Following last month's report on Digicon '85, we bring you news on another event that took place in Vancouver just a couple of days later — the International Computer Music Conference.

'They're living in a world of their own!' said one perplexed Digicon survivor who had decided to stay on for the International Computer Music Conference (ICMC), and had just sat through a particularly taxing paper entitled 'Semiotics and Computer Music Composition'.

Whilst it's true that a double PhD in Digital Signal Processing and Computer Musicology would have gone some way toward helping many visitors decipher some of the presentations at this, the second of the two Vancouver conferences, it's also true to say that, if you persevered, there was much of interest mixed in between the impenetrable algorithms and incessant verbiage.

If anything, the Vancouver ICMC was one of the most productive yet, partly because it came at a time when there was a definite movement of academics away from the university research labs and into commercial positions. Part of the reason for this is shortage of research funds, but there are now some lucrative job offers from commercial concerns, who are becoming increasingly aware that 'top brains' are essential to have onboard if they are to keep up with competition from other firms. At Digicon, both Roger Linn and Bob Moog acknowledged the necessity for employing highest-level technologists in the development of the Linn 9000 and Kurzweil 250 respectively.

And it isn't just a case of academics going into the marketplace. There's also been a substantial increase in the numbers of engineers and designers from commercial companies attending the conference — so the exchange has worked both ways. This year, companies such as Eventide, E-mu, Apple, Roland and Yamaha successfully infiltrated what has been, in the past, an almost exclusively academic forum.

One of the effects of this new dialogue is that even the established 'centres of excellence' for large-scale computer music research and development such as IRCAM, MIT and Stanford have caught the commercial bug, and are getting more involved with smaller, streamlined systems, often scaled-down versions of their larger installations. IRCAM, for instance, is making its CHANTES and FORMES programs available as micro-based packages, and a lot of research has been going on with relatively run-of-the-mill machines like Macintosh computers and DX7s.

Many of the papers at ICMC dealt specifically with commercial or potentially commercial systems. In the case of the latter, it seems the conference is becoming very much a meeting place for people to negotiate ways of crossing the threshold from theory to reality.

One such paper was presented by Mark Lentczer of Apple. He described a system called the Sound Kit, which is essentially a software package for manipulating and editing several individual sampled sounds and assembling them into a composite musical structure. 'New' terms such as paste, copy, cut, zoom, merge, undo, and redo are used to describe the processes used in the system, all of which are undertaken on the Macintosh via its powerful graphics capabilities. The fact that Lentczer didn't talk about any hardware suggests that the system is yet to be formed into a commercially viable product. If this is indeed the case, he was certainly in the company of the right people: it shouldn't be too long before a rather interesting multi-sampling sound-manipulator becomes available for the Macintosh. And when it is available, it should bring with it the possibility of digital versions of musique concrete-type tape manipulation techniques, where several sampled sounds are processed and interrelated at the same time.

Another Mac-based system discussed at the conference was the Sound Designer package for the Emulator II, reviewed elsewhere this issue by Paul Wiffen. From what I've seen, I'd say the system amounts to an impressive low-cost Fairlight alternative: the paper describing it — presented by E-mu's Dana Massie — was little more than a cleverly camouflaged sales pitch. There were to be several more of these during the conference, and that's not a criticism, just a comment on the way things are going. If anything, I'd say the fact that designers of commercially-available systems have a forum in which to discuss them in detail to a technically-literate audience is A Good Thing.

Gesture mapping techniques and sensors for playing computers with expressive control were some of Vancouver ICMC's most dominant themes. Many of the ideas involving highly touch-sensitive interactive graphics surfaces and body sensors will, I'm sure, find their way into commercial products in the not-too-distant future. We'd already seen Bill Buxton's touch tablet system at Digicon, but at ICMC a rival system was presented by Eric Johnstone from McGill University. It's called the Rolky, and it consists of a glass surface illuminated by a cathode ray beam. When a finger touches the surface, the conditions for internal reflections of the rays are disturbed, causing a bright spot at the finger's location. This spot is recorded on a TV camera and digitised. It is then analysed to extract spot co-ordinates, size, angle, and elongations, all of which help to define how the surface was touched. Polytouch (ie. multi-spot analysis) is also possible, and this allows users to use the whole hand or both hands to 'touch a computer'. The system is said to have extremely high resolution, enabling the playing of a connected synthesiser with violin-type sensitivity. We shall see.

Another system providing information generated by finger/hand/arm gestures is 'The Hands', a product of Steim Studios in Holland. A paper describing it was presented at the conference by Aad Te Bokkel, a software engineer at Steim. 'The Hands' is a set of remote MIDI controllers consisting of two aluminium ergonomically-shaped plates, with sensors, potentiometers and switches strapped under the hands of a 'performer'. Analogue information generated by finger movements and changes in hand/arm positions is scanned by a microprocessor with an onboard MIDI encoding program. The MIDI codes that are generated can then be sent to any MIDI instrument. The user can define which movements control each of the MIDI data parameters, and the switches and pots can be programmed to act as remote controllers for synths or other MIDI devices connected to the system. They could, for instance, represent pitchbend and modulation wheel controls, preset voice switches, or even the status of a DX7's internal operators. All in all, an impressive and extremely versatile system for movement-generated music - but it isn't the only innovation the Dutch were showing at ICMC.

Steim are also working on a 'MIDI Assistant' intended to solve MIDI incompatibility problems and enable code reprogramming to change the meaning of MIDI — echoes of Digicon's reverse pitching and pitch-to-patching here, I'm afraid.

Another Steim engineer hy the name of Joe Ryan is busy modifying Yamaha's CX5 to enable microtonal intonation on each of the eight voices, so it seems Steim Studios are just about the busiest champions of small, affordable systems currently researching.

A couple of ICMC papers dealt with computer tracking of live instruments.

Barry Vercoe of MIT presented a paper called 'Training the Synthetic Performer', which described a system in which data representing score information is prepared and fed as control processes to a host processor.

Through pitch-detection devices, the processor can follow a live performance and then trigger the controls necessary for a synthetic accompaniment. Intelligent decisions can be made by the processor with regard to tempo and pitch errors by the live performer, and more interestingly, the system also has the ability to actually 'learn' during rehearsals to allow for a performer's particular phrasing or style in interpreting the score. Thus, it learns to alter its own score data to match the live performer's playing technique. Who said computers aren't friendly?

Roger Dannenberg also presented a paper on computer accompaniment, but concentrated on the problems of tracking polyphonic performances (Vercoe's system is essentially monophonic). Again, 'score' information must be fed in advance, and the system employs several matching algorithms after grouping notes into sets of events.

The natural extension of these separate systems would be a composite one incorporating Vercoe's learning capabilities and Dannenberg's polyphonic tracking, but nothing exists along these lines yet. And if you think all this sounds too scientific to have any relevance to music, some possible applications of computer tracking techniques are surprisingly close to home.

Take electro-acoustic concerts, where tapes and orchestra often co-exist in an uneasy 'fixed time'. A new situation could well arise in which the musicians and/or conductor can dictate the pace of a piece of music, and any required tape accompaniment would simply follow humbly along.

Then there's conventional rock music, where it should become possible for backing tapes and sequencers to follow key musicians, rather than the other way round.

As far as research is concerned, a lot of the ideas (or 'heuristics') for computer tracking have developed as a result of the influence Artificial Intelligence has had on the computer music community. Musicians have already benefitted from the way AI has been used in the research and development of commercial products such as the Kurzweil 250.

One area that AI features prominently is that of computer-aided musicology, in which styles and techniques of past composers are analysed in great detail and artificially-intelligent models are developed to enable generation of 'new pieces' in the style of those composers, many of whom are now long dead.

Now, some new music composers have 'borrowed' this idea and used it for their own devious ends. This borrowing usually involves putting a spanner in the works in one way or another, like introducing a random variable to upset a model or programming a carefully-engineered deconstruction. Letting the AI model add to its own knowledge base (ie. software writing its own software) is another interesting possibility. Whatever the nature of the 'spanner', its introduction can result in some sparklingly original and thought-provoking music. There were several examples of this at the conference — though I shan't name names in case the composers concerned object to my description of their compositional process...

All in all, the music at Vancouver ICMC reflected a noticeable change of direction, itself probably related to the developing academic-commercial technology dialogue. This dialogue is still not as widespread as many would like it to be, and I doubt any of the music performed at the conference will make the Top 40. But there was a marked tendency toward a more engaging, more accessible, and in some ways more physical music that contrasted starkly with the cerebral, esoteric style of so much early computer music.

As an example, composer Morton Subotnick, an electronic music pioneer with works such as 'Silver Apples of the Moon' under his belt, professed to being extremely optimistic and excited by current developments. He has recently been working exclusively with the Yamaha QX1/TX816 system, and his latest work was actually premiered at Digicon as an FM tape piece. Along with other composers such as David Keane and Bill Buxton, Subotnick claims to he heralding a new age for contemporary music, in which computer technology will break down the polarisation between academic and non-academic music. The intention, it seems, is to create something akin to a Digital Folk Art form which listener and composer unite in a shared experience. A little unrealistic, maybe, but wouldn't it be good if even a fraction of this barrier-breaking actually took place?

Of all the tape pieces played at ICMC, Paul Dolden's 'Veils' was particularly memorable. Ironically, this didn't actually use computers to generate any sound, but was based instead on computer music research into textural transformation. The sheer physical power of the piece's mass, textural movement through endless layered timbre manipulations was quite stunning. Most of it was assembled on a Tascam eight-track using carefully recorded real sounds, so it's the kind of thing almost anyone could do, no matter how ugly their bank balance.

Other interesting tape pieces included David Keane's 'Elektronies Mozaik', an engaging rhythmic romp with the DX7; Denis Smalley's impressively spatial 'Tides', which was divided into two performances, one out of doors, the other in; Paul Berg's wonderfully painful string quartet, which used distorted plucked string algorithms; and Barry Truax' 'Solar Eclipse'. Truax was, in fact, the host of the conference, and his piece had a quality of involvement and accessibility that was common to a lot of the conference's music.

The live music was less impressive. Several pieces used interesting 'mad-professor' invented instruments, and there was a lot of MIDI interfacing. But overall, it was a case of too much technique, not enough content.

Richard Boulanger's 'Book of Dreams' was a case in point. It utilised Max Matthews' electronic violin, with each string MIDI-linked to a separate synthesiser. It made some interesting noises, but not much else.

David Rosenboom's 'Zone of Influence' had more going for it: a relentless but enjoyable interaction between Buchla synth and percussionist. This was followed by an extra-curricular happening after the official conference had ended, in which Donald Buchla (who'd been present at some of ICMC but kept a low profile) performed on his own digital 400 synth together with Rosenboom on the Buchla Touche. The duo presented us with a piece that was as good as, if not better than, any of the live music at the conference itself, essentially a controlled improvisation full of interesting rhythmic cascades and unexpected (to the audience, if not to the performers themselves) twists and turns.

Rosenboom, incidentally, was the man responsible for 'Brain Music for John & Yoko', which was performed on them on American TV in 1972.

Your intrepid reporter was fortunate enough to meet up with Alex Douglas of CLEM contact list fame. As well as taking me up to see his extremely impressive 'nerve centre', he also invited me along to the Alien Soundtracks radio show he presents - along with others — every Saturday evening. Vancouver is blessed with several electronic music radio shows, a fact which, for a city no bigger than the likes of Manchester or Sheffield, certainly puts British endeavours to shame. Alien Soundtracks presents a good cross-section of mainly European (including several British) acts, and I found myself becoming extremely envious. Fancy going all the way to Vancouver to hear stuff that should be on the radio over here...

Next year's computer music conference is closer to home. It's to be held in The Hague, Netherlands, on October 20-24. Conference organisers are the Computer Music Association (CMA), who can be contacted at (Contact Details). Members receive a regular newsletter called Array that's currently undergoing the same transformation as the conference itself, ie. a growing interest in smaller, affordable systems and the inclusion of useful information on MIDI, complete with a MIDI agony column and an information contact list. Members also receive a 10% discount to the International MIDI Association.