Composing: How and Why? An E-Text


We have decided to formally launch the web presence of Tonality Systems Press by presenting an e-text about composing. We’re serialising it in installments as a blog, posting a chapter every Friday, after which it will then be released as a Kindle e-book and PDF.

This work has a lot to say about using the computer as an aid to what a composer does. It focuses not so much on the end product – the final score and the recording – but on how composed music comes into being. The computer is viewed as a mediator between what a composer imagines and how that imaginative idea becomes thematic material and a structure, and thence a composition.

You’ll discover ways that may enable you to work productively and imaginatively using the computer as an aid to composing. Most of it is not specific to any particular technology or software, but the code examples do make reference to the new and innovative parametric composing environment Opusmodus.

What’s presented here in terms of the basic concepts of computer-aided composing (CAC) is hardly new. Lejaren Hiller and Leonard Isaacson, a composer and industrial chemist, laid down principles and approaches to using computation as a composing aid some 50 years ago. Their book Experimental Music still speaks much good sense, introducing ways of handling musical events (rather than sonic objects) that are still very relevant today. Their focus, as composers on working with a computer, was an empirical one: they wanted to  compose a string quartet.

From The Illiac Quartet, Experiment 3.

There is no one type of composer. Everyone works differently, has different objectives and seeks different outcomes. We all take different pathways through the creative process. Some of us have been fortunate to have had guides and mentors who have shared aspects of their working techniques and musical literature that, to varying extents, have served as models or examples. But mostly, what we do is what we’ve learnt to do ourselves: our teachers have been critical friends, helping us produce our own work in the best way possible, so that it can be realised successfully by our performers and make its point with the listener. We also learn from what we hear and, if we’re fortunate, what we’ve able to see and study in notated scores.

So much of what composers do is wrapped up in sorting out what Paul Hindemith referred to as the conditions surrounding any performance. Conditions include a profile of the skill and style of our performers, the space of the performance, the requirement of the music’s duration, of appropriateness and rightness for performers and audience on a specific occasion and within a specific programme. These are the things a composer usually considers, part of that dreaming and imagining that will take the composer towards the first notes of a composition, and hopefully to the final bars.

Computer-assisted Composing offers a surprisingly varied cornucopia of ways to prepare, compose and produce compositions rendered into sound and score. Whilst the emphasis here is firmly on music for voices and instruments there are many strategies and ways of thinking that transcend styles and media. Composers working in music for games, installations and internet media could well find much of the text stimulating and useful.


The e-book text is in 25 sections. It’s not a ‘have to read from the start’ kind of text. It’s written to be browseable, to be dipped into.  In short, it has an open-form. It can be read from start to finish of course, and there are certain advantages in doing so. But to begin with, read what appeals and excites the curiosity. It is absolutely not a manual or tutor. It introduces ideas and concepts, explaining these with short examples in music notation and program code.

The idea is to introduce the composer to the world of computer-aided composition with script-based programming, and weighted towards composing music for human or simulated instrumental performance. Whilst visual-programming remains very much in the ascendent, born out of IRCAM developments such as MAX , PWGL and OpenMusic, it favours discrete experimentation at the pre-composition stage rather than serving a composing continuum, that is, the journey from first thoughts to finished notated score and reference recording.

ompatchVisual programming in IRCAM’s OpenMusic.

Composing either way can be a most productive activity: to grow ideas through experimentation, through saying ‘what if’. Many composers use CAC systems simply to experiment and to create material that they can then work into a growing composition in a traditional way.

Composers using visual-programming systems often build their own libraries of material from which they can pull possibilities to furnish musical situations. The ‘patch’ from the example below is a PWGL user-library by the Finnish composer Magnus Lindberg. This library is specialized in various rhythm manipulations, including quantizing. The patch displays in the Score-Editor a quantized result of interpolating rhythmic gestures.

pwglpatA PWGL user library.

Creating a whole score inside a CAC environment is another matter altogether, and, as yet, relatively few composers have the patience, time and commitment to work towards this end. But for those who have and can, and – as CAC systems develop and improve – the results can be efficient and deeply satisfying. On the whole, it is the script-based systems that achieve this composing continuum. The example below is, in fact, a complete score-file for a short canon after the Telemann’s canonic sonatas. The source material for pitch is a modulated sine-wave with rhythm created from the generation of binary integers.

canonpngA canon composed in Opusmodus.

However, visual-programming is very much the approach of choice for those composers for whom the physical world of sound has become the source of musical material for instrumental composition. Spectralism is a good example of this trend, requiring computer-analysis of live sound and effective mapping of digitally-recorded sound into notational forms that will involve the presentation of micro-tunings. This means that for a CAC system to handle spectral composition it has to be able input and export digital sound files. In our text, we’ll be concentrating wholly on the  symbolic and the abstract parametric nature of traditional art and concert music and not acoustical phenomena.

In contrast with spectral composition this text focuses on another recent movement in digital creativity: parametric design, an approach that has already made a big impact in architecture and is illustrated wonderfully in the work of Zaha Hadid and Patrick Schumacher. Here’s an image of the proposed Hoxton Square Gallery building approved for construction in December 2013.

param Parametric thinking in music became an important presence in the 1960s through the music of Stockausen, Cage and Xenakis, and remains of lasting influence in today’s music. Such thinking began to evolve before computers were readily accessible to composers, but with personal computers those who did start to use computation found they had to make clear separation between, for example, pitch and rhythm, and that in doing so all kinds of new approaches to designing and making music became possible.

Much of the text is organised into sections profiling distinct parameters and also how such parametric elements combine and work together. Pitch, Rhythm, Dynamics, Ornamentation, Timbre, Orchestration, are all featured. So too are musical structures and processes of traditional kinds (transposition, inversion, retrograde, variation, heterophonic and contrapuntal forms, collage and open-form structuring). These are presented  alongside the more experimental approaches that come from the growing interest and awareness of the potential for composers of evolutionary and natural phenomena, mathematics and scientific worlds (such as fractals, wave-forms, boolean maths cellular automata).

Although the text is accompanied by very short examples of script-based programming (taken from the new Opusmodus composition environment), what is discussed is not system dependent. Indeed, many of the approaches shown could be done with pencil and paper (only rather slower!!).

The text is rich in examples in staff notation many of them taken directly as ‘snippets’ from the notation output of Opusmodus. There are short pieces of code in Common Lisp, Opusmodus, Symbolic Composer and Processing. You’ll find copious links to YouTube of music from the contemporary repertoire cited in the text – including pieces by Morton Feldman, Gyorgy Ligeti and Brian Ferneyhough. Finally,  the Author’s unique web archive is a valuable source of reference to many of the scores and projects described throughout the ebook. This  archive contains downloadable scores, recordings, score-scripts, annotations and background interpretation from over 25 years of working with CAC systems.

A Word about the How and Why

A number of readers may question the title, and perhaps rightly. The business of how seems straightforward, the why less so. The subtitle of the book claims it to be: An Introduction to Computer-Assisted Composition with Script-Based Programming. The text deals with how such programming might be approached as much as why such an approach might be considered. There’s also a secondary answer to this that refers to the description Matthew Lipman, an educational theorist, has made about thinking. He suggests there are two kinds of thinking: the algorithmic and the heuristic. He says, ‘ the touchstone of the first is method, the second is results. An algorithm provides a track or procedure along which our thinking can move. Methodical thinking embraces logic, and by employing logic we might arrive at correct or ‘true’ conclusions.’

‘The heuristic approach is more anarchic. It’s not fastidious about method. It’s what Edward de Bono has in mind when he distinguishes between vertical and lateral thinking.  Lateral thinking de Bono says ‘is successful, productive, creative thinking, whereas vertical thinking is mechanical, unimaginative and uncreative.’

Personal experience suggests that working with Computer-Assisted Composition (CAC), a composer needs both kinds of thinking to produce successful music; hence the book’s title. We need the How and the Why. It’s useful to follow, or at least be aware of method, but equally we need to be critically sensitive to the consequences of our results.

It’s for this reason the book attempts to balance musical making with examining the musical result, the finished score, performance or recording.  Often the music examples the text offers and links to are not composed with computer-assisted means but are there as a models of successful outcomes of musical thinking. This would include music cited by Gyorgy Ligeti and Franco Donatoni, two composers whose music is so rich in the kind of processes and outcomes that can now be achieved with CAC, and which influenced CAC function design. Donatoni’s  extended solo for clarinet titled Clair is a fine example.

References, Links and Further Reading
Nigel Morgan – Extensive web archive with many computer-assisted compositions.
Opusmodus – Parametric software for music composition.
The Illiac Suite – A summary of Hillier and Isaacson’s work by Alex Di Nunzio.
Experimental Music – The full text of Hillier and Isaacson’s work on
Algorithmic Composition – An accessible summary of techniques past and present by Karlheinz Essl.

Leave a Reply