Evolution of Harmonic Language: Overview
[Beginnings]I was interested in the sound of timbres which contained microtonal material *eg; creaks, bells* I wanted a formalised approach to using microtones, the microtonal theorists' approach was based too much on traditional thinking about melodies, motives etc.
[Electroacoustic]Reading Wishart's On Sonic Art and Smalley's Spectromorphology was a more stimulating approach to both form and harmony. The notion of harmonic change as a gradua process was something which in early works I experimented with, but found that 12tet pitch material was simply too clunky for this: I couldn't achieve the seamless and gradual changes that I was after and didn't wish to abandon real instruments for electroacoustic composition.
[Spectral]Discovering spectral music, especially Grisey, was the key that I needed. The techniques which I assimilated most thoroughly were spectral modeling and FM synthesis, techniques which I have evolved my own implementations of and which through this have become a cornerstone of my musical language. This evolution of language took the form of altering how I thought about music at the fundamental level, the point of musical conception gradually shifted from a note-based idea to sound-based idea. It was still necessary to write the music as notation in order to communicate with players, but along the way this became less about notation *codification* of my music and more like translation of my music into something which players could understand *more readily than my lists of frequencies*.
[Breaking with 'notes']'Frequencial harmony, is Tristan Murail's term for this way of conceptualising the music not in terms of discrete note-objects but as an infinite continuum of possible frequencies. This move from the discrete to the continuum brought about a decisive break with 'notes'. The traditional compositional devices and structures of Western Art Music are fundamentally based on the 'note' as the atomic unit, in a Heisenbergian world it is *uninteresting* to work with the 'note', a gross simplification of the continual possibilities of the frequency spectrum. *blue-notebook reference to postmodernism/Kramer? - 'we become aware that objects are not so much “in the world” as products of perspective' + Eco – 'promotes an ideal of an ordered world which does not really exist, avante-garde attempt to tell it how it is.'(Eco – open work p.141). The notion of smoothly evolving systems has been largely debunked in favour of growth spurts, motion between fixed energy states, punctuated eqilibrium etc: I favour this reality, surely discrete notes are part of this, or are 12tet notes an artificial system imposed on the 'natural' order of partials (don't get sucked into spectral mysticism) *
Combining the process-based ideas of Smalley and Wishart with frequencial harmony and Spectral approaches to time led to me working more closely with process-based music. I had attempted to use processes from minimalism in my earlier music, but again found the clunkiness of the notes to be unsatisfactory.
All this was combined with a growing interest in free-improvisation, something which I began largely as an opportunity to explore the possibilities of extended techniques on any instrument which I could get my hands on. Through improvisation I explored new textures and timbres, building up a palette of sounds.
Summary of Harmonic Archetypes
In my portfolio, the basic harmonic language used for each piece can be broadly summed up as follows:
This has all led to a method which can be separated into two areas in any given composition, an Environment and a set of Interactions. The Environment is the specific set of sounds/timbres used in the piece, this arises from the choice of instrumentation, and the harmonic language archetype used: these may or may not be part of the initial conception. The Interactions are to form of the piece, how the various agents interact to produce tension and release, where pattern formation lies: this is often the initial concept of a piece.
[Definition] To avoid using the term harmony. This section is concerned with the generation of vertical elements without respect to time (the horizontal). Various processes are used to generate this material, from piece to piece these processes are applied with varying levels of strictness.
[Two examples of relative strictness]
Marx uses a simple arithmetic process derived from frequency modulation synthesis (FM synth) to generate pitches which are then applied in a strict manner, on the other hand uses pitches which are derived strictly but applied with a high degree of indeterminacy, the two layers of which are not immediately apparent from the notation.
In Marx, the indeterminacy is largely horizontal: the vertical is strongly generated in the synthesiser part, but the viola d'amore and vocal parts, while strongly notated, are likely in performance to have a more loose intonation which will beat against the synthesiser pitches. Because the synthesiser pitches are generated from theoretical frequency modulation between the viola and vocal parts, there will inevitably be a clash between the frequencially exact synthesiser and the approximations of the viola and vocals (no matter how much they strive for accuracy they can never match the digital accuracy of the synthesiser over sub-semitonal ranges of intonation).
In Whitewater II the pitches are derived from spectral analysis of saxophone multiphonics, subsequent notation includes accidentals tempered to the quarter-tone, but with additional accidentals which imply a further sub-quarter-tone inflection [what? - and show examples]. Wind instruments have inexact tuning at the best of times, non-12tet pitches are even less exact between individual instruments/players, this adds a layer of indeterminacy in that pitches notated the same may not be played in tune by two players. In most music there is a tacit understanding among the players that pitches of the same notation should be in tune, but the score of demands that most of the players detune their instruments by an unspecified amount of around a quarter-tone, breaking this understanding. The ultimate aim of this indeterminacy is twofold; (1) to foster a rough vertical timbre/texture with high density of acoustic beating, (2) to create structural ambiguity in terms of pitch across time: to disorient the listener with respect to the identity of temporally distinct (or indistinct...) sections (in a rarefied soundworld such as this, pitch forms a strong identifying marker/element).
Indeterminacy here can be defined as an unwillingness to specify detail sufficient for players to have an objective reality where a pitch is in-tune, or 'correct'. The objectives, as mentioned in part above, have to do with the music's aesthetic on a musical and socio-political way. The vertical macro-identity of the piece is an emergent result of the local interactions of imprecise tuning, generating a texture which is dense and beat-filled. The socio-political aspect involves the individuality of each player being a contribution to the whole, but without a governing intonation: each player chooses their own interpretation of the accidentals – it's a tiny window of free choice, but choice nonetheless.
'Strictness' in this context means that there is no attempt to alter the pitch material in accordance to taste or context, other than a freedom in choosing which (if any) of the set of derived pitches to use in any given chord/field/simultaneity.
In relation to actual FM synthesis there is a wide degree of latitude, in that the amplitude values which would be generated by true FM synthesis are here largely ignored, taken into account only in a metaphorical way where more distant side-bands are generally accorded lower amplitudes: but not necessarily, occasionally compositional choices may intervene to artificially increase focus on a ordinarily lower weighted pitch.
[Processes]
The processes involved in generating vertical material can be separated into strict processes of pitch generation and free(quasi-intuitive) processes to package this data into strings of material fit for composition.
Data is the numbers generated by the processes, usually this is sets of frequency values (with amplitude values sometimes strongly generated [really? - examples] and sometimes implicit from the generating method: where FM synthesis implies amplitudes related to side-band level).
Material is the next level, applying contextual filters to the data to make it usable in moment-to-moment composition. Material may be the final vertical level, ready for embedding into the horizontal aspect, or there may be further levels of vertical organisation (or disorganisation) and parsing/filtering.
[FM synthesis]
While the process I use is not frequency modulation synthesis in any real sense, I will use the term FM synth throughout this document to refer to the below described arithmetic process which I use. The process I have adapted from the techniques of the French Spectral school, most notably Gerard Grisey and Tristan Murail, via the writings of these composers and that of Julian Anderson.
Harmony is derived by calculating summation tones and difference tones between two pitches, according the formula: sum tones {x = f1 + f2, 2f1 + f2, 2f1 + 2f2, 3f1 + 2f2...nf1 + nf2} and difference tones {x = f1 - f2, 2f1 - f2, 2f1 - 2f2, 3f1 - 2f2...nf1 - nf2}
The pitch generating process here is strong, leading to low levels of vertical indeterminacy, in Marx the process is taken to its strongest level as the synthesiser allows specific frequencies to be played, in other pieces the pitches have been approximated to 8th tones. Marx relates to Primes in creating an opposition between perfectly accurate pitches generated by computer (synthesiser) and human generated pitches which strive for intonational accuracy to be within a certain soundworld. The main difference being that Primes uses a harmonic spectrum (only the prime-numbered partials thereof but harmonic nonetheless) while Marx uses inharmonic spectra generated from the FM synthesis. In my earlier work it would have been more important for players to aim for pitch accuracy but with the understanding that they could not succeed and that beats would be generated from their striving, as well as a macroform built of relative successes and failures with respect to achieving the perfect intonation. As my work has developed I have realised that this has become less important for me in dense textures where beating/density are the functions of the harmony.
[Spectral Modelling]
Technique derived also from the Spectralists, as above. Creating a model of a given sound by using spectral analysis to identify the most perceptually important partials (a subjective process, where the definition of 'important' is subject to the composer's taste) and using them to form a simplified model which may then be orchestrated. The temporal aspect of the spectral analysis is applied more intuitively to the model.
Vertical indeterminacy can be strong with this process. My personal analysis of inharmonic sources such as bells and other metalophones has led me to the conclusion that the richness of such sources derives in part from the superimposition of several simpler spectra which are out of tune with each other at a microtonal level [check this]: this results in two phenomena which areassociated with such sources, acoustical beating, and tonal-centre ambiguity. These characteristics hold a fascination for me and as such are strongly reperesented in my musical language.
[Beats] Focusing on pitches less than c.20hz apart to produce audible beating. This archetype often uses poles of beating and non-beating to generate structural tension and formal space. Many pieces which are formally based on one of the archetypes below will also incorporate beats at a local level of harmonic language. Thus 'beats' can be considered as a fundamental timbre archetype rather than a specific compositional archetype.