ANALYSIS AND LISTENING PROCESSES

ELECTRO-ACOUSTIC MUSIC: ANALYSIS AND LISTENING PROCESSES
by Lelio Camilleri
  
INTRODUCTION
The introduction of technology in its various phases of development from the end of the 194O ‘s up to today has brought about not only an enrichment of the sonoral palettte, it has also brought about a great deal of theoretical reflection on how to classify the sounds of the works thus created and how to analyze them. Thus, a problem arose, not merely regarding the study of music using technological instruments, but about all sonoral phenomena which cannot be described by means of the lexicon and methodology of existing theories, however advanced they may be. The question of analyzing electroacoustic music is not a case unto itself. Indeed, even keeping in mind its specific characteristics, it involves other repertoires: for example, popular music, some contemporary instrumental music and other kinds of music which extend the sound landscape in an unconventional way. Since electroacoustic music includes a sound landscape that is extremely vast, I cannot but agree with Smalley’s provocative statement (Smalley, 1991a) that describes instrumental music as a sub-set of electroacoustic music; thus the methodology of analysis of electroacoustical music can make an important contribution to the problematics of the analysis of instrumental music as well.

The purpose of this article is to reflect on some of the theoretical-analytical approaches to the problem of the study of electronic music, emphasizing the most important aspects and their possible development. These approaches may be seen as a segment of music theory, that of the analysis of music without notation, in continuous development. Obviously, when we speak of music without notation, we risk creating a certain confusion with other musical fields – ethnomusicology for example, in which there often can be found problems of unconsolidated or inexistent notation. Moreover, one might object that in this field there already exist a great number of analytical studies and theories. The problem, however, is not the same, also because, although the context is quite different, several basic concepts of existing analytic theories may be useful in analyzing this music in which, among other things, we come face to face with sound phenomena not so very different from those of classical music. For example, the concept of pitch may be widely used, as well as that of the scale, to cite only two definitions.

The creation of analytic strategies for electro-acoustic music must regard three aspects:
1) a reflection on the musical language, its potentialities and its ties with the natural sonoral world;
2) the relationship between .psycho-acoustic properties of sound phenomena and their description;
3) the creation of an analytical lexicon. Naturally, these three aspects are theorized and framed in different ways in the various analytic tendencies.

It is for just this reason that I shall refer to the distinct analytic approaches to the music of the electroacoustic medium, especially concerning the first two problems. Before confronting these aspects in their specificity and in the interpretations furnished by several theories, it is necessary to throw some light on problems that the analysis of electroacoustic music evidences. First of all, there is the problem of the neutral level, (niveau neutre), understood as the level limited to the musical text as it is thus understood by Nattiez. The analytical approach to electroacoustic music cannot be other than aestethic/perceptive/cognitive, since no consolidated correlation exists between graphic representation and sound text (as there does for instrumental music); the only text that we can analyze is the sound text. It is my opinion that this characteristic represents a strength and not a weakness of this musical genre. Concentrating solely on the sound material without any written reference avoids the risk of carrying out analyses based on only one aspect (that of the relationships among pitches, for example) which often don’t consider what we are actually listening to. To use a lapidary and provocative phrase, the neutral level in electroacoustic music finds its confirmation of its inapplicability. In this case it might be objected that since no written text exists, fixed in a score, but only listening to the electroacoustic musical text as the object of analysis, this might orient us toward unique, single and subjective readings. Even if this is partly true, also in traditional music texts, the variations that occur are limited by the srtucture of our perceptual-cognitive faculties as well as by cultural factors (Smalley, 1991b).

An approach of this type is used in analyzing sound objects in electroacoustic musical texts, starting from the very correlation of sample data obtained by presenting the sonoral material in question to different subjects. The use of the term sound object is not accidental; it provides a useful definition denoting a more complex event that any single musical element with its properties of frequency, timbre and duration. Later, we shall examine the typological and morphological problems of this class of sounds. Another aspect of interest is that the analysis of electroacoustic music does not coincide, as often happens also in the analysis of a great deal of instrumental music, with an examination of the compositional process of the work. Even if information on production strategies can sometimes be a useful instrument in corroborating the analytical information gathered from a hearing of the piece, it is not a main part of the analysis itself. How a composer has built sounds and mounted them in a formal articulation is only useful in emphasizing and correlating some physical characteristics of the sound objects which make up the work. Nothing more. This sort of warning, which in itself might seem banal, is due to the fact that, because of problems in defining sound objects and the typology of formal structuralization, many analyses often fall back on a reading of compositional processes, one of the few written sources available about an electroacoustic work. It has been stated that not the score but the scores, in much of synthetic music for example, represent an important source because they contain sonological information on the work itself. This statement seems a bit rash to me because scores contain acoustic data which often do not coincide with the performance of the sound phenomena as our ears perceive it. If we examine one of Shepard’s sounds (1982), a reading of its physical structure tells us nothing about the sound image we perceive (1).

In the second place, the analysis of electronic music must not correspond to the creation of aesthetic criteria in order to justify the existence of the electroacoustic genre. Often, and this is also due to a part of Schaeffer’s heredity, definitions of analytical frameworks also contain considerations of aesthetic character which have nothing to do with analytic methodology and, even more than that, weaken it. Even more so because one cannot consider the analysis of electroacoustic music as the investigative instrument of a single and homogeneous genre. Often, the only characteristic that many types of electroacoustic music have in common is their use of technology; they are quite different in their stylistic aspects. An analytical framework should furnish, in its various stylistic aspects, instruments with which to examine music demonstrating exceedingly complex sound phenomena which are inexplicable by means of theories used for instrumental music. This does not mean that analysis and aesthetics are two fields of study having nothing to do with one another. My present working hypothesis is that – given the embryonic state of the analysis of electroacoustic music- it is more important to focus attention on the aspects inherent to its structural characteristics.

ELECTRO-ACOUSTIC MUSICAL LANGUAGE: NATURAL AND MUSICAL RELATIONSCHIP
One important aspect of electroacoustic music is its acousmatic character. This definition derives from the lack of information which the listerner possesses regarding the source which produces these sounds. This allows us to place the sound material in relation to meaningful elements which may be realized according to the morphological characteristics of the sound objects heard. In addition, electroacoustic music is able to dip into a vast sound landscape, in which we can place at opposite poles natural sounds without tranformation and totally syntetic sounds. But this is only a part of the question.

The most important aspect is that we can abstract several characteristics of the micro-structure of natural sounds and make them belong to sound objects realized in studio, as for example, the utilization of concrete sound sources in order to use them for narrative purposes or to emphasize articulative properties. Thus, the signification process which the listener puts into action at the highest level of analytic practice concerns relationships that are quite different because sound material possesses properties that direct in two opposite directions, quite differently from traditional sound material. This process is called source bonding (Smalley, 1991b) and its use can be noticed in the development of sound material in electroacoustic music using the definitions mimetic or aural (Emmerson, 1987). The accentuation of one or the use of both in different ways characterizes some of the structural properties of the work: not only may they be used as a signification level, but also to model some properties of sound objects and their development. These aspects should not cause to forget another listening situation – defined by Schaeffer – which plays an important role in comprehending an electroacoustic work. This is called reduced listening (ècoute rèduite) and is, in other words, the disposition of the listener, in virtue of which his attention is focused exclusively on the sound object itself with no reference to the source causing its production.

All of these definitions are important since, as has already been mentioned, the starting point of the analysis of electroacoustic music is listening to it; the receptive attitude of the listener influences the work of analysis because these properties are not merely associative but are part of the structural potentiality of the electroacoustic medium. Among other things, one can imagine rather clearly that during listening and analytic pratice, perceptual-cognitive changes may occur. So, in some cases, both a referential reading (referring to several natural properties of the sonoral object) and a musical one (with reference to the intrinsic properties of the sound material) are equally valid and interconnected.

ANALYTIC TENDENCIES FOR A DEFINITION OF SONORAL OBJECTS AND THEIR STUCTURING
The analysis of sound objects and their articulation within an electroacoustic work can be divided into several tendencies which, for simplicity’s sake, I shall group into three main fields. Briefly, I intend to explain these approaches and shed light on their salient points. Contributions on the analysis of electroacoustic music are extremely few (Delalande, 1972, 1986; Stroppa, 1984) and often insufficiently systematic when faced with problems of analyzing the text. In this explanation I shall linger on three analytic methodologies which may be seen as interdependent and which, in my opinion, represent important contributions in this field. As may be noticed, one of the main problems is finding an analytic lexicon suitable for describing sonoral phenomena of the kind used in electroacoustic music (2).

SCHAEFFER AND SOUND OBJECTS
Any reflection on the theoretic and applied contributions in electroacoustic music must begin with the work of Pierre Schaeffer (1952,1966). Apart from temporal reasons – Schaeffer was one of the first to devote himself to music produced by electroacoustic means and one of the first to undertake a theoretical reflection on it – he is of interest because of several of his formulations which are still pertinent today. In his two books, A la Recerche d’une Musique Concrète and Traitè des Objects Musicaux, Schaeffer elaborates a theoretical framework, within which he attempts to confront both terminological and methodological problems.
This undertaking, later continued by other French scholars (Chion, 1983. Delalande, 1987), contains both compositive and aesthetic guidelines but also a series of analytically pertinent criteria which tackle the problem of perception of sound objects from an esthesic point of view. Schaeffer classifies sound objects according to typological and morphological criteria; the former indicate the general types of these objects, while the latter describe their characteristics (3).

Essentially, Schaeffer indicates three types of sound objects: continuous,iterative and impulsive. Another of Schaeffer’s important formulations concerns the three plans of reference (plans de reference) through which the “sound objects” are described and classified according to therir properties.
These three plans are:
1) the melodic or texture plan, the evolution of pitch in time;
2) the dynamic or form plan, the parameters of intensity in time;
3) the harmonic or timbral plan, the relationships between the preceding parameters as represented in their spectral components.

Each reference plan presents several systems of classification according to the type of melodic, dynamic and timbral movement. In addition, four classes of criteria are formulated: criteria of material, maintainment, form and variation; these are correlated in turn to the morphological characteristics of the sound objects. Other formulations concern material classification of the sound objects, their temporal length and center of interest, and musical classification, their nature and complexity. It is interesting to notice how these two types of classification make use of terms like segment, cell and group, terminologies which can be found in the vocabulary of the theory of segmantation of musical texts. This aspect allows us a glimpse of how some formulations belonging to this theoretical framework may be used in the segmentation of an electroacoustic music text. Unfortunately, it is to segmentation (or even better, to macro-segmentation) itself that Schaeffer and his followers do not pay enough attention.

Only one branch of Schaefferian theory, aural analysis (Thoresen, 1983; Delalande, 1989), seeks to define, more or less systematically, the criteria of segmentation in musical texts. Aural analysis (Thoresen, 1983) theorizes two levels of segmentation: time fields, concerning the significant segments into which a text may be hierarchicalliy divided, and layers, which in turn are hierarchic themselves and regard the content of the text’s timbral mass. Each aspect possesses a series of connective modes and parametric values. Another type of aural analysis is that which Delalande (1989) calls “analysis of listening conduits”, a kind of esthesic approach, which seeks to define segmentation in sound objects of a certain temporal length, the grouping of which is due to the presence of several specific traits. It is important to notice that, at least at the present time, both aural analysis and that of listening conduits are mainly employed for the analysis of instrumental music (4).

Even if the theoretical reflection put into practice by Schaeffer – and proliferated in other directions – contains aspects of great importance, it also presents many problems. First of all, this reflection derives principally from a type of musical context: concrete music (musique concréte); its descriptive criteria are too strongly influenced by ties with very precise sound phenomena, and in some cases they are lacking in generalization.

That this lack reveals limits becomes evident when these criteria must be applied in describing different sound phenomena, especially those obtained with technologically advanced instruments. I am referring to the great quantity of music employing hardware and software instruments, which permit the synthesis of sounds and their structuring within the composition. Another problem lies in the purpose of the treatment: too close to the compositive aspect – for example the formulations on manipulation and transformation – and too slanted toward sustaining aesthetic considerations (5).

In any case, Schaeffer’s work offers important aspects to anyone desirous of taking on the task of fixinf or theorizing analytic criteria for electroacoustic music.

THE SPECTRO-MORPHOLOGICAL APPROACH
An in-depth examination of Schaeffer’s theoretical affirmations comes from Denis Smalley, the New Zealand composer (1987). Smalley’s examination becomes a systematic re-formulation and enlargement of Schaeffer’s affirmations, preserving some of their original characteristics and conferring generality on many aspects. The bond between the two approaches is evident, especially in the relationship between spectral typologies and morphologies. In Smalley’s definition, however, these terms are more exhaustive; they can be applied to different musical contexts and include other categories such as that of motion.

The theoretical framework of spectro-morphology is articulated mainly in four parts: the typology of the spectra, morphology, motion and structuring processes. Each part contains classifications including a lexicon .flexibly suitable to descriptions of sound phenomena. One vitally important aspect of spectral typologies is their explicit relationship with morphological classifications. Especially relevant is the definition of the noise-note continuum subdivided into three principal elements: the noise, the node (an event having a more complex texture than a single pitch) and the note, which is in turn subdivided into note, harmonic spectrum and inharmonic spectrum. This classification of the minimal elements in the electro-acoustic lexicon represents a good starting point, in spite of some less well-defined points, as for example, the typologies “harmonic spectrum” and “inharmonic spectrum”.

Actually, each of these may include an infinity of sound objects that are each quite different from one another. Concerning this problem, we shall later see that a model is needed which allows us to realize a lexicon of minimal sound elements starting from their acoustic and perceptive properties. An integration of definitions belonging to other approaches will reveal its usefulness in better formulating these typologies. As far as morphological aspects are concerned, Smalley goes far beyond Schaeffer’s definitions, designating four types (attack-impulse, closed attack-decay, open attack-decay and graduated continual) and creating a series of models obtained through transformations like retrogradation. Since the elements present in works of electroacoustic music are quite often dynamic objects, i.e. made up of several morphological models, there enters the concept of morphological string, a string composed of two or more different morphological models. I want to emphasize just how useful these morphological strings are, both for an analysis by reduction and for segmenting a sonoral text into macro-elements. Perhaps the most complex portion of theory concerns the concept of motion (6).

It is classified into six categories (bi-directional, uni-directional, linear, curved-linear, reciprocal and centric/cyclic), correlated among themselves by diverse degrees of affinity and subdivided into subcategories. The definition of categories of motion is of great importance since these categories can be related to single sound objects as well as to sections of a work. Smalley indicates this difference with the terms texture and gesture, to indicate whether these categories refer to the internal structure of a sonoral object or whether the reference is to a movement of greater breadth, a section of a work. This theoretical framework is further integrated by other formulations concerning structural functions, descriptions of formal processes, and other aspects such as functional chains, structural relationships and elements describing the spatialization of the sound material.

I believe that description, brief though it may be, has emphasized the structure of the theory and accented its most important characteristics, at least for purposes of analysis. The creation of lexicon of sound objects and their structuring represents an important and sufficiently eleborated part of this theory, for example the relationships between “structure of objects” and “motion of formal structures”. Besides, the generalization of many definitions leads us to believe that this theory is an excellent and advanced starting point for the construction of an analytical theory of electroacoustic music.

THE PERCPETUAL-COGNITIVE AND ESTHESIC-COGNITIVE APPROACHES
The definitions explaining these approaches point to their principal characteristic: they are based on the comprehension and explication of the mental processes underlyng musical perception and cognition, with an eye toward listening processes. In spite of the growing number of studies on musical psychology, few have been oriented toward the perception and cognition of complex sound structures like those of electroacoustic music, except for studies on timbre and its aspects. Some studies, for example those of McAdams (1983), present interisting elemts and have provided the starting point for the creation of models for the analysis of electroacoustic music like that formulated by Doati (1991). One important concept in McAdams’ theory is represented by auditory images and their properties of “fusion” and “fission”.

Coherent auditory images are classified in four types:
1) harmonic spectra,
2) spectra with shifted harmonics,
3) spectra with stretched harmonics,
4) spectra with compressed harmonics.

The latter three typologies, describable though precise models, furnish a useful instrument for the calssification of spectra in a category that is too general, like that of inharmonic spectra. McAdams’ formulations are not only interesting for their definition of the concepts of “fusion” and “fission”, but also because they are based on an explicit correlation between acoustic and perceptual properties. Naturally missing are indications for research into structures at high level, those, that is, which articulate the formal design. Doati (1991) has proposed a model based on McAdams’ assumptions, which is articulated in five phases: the reading of the acoustic surface, the organization of coherent auditory images, the segmentation and extraction of a lexicon, the construction of structural relationships and the attainment of a musical dialogue.

Even if an approach of this type touches on the main aspects of the analysis of electroacoustic music, it does not well define, for example, the types of structural relationschips, the type of segmentation and its levels. Some of these characteristics derive, in my view, from the fact that this model is mainly oriented toward defining compositional strategies. This approach, here called perceptual-cognitive, offers, as we have seen, many interesting elements for defining minimal elements (for example, the relationship between physical and perceptual properties), but it does not furnish sufficient elements for the construction of analytical strategies on the formal structure. Another approach which refers explicitly to the analysis of esthesic-cognitive strategies is the one developed by Giomi and Ligabue (1991). The principal theoretical assumption of this methodology is that of considering the sound object as a unit of first articulation made up of units of second which act as distinctive traits.

The categories of traits used are a selected and integrated class of those proposed by Schaeffer. A series of multiple listenings to electroacoustic woks is used to verify convergencies and discrepancies in the listening process and also in the process of identification of sound objects in relation to the methodological assumptions illustrated above. Thus, a classification can be compiled of the pertinent traits of the sound objects and their possible variations, putting in order and integrating the classification attained by Schaeffer in his Traitè. CONCLUSIONS An examination of problems facing the analysis of electroacoustic music and a brief treatment of the various analytical tendencies allows us to draw a precise picture of the present state of this field. In conclusion, several considerations come to the fore, none of which is conclusive, given the extreme fluidity of the situation, for the possible orientations which, in my opinion, should be followed in creating a theory and relative analytical model of electroacoustic music. Each of the contributions we have examined contains interesting aspects, even if at various degrees and levels; the concept of sound object and some musical classifications in Schaeffer; Smalley’s structure of theory and typologies; some hypotheses of models of segmentation and fusion of sound objects in the theory based on perceptual processes; and some aspects of semiotic methodology in defining sound objects in reference to the esthesic processes (7).

The problem of creating a lexicon for electroacoustic music, and not only for that, can be found, I believe, in an integration between the definitions of spectro-morphological theory and the models for the fusion of auditory images proposed by McAdams. One practical example could be the integration of the typologies of inharmonic spectra formulated by McAdams with that defining the inharmonic spectrum as indicated by Smalley. In this case, models based on theories which place acoustical elements in relation to perceptual elements are already sufficiently reliable in describing sound elements.

If we go beyond the problems of lexicon and of minimal elements in the sound text, we find ourselves facing problems such as segmentation and srtuctural relationships. I believe, for example, that a theory starting from cognitive basis like that of Lerdahl and Jackendoff (1983), can offer interesting suggestions, even if it does not specifically deal with problems related to this type of sound phenomena. Morphological strings and the definition of typologies of motion together with the elements that make it up, can already furnisch a basis for a direct application on electroacoustic texts. A piece of electroacoustic music can be examined using these definitions, for example the typologies of motion, in a way that underlines several characteristics, both of its global form and of the structure of the single sections. Focusing attention on the typology of motion should not be considered uncommon since it represents one of the aspects that characterize a great deal of electroacoustic music (8).

Another hypothesis is that commencing with the rules of segmentation, for example “elision” and “superimposition” of Lerdahl and Jackendoff’s theory, in order to apply them in the analysis of structural groupings of the sections composed of sound objects. In conclusion, I believe that the analysis of electroacoustic music possesses the theoretical bases needed to begin the work of systematization. We can go beyond the phase of pure and simple segmentation and the description of sound objects and their relationships and develop and apply concepts concerning the hierarchy of structural elements, motion and directionality of structures, and the definition of a lexicon for the various levels of analytical description.

The lack of consolidated notation is not a handicap at all, on the contrary, it is a strength . With a rather provocative (but not so very) statement, I would like to say that a theory and an analytical model exclusively focused on the study of the sound text represent a real point of contact between musical theory and the modelling of perceptual and cognitive masical strategies. By possesssing these properties, the analysis of electroacoustic music is a field of study full of fascination, even if it is difficult and complex.
NOTES
(1) A tipical example of one of Shepard’s sounds is the infinite glissando. An effect produced by its spectral content which emphasizes how we differently perceive the pitch and the croma. In any case, a phisical reading of its sound would give us no information concerning the effect it produces.
(2) This represents a similar problem for another musical genre based on technological instruments, pop music. I quote part of an example found in Middleton (1990), regarding an attempt at describing some types of sonority:”…a “dry” acoustic guitar tracks is mixed “un front” while the drummer asks for “wet, muffled” sound and the producer wants the Fender piano to be “heavier” and more “explosive”, not so “rickety tin can””.
(3)An important text for getting one’ bearings in Schaeffer’s Traitè is Michel Chion’s Guide, which contains a rational annotation of key concepts and summarizing tables on the classification with relative exlpanations.
(4) It would be interesting, especially in the analysis of contemporary instrumental music, to analyze the structure of some sonoral situations by rferring to their morphological structure without using traditional musical concepts.
(5) Another of the aspects of the Traitè is the furnishing of a basis for the justification of the existence of electroacoustic music.
(6) In this context these terms are defined as modulations: vibrato, portamento, and random vibrato or shimmer.
(7) Both this approach and Smalley’s already offer a development and integration od Scaffer’s theory, as can be noticed from their description.
(8) Some of the pieces which could be analyzed from this point of wiev are Vortex by Smalley (see Levis, 1983a, 1983b), Kontakte be Stockhausen, Parmegiani’s La Creation du Mond, and Terminus II by Koenig.
REFERENCES
BELKIN, A., 1988. “Orchestration, Perception, and Musical Time”, Computer Music Journals, 12 (2)
CAMILLERI, Lelio, 1991. “Musica senza notazione: riflessioni sull’analisi della musica elettronica”, Analisi Musicale, ed. by M. Baroni and R. Dalmonte (Milan: Unicopli)
CHION, Michel, 1983. Guide des Objects Sonores (Paris: INA/Buchet-Castel)
CHION, Michel, 1988. “Du Son à la Chose: Hypothèses sur l’Object Sonore”, Analyse Musicale, 11.
DELALANDE, Francoise, 1972. “L’Analyse des Musiques Electro-acoustiques”, Musique en jeu, 8.
DELALANDE, Francoise, 1987. “L’Analyse des Conduites Musicale: une Etape du programme Sèmiologique?”, Semeiotica, 66 (1/3).
DOATI, Roberto, 1989. “Analisi dell’opera Glissandi di György Ligeti”, in Analisi Musicale, ed. by M. Baroni and R. Dalmonte (Milan: Unicopli).
EMMERSON, Simon (ed.), 1987. The Language of Electroacoustic Music (London: Macmillian).
EMMERSON, Simon, 1987. “The relation of Language to Materials”, in The Language of Electroacoustic Music, Emmerson, S. “Polarised Listening Strategies for Electroacoustic Music”, paper presented at the II European Convention of Musical Analysis, Trento. (ed.), (London: Macmillian).
GIOMI, Francesco, and LIGABUE, Marco, 1991. “Un approccio estesico-cognitivo alla descrizione dell’Object sonore”, paper presented at the II European Convention of Musical analysis, Trento.
HEIKINHEIMO, Seppo, 1972. The Electronic Music of Karlheinz Stockhausen (Helsinki: Acta Musicologica Fennica).
TEN HOPPEN, Christiane, 1991. JANZEN, Thomas, 1986. “Aestethic Appeal in Computer Music”, Computer Music Journal, 10 (3).
LERDAHL, Fred, JACKENDOFF, Ray, 1983. A Generative Theory of Tonal Music (Cambridge MA: The MIT Press).
LEWIS, Andrew, 1983a. “Motion and the Analysis of Electro-Acoustic Music: Denis Smalley’s Vortex (I)”, Electro-Acoustic Music, 3 (3). LEWIS, Andrew, 1983b. “Motion and the Analysis of Electro-Acoustic Music: Denis Smalley’s Vortex (II)”, Electro-Acoustic Music, 3 (4). MANNING, Peter, 1987. Electronic and Computer Music (Oxford: Oxford University Press).
MIDDLETON, Richard, 1990. .Studying Popular Music (Milton Keynes: Open University Press).
McADAMS, Stephen, 1983. “Spectral Fusion and the Creation of Auditory Images”, in Music, Mind, and Brain, Clynes M. (ed.) (New York: Plenum Press).
SCHAFFER, Pierre, 1952. A la Recherche d’une Musique Concrète (Paris: Editios du Seuil).
SCHAFFER, Pierre, 1966. Traité des Objects Musicaux (Paris: Editios du Seuil).
SHEPARD, R., 1982. “Structural Representation of Musical Pitch”, in D. Deutsch (ed.), Psychology of Music (New York: Academic Press). SMALLEY, Denis, 1987. “Spectro-morphology and Structuring Processes”, in The Language of Electroacoutic Music, Emmerson S. (ed.), (London: Macmillian).
SMALLEY, Denis, 1991a. Personal communication.
SMALLEY, Denis, 1991b. “Can Electro-Acoustic Music Be Analyzed?”, paper presented at the II European Convention of Musical Analysis, Trento.
STROPPA, Marco, 1984. “Sull’analisi della musica elettronica”, Bollettino Limb, 4.
THORENSEN, Lasse, 1983. Layers and their Functions, Unpublished manuscript.

Download the article
L. Camilleri
Electro-acoustic music:analysis and listening processes

© 2005 sonus publishing/sonusedizioni

합성이론 9장 : VCA 에 대하여

합성이론 9: VCA 개요


제가 VCA에 대해서 언급했었던 기억이 납니다. Voltage Controlled Amplifier,  ‘VCA’, 라 불리는 이것은 analogue synthesizer 에서 아주 중요한 역할을 해 줍니다. Digital로 컨트롤되는 analogue synth는 ‘DCA’라는 것을 대신 사용하게 됩니다. 반면 pure digital synth와 sampler들은 time-variant amplifiers인’TVA’, 을 가지게 됩니다. 따라서 이번에는 VCAs에 초점을 맞추어서 여러분이 원하는 소리를 만드는데 보탬이 되어 보도록 하지요.

1. 두가지 형태의 VCA?
첫째로 우리는 참 혼동, 또는 착각하기 쉬운, audio signal chain에서의 amplifiers와 Voltages를 컨트롤하기 위한 것과 차이를 알아야 합니다. 합성이론 3장에서 보았었는데 여기 위의 그림 1을 보면, tone generator 와 amplifier로 구성된 단순한sound generator를 볼 수 있습니다. 분명 이 그림에서 amplifier의 역할은 hi-fi나 car stereo에서 하는 amp와 비슷한 역할을 하겠지요. 이것은 신호를 더 크게 해주겠지요.

Amplifier란 무엇일까요? 만약에 여러분이 기타리스트라면 아마 이것은 소리를 증폭하기 위해 distortion이나 다른 effect를 만드는 것이라고 말할 수도 있겠지요. 만약에 여러분이hi-fi 에 관심이 많은 사람이라면 이것을 거대하고 제어하기 쉬운 값비싼 power supply라고도 말할 수 있지 않을까요? 그러나 엔지니어라면, amplifier은 입력신호의 amplitude A1 이 모양의 변화 없이 A2로 변화시키는 장치라고 말할 것입니다.

하지만 분명하게 중요한 것은 A2 가 A1보다 크다는 것입니다. 즉, 신호는 전보다 커진다는 것이죠. 하지만 만약 그 A2 가 더 작다면 소리는 더 조용해지겠죠. 단순하죠? 그러나 이렇게 간단한 문제로 치부할 것은 아니랍니다. 우리가 알아야 하는 건 ‘어떻게 하는데?’입니다. 이것을 아주 단순하게 정의해보지요. A2 에서 A1 의 비율을 계산하고 그것을 amplifier의 ‘Gain’ 이라고 해봅니다.(Equation 1). 예를 들어 A2 가 A1의 두 배라면, 그 비율은 ‘2’ 가 되고 Gain은 2가 되겠지요. 하지만 이렇게 2배가 한다고 2배로 소리가 커지는 것이 아닙니다. 우리의 귀는 이런 방식으로 소리를 처리하지 않습니다.

하지만 지금은 단순하게 그냥 라디오에의 volume control 에 대해서 생각해봅시다. 여러분이 소리의 볼륨이 증가하고 감소한다고 이야기 할 때에 Gain이 조정되는 것임에 분명합니다. Figure 2는 이런 조절의 예를 보여줍니다.

이 그림에서 receiver회로는 low-amplitude signal을 만들며 이것은 직접 preamplifier를 통과합니다. 이것은 signal을 ‘line level’로 증폭합니다. 출력은 이때에volume control을 통과합니다. volume 손잡이가 시계방향으로 끝까지 올라가면 signal은 변화가 없고 line level에서 the power amplifier로 통과합니다. 그리고 시그널은 그 레벨로 증폭됩니다.
정리해볼게요.

1.preamp로 거치는 low-amplitude signal생성
2.signal을 line lever까지 증폭
3.volume control로 감
4.볼륨을 최대치로 해도 시그널은 영향을 받지 않고 line level에 통과
5.power amp로 간다
6.그 레벨로 소리 증폭

그러나 여기서 volume control을 버리면 소리가 줄어들게 됩니다. 그 이유는 위의 3번에서 volume control로 쓰였던 potentiometer가 signal의 양을 점차 줄이면서 power amp로 가게 됩니다. 이 볼륨 손잡이는 amplifier의 하나로 정의되어야 합니다. 이것의 Gain은 1을 넘지 않고 off일 때는 Out/Input비율은 0이 되는 것, 즉, Maximum일 때는 그 비율이 1이 된다는 말입니다.
따라서! Gain은 언제나 0과 1에 놓인다는 사실을 알았습니다. 그리고 이것을 Attenuator(감쇄기)라고 합니다.

중요한점은 하나 이상의 amp와 attenuator를 연결하여 사용할 때 전체 system의 Gain은 각각의 Gain을 곱하여 나오게 됩니다.
(pre-amp Gain * attenuator Gain * power Gain)

,2. 더 개선된 회로도 만들기
위의 그림 2는 이해하기가 제법 쉽지요. 하지만 그 회로는 단순하게 예를 들어 설명하기 위한 것이라서 썩 좋다 볼 수는 없겠습니다. 왜냐하면 audio signal 이 그냥 attenuator를 통과하는 것에 불과하니까요. 사실상 이런 회로는 distortion과 같은 부작용을 일으키기도 합니다. 이런 것을 피하기 위하여 더 나은 회로도를 작성해 볼 필요가 있습니다.

Figure 3이 바로 그런 것입니다. 이것은 같은 요소로 이루어져있지만 volume control 손잡이가 지금 voltage source를 줄여서 preamp의 Gain을 조절합니다. 다시 말해서 우리는 preamp를 Voltage Controlled Amplifier로 정의 내릴 수 있다는 것이 됩니다. 그리고 그audio signal 는 더 이상 직접적으로 volume control을 거치는 것이 아니라는 이야기입니다. 일종의 안전장치같이 말이에요.

3. Radio 에서 Synthesizer까지
이제 라디오에서 단순한 analogue synthesizer까지 가보는 겁니다. 분명, receiver 회로 에서 만들어지는signal은 어떤 것이든 가능하겠죠? 마이크로 들어오는 연설이 될 수도, Beethoven심포니 일수도, saw tooth wave일수도 있다는 말입니다. ‘receiver 회로’ 를 ‘tone generator’라는 말로 바꾸어 보는 것은 어떨까요? 그리고, 그 다음에 그림에 있는 power amplifier를 생각해봅시다.  몇몇의 합성기기에 있는built-in speakers들은 이런 power amp를 가지고 있지요. 그러나 대부분은 외부장비를 사용해서 쓰지요. 그 결과로 우리의 block diagram에서 ‘power amplifier’가 빠져있습니다.

지난번 합성이론 3장에서 우리는 volume control을 controller circuit로 이야기 했던 기억이 나시는지요. 위의 그림 4에서는 지난번처럼 했던 것과 같은audio signal을 보여줍니다. 그러나 preamplifier (줄여서’Amplifier’) 는 contour generator에 의해 조절되고 trigger에 의해 trigger됩니다. 이렇게 Figure 4 는 Figure 3는 아주 달라 보입니다. 그러나 이것은 generator, amplifier, volume control의 관계상에서 다른 점이 없습니다. 그럼 자세히 살펴보지요.

Tone Generator 가 ±2V로 시그널을 만들어 낸다고 해봅시다. 또한 contour generator 가 0V 에서 +5V 로 ADSR envelope을 만들었다고 해 봅니다. 그리고 VCA 가CV input에 0V가 올 때 아무것도 출력하지 않는다고 가정하면, 그것은, CV 입력이 +5V일 때에 ±10V의 maximum audio signal amplitude을 출력합니다. 즉, 그 amplifier는 maximum Gain 이 5 (G=10V/2V) 이며, minimum Gain 은 0 (G=0V/2V)이라는 이야기가 됩니다. 그리고 마지막으로 amplifier의 response 은 ‘선형적’입니다. 예를 들어CV input이 1V 은 Gain 2; CV input 이 2V면 Gain of 4를 만들어낸다. 이것은 즉, audio signal Gain 의 양은 amplifier의 CV input에 적용되는 contour 의 level에 비례한다는 것. 복잡해 보이지만 Figure 5를 보면 더 분명해집니다.

4. Initial Gain
이만하면 충분하지 않냐구요. 않습니다. 하하..여러분 synthesizer에 판낼에  ‘Initial Gain’ 또는 ‘VCA Gain’라고 표시된거 본적 있으신가요? 이것이 바로initial CV, 또는 ‘offset’을 Contour Generator 가 만드는CV 에 첨가하는 것입니다. (Figure 6,).

그래서 예를들면 우리가 Figure 5 에서Initial Gain +3V을 더했다면, 우리가 얻는 것이 바로Figure 7이 됩니다. 이 offset은 즉각적인 효과를 가집니다. —  VCA 는 항상 0보다 큰 Gain을 만들어 내라고 말합니다. VCA 에 보여진 CV 는 계속적인+3V offset을 가집니다.

5.숨좀 돌리고...
지금까지 본 것을 가만 생각해보면 VCA 는 끝이 없이 증가할 수 있는 것처럼 보여집니다. 다시 말하면 여러분이 얼마만큼 증폭을 원하는지에 상관없이 이놈이 계속 증폭시켜줄 것처럼 보입니다. 하지만, 물론 이것은 불가능하지요. 만약에 능력 이상의 것을 요구하면 이것도 특정 방식의 distortion을 만들어냅니다.

다시 Figure 5를 봅니다. VCA 의 maximum output 을±10V로 했었지요? (CV 가 +5V일 때) 즉 우리가 만약에 0V 을 +8V contour를 동반한 +5V ADSR contour로 대체하면(Figure 7)? 분명 그 contour의peak 에서, VCA 는 시도는 하겠지만 ±16V 를 만드는것에는 실패하게 됩니다. ±10V이상을 넘지 못하기 때문에, signal 은’clipped’됩니다.(Figure 8)
ADSR contour의 주기 동안에 발생하는 출력 wave를 잘 살펴보면 우리는 원래 입력 시그널의 sawtooth shape 모양이 더 이상 출력에서 찾아볼 수 없음을 알게 됩니다. 그 Wave의 ‘정상’ 은±10V 제한위로 증폭하지 못하는 amplifier에 의한 Attack 과 Decay 단계에서 다시 구정되고 그 결과는 harsh distortion (‘clipping distortion’) 을 가집니다.

6.VCA의 또다른 사용!
지금까지 이야기한 VCA 는 모두audio signal path에 있습니다. 그러나 현실적으로 VCA의 대다수는 여기에 쓰이지 않습니다. 그것은 sunthesizer내에서 control voltage path로 쓰입니다.

Figure 6을 다시 봅니다. 이것이 Attack 의 끝에서maximum +5V를 가진ADSR contour를 출력한다는 것을 기억하시지요. 지난달 이야기했던 것으로 되돌아가 봅니다. analogue contour generators의 다수가 Attack끝에 level을 더 이상 제어하지 않았던 것 기억하시나요?( A, D, S, R의 setting과 관계없이), 그Attack Level은 언제나 +5V 입니다.(또는 장치의 성능에 따라 다름). 우리가 이미 이야기했듯, 이것은5의 gain을 VCA에서 만들어 냅니다. 그러나 우리가 드라마틱하게 signal에 영향을 미치고 싶지 않을 경우에는 어떻게 해야 할까요?
Figure 9를 봅니다. 보다시피VCA 이 control signal path에 있습니다. 이것은 CV에 의해 제어되고 그것은 또한attenuator에 의해 제어됩니다. 이VCA 는 Gain을 (attenuator에 위치에 의해 결정되는) ADSR contour에 적용시켜서 여러분은 그것을 shape의 변화 없이 attenuate/amplify 할 수 있게 되는 것입니다.

그 결과는 만약, ADSP이 단지 signal amplifier 에 영향을 주는CV라면 별로 중요하지 않을 것입니다. 결국, 그amplitude를 줄이는 것은 synth 의 최종 출력을 줄이는 것과 별로 다를 것이 없겠지요. 단순히 외부 amplifier를 줄이는 것과 같이 말이지요. 하지만 그 ADSR 는 단순한 CV가 아닙니다. 그림 9에있는 Initial Gain처럼 여러분은 LFO또는 다른controller들로 audio signal amplifier의 action을 변화시킬 수 있습니다. 이 경우에 ADSR 은 initial level 과 다른 modifier와 관련된 signal을 변화시키는 ‘범위’를 변화시킵니다.

이것도 복잡합니다. 더 분명하게 하기 위해 다른 예를 보지요. Figure 10 은 Figure 9와 비슷하지만, CV들이 amplifier 가 아닌 low-pass filter를 제어하고 있습니다. 보시다시피,  꼭대기에 있는fader 는 지금 filter의initial level이고, 이것을 ‘cutoff frequency’라 부릅니다. 그리고VCA 는 여기에 적용되는contour를 제어합니다. 분명히 여러분은filter가 여러분이 만드는 소리마다 전부 적용되는 것을 원하지 않을것이지요. 따라서 대부분의 synth 는 여러분이 VCA를 사용하여 그 contour를 attenuate 하게 해 줍니다. 보여진것 처럼요.

7. Real VCAs
VCA에 의해 연주될때의 중요점을 우리가 sunthesis를 생각할때에 줄곳 지나치게 되게 됩니다.참으로 우리가 많은 analogue synths의control panel들을 보면 ‘Amplifier’, ‘VCA’라고 쓰여진 것이ADSR contour generator에 속해있고 아마도 envelope level control ( ‘Amount’) 또는 Initial Level control 로 사용되는 것을 볼수 있습니다. 이것이 바로 많은 경험이 없는 사람들이 그것을 contour generator (CV path의 VCA) 또는 audio signal amplifier 와 구분하는데 힘들어 하는 점입니다. 이와 비슷하게, filter senction은 종종 second contour generator 와 또 다른 ‘Amount’ control 을 종종 포함할 것입니다. 이말인 즉슨 물론, VCF section 에도 VCA들이 있다는것입니다.

Figure 11 은 4개의 input과, 5개의 손잡이, 그리고 하나의 출력을 가집니다. 2개의signal input(SIG 1 IN, SIG 2 IN) 이 있고 그것은 module에 signal mixer가 있다는 것을 의미하기도 합니다. 또한 두 개의CV input이 있고 그것은VCA의 amplification을 조절하며, 이 말은 그 안에 CV mixer가 있다는 것을 의미합니다. 그CV input들은CV1-IN LIN 와CV2-IN LOG라고 표시되어있습니다. 각각 입력은 level control이 서로 연관 지어져 있으며 Initial Level control은 우리가 위에서 이야기한것과 같습니다. Figure 12처럼요.

우리는 이것을 단순하게 VCA라고 하지는 않습니다.

8…그리고 마지막으로
끝내기 전에 한가지 더 생각해 볼 것이 있습니다. 우리는 여태껏 main signal path를audio signal path로서만 생각해 왔습니다. 그러나 우리가 그림 4, 5, 6, 8, 9, 10, 12에서 CV들을 사용하는 것을 멈추면?

VCA의 가장 흔한 사용목적 중 하나는 또 다른 CV를 사용해서 CV의 Action을 변화시키는 것에 있습니다.using other CVs;
그리고 또 중요한 것은

반드시 audio signal의 amplifier와 control voltages를 변화하는것에 사용되는것들과의 구분을 두어야 한다는 것입니다.(맨처음에 말했듯이)
그리고 여러분이 하나 이상의amplifier 또는 attenuator를 가질 때 마다, 단순하게 그 각각의 Gain들을 함께 여러 개 사용하면, 언제든지 전채의 시스템의 Gain을 계산할 수 있다는 것입니다.
이것이 자유롭게 patch될 수 있는 synth가 아주 강력한가에 대한 하나의 예가 될 수 있겠습니다. 따라서 여러분이 다른 또다른 신호를 이용하여서 아주 다이나믹하게 어떤 신호든 제어하고 변화시킬 수 있다는 것입니다.