22 – Strategies

Towards the composing continuum : the IAMUS project : the desire for aesthetic control : starting the composing journey – the conditions for creativity : playful improvisation : verbal thinking and planning : After Hindemith – the 4-stage approach of Paul Hindemith : the diary evidence of the start of a piano trio : combining code with hand-crafted material : Strategic thinking in a fully-coded work : Objects of Curiosity II(The Prisoner) - a large-scale piece for double bass and string orchestra : how musical decision drive code solutions and structures : general strategic advice : strategic visualisation in large ensembles groups : aesthetic ‘honing’ : Xenakis and Barbaud : being realistic about the coding of musical detail.

This sequence of chapters has been assembled to demonstrate the possibilities and advantages of script-based programming.  For so many composers the computer is already a partner in the creative process, at some point or another, but it is still rare at this level of engagement during the onset of the composing act. Many researchers and developers continue to see the goal of Computer-Assisted Composing as moving towards a seamless continuum from pre-composition, those first thoughts and nebulous ideas, to the fully- finished score.

There are already fledgling systems that begin to offer something close to such a continuum, notably the IAMUS computer cluster. Sadly, the existing papers available on this academic development don’t suggest how professional composers might align and integrate their work, style and preoccupations with the attributes of this geonomic system. It would appear that the Melonomics’ technology the computer cluster uses can hold a library of existing compositions by one composer (or many) and from this generate further works. But this facility bypasses what most composers might desire from a partnership with a computer and its software, which is not about automation of the composing process but having a degree of personal control that allows either for improvisation and experimentation or as a means for calculation and realising complex formulations and ideas.

The recent Composers’ Rooms series on BBC Radio 3 during 2014 (and archived on the web) provides a valuable archive of anecdotal evidence on how composers work and think. Some composers begin compositional journeys with verbal and graphic descriptions before any sound is sought. For others the direct engagement with sound and music, usually with a keyboard, is where composition invariably begins.

What the early chapters of this ebook suggest is that a composer can begin from a tiny piece of coding that ‘search’ “Papa” Haydn is said to have described when he went down on his knees to pray for a successful day  of composition at the harpsichord. The aspect of playful and exploratory improvisation with a script-based programming environment can be as effective and illuminating as interacting with a piano or keyboard. But once a composer has begun to grasp the possibilities, such software can provide interaction with a degree of verbal thinking and planning and be equally successful. This is what many of the scores discussed in Part 2 of Composing How and Why demonstrate.

Many strategies for composing with script-based programming have already been mentioned whilst individual parametric elements have been focused upon, and the genesis of completed scores discussed.  Perhaps now is a good time for the reader to review these in a more general way.

22-1One score that has so far not been included in the author’s resume of past work is a piano trio titled After Hindemith. In terms of describing how script-based programming can contribute to thinking and reflecting about music composition this score has an intriguing backdrop. It was composed against a period of research into Reflection about Intention in the Act of Music Composition. The objective of this research, undertaken with Professor John Cook, was in the field of artificial intelligence in education: to build a guided tutor suitable for composers studying in university level. As part of my role as the composer-tutor the AI system would model,  I investigated what literature existed on the teaching of music composition. During this literature search I happened upon the experience of the composer Paul Hindemith immediately prior to and during his tenure as professor of music at Harvard in the early 1950s. Although Hindemith gave a celebrated lecture series, which became his book The Composer’s World, it was the records kept by his students of his advice on the composing act that I found most valuable. What follows comes from Eckhart Richter’s A Glimpse into the Workshop of Paul Hindemith where he describes Hindemith’s working method as presented to a class at Harvard in 1951. Hindemith also ‘converted’ this procedure into a guide for listening in a lecture he gave at the University of Zurich in 1955.The working method is outlines thus:

1. The general determination of the character, medium, and the basic purpose of the piece, as well as its expressive character, and even place of performance.

2. A master plan of formal design, including the overall shape, the number and character of sections, changes in mode and tempo, rhythmic character, texture and the degree of activity, the gauge being the amount of effort the listener must expend to comprehend.

3. Then ‘came the tonal layout in which the basic tonalities of each section and their relative degrees of tonal stability and complexity, as well as the modulations, were mapped by means of a diagram’.

4. Specific thematic material.

It was this guide that was tested in the composition After Hindemith (1998). Here is an entry from the author’s notebook / diary during the composition:

 After an initial ‘imagined’ statement teasing out a ‘tonality’ pattern from a 3 note figure the combination of a tonality statement of 4 ascending minor triads (in arpeggio) and the octotonic scale (diminished1) I applied the idea of drafting (on graph paper) a rhythmic and instrumental sketch. The attraction of this is that is can be quick to do and it’s good then to have something concrete to ‘flesh out’. This process certainly enabled me to ‘get started’.

22-2The temptation to continue using Symbolic Composer (SCOM)  at this stage was pretty great, partly because I felt it was important to re-establish my working relationship with it. In some respects my initial inner design for the tonality reflected my developed practice with SCOM. The ‘tonality first’ mode of thinking that SCOM promotes is certainly in line with the Hindemith model! In the SCOM manual the notion of tonality is presented in a way that can be applied to (almost) any compositional situation: whatever you imagine that resides pitch-wise on the lattice of the chromatic scale (no matter what you ‘touch’ on a keyboard is ‘in a tonality’ or can be a ‘tonality itself’).

The code itself is based around a single function-based expression that ‘collects’ a list of 20 symbol lists between three and six symbols (This function is known as gen-loop in Opusmodus). These symbol lists are derived from variants of the initial arpeggio shape / chord ‘((a b c) (a b) (c b) (a c)) and symbols comprising the remaining arpeggio shapes / chords which are picked – 1 to 3 in number at a time – appended and shuffled together.

Here are versions of this expression in SCOM and Opusmodus:

22-322-4What is happening here is that the initial arpeggio / tonality is having added to it symbols from the other three arpeggio positions. Note that each symbol list is ‘grounded’ in at least two symbols from the initial arpeggio. And from this, symbol output rhythm and the ‘play’ of instruments are derived; each symbol list is also a separate tonality ‘zone’.

Clearly, looking at the whole code, the indeterminate element – or this mode of working that proposes possibilities through controlled randomization – provides a playful, improvisatory ‘what if’ component within the scheme of the composition. Quite radical ‘what ifs’ can be proposed without damage to the composer’s vision, or, more significantly, without taking inordinate amounts of time to realise.

This anecdotal record of how the opening of a piano trio came into being demonstrates a mixture of composing strategies, from the ‘sketched and hand-crafted’ opening to the preparation of tonalities, to creating a development section working in code. To look in more detail at the whole composition an annotation of After Hindemith can be downloaded from the composer’s web archive.

22-5Above is shown the link between a brief chaconne-like section and the second  and fully-coded section. Almost the same code-expression is used at letter C as it was at letter A1. Notice the tonality has now changed:

22-6The third coded section is a variant on the second and remains centered on Eb:

22-7The fourth coded section now adopts the tonality of the f# minor and has a quite different organisation of material than the earlier sections:


Suffice to say that despite using Hindemith’s strategic thinking the piano trio sounds very little like the music of Hindemith, and its title simply reflects the fact that the composer, despite being inordinately prolific, never actually composed a piano trio.

Another significant composition has also been omitted from earlier discussion or example. Objects of Curiosity II (The Prisoner) is the second in a trilogy of pieces for strings that focus on the work and achievements of architects. In this instance the architect is Clough William-Ellis, creator of the Italianate village Portmeirion, but expressed through the drama of the pilot episode of the 1960s TV drama The Prisoner. This extended composition for solo double bass and string orchestra is in a single movement of fourteen sections, using a wide variety of strategies to fulfill a busy scenario of action. The composition is written entirely in code and is contained in twenty-three separate files. It was completed in 2007 during a period as composer in residence at the Interdisciplinary Centre for Computer Music Research, Plymouth University.

Strategically this extended piece needed careful preparation, but that was only after an initial Introduction was attempted scored for double bass and piano. The music here presents an algorithmically-derived chord sequence of six 4-note chords, each chord to represent one character in the complex scenario. The chords are created here in right and left hand expressions:

22-922-10What a composer has to allow for here is a proto-score that will enable that rough prototyping of six chords: a combination (if possible) of audio playback and notated display. In such instances a test-score definition for keyboard instrument can be invaluable, where by changing the random seed or the weighting of the content of the chromatic scale (as above) different outcomes and emphasis can be achieved, and achieved very quickly. Because of the string orchestra scoring of The Prisoner the six chords were tested against sampled string sounds as well as a piano sound before a decision could be made.

With chords in place the construction of the introductory section could get under way. What was required by the scenario called for something frenetic and exciting that had a fast and urgent harmonic rhythm with swift changes of chord. We’ve seen the use of substitute-map in the chapter Harmony & Chords. Here is the same mechanism being used in Symbolic Composer code with an identical function e-substitute. The resultant chord sequence was altered still further by expanding the chord intervals of certain chords. Forty-eight chords in all were produced with these expressions:


As the pace of the harmonic rhythm begins to slow down there is a section where the six-chord collection makes an appearance ‘in sequence’. See below:22-13Once these chords were established their individual tonalities could be created using the pattern-to-scales function introduced in the previous chapter. This strategy is able to turn randomly-selected combinations of chords  into scales and a quite different aspect and potential is revealed:


Each of these ‘chord scales’ appears as important elements, often starting points, within the fourteen sections of the work. A good example of this is the way the third chord associated with the character of the Butler takes the Prisoner on a tour of the village. His music is a 12-bar blues starting with the E Major seventh arpeggio seen in the third bar of the example above and the first bar below.


To get a broader picture of the strategies involved in this introductory section download the score-files from the composer’s web archive.

When composing with staff notation on manuscript (virtual or paper)  there is no trace of which part of the imagination or personal technical library  the pitches and rhythms have come from.  Script-based programming however does record the ‘making’ ; it leaves a trace, and a trace that can be reused – as many examples in this ebook testify. However, it is all too easy in the heat of writing code to a) forget to ‘comment’ on complex expressions and , b) to save stages of the process in separate files.

Once some experience is gained putting together more complex expressions there is something to be said for continuing to keep expressions short and focused on a single function. This example from an earlier chapter could easily be simplified and the different stages clarified (as below), obviating the need to ‘comment’.

22-17Remember, unless a very specific quality is desired, musical output does need to breath, and some instruments have to! This is where attention to rests and pauses in streams of code benefit from strategies to break up the music into phrases.

The beat/space concept and the idea of swallowing pitches when rests and pauses appear remains a powerful strategy to be aware of and when necessary adopt, particularly in instrumental music where hocketing and antiphonal effects are required.

Using library locations to store material that might be used to randomly select rhythmic patterns can significantly reduce the amount of written text on a score-script and aid that visual scanning of code referred to in the chapters on writing for orchestra and large ensemble.

Graphic visualisation and plotting can be valuable strategies to secure overall perceptions of structure,  particularly when streams of data are being used or a large number of instruments are being scored, as in this timesheet for a large percussion ensemble used in Seven Magical Preludes:

22-18Above is a complete timesheet visualisation of the 4th Prelude and below  the final section of  the 4th Prelude from letter L in the full-score:


One aspect that needs to be included in any discussion of strategy in script-based programming is that of what composer Iannis Xenakis referred to as the aesthetic honing of algorithmic output. Close study of the code published in his book Formalised Music (pdf) shows that his compositions were often ‘massaged’ or aesthetically honed as they came to be realised in musical notation. One such piece XAS for saxophone quartet is a fine example of a consistent approach taken by the composer to produce a ‘musical’ result from what in mathematical terms appears exciting, though intractable, material.

Working with the output of computation as a means towards music composition can so easily become a seductive pleasure. It can present an ‘automatic’ quality which can, without care and a critical filtering, result in difficulties for both performer and listener. To experience this phenomenon, that of blind acceptance of untempered computer output, the music of Pierre Barbaud must serve as a possible warning. A contemporary of Xenakis, Barbaud helped establish the Institute de Recherche de Informatique et Automatique (INRIA), a forerunner of IRCAM. His work was predominantly composed for human performance and as such he severely tempered the input of formulations. In other words he did not allow or seek the musical flights and experiments found in the scores of Xenakis. He played safe.

In our own relationship with creating music there is much to said to being realistic about the ‘Why’ of the process. Whilst there can be aesthetic pleasure gained from producing coding that attempts to reproduce fine detail of sound and human musicality, when composing for human performance this may be time and effort wasted, despite the need for precise textural detail so often required by composers. What has been focused upon throughout this book is practical composition for real performers, ensembles and orchestras: to realise in performance and recording. Most professional composers work to commission, which in itself demands the observation of precise conditions of instrumentation, duration, performance location and stylistic considerations, and there are very often severe time constraints surrounding the composing act.  Simulation of performance certainly has its role, and is necessary and useful in certain circumstances, film and media music being one, another being an invaluable rehearsal aid. That said, script-based programming remains an ideal medium to realise imagined music into notation and retain something of the poiesis of invention, that physical trace of the creative act.

Be sure, the analysis and awareness of that mysterious process of decision-making that music composition generates can be of significant value to the composer as a personal trajectory of creativity develops and becomes a body of composed work.

Links and Resources
Pierre Barbaud – Association Pierre Barbaud
Paul Hindemith – The Poetics of Music
Nigel Morgan – After Hindemith, Objects of Curiosity II (The Prisoner), Seven Magical Preludes
Radio 3 – Composers’ Rooms
Iannis Xenakis – Formalised Music (pdf), XAS
University of Malaga – Melomics/IAMUS
UWE Bristol – Professor John Cook

Leave a Reply