|
Stephen Travis Pope
Computer Music Journal
I have prepared a series of reference documents for the field of
computer music, including a bibliography, a diskography, a taxonomy,
and a list or electronic network resources. These have appeared over
the last year as Editor's Notes in Computer Music Journal. The note
below is a "rough cut" at a comprehensive taxonomy of the field. It
appeared in Computer Music Journal 18:1 (Spring, 1994) along with
several of the responses printed below. All four of these reference
documents are to be viewed as "works in progress" or--more
correctly--as on-going community discussions taking place in letters to
the editor of Computer Music Journal. The documents are also available
in electronic form by Internet ftp file transfer from the server
computer MITPress.MIT.edu in the directory pub/computer-music-journal.
There are various reasons for wanting a formal taxonomy of a field such
as computer music. First, as it is rapidly developing and evolving, a
taxonomy always represents a relatively subjective weighting of the
sub-fields at one point in time. This is appearent in the taxonomy that
will be introduced below. Secondly, the process of development forces
one to focus on the subtle differences between the defining keywords of
the field, and may help to clarify their usage. Lastly, of course,
having such a taxonomy can be useful for categorizing work in the
field, although projects and publications very rarely fit "neatly" into
exactly one category. More often, they are described by a set of two or
three topics, so such a taxonomy cannot be consdered strict or binding.
There are several sources for a possible taxonomy of computer music
(Pope 1993). These include the tables of contents of the Proceedings of
the International Computer Music Conferences and the excellent
bibliography Computer Applications in Music, by Deta S. Davis. Three of
the central contributors to the field--Roger Dannenberg, D. Gareth Loy,
and Bruce Pennycook--have also at various times prepared subject
keyword lists or taxonomies. These will all be presented and discussed
below, and a new multi-layered taxonomy of keywords will be proposed.
The top "layer" of Roger Dannenberg's informal taxonomy (which he
derived by hand from his library), presents perhaps the tersest list of
high-level categories for work in our field (Dannenberg 1993). His
items reflect the differences between the different types of software
systems used in computer music, and include single topics for music
understanding and music composition. It is also interesting to note
that he separates human computer interaction from music representation
and real-time systems.
Music understanding
Signal processing
Music representation
Music composition
Real-time systems
Human computer interaction
Acoustics
Education
Miscellaneous
Figure 1: Top-level topics of Roger Dannenberg's 1993 taxonomy
Dannenberg's hierarchy continues to a finer level of detail, with
subtopics representing the most important keywords found in any
abstract in our literature. D. Gareth Loy derived a list of keywords
from his database of the Proceedings of the International Computer
Music Conferences as part of a report for the board of the
International Computer Music Association (Loy 1993). Musical topics are
more finely delineated in the topics composition research, music
languages and notation, music theory and analysis, and performance
research. Signal processing is also differentiated from synthesis
techniques, and music printing from music notation.
Acoustics and psychoacoustics
Composition research
Computer-aided music instruction
Music languages and notation
Music networking
Music printing
Music theory and analysis
Performance interfaces
Performance research
Real-time interactive composition and performance systems
Signal processing
Studio report
Synthesis hardware design
Synthesis techniques
Tutorials
Figure 2: D. Gareth Loy's 1993 keyword list
The chapter outline of Computer Applications in Music: A Bibliography
(in two volumes) by Deta S. Davis introduces a medium-grained topic
list, which further subdivides the topics related to sound synthesis
and generation, and aesthetics and composition (Davis 1988, 1992). The
topic of digital audio appears and musicological applications are
mentioned for the first time. Aesthetics is included, and composition
separated from compositions, but sociology and performance practice are
not present.
Aesthetics
Composition
Compositions
Computers in Music Education
Conferences
Digital audio
Digital Signal Processing
Electronic and Pipe organs
Micro- and mini-computers
MIDI
Music industry
Music printing and transcription
Musical instruments
Musicological and analytic applications
Programming languages and software systems
Psychology and psychoacoustics
Reference, music library applications
Sound generation hardware
Sound generation software
Sound synthesis techniques
Sound generation in real time
Spatial simulation and room acoustics
Speech
Studios
General
Figure 3: Chapter outline of Deta Davis's "Computer Applications in Music"
As part of Bruce Pennycook's recent database of the ICMC Proceedings,
the following comprehensive topic list emerged (Pennycook 1993).
Several important but unclear issues come up here, including artificial
intelligence and music, and the history of electroacoustic music.
Diffusion is associated with sonorization (sonification), and
interactive performance systems differentiated from real-time hardware
and software (and from MIDI applications).
Artificial intelligence in music
Aesthetics, philosophy and criticism
Acoustics of musical instruments and voice
Audio analysis and re-synthesis
Audio hardware design
Audio signal processing techniques
Composition systems and techniques
Diffusion, sonorization
History of electroacoustic music
Interactive performance systems
Machine recognition of audio signals
Machine recognition of music data
MIDI applications
Miscellaneous
Music analysis
Music data structures and representations
Music education
Music grammars
Music and graphics
Music languages
Music printing
Music workstations
Optical music recognition
Performance interfaces
Psychoacoustics, perception, cognition
Real-time hardware
Real-time software
Room acoustics
Sound synthesis languages
Sound synthesis methods
Studio report
Figure 4: Keywords used in Bruce Pennycook's database of the ICMC Proceedings
In preparing a hierarchical taxonomy from these lists, several main
topics suggest themselves as the highest-level categories. The first
level of the list should reflect the spectrum ranging from music
theory, through psychology, to computer music engineering.
1. Music theory, composition, and performance
2. Musical acoustics, (psychoacoustics,) perception, and cognition
3. Musical (signal and event) representation and notation
4. Digital (control and sound) signal synthesis and processing
5. Hardware for computer music (instruments and tools)
6. Computers in music education, and computer music education
7. Computer music literature, history, and sources
Figure 5: Proposed top-level topics for a taxonomy of computer music
Expanding this one more level to map the major topics from the other
lists discussed above might result in an outline that subdivides these
items as shown next.
1. Music theory, composition, and performance
1.1. Music theory, sociology, and aesthetics
1.2. Composition of electroacoustic music
1.3. Algorithmic and computer-aided composition
1.4. Performance situations and interfaces
2. Musical acoustics, psychoacoustics, perception, and cognition
2.1. Musical acoustics and psychoacoustics
2.2. Music perception and psychology
2.3. Music understanding and cognition
3. Musical signal and event representation and notation
3.1. Models of signals and events
3.2. Musical event description languages
3.3. Musical signal description languages
3.4. Music notation and printing tools
4. Digital control and sound signal synthesis and processing
4.1. Sound synthesis methods
4.2. Time- and frequency-domain signal processing
4.3. Sound spatialization and localization
4.4. Machine recognition of signals and events
4.5. Real-time processing and scheduling
4.6. MIDI and control processing
5. Hardware support for computer music instruments and tools
5.1. Hardware for DSP and digital audio
5.2. Computer music workstations
5.2. Input/Output devices for music
6. Computers in music education, and computer music education
6.1. Computers in music education
6.2. Computer music education
7. Computer music literature, history, and sources
7.1. Bibliographies/diskographies
7.2. Studio reports
7.3. Descriptions of compositions
7.4. History of electroacoustic music
Figure 6: Extension of the proposed taxonomy to two levels
The most extensive taxonomy includes almost all of the topics from all
four authors' lists. Its "weighting" or "focus" reflects the average of
four current views of the most important facets of the computer music
literature.
1. Music theory, composition, and performance
1.1. Music theory, sociology, and aesthetics
1.1.1. Music theory and analysis
1.1.2. Temperament and tuning systems
1.1.3. New musical aesthetics and sociologies
1.2. Composition of electroacoustic music
1.2.1. Sound and composition models and notations
1.2.2. Models of the composition and performance processes
1.2.3. Sound design and processing
1.2.4. Realization and production techniques
1.2.5. "Aural rendering" or "sonification" of scientific data
1.3. Algorithmic and computer-aided composition
1.3.1. Compositional algorithms and languages
1.3.2. Composition systems for score or sound synthesis
1.3.3. Artificial Intelligence and composition
1.4. Performance situations and interfaces
1.4.1. Performing and conducting
1.4.2. Gesture recognition and interfaces
1.4.3. Score following in performance
1.4.4. Expression representation and analysis
2. Musical acoustics, psychoacoustics, perception, and cognition
2.1. Musical acoustics and psychoacoustics
2.1.1. Acoustics of musical instruments and voice
2.1.2. Psychoacoustics
2.1.3. Room and spatial acoustics
2.2. Music perception
2.2.1. Physiology of hearing
2.2.2. Pitch identification
2.2.3. Rhythm identification
2.2.4. Timbre perception
2.3. Music understanding and cognition
2.3.1. Rhythm understanding
2.3.2. Key and scale recognition
2.3.3. Higher-level structures
3. Musical signal and event representation and notation
3.1. Models of signals and events
3.1.1. Language systems
3.1.2. Encodings and file formats
3.1.3. Graphical notation systems
3.2. Musical event description languages
3.2.1. Note-list formats
3.2.2. Music input languages
3.2.3. Music programming languages
3.3. Musical signal description languages
3.3.1. Signal models and descriptions
3.3.2. Software synthesis languages
3.4. Music notation and printing tools
3.4.1. Transcription or performance
3.4.2. Optical recognition of scores
4. Digital control and sound signal synthesis and processing
4.1. Sound synthesis methods
4.1.1. Additive sound synthesis methods
4.1.2. Subtractive sound synthesis methods
4.1.3. Nonlinear sound synthesis methods
4.1.4. Physical models of acoustical systems
4.1.4.X. Various types of physical models
4.1.5. Other synthesis methods
4.1.6. Analysis and resynthesis systems
4.2. Time- and frequency-domain signal processing
4.2.1. Software architecture
4.2.2. Time domain model synthesis
4.2.3. Frequency domain model synthesis
4.2.4. Ad hoc synthesis techniques
4.2.5. Effects and filters
4.3. Sound spatialization and localization
4.4. Machine recognition of signals and events
4.5. Real-time processing and scheduling
4.5.1. Real-time scheduling
4.5.2. Real-time languages
4.5.3. Hardware architectures
4.6. MIDI and control processing
5. Hardware support for computer music
5.1. Hardware for DSP and digital audio
5.2. Computer music workstations
5.2. Input/Output devices for music
6. Computers in music education and computer music education
6.1. Computers in music education
6.2. Computer music education
7. Computer music literature and sources
7.1. Bibliographies/diskographies
7.2. Studio reports
7.3. Descriptions of compositions
7.4. History of electroacoustic music
Figure 7: A taxonomy that encompasses all of the keywords presented in the discussion
The note above was circulated via electronic networks In December
1993; several of the responses are included below.
References
Dannenberg, R. B. 1993. Private email communication of 9 August, 1993.
Davis, D. 1988. Computer Applications in Music: A Bibliography (with 1992 supplement). Madison, Wisconsin: A-R Editions.
Loy, D. G. 1993. Private email communication of 28 January, 1993.
Pennycook, B. R. 1993. Private email communication of 21 October, 1993.
Pope, S. T. 1993. "A Computer Music Library." Computer Music Journal 17(3): 3-6.
Responses
One observation I made was that many papers span more than one topic,
so I question the effort to produce a fine-grained taxonomy. I think
the one-level list of topics that Bruce Pennycook and the ICMA have
generated is a good one, although not fixed in stone. I have the
concern that there are already too many lists. I would be much happier
to go along with the ICMA scheme. For this reason, I did not submit my
taxonomy article for publication, and I plan to arrange my personal
library to match the ICMA taxonomy. I think the ICMA scheme, if
adopted by others, will be of great benefit.
Roger Dannenberg
Pittsburgh, Pennsylvania USA
One question I would ask is, what is the intended purpose of this
taxonomy, as this is not stated in the proposal? It seems to me that
the stated form of organization would serve admirably as the basis for
a curriculum on computer music. Given the current explosion of interest
in using personal computers as musical instruments, such a contribution
is timely. Further, the hierarchical organization is a good starting
point for identifying areas where further research is most needed.
One area which I am not yet convinced is best served by this taxonomy
is the composition aspect of computer music. I have developed numerous
taxonomies and systematic organizations of various aspects of the
field, and have always been overwhelmed by the complexity of possible
interrelationships, often quite idiosyncratic, such that I ultimately
find myself mapping out my own intuitions - an act which is
accomplished much more efficiently and gratifyingly through the
compositional act itself. This is no more than my personal experience;
I do not expect it to be universal. Indeed I am curious and interested
to hear of the successes of other musicians and composers who have
incorporated systematic studies of computer music into their
compositional cosmologies.
I have one other apprehension about the proposed organization.
Theoretical work can have a drastic impact on the practice of an art.
Consider the impact of temperment theory on the evolution of tonality.
Consider how sequencer programs have shaped various forms of popular
music. One danger I see in the proposed taxonomy is that we may assume
a false sense of the degree of integration in computer music. Surely
there is much more conceptual overlap between species counterpoint and
Indian classical music than between the sociology of computer music and
digital filter design theory. This is not a flippant remark. There is a
tendency in universities to treat the humanities as second class
citizens: the sciences are rigorous, hard (well funded); the arts are
intuitive and soft. Of course we know better than to adopt such gross
generalizations. But if the field of computer music is to develop as a
hybrid, with aspects of both humanistic and scientific thought, there
is a significant possibility that the humanistic side will insensibly
be moved to a backseat role, given its locus in academia. This would be
unfortunate as there are many pressing aesthetic and sociological (dare
I say political?) issues which need to be addressed if computer music
is to fulfill its destiny in the information age of enhancing the lives
of a great many people participating actively, rather than to remain
the preserve of the few, as a sterile, "Fourth Viennese" school with
little interest or appeal to non-specialists.
Eric Lyon
Keio, Japan
I saw the Editor's Note you prepared for Computer Music Journal 18:1
on electronic mail news. Such taxonomies are bound to serve a purpose
only temporarily and tend to make the field seem more conservative than
necessary, I feel. Nevertheless it always is appealing to be able to
take a look at a field and be able to classify all of its activities,
and I guess computer music needs this.
Axel Mulder
Burnaby, British Columbia Canada
I see three areas in which the proposed taxonomy needs work:
collaboration (and networking), and physical and mathematical
underpinnings. The taxonomy seems to tacitly assume (particularly in
sections 1, 4, and 7) that computer music is a field generally
populated with individuals working by themselves, using tools that
stand by themselves. I suppose this reflects the current literature
(which is, after all, the taxonomy's purpose); if so, it indicates to
me a gap that needs to be addressed. Perhaps this taxonomy might serve
not only as a guide to the conceptual space explored thus far, but also
to indicate new areas of (published) exploration.
Section 2.1 would benefit from explicit inclusion of sonic physics in
general, not just as it applies to the most common contemporary musical
sources, media, and targets of sound.
Section 3 could use a section to specifically address algorithms and
their mathematical foundations. Other sections might benefit from
explicit mention of algorithmic and mathematical considerations
(efficiency, tractability), in the design of formats, models,
languages, etc., again in different domains (performance, analysis, etc.).
Specifically, I might re-factor your taxonomy to end up with a
hierarchy more like the following.
1. Music theory, composition, and performance
. . .
1.3. Collaborative composition
1.3.1. Inter-human, algorithmic, and computer-aided
composition
1.3.1.1. --in non-real-time (as an activity
distinct from performance)
1.3.1.2. --in real-time (e.g., improvisation)
[Algorithmic and computer-aided composition may be
considered "collaborating with computers", if you will. ]
. . .
2. Musical acoustics, psychoacoustics, perception, and cognition
2.1. Musical acoustics and psychoacoustics
2.1.1. The physics of sound
2.1.1.1. Acoustics of musical instruments and voice
. . .
3. Musical signal and event representation and notation
3.1. Algorithmic and mathematical considerations
. . . [insertion renumbering]
4. Digital control and sound signal synthesis and processing
. . .
4.6. Music networking
4.6.1. MIDI and control processing technologies
4.6.2. Compositional and performing considerations
. . .
7. Computer music literature and sources
7.1. Archives (particularly on-line archives)
7.1.1. Bibliographies/diskographies
7.1.2. Sources
. . .
As with most structurings of conceptual space, it's challenging to
find a "maximally orthogonal" (not to mention "culturally appropriate")
mapping; the space is large and diverse, the importance (and
existence!) of each segment varies with each practitioner, and links
between segments abound. I think the proposed taxonomy does a good job.
Craig Latta
Sunnyvale, California USA
Generations of practitioners and many years of specialization have
been directed towards the creation of a viable field of computer music,
representing enormous human resources. Devising a comprehensive
taxonomy of computer music is problematic because of the tendency to
view the totality as a sum of independent fields, and to express
relationships and process in static representations. Stephen Pope's "A
Taxonomy of Computer Music" reveals a great deal as a representation of
a diverse constellation of fields and interests. In addition to his
effort to view the music technology field in its entirety, Mr. Pope
approaches the development of a taxonomy as a process.
A brief survey of the top levels of the topic lists developed by D.
Gareth Loy, Roger Dannenberg, Deta Davis and Bruce Pennycook reveals
the wide scope of the computer music field. At this level it is easy to
appreciate the characteristics that distinguish the categories.
Distinction among aesthetics, hardware design, and representation flow
naturally enough to create a reasonable system of classification and
orientation. As one expands from the flat to the hierarchical
representation, the connections among the fields and categories become
more evident. The lattice of relationships in the second level of the
hierarchical model can be seen in just a few examples. (A) (2.2) Music
perception and psychology, under musical acoustics, psychoacoustics,
perception, and cognition is closely related to (1.1) music theory,
sociology and aesthetics, under music theory, composition and
performance. (B) (4.6) MIDI and control processing, under digital
control and sound signal synthesis and processing has some obvious
overlap with (3.2) musical event description languages, under musical
signal and event representation and notation. (C) (1.2) Composition of
electroacoustic music, under music theory, composition and performance
sounds very similar to (7.3) descriptions of compositions, under
computer music literature, history and sources.
Expansion to the third level invites one to traverse a variety of
threads through the field, reinforcing the lattice of connections which
cross structural divisions in the hierarchy. This suggests that one
develop a global perspective which links new musical aesthetics and
sociologies (1.1.3) with timbre perception (2.2.4), or software
synthesis languages (3.3.2) and computers in music education (6.2), for
none of these categories exist or evolve independently. One might
further wonder what the taxonomy would look like if it were to be
reconstructed on different terms. For example, it would be particularly
revealing to examine category (4) Hardware for computer music
(instruments and tools) as a function of category (1) music theory,
composition and performance, and not as an independent topic, sharing
the same degree position in the hierarchy. One might gain further
valuable insights by traversing the hierarchy in the other direction,
viewing the flat top level as a reflection of categories within some
broader perspective, perhaps music, creativity, communication or
philosophy. This exercise is not designed to contest the specific
categorization represented in this taxonomy, or to refute its relevance
in the field. The intention is to demonstrate that a taxonomy that
encapsulates complex issues in a static, linear structure does not
offer the lattice of connectivity that is fundamental to developing a
comprehensive understanding of the topics.
During this century in music Edgar Varese, Harry Partch, Gyorgi
Ligeti, John Cage, Iannis Xenakis, Karlheinz Stockhausen, and many
others challenged world views--not simply musical views. Their work
points toward a sonic realm and an expressive language that we have yet
to address in a comprehensive and integrated fashion. The development
of an historical perspective of computer music is particularly
meaningful in the context of the evolution of music, incorporating the
development of music's instruments and creative resources within the
larger domain. The perspective gained reveals an evolution that links
technological development directly with communication and human
expression, leaving artifacts that embody the views of the society. An
aesthetics or sociology sub-topic in a hierarchical representation
fails to establish the activity as a product of music making or self
expression within the society. Performance practice, education,
input/output, composition, and theoretical analysis do not exist
independently, and cannot effectively be represented in isolation.
Instrument development does not take place outside of consideration for
compositional processes, performance practice and economics. The
instruments and musical characteristics that evolve in a society are
intricately integrated into its world view, a fact which becomes
evident in exploring the music of different cultures and the
instruments designed to create it. A comparison of the instruments and
music of Bali, India and Europe provides support for this view. A
"world view" permeates the hierarchy of the taxonomy.
I envision a top level in the "Taxonomy of Computer Music" that
reflects world view, and a multi-dimensional and configurable modeling
resource that facilitates navigation among multiple world views. The
taxonomy becomes a dynamic, interactive world view, supported by
technological resources that reflect the orientation which governs the
development and application of the technology supporting creative expression.
Readers with comments, suggestions, pointers to resources of the kind
described, or an interest in working on the taxonomy are invited to
respond to me at Craig@WELL.sf.ca.us and to Stephen Travis Pope at
CMJ@CNMAT.Berkeley.edu.
Craig R. Harris
San Francisco, California USA
|
