19 – Composing for Orchestra I

Composers and orchestral composition : the role script-based programming has played in the composition of orchestral music : a survey of orchestral and large ensemble composition of one composer using CAC : Heartstonefor wind, piano and percussion : L-Systems and visual beat /space scoring : two BBC commissions for code-based scores realised in distributed performance : reintroducing the Baroque continuo into the orchestra : Quatuor des Timbres - computer-guided structuring and orchestration : Self-Portrait- an ensemble score in Open-FORM.

Composing for the orchestra, though regularly and seriously challenged as a culturally necessary activity for the composer, remains an imperative. The musical establishment holds onto the aesthetically pleasing phenomenon of an ensemble of musicians playing pre-composed music in a largely synchronous fashion, usually under the direction of a conductor. Traditional forms of notation continue to be required to promote and pass on performance practice inherited and largely stabilised from several centuries of education and tradition. Today’s music is sadly not today’s music, but music of the 17th to 20th centuries. Even areas of newly-minted music are required to employ musical techniques of the past, most notably in music for film and TV. Whilst there have been since the early years of the 20th century brave experiments in orchestral music, think of scores by Varese, Xenakis, Penderecki, Grisey, Harvey, Haas, composers have to wear the straight-jacket of the traditional symphony orchestra.

The down-sizing of the late 19th century orchestral medium began in the hands of Stravinsky, Schoenberg and Webern during the inter-war years and continued post 1945 with what is now a prevalent one-to-part chamber group adopted by ensembles such as Ensemble Modern, Bang on a Can and the London Sinfonietta. There are relatively few composers who have held out against writing for the large orchestra required to play the 20th century masterworks such as Stravinsky’s The Rite of Spring, a work which has spawned 21st century compositions such as Asyla by Thomas Adès, Harmonielehre by John Adams, Feria by Magnus Lindberg and Earth Dances by Harrison Birtwistle. Composing for orchestra remains the standard test and the goal of the emerging composer. There are few exceptions that come to mind, but those who have side-stepped this cultural requirement are amongst the most interesting and innovative voices today: Christopher Fox and Michael Van der Aa are two that spring to mind.

It was the novelist Herman Hesse who in his prophetic book The Glass Bead Game made one of the earliest pronouncements of the demise of the symphony orchestra. He foresaw the rise of the Early Music Movement and what is now a continuing fascination with authenticity and performance practice. He likened the state of music to that in ancient China when the condition of music was considered as a barometer of a healthy imperial state. If music became too complex (too chromatic and loaded with instrumental effects) its artefacts, instruments and scores were destroyed and music returned to its pentatonic simplicity. In some sense this juncture has occurred in our time with the rise of minimalism, although this hasn’t stopped the existence, indeed the development, of music of extreme complexity. It is with spectralism that there has been a merging of the two: a degree of minimalism within those parameters of traditional pitch, harmony and rhythm pitted against the demanding microtonal complexity of the physical world of sounds that has eschewed equal temperament. The exotic in sound, created with new instruments or extended playing techniques, continues to have a presence within orchestral music.

The role script-based programming has played in the composition of orchestral music is there to be seen and heard in the music of Iannis Xenakis, Milton Babbitt, Jonathan Harvey, and more recently with Magnus Lindberg, Tod Machover and the intriguing experiments of Iamus and the composer Gustavo Diaz-Jerez. Whilst there certainly are and will be scores that benefit in some form or another from the interaction or mediation with computer assisted composition the orchestral medium poses all kinds of problems for the composer, not least that of a continuum that might represent the whole work of a composed piece. Despite developments in software representation the sheer size and complex state of an orchestral composition remains a problem for all but the most determined of composers.

What follows is an account of one composer’s approach to the issue of computer-assisted composition in the orchestral realm. It reviews a series of works spanning a 20-year period and includes a variety of different approaches, and with combinations and differing sizes of ensemble. It tackles what the composer felt to be a very necessary issue and opportunity: bringing together instrumental scoring with the composition of pitch and rhythm.

The chapter Every Composer is Different mentions the composition Heartstone for wind orchestra, piano and percussion. Hindsight now reveals that with such limited experience of script-based programming the composition of this work would probably not have been undertaken, but for the fact that features and possibilities in the software employed (Symbolic Composer) seemed to fit with the composer’s search for a solution to the ‘problem’ of this project. This problem was in regard to the exploration of harmonic rhythm with a large ensemble whilst using relatively straight-forward rhythmic structures. The software also offered a visual-approach to organising scoring (and harmonic change) suited to the requirements of a large ensemble. This approach, what has been termed in previous chapters the beat/space mechanism, enables the composer to produce in a code script a visual paradigm similar to that perceived on a full orchestral score, an image that clearly shows where instruments play and where they do not. Remember that opening page from the 2nd Symphony of Brahms:

19-1The image above is a powerful indicator for the composer of the play of instrumental activity and textual colour in a composition. In Heartstone beat/space notation is employed to give a similar impression but without showing the individual notes that fill the beat space. This is demonstrated in the code illustration below. The kind of code notation used here has been discussed previously in the chapters Starting with Rhythm and Sextets. However, this extended score format also indicates where tonality change occurs. This is shown in the top third of the score by the changes and tonal indications marked on the time space by dots. This image below is taken from a detailed annotation of the score created for presentation of an academic paper featuring the composition at ECAL 2007 in Lisbon.


What this image from the code reveals is a design feature that required a fixed resolution of the beat / space. Here that space is given as 2/2, and because of the composer’s relative inexperience with the software language at the time of composition it appeared that a variable resolution could not easily be introduced.

Although the notated score Heartstone was substantially revised in 2008 the original composition created in code in 1993 has remained largely unchanged. Looking at its script-based composition in the light of nearly a dozen subsequent orchestral and large ensemble works, and in the intervening twenty years, it is clear that the script itself is both economical and manageable as a document that gives a trace of the workings of the piece. What the script shows is that the music is the result of responding to the major features available in the software that engaged the composer with their novelty and their seeming potential for large-scale structuring. Bear in mind that the work is in six movements and lasts about 25 minutes.

These features, which can be recognised clearly in the published annotation, can be summarised as: the creation of original / unique tonalities upon which to map symbolic data (first and final movements); the ability to create a counterpoint of tonalities (as in the first movement); the use of fractal algorithms such as L-Systems to generate pitch and rhythmic data; employing libraries for storing pre-defined pitch, rhythm and tonality data (this significantly reduces the content mass within the score-script); visualisation and control of activity, density and silence with a beat / space notation.

In many ways Heartstone remains the most speculative of all the orchestral and large ensemble works produced by this composer in a 20 year period, but is none the less effective for being so. In later pieces, particularly the Six Concertos series, the composer sought to satisfy a personal wish to integrate decisions about and implementation of orchestration and scoring into computer-assisted composition. There was also the desire to be able to respond at a bar-to-bar level of composition to streams of generated data rather than, as in Heartstone, setting up a mechanism at the opening of each movement which was allowed to run its course unhindered.

Before moving on to the music that followed Heartstone there was a further off-shoot of this experimental orchestral score : a chamber work composed in 2008 for two instruments that uses Heartstone’s score-script data.

Treeness for viola and chamber organ was written as a result of the publication of the composer’s contribution to a book of articles titled A-Life for Music: Music and Computer Models of Living Systems edited by Professor Eduardo Miranda. The chapter titled Transformation and Mapping of L-Systems Data in the Composition of a Large-Scale Piece of Music is a detailed discussion of the Heartstone score. The book was published with a CD, but recording Heartstone in its original form proved too expensive, so a chamber version was suggested, a version which shows effectively how a quite different score might be reinvented from the original score-files. The title Treeness is born out of the composer’s concluding observations in his chapter on the composition of Heartstone that ‘the beautiful tree-like structures made possible with L-Systems speak of the Gestalt of tree, a treeness more intense and exact than the living object, and able, because of this exactness to be embedded as a structural mechanism to play in the musical imagination.’

The onward path from the composition of Heartstone towards music for the standard orchestral forces found in a symphony orchestra was predicated upon the composer’s appointment in 2000 to one of the BBC’s resident orchestras, the BBC National Orchestra of Wales. Working as a ‘project composer’ over a period of 5 years his brief was to give orchestral players a window on new technologies in digital audio recording, the Internet, use of digital images and video, electroacoustic composition, real-time improvisation with computers and algorithmic music. His first project required the live link up across the principality of Wales with the orchestra and two performing groups in different locations. The live link was managed by ISDN communication although during the production process use of the internet was explored and found (in 1999) to be to slow for synchronous performance.

This project provided an opportunity to experiment with the reintroduction (from 18C practice) of a continuo group into the orchestra (albeit the continuo playing electric and MIDI instruments). In fact individual members of the continuo group led each of the three ensembles in performance, but like the continuo group of the Baroque period had rehearsed as a collective entity and ‘knew’ the workings of the whole score. The resulting piece titled Conversations in Colour has already been discussed in the chapter Being Harmonious for its script-based score developing its entire material from the text of words used by the artist Josef Albers to describe the affect of colour.


With the tonality in the scale of C chromatic the text(me la me la nc ho nc ho), when converted to pitches creates the left and right hand parts found in the opening bars of the electric piano part. Notice how the three ensembles are organised with piano leading the SATB choir, the bass leading the wind and string players, and the mallet instrument (A MIDI Kat) leading the percussion.

The five movements of Conversations in Colour were created in code scored for the 4-piece continuo group. The remainder of the scoring was conceived ‘by hand’ on a score writer composed on top of the continuo part.

Shortly after Conversations in Colour a commission from a festival in Harstad, Norway offered an opportunity to devised a more elaborate companion piece to this work, it being in a similar disposition for distributed performance, again for three ensembles in three different locations.


Schizophonia is a fifty-minute work entirely written in code for a large wind ensemble, a string orchestra and choir with four solo singers. The continuo group formed for Conversations in Colour acted as leaders and the score began life with experimental music for this 4-piece ensemble. As with Heartstone the algorithmic devices employed in the instrumental scoring often reflect standard composing techniques: transposition, inversion, interleaving, retrograde. But in Schizophonia the interplay between ensembles often resembled that found in the multi-choir antiphonal music of Gabrieli and the Venetian masters. The music has two extensive 15 minute toccatas for the full ensemble between which movements for a capella choir, a cabaret choir with continuo and an electroacoustic finale are interleaved. The source material comes from two sets of patterns (186 -192) and (231 -236) contained in the Slonimsky Thesaurus. What is particularly unique to this composition is the way the text (by Sean Cubitt from his book Digital Aesthetics) ‘generates’ the vocal music, the composer creating a series of innovative functions to make this possible. This aspect will be dealt with in detail in a later chapter Text into Music.

Whilst the experience of creating and realising this work was invaluable it didn’t address some of the fundamental concerns surrounding composing for orchestra with script-based programming. This process began immediately following Schizophonia with the composition of a series of chamber pieces exploring the management of timbre and structure.

Quatuor des Timbres gave rise to a number of technical solutions in coding without which later works might have been impossible. Milton Babbitt’s Composition for Four Instruments (already featured in this text in the chapter titled Sextets) and Ixion by Morton Feldman provided valuable models in this respect. The opening of Ixion for eleven instruments is shown below:


The idea of Feldman’s box notation is disarmingly simple and is fun to recreate in code with a script-based system. Here is the score recreated in Symbolic Composer. A generating system built around a def-neuron expression was devised to realised the score algorithmically.

19-7What the score and table show above are lists of attack-values occurring within a time space, a space that can have a set duration (indicated by a conductor or pre-defined metronome value) or a duration that can be varied as the ensemble to conductor wishes. The number of attacks can be varied by the addition of rests and rhythmic values as this table below shows:


The neural definitions shown above were created for IXION but also used in the second movement of Quatuor titled Gioco dell Coppie (Game of Pairs), the name given by Bela Bartok to the second movement of his Concerto for Orchestra. In Quatour the instrumental scoring is devised (through computation) so that this pairing happens continuously, though its content is constantly changing.


It was in Quatuor des Timbres that the seeds of algorithmic selection of scoring were sown. The background to this innovation will be explained in some detail in the following chapter as part of the discussion of the Six Concertos for self-directed orchestra. For now the basic mechanism will be presented. The abbreviations wd, br, pc, st denote woodwind, brass, percussion, strings.


Compare now part of the list of output-2 against the scoring in the first three bars of the music example. The function gen-collect in Symbolic Composer is equivalent to gen-eval in Opusmodus. The function power-set creates a list of every possible combination of the instrumentation list:

19-11With such scoring information a visual paradigm can be created be turning the incidence of an instrument’s occurrence into a string of symbols used in the beat / space notation already shown earlier. As before, look at this ‘timesheet’ in relation to the notated excerpt of Gioco dell Coppie:

19-12Now, with rhythm, structure and orchestration in place, pitch is gathered from segmenting or partitioning a stream of white-noise data converted to symbols a to l, which are mapped to the 12 degrees of the chromatic scale in C:


This stream of 128 symbols is processed with functions called seq and gen-seq. This only appears in Symbolic Composer and is linked to its unique neural expert architecture. It enables in this instance a stream of pitches to be applied sequentially and simultaneously across a group of instruments. The code example begins with an expression that shows the woodwind part gaining its attack-values. The second expression using seq and gen-seq shows how the pitch stream is matched to the attack-value:


Looking again at the music example, or better still the whole score of the piece, the pitch material remains identical in each part, although placed in different octaves and playing in different rhythms.

Whilst Quatuor des Timbres tackled the issue of automated scoring the next chamber ensemble added the continuo group to the instruments of Quatuor making a septet. Self-Portrait for seven musicians moves closer to a code structure that begins to anticipate a working environment for composing orchestral music. It’s an amalgam of experiments, not least in its adoption of the aesthetic principles of Open-Form, an aspect of structuring musical material which begins to appear in the composer’s work of the mid 1980s, in particular the composer’s score for variable ensemble Metanoia. In Self-Portrait the difficulties of presenting notation inherent in the early experiments with Open-Form (Boulez, Pousseur and Stockhausen) are gracefully overcome by the use of networked computer displays instead of printed scores.

In the context of this chapter the step forward found in the score of Self-Portrait was to configure the timesheet, that visualisation of instrumental action and silence, with variable resolutions of the beat/space. The timesheet was made to accept strings of different metres giving the resultant music a wholly different aspect, either rhythmically incisive or an attack-free flow of material. 


The musical starting point for Self-Portrait was an investigation of the potential of the two positions of the whole-tone scale.

19-20One function able to further this process was pitch-transform. If variable w1a is an extended pattern using the first position of the whole-tone scale and w1b the extended pattern of the second position, pitch-transform can produce a controllable transformation between one list and another:


This is a further instance of how it is possible with script-based programming to start the composition task from any parameter.

With these experiments came the realisation that the resultant score-scripts were becoming complex and difficult to read (and therefore to debug). Take a look at the score-annotations of Heartstone, and then Self-Portrait. To extend composition with scripting to the demands of composing a work for full symphony orchestra required a further step. This was achieved in part through two further compositions, the string quartet Le Jardin Sec and a work for 3-manual organ, a reinvention of Bach’s Piece d’Orgue BWV 572. In viewing the annotations for the code script of the reinvention of Piece d’Orgue it is possible to see the origin of the quite radical solution to the composing for orchestra with script-based programming, a solution that made possible the composition of a sequence of Six Concertos for self-directed orchestra. This is the subject of the next chapter of Composing How and Why.

Links and Resources
Sean Cubitt – Digital Aesthetics
Morton Feldman – Ixion
Eduardo Miranda (ed.) – A-Life for Music
Nigel Morgan – Heartstone, Treeness, Conversations in Colour, Schizophonia, Quatuour des Timbres, Self Portrait, Metanoia, Piece d’Orgue, Six Concertos

Leave a Reply