Page  467 ï~~A Systematic Approach to the Analysis of Music for Tape Rajmil Fischman Keele University, UK Tel 44-1782-583297/Email: ABSTRACT: This paper discusses a methodological approach to the analysis of tape music which draws concepts and techniques belonging to two main disciplines: Music Semiology and Spectromorphology. The former provides the concept and method of semiological neutral inference - as applied by J. J. Nattiez - by adopting a 'bottom-up' approach in order to find paradigmatic axes and determine hierarchical structure. The latter - developed by D. Smalley - offers a variety of analytical tools for the characterisation of sound and timbre and introduces the idea of multilevel functionality of the three temporal phases. Introduction Undoubtedly, the tradition of music for tape (whether it is called electroacoustic or acousmatic, etc.) has produced a significant repertoire during the last fifty years. However, when one compares theoretical work in this area to that concerned with contemporary western instrumental' music, relatively very little has been done to develop systematic tools for analysis of works for the medium. Although some of the tools applied to instrumental music can still be useful and effective in the case of certain tape pieces, there are various issues particular to the latter genre which require a complementary - sometimes alternative - systematic approach. In the first place, except in extreme cases, most instrumental music is presented as a set of instructions in a prescriptive score. In a typical semiotic process, this score corresponds to the trace, which then has to undergo another semiotic level - corresponding to the activity of the performer - in order to reach the aural realisation of the piece. This process is normally bypassed in the case of tape music: even if there is a score, this is usually rather descriptive, except, perhaps for any diffusion indications. However, as much as the process of diffusion is a vital part of the performance of music for tape, its role is rather different from that of the actual realisation of the piece by instrumentalists: in a piece for tape, the sound is already there, whereas an instrumental performer must convert an abstract set of instructions into sound 'from scratch'. Thus, as a result of the existence of a prescriptive score the analyst is provided with an abstraction of the piece outside real-time which must carry enough information for its realisation, at least within a particular cultural context based on specific performance practice. Inevitably, this out-of-real-time trace is bound to aid the clarification of structure and can even shed some light on the poietic processes (for example, the inference of a series in twelve-tone music). In the second place, tape works may include sounds which require an extension of the categories used for instrumental music. For example, the concept of timbre may have to be extended beyond the idea of interaction and combination of a given set of instrumental sounds to a multi-dimensional continuum. Within this continuum pieces for tape may articulate a subset common to most instrumental music, in 'For practical reasons, and to avoid unnecessary repetition, the term 'instrmmental' will also apply here to the human voice and to electronic devices which are compositionally treated in the same way as non-electronic western instruments. I C MC P R OCE E DI NG S 199546 467

Page  468 ï~~which case, instrumental analytical techniques may still be appropriate. However, the latter lose their effectiveness when this coincidence does not take place. It is therefore necessary to resort to theoretical tools which can solve this problem. As will be shown below, spectro-morphology provides a powerfiul environment and vocabulary which serves this purpose. Finally, the discourse of tape music may function at a mimetic level, imitating 'not only nature but also aspects of human culture not usually associated directly with musical material (Emmerson, S.)'. The fact that physical objects may be recognised adds an extra dimension to structure, corresponding to what Wishart has termed 'metaphor': a piece of music may use recognisable objects to articulate a structure based on the inter-relations of these objects not only at an aural level but rather because of the physical objects and landscapes they represent. Analysis of pieces articulating mimetic discourse requires tools emanating directly from the human capacity for sound recognition. The Analysis Problem According to Nattiez (1990), music analysis can be articulated in the poietic, esthesic and neutral dimensions. In essence, the issues encountered in the first two are roughly common to all music: the poietic level depends on the availability of information regarding the composer's strategies, state of mind, etc.; the esthesic dimension is related to research in music perception. Both of these are beyond the scope of this paper. However, analysis at the neutral level is directly dependent on the actual musical material and its physical articulation; therefore, it is certainly relevant to the current discussion. Nattiez's approach to neutral analysis involves the partition of a work 'according to abstract paradigmatic axes, that is axes which group together identical or equivalent units from an explicitly stated point of view (1982)'. The question is what criteria can one apply in order to establish a paradigm? In order to answer this, it may be helpful to revise the points raised in the previous section, where it was stated that musical discourse can function at various levels. It was first mentioned that, in certain instances, a tape piece may be articulated within a subset of all of sound properties which coincides, more or less, with that of instrumental music. In this case, paradigms used in the latter may be applicable. For example, concepts such as pitch-class organisation may be successfully - although not exclusively - applied to the analysis of Jonathan Harvey's (1980) Mortuos Plango, Vivos Voco (Griffiths, P.). It was also stated that tape music may work at a mimetic level. In this case, it is a matter of establishing relationships between recognisable objects: 'by articulating the relationships between the sound-images we could develop not only sonic structures... but a whole area of metaphorical discourse (Wishart, T)'. Although, it must be admitted, the establishment of structural relationships between sound-objects is by no means trivial, 'all listeners have considerable practice at the concrete aspect of daily life, while an abstract approach needs to be acquired' (Smalley, D.). It is then reasonable to assume that the skills required for analysis at a mimetic level are already in place in the listener's psyche - albeit in an embryonic state - and only need to be developed. However, this still leaves works which do not employ mimetic discourse and encompass a rather different subset of timbre than instrumental music. Furthermore, 'regardless of any other levels of paradigmatic coherence, music can work at an ultimate abstract level: the structuring of time 468 8ICMC PROCEEDINGS 1995

Page  469 ï~~(Fischman, R)'. This is usually achieved by articulating abstract characteristics of sound; therefore, a theoretical model which can describe these structural characteristics may provide the right tools for analysis at an abstract level - the theory of spectro-morphology (Smalley, D.) provides such a model. Spectro-morphology In addition to its exhaustive typology and classification of the various aspects of timbre, spectromorphology is ideally suited to analysis at the neutral semiological level, since it advocates that conclusions must be arrived at by direct aural evaluation of the material. While an exhaustive description of spectro-morphology is already well documented by Smalley, it may be useful at this stage to summarise those aspects which facilitate abstract analysis: 1. It provides a set of continua which can be used as criteria for the classification of sounds by comparing their intrinsic characteristics. The main spectral continua consist of the pitch-effluvium, attack-effluvium and gesture-texture. 2. It provides an exhaustive set of basic morphological archetypes characterised by three temporal phases: onset, continuation and termination. 3. It provides a strategy for the formation of structure by means of a bottom-up process: spectral components are combined to form the basic meaningful units characterised by the three temporal phases. The combination of these basic units then produces a directional tendency at a higher level (which can again be characterised according to the temporal phases) and so on, until the last hierarchical level encompasses the attack, continuation and termination for the whole piece (see figure 1). 4. It provides typologies for the behaviour and interaction of sounds within the different spectra and in physical space. 5. It provides a terminology essential in the establishment of a vocabulary for the articulation of analytical discourse. Figure 1. Bottom-up formation of structure level N structure (experiential musical time and motion) T level 0 directional tendencies (combination of morphologies, combinations of combinations, etc.) spectral types morphologies (combination of spectral components) (basic temporal shapes) \ pitches (spectral components) Methodology The aims of the analytical approach are listed below: I. to establish structure by autrally identifying the hierarchical levels according to the three temporal phases, as explained above. I CM C P ROC E E DIN G S 199546 469

Page  470 ï~~2. to establish sonic relationships along and across hierarchical levels identifying paradigmatic axes. 3. to establish general directional tendencies within the various spectral continua along and across hierarchical levels. A strategy for the achievement of the aims above may consist of the following steps: 1. Determination of the basic meaningful sonic units and classification of these into families according to spectro-morphological characteristics: pitch-effluvium, attack-effiuvium, gesturetexture and morphological archetypes. 2. Assignment of a symbol to each of the sound families and production of a graphic score. 3. Segmentation of the piece according to hierarchical levels. A possible implementation may assign two numbers to each segment: the first number corresponds to the hierarchical level (0 corresponds to the bottom of the hierarchy) and the second corresponds to the temporal appearance of the section. Hierarchical level 0 may be reserved for combinations of sounds which, while being morphologically complete, may not, on their own, create directional tendencies. 4. Description of each level, pointing out relationships and identifying paradigms and summarising at the end of relatively higher hierarchical sections (e.g. level 3 upwards). 5. Overall summary comprising the following issues: * description of the main paradigms and listing of paradigmatic axes. " summary of overall directional tendencies for the higher hierarchical levels (e.g. level 3 upwards). " summary of overall specro-morphological tendencies in the various continua (pitch-effluvium, attack-effiuvium, gesture-texture) as well as within spatial articulation. Example: Analysis of CROSSTALK In this case, examples taken from an analysis of CROSSTALK (Vaughan, M.) are used to illustrate the various steps described above: 1. Table 1 contains the sound family classification. 2. Figure 2 contains the graphic score for the first minute of the piece resulting from the above classification. 3. The segmentation into hierarchical levels (0 to 5) is also shown in the score (figure 2). Note that from level 1 upwards, all sections are labelled according to their temporal function. Also note how combinations of onset-termination or onset-continuant-termination constitute the next hierarchical level. 470 0IC MC P R OC EE DI N G S 1995

Page  471 ï~~Table 1. CROSSTALK - soundfamilies Sound Graphic Morphology Pitch Attack Gesture- Brief family symbol continuum continuum texture description continuum A closed noise discrete gesture noise burst attack-decay__ B attack node discrete gesture more or less impulses 1 metallic blows noise C graduated noise effluvium texture continuant D graduated inharmonic effluvium texture gliding sound, continuant note inhannonic, f sometimes 'Ge. begins with **,." node iteration 'e';s" which fuses *, _____into effluvium E attack-decay harmonic note effluvium texture/ continuous -texture setting pitched drone, sometimes glissando F attack-decay note proper discrete ostinato pitch belonging to a mode or scale 4. Description of each level. This can seconds of CROSSTALK: be exemplified by the following description of the first 16 Level 5: Section [5.1] (0)2 (0:00 - 1:42) This section is the onset of the whole piece. It consists of two subsections: onset [4.1] and termination [4.2]. Level 4: Section [4.1] (0) (0:00 - 0:53) Section [4.1] has an onset [3.1], a continuant [3.2] and a termination [3.3]. Level 3: Section [3.1] (0) (0:00 - 0:23) Section [3.1] has an onset [2.1], a continuant [2.2] and a termination [2.3 ]. Level 2: Section [2.1] (0) (0:00 - 0:08) " [0.1] (=[1.1]) is used to begin the whole piece. It consists of a strong gesture based on sound family B, followed by a sound with grain functioning as a centred/pivoted setting. The initial gesture does not have an immediate continuation at levels 0 to 2 but, as it will be shown later, it functions within a paradigmatic axis at levels 3 to 5. It will henceforth be referred to as cadential paradigm 1. 2 0: onset. C: continuant. T: termination I C M C PROCEED I N G S 19954 471

Page  472 ï~~00 Â~Â~ o. 472 I C M C P RO C EE D I N G S 1995

Page  473 ï~~" [0.21 provides the next seminal gesture: counterpoint between bursts. This is realised by an initial ascent, panned right -> left, which is developed as [3.1] unfolds. Ascending and descending motion throughout the piece can be traced to this initial gesture. " [0.3] is the reverse of [0.2] both in pitch and in spatial location. In fact, [0.2] and [0.31 set two simple basic trends that are reflected at other levels, namely, ascent and descent. In this case, these are produced by contrasting two discrete sounds; however, as the piece unfolds, they are also realised using continuous sounds which change pitch gradually. Level 2: Section [2.2] (C) (0:08 - 0:16) [2.2] is a developed inverse of [2.1], with the following characteristics: * The direction of [1.2] (ascent-descent) is reversed in the relationship between [1.3] and [1.5] (descent-ascent). * A continuant is added. Therefore, becomes [2.1] [2.2] 0 [1.1] ascent 0 T [1.2] descent C [1.3] [1.4] descent undulation (centred) T [1.5] ascent 5. Overall summary. This can be exemplified by the following description of one of the main axes in the piece: The ascent-descent paradigm is initiated by [0.2] and [0.3]. This is then developed at all levels, from simple morphological gestures to the general directional tendencies of level 5. Furthermore, some of the realisation of this paradigm cuts through divisions between levels and sound families. The axis is listed below: Section [0.2] - [0.3] [2.10]-[2.12] ([3.4]) [1.37] - [1.38] [1.41] [1.43] [1.60] [1.61] [1.67] [0.134] (-[0.38]) [1.87]-[1.90] [2.38]-[2.39] [0.138] [0.140] [1.96] [2.41]-[2.42] Sound Family and comments A A+B D A+D (incomplete) C+D (incomplete) D (incomplete) C+A (extended) E (incomplete - pitch compression) A+D (incomplete) F (G#-D above, G#-D# below) various sf. A+D (sf. A predominant) sf. A+D (sf. D predominant) sf.E sf. D+B+E (sf. D and E predominant) ICMC PROCEEDINGS 199547 473

Page  474 ï~~[0.144]-[0.147] [0.1461-[0.1471 [2.43]-[3.151 [0.152]-([0.153])-[0.154] [0.155-0.156] [0.158]-([0.159])-[0.160] [1.34]-[1.106] [1.106] -[0.164] [0.164] [0.4]-[1.110] sf. B+D (sf. D predominant) sf.A sf. B+D (D predominant) in ascent - sf. E in descent sf. D+A (redundant gesture) sf. D+A (missing gesture) sf. D+A (redundant gesture) inverted. sf. B+D (B predominant). sf. B+D (B predominant) in [1.106] - sf. A in [0.164] sf. A inverted sf. A in ascent - sf. E in descent Conclusion The discussion above presented a systematic methodology for the analysis and determination of structure in music for tape based on the properties of sound, as defined by spectro-morphology, and on the organisation of discourse into paradigmatic axes. A set of strategies for the implementation of this methodology was discussed and illustrated by examples resulting from its application to a specific piece from the repertoire. Obviously, the methodology is not exhaustive: it may be necessary to complement it with analysis of the mimetic aspects of a piece and it may also be necessary to consider the poietic and esthesic dimensions. It would also be foolish to claim that this is a universal tool which can be applied indiscriminately to any work in the medium. Nevertheless, it is hoped that it can offer a reasonably rigorous approach - comparable to those existing for instrumental music of any period - which may cater for the particular characteristics of the materials, discourse and syntax of tape work. References Emmerson, S. (1986) Fischman, R (1994) Griffiths P. (1984) Harvey, J (1980) Nattiez, J. J. (1982) Nattiez, J. J. (1990) Smalley, D. (1986) Vaughan, M. (1988) Wishart, T. (1985) The Relation of Language to Materials. In: Emmerson S. (Ed.) The Language of Electroacoustic Music. Basingstoke: Macmillan. Music for the Masses. In: Leman M. & Berg P. (Ed.). Journal of New Music Research. Amsterdam: Swets & Zeitlinger. pp 245-264. Three Works by Jonathan Harvey. In: Machover, T. (Ed.). Contemporary Music review. 1/1, pp 87-110. Mortuos Plango, Vivos Voco. Compact Disk WER 2025 2. Wergo, 1990. trans. Barry, A., Varese's "Density 21.5": A study in semiological analysis. In: Music Analysis 1/3, 243-340. trans. Abbate, C., Music and Discourse. Towards a Semiology of Music. New Jersey: Princeton University Press. Spectro-morphology and Structuring Processes. In: Emmerson S. (Ed.) The Language of Electroacoustic Music. Basingstoke: Macmillan. CROSSTALK. Compact disc. OhM 001. Overhear, 1989. On Sonic Art. York: Imagineering Press. 474 I CM C PROCEEDINGS 1995