Transcript Musicology, Music Cognition and Musical Similarity

Musicology, Musical Similarity
and Music Cognition
ISMIR Graduate School, Barcelona 2004
Musicology 1
Frans Wiering, ICS, Utrecht University
Outline





Introductions
Musicology and musical similarity
Music perception and cognition
Lecture plan
Assignment(s)
Introducing myself

Studies:




biology (1975-79)
musicology (1978-86)
PhD musicology (1995)
Employment

Musicology (1985-88; 1989-93)


Computer and Humanities,
Utrecht University (1988-89; 19941998)



computing and musicology
Online Italian treatises (TMI)
Computing and Information
Sciences, Utrecht University
(1999-)




Polyphonic modality (history of
music theory)
Electronic Document Technology
Music Retrieval (Orpheus project)
we have a vacant PhD position
Amateur musician

(mainly) singer, conductor
Introducing you (some questions)











who listens to music?
who plays a musical instrument?
who sings (outside the bathroom)?
who is familiar with basics of music notation?
who is familiar with advanced music notation?
who has created new music?
who is a computer scientist?
who is a musicologist?
who is a cognitive scientist?
who are the others?
things you want to hear about from me?
Musicology

What is musicology?




Aim of musicology


scholarly study of music (New Grove)
science of musical content processing (Leman 2003)
long history (Greeks) and tradition/bias (focus on classical
music scores)
providing reliable answers to questions about music
Central question here:


what is musical similarity
no simple answer, but musicology can help to find partial
answers
Aims of my lectures
introducing the ‘language of musicology’
1.


practice vocabulary
2.

assignments
understand relevance of musicology to music retrieval
3.



4.
can be quite different from the language of end users, fans and even musicians
As questions of I use expressions you don’t know/understand
knowledge about music
interesting (complex) retrieval questions
the bias(ses) of musicology
pointers to literature

The New Grove Dictionary of Music and Musicians,
(www.grovemusic.com)

very good source, but written in the language of musicology



trumpet: A lip-vibrated aereophone
your University may have a subscription
Marc Leman. Foundations of Musicology as Content Processing
Science. Journal of Music and Meaning 1 (2003)
(www.musicandmeaning.net)
What is music?

some answers

product
What is music?

some answers



product
music notation
sound
What is music?

some answers





product
music notation
sound
activity
example of musical
similarity within one
piece of music
More examples of musical similarity

Questions to ask about similarity:




Example 1 (play 2)




Mozart, Ah, vous dirai-je, Maman
variation on well-known song
Example 3




Beethoven 5 and Walter Murphy
reusing chunks
Example 2 (5.20)


what is similar
what is different
how would you characterise the relationship
Purcell, Overture Dido and Aeneas
Bach, Nun komm der Heiden Heiland
genre similarity
Example 3a


Ravel, Le paon
similarity creates musical meaning
Retrieval tasks
genre
artist
work
instance
industry
copyright and royalties
plagiarism
recommendation
sounds as
consumer
mood
emotion
style
performer
professional
feature
composer
intertextuality
identification
source
Retrieval tasks
genre
artist
work
instance
mood
sounds as
style
feature
recommendation
intertextuality
emotion
performer
notation
composer
identification
plagiarism
audio
copyright and royalties
source
Remarks




many kinds of musical
similarity
concepts and vocabulary
needed to describe these
musicology provides
these
returns in assignments
Aside: Ground Truth Experiment


For Orpheus Project (Utrecht University)
Why



When


Tuesday, Friday, 6-8
Required knowledge


have queries evaluated by domain experts
evaluate performance of different methods
compare short melodic fragments, in notation
Many thanks
A closer look at musicology

scholarly study of music (New Grove)


traditional components: theory, history
modern division:

historical musicology



systematic musicology



masterworks, editions, styles, historical change
representative research themes: influences, intertextuality
aesthetics, theory, analysis, acoustics, etc.
representative research themes: basic elements of music
(MIR: ‘features’), construction of musical works
ethnomusicology


world music, music and (traditional) culture
representative research themes: place of music in ritual,
transmission and change
The bias of (traditional) musicology

Western art music





speculative nature of theorizing




scores
works ‘themselves’
creation and reproduction of these
things rather than processes (role of time)
based on personal observation/intuition
(extreme) elaboration
no testing
Example: theories of Heinrich Schenker (1868-1935)

principle of voice leading (stepwise movement of melody) elaborated to
its extreme
Reactions

Musical grammars




Music cognition and perception




influence from linguistics: music as language
inspired by Chomsky and Schenker
best-known example: Lerdahl and Jackendoff: A Generative Theory of
Tonal Music (1986)
focus on the mind rather than the notes
experimental approach
my personal favourite: Bob Snyder: Music and Memory (2000)
Computing and musicology





corpus creation
formalization of analysis methods
implementation of musical grammars
example of musical grammar: Kemal Ebcioglu, ‘Bach’ chorales

two examples

another kind of musical similarity: style
Center for Computer Assisted Research in the Humanities

www.ccarh.org

journal: Computing in Musicology
Music and meaning

evidently, music gives rise to very specific meanings




not just the text of the song
hard to express in language
similarity=same meaning?
Theories of meaning


traditional: related to musical features
Leonard Meyer, Emotion and Meaning in Music, 1956:
musical meaning derives from arousing, frustrating and
fulfilling of expectations. Expectations are culturally
conditioned.


links to Gestalt theory and cognitive musicology
Eugene Narmour: Implication-realization model (1977,
1991)

several computational applications
Music perception and cognition





study of processes that take place
 in the ear during perception
 in the brain after perception
result: perceptual and mental models of music
models describe how we react to music, for example:
 how we store music
 what meaning we assign to music
 what it is we reproduce when remembering music or perform
music
 when we consider music as similar, and when not
elements of music cognition will occur in most of my lectures
 start here with basic model after Snyder, Music and Memory
(2000)
caution: not my expertise
Memory

Source: Snyder (2000)



early processing



event fusion
boundary creation
short term memory




(simplified)
example: melody
chunking
time-aware
rehearsal
long term memory




broad categories
schemas: groups of
associations that represent
musical culture
time-awareness lost,
order can be reconstructed
by following association
Categorisation


Categories are a means of reducing
information overload
perceptual categories: arise during
perception (e.g. pitch discrimination)



not dependent on culture
short-lived (STM)
conceptual categories



formed by training, dependent on culture
e.g. scales
long-lived (LTM)
So what happens when listening to music




After early processing: pitch events:
 have pitch, duration, loudness, timbre
 grouping
primary parameters are treated as categories
 e.g. pitch, rhythm, harmony
secondary parameters are not categories:
 tempo, duration, timbre
(characteristics of) chunks are remembered, not individual notes
 ‘holistic processing’
 musical meaning not in single notes/intervals

connections between chunks (‘associations’) are easily lost
music notation can be seen as a way of recording conceptual
categories + performance instructions


explains difficulties with reproducing music
Summary

musicology







music theory
music and language, musical grammars
computing and musicology
music cognition and perception
what is music
different levels of musical similarity
music and the mind


categories
why are notes interesting
Lecture plan

1.
2.
3.
4.
5.
6.




outline
[musicology, music cognition and musical similarity]
music notation and encoding
melodic feature and retrieval 1
melodic feature and retrieval 2
harmony and tonality
high-level features
assignments
develop/help others develop skills in music description
will be given at the end of the lecture
2-3 students (volunteers) will give a presentation of max. 5
minutes at the beginning of next lecture


hopefully, many other styles than my examples
others ask questions
An exercise in music description


Play 2 short fragments you consider similar
Questions to ask yourself:





what precisely makes them similar
what is different
how would you characterise the relationship of the 2
fragments
5 minutes
Other assignment:

make sure you are familiar with basics of music notation


http://datadragon.com/education/reading/
http://library.thinkquest.org/15413/theory/theory.htm