're theoretically endless, generating new stuff as they go, new combinations. I always wanted that kind of music - not only Discreet Music but the things that followed it like Music For Airports - to be endless pieces. I saw them more like paintings, just things that stayed in place, than compositions, things that had a structure to them. I was always looking for creating, not a recording of the results of the generative process, but creating a generative machine itself. This lead to the use of Tim Cole's Koan software, a program he had hunted for in research centres in Stanford and Palo Alto but had failed to find.

The desire to make a music that exists in a state of being, theoretically without beginning or end, is paralleled by Evan Parker's interest in relatively long forms and their relationship to improvising. What happens when you work with the longest elements? asks Parker. Maybe you're not improvising anymore. You're just remembering. That dialectic, at the core of his music, contributes to the subjective impression in the listener that something is alive and growing, like a time-lapse photograph of plant growth, one of the creatures grown in the 'garden of unearthly delights' by William Latham's computational breeding program or the volatile communities generated by Conway's Game Of Life.

The observation of nature, either through bio-acoustic study, environmental sound recording or ecology, has led some musicians to the creation of emergent systems based on non-human source material. Mamoru Fujieda, for example, wired up plants using a Plantron interface devised by botanist Yuji Dogane. The data collected by electrodes recording changes to the surface electric potential of the plant leaves was converted to MIDI and then transformed into melodic patterns using MAX, the graphical music programming environment developed by Miller Puckette and other authors at IRCAM in 1986.

While Fujieda translates plant activity firmly into the human sphere, Michael Prime's work is more of an intuitive mapping of the interface between humans and non-human species. As in Fujieda's Pattern Of Plants, Prime, member of London Improv group Morphogenesis, uses a bio-activity translator. This controls oscillators which are used as sound sources. His L-Fields, a work for hallucinogenic plants, is named after studies in voltage potential made in the 1930s and '40s by a Yale scientist, Dr Harold S. Burr. According to Prime, speaking in an interview with Francois Couture of Québec radio: He had several local trees connected to voltage meters for a period of years, ,and discovered that their voltage potentials varied not only with periods of light and dark, but also with the cycles of the moon, magnetic storms and sunspots. The fields of humans varied not just with these natural rhythms, but also according to mental state, health, presence of cancer, et-cetera. He finally postulated that these fields were not just a pattern produced by living organisms, but were also the morphogenic blueprint that controlled their development.

Prime describes his use of a bio-activity translator as occupying a kind of hinterland between composition, improvisation and process/generative music. One of his inspirations is the musical use of human brain-waves explored by Alvin Lucier, Richard Teitelbaum and David Rosenboom in the late 1960s, another version of generative music that relates to speculations made by Evan Parker about the role of the left brain/right brain activity during his solo performances. In a sense, Prime simply plugs into biological activity and during the period in which he is plugged in, the unpredictable and inevitably mysterious signals given off by plants both create and are folded into Prime's soundscape. The intricacy and alien beauty of bio-acoustic feedback systems such as the hunting relationship between bats and moths - the bat tracking moths with ultrasonic pulses, the moths using using evasive flying measures whenever they hear ultrasound - can suggest new ways of 'growing' music. Pieces like Chaos & The Emergent Mind Of The Pond, created by sound recordist and composer David Dunn in 1990, are illustrations of the way in which 'shaped' soundscapes can become a category of found art that links generative work of all kinds.

In his book, Why Do Whales And Children Sing?, Dunn quotes the anthropologist and musician Steven Feld, whose research and recording among the Kaluli people of Papua Nw Guinea and the rainforest in which they live has drawn new maps of the relationship between favoured sound patterns, aesthetic preferences and social relations. Steven Feld describes the New Guinea rainforest as a world of co-ordinated alarm clocks, writes Dunn, an intersection of millions of simultaneous cycles all refusing to ever start or stop at the same point. In such books as Music Grooves, co-authored with Charles Keil, Feld has written extensively about valued sonic qualities among the Kaluli, including ...interaction of patterned and random sounds; playful accelerations, lengthenings and shortenings; and the fission and fusion of sound shapes and phrases into what electro-acoustic composer Edgard Varése called the 'shingling' of sound layers across pitch space.

Feld's observation of simultaneous cycles working out of phase, or the Kaluli love of in-sync, out-of-phase patterning recalls Brian Eno's enthusiasm for Terry Riley's In C, It's Gonna Rain and Paragraph 7 of The Great Learning. One of the most enthralling examples of this phenomenon can be heard when large groups of frogs are calling, each frog responding to another, calls sometimes falling in perfect synchronisation, moving in and out of phase, then falling suddenly silent for reasons a human can't divine. David Dunn has extrapolated from his recordings of this emergent mind to develop a series of real-time multi-channel electro-acoustic performances and installations for live computers.

They explore the global behaviour of hyper-chaotic analogue circuits modelled in the digital domain, he tells me, via email from Mexico. These circuits exhibit an immense range of sonic behaviour, all generated from the equivalent of three sine-wave oscillators linked together in a feedback path that exhibits two of the essential traits of a chaotic system, non-linearity and high sensitivity to initial conditions. The emergent complexity results from the dynamical attributes of cross-coupled chaotic states interacting in a multi-dimensional phase space. My role as composer/performer of this 'chaos' instrument is to explore various regions of these behaviours in a manner analogous to the exploration of a physical terrain. While I can influence the complex sonic behaviours, I cannot control them beyond a certain level of mere perturbation, the amount of which is constantly changing. The experience is often tantamount to surfing the edge of a tide of sound that has its own intrinsic momentum.

The compositions are to be regarded as improvisatory in structure but based upon a prescribed set of zones where particular chaotic behaviours reside. The opening and closing of virtual switches determines various combinations of structural coupling between distinct chaotic circuits, allowing different self-organising behaviours to arise. The composition is a charting of transitions between these different zones that arise from a fundamental generative structure and its behavioural diversity, much like a genetic code.

My fascination for these sounds has less to do with the underlying mathematics, or the current fashion of applying complexity science to music, than with the similarity these sounds have to natural sound environments where the same dynamical properties might be operating. These sounds excite me because they are so physically reminiscent of the global sound behaviours that emerge from natural habitats such as swamps, forests and oceans.

My main question on generative music is: can we trust machines to create for us? asks David Rothenburg, musician and author of Hand's End: Technology And The Limits Of Nature. The life's work of John Cage could be interpreted as that question almost in reverse: can we trust humans to create music? Through the influence of books as much as anything else - the oracular hexagrams of the the I Ching, James Joyce's Finnegan's Wake and the writings of Gertrude Stein - Cage arrived at The Music Of Changes in 1951, a composition he described in Musicage, his conversations with Joan Retallack, as where the process of composing was changed from making choices to asking questions.

Although Cage's ghost is present almost anywhere we care to look, his philosophy of non-intentionality has become a resource, rather than a way of lie, for many musicians currently working with electronic media. As a member of the think-tank (also including architect Paul Shepheard, landscape architect Georgina Livingston, digital sound artist Joel Ryan and Brian Eno) that offered guidance to Jem Finer in his development of the Longplayer project, I remember a phase during which Finer considered using a segment of Cage's prepared piano music as the source material to feed through SuperCollider, the real-time sound synthesis program developed by James McCartney. The intention of Longplayer was to generate a piece of music that would last for a thousand years, using SuperCollider's capacity to loop small segments of music and gradually move the start point of the loop, with each new loop applying the same process to itself to create a nest of loops, all working within the differing boundaries of its parent loop to create constant evolution. Fascinating, but though informed by Cage, perhaps not a particularly Cageian way to compose.

Yoshio Machida is a musician and visual artist living in Tokyo. His 1999 album, Hypernatural, was supplemented by the release of two floppy disks that remixed the CD tracks with Sseyo's Koan program. I like to enjoy 'texture' and I was looking for some generator that makes the texture of a soundscape, Machida writes as an explanation of how he came to use Koan. I like to see the changing view of a landscape through a window of a train or an airplane. Especially, the view from an airplane is wonderful. At first I was using ALPS. This is Japanese Mac-based software. It's very simple but this can generate some phrases by itself. Then I knew Koan through some info about Eno. I thought, 'Oh, that's what I want!'. These softwares generate some not-perfect-things by themselves.

These textures are unique for me. I want to get some dynamics with not only my own power but also the other power, and I want to keep a balance of instinct and thought. That's my way. Hypernatural sounds aren't always perfectly in chaos. It's an amorphous thing, not a crystallised thing. I think what is important is the thing between sound and listener, not only sound. So this software helps me to make 'amorphous things'. For me, this is similar to field recording sounds and some analogue sounds. John Cage said he can enjoy listening to all sounds, except the sound with someone's intention. I feel so, too.

Issues of intentionality, linearity and the model of active composer and passive listener are being challenged by software and software users, yet held in place by the dominant carrier of music, the compact disc. Our minds have become nodes in the expanding space of the Internet, wrote Kim Cascone for the liner notes to Selected Random Works, released on Ritornell, connecting freely with other nodes in a rhizomatic manner. Comparing this fluidic, smooth space with the linear space of the audio compact disc, we find that a linear model has been imposed onto an inherently non-linear medium.

Live streaming, installations, MIDI files and the release of authored software, rather than finished product, offer ways around this contradiction, though the effect at the moment can feel and sound like the aimless exploration of a huge choice of possibilities, something like the experiments of the 1960s when the excitements of process and change could obscure the imperatives of making music that was worth a second listen. The limitation of the range of possible parameters changes focus within the music, giving a distinct character, says Olivier Alary, whose Rephlex release as Ensemble, Sketch Proposals, was one of the most promising albums of last year. The output of such a system can be familiar and different at the same time, but never repetitive. It could be compared to the weather or the movement of clouds in the sky.

There is a significant difference between software programs such as Logic Audio or Cubase, basically emulations of the recording studio, and more open applications such as MAX/MSP, Cloud Generator developed by Chris Roads and John Alexander, or interesting curiosities such as Akira Rabelais's Argeïphontes Lyre (elliptically explained to me by Rabelais by means of a lengthy chunk of Greek mythology). Composers who have devoted a lifetime to compositional methods that go beyond the customary means of committing sound to tape, its equivalents or emulations, are increasingly important in this shifting field: Iannis Xenakis, for example, for his theory of stochastic processes, derived from mathematician Jacques Bernouilli's 'law of large numbers' or the cybernetic and entropic compositions of Roland Kayn.

Their music raises questions of authority, though not necessarily the more fashionable subject of authorship. For Autechre, digital music offers greater possibilities for making their music sound the way they want it. You probably know already that sound design using digital gear takes more maths than using analogue gear, they explain via email. You have to put the right numbers in to get the numbers you need out.. Computers just speed up the process for us. They make the maths easier. They also paradoxically provide more opportunity to claim responsibility for the resulting music, as the user has to define so much in order to get personally satisfying results, depending how much they like to get their hands dirty. This can be applied to sequencing and sound design, though thanks to clever high level developing environments the boundaries between the two stages are disappearing. The way computers can help make things easier and faster sometimes means we're pursuing ideas that maybe would have been abandoned previously. Keeping the vibe has always been important and rather than sitting with a calculator working out exactly when two beats are going to realign, we can get on with being inspired. Also, being limited is more interesting when you've set the limits yourselves - or for each other.

For Markus Popp of Oval, one of the most important factors in this recent trajectory is the presentation of his Oval Process software, developed with Richard Ross, as an interactive installation object. That is this tangible interface, he says, speaking from his studio in Berlin, declaring the interface public domain and just handing it over to the audience or whoever is present at the given time of the exhibition or wherever the unit is on display. This is one aspect of it, and the other aspect, which might even be considered the stronger statement is, of course, the available audio content which is on my CD, which is a quite vigourous statement against the typical productivity work flow in music.

He describes his most recent CDs - last year's Ovalprocess and now Commers - as the tangible front end of an attempt to introduce an alternative rhetoric to the production of electronic music. At the same time, Oval Process is a statement to encourage non-expert audiences. It is immediately visible for the user whether it's cool or crap, he says. I never considered this would be ground-breaking in the sense that it would hand over the interface, he continues. It's just a simple gesture. It's just the defining principle of something like an info terminal at the airport. You hand over the interface, otherwise the info terminal wouldn't be useable. A very simple question: how would Oval be like an info terminal? I would happily take the risk of making Oval Process just this box.

Like Japanese paper buried underground, the final organisation of music is relinquished by its maker, though the elements remain intact. Popp seems to interpret the current situation in music as a moment for making statements that jump out of established historical frameworks, for when people are confronted my music designed to grow and evolve beyond the composer's intentions or even understanding, the old science fiction anxieties still recur. I understand that people are easily led when all they see is this designed object which provides the tangible means of approaching this type of Oval music in this very strict and impersonal way, he says, just being a sound installation doesn't require any assistance and is completely interactive. So, of course, people are easily led to assume that this is the Kraftwerk approach driven to the extreme, just placing this object somewhere which would eventually replace myself as the protagonist in Oval music which has never been intended. It's very simple, he concludes. I just want to create a contemporary statement.