Transcript The demise of number ratios in music theory
Intervals as distances, not ratios: Evidence from tuning and intonation
SysMus Graz
Richard Parncutt
Centre for Systematic Musicology University of Graz, Austria
Graham Hair
Department of Contemporary Arts, Manchester Metropolitan University, UK ISPS, 28-31 August 2013, Vienna International Symposium on Performance Science
Abstract
Many music theorists and psychologists assume a direct link between musical intervals and number ratios. But Pythagorean ratios (M3=61:84) involve implausibly large numbers, and just-tuned music (M3=4:5) only works if scale steps shift from one sonority to the next. We know of no empirical evidence that the brain perceives musical intervals as frequency ratios. Modern empirical studies show that performance intonation depends on octave stretch, the solo-accompaniment relationship, emotion, temporal context, tempo, and vibrato. Just intonation is occasionally approached in the special case of slow tempo and no vibrato, but the reason is to minimize roughness and beating - not to approach ratios. Theoretically, intonation is related to consonance and dissonance, which depends on roughness, harmonicity, familiarity, and local/global context. By composing and performing music in 19-tone equal temperament (19ET), the second author is investigating how long it takes singers to learn to divide a P4 (505 cents) into eight roughly equal steps of 63 cents, or a M2 (189 cents) into three; and whether the resultant intonation is closer to 19ET or 12ET. Given that the average size of an interval depends on both acoustics (nature) and culture (nurture), it may be possible to establish a sustainable 19ET performance community.
Boethius
Italian philosopher, early 6th century
“But since the nete synemmenon to the mese (3,456 to 4,608) holds a sesquitertian ratio -- that is, a diatessaron -- whereas the trite synemmenon to the nete synemmenon (4,374 to 3,456) holds the ratio of two tones....”
The major third interval (M3)
perceptual category
“Pythagorean tuning”
reflects motion tendencies (leading tone rises) emphasizes difference between major and minor
“Just tuning”
minimizes beats between almost-coincident harmonics - only if spectra are harmonic and steady (slow, non-vibrato)
The difference
81/80 = 22 cents = “syntonic comma” Much smaller than category width of M3 = 100 cents
The major scale in 3 tuning systems
ratios and cents
Scale step ^2
12ET* Pythag orean
200 8:9 204
Just**
8:9 204
^3 ^4
400 500 64:81 3:4
408
498 4:5
386
3:4 498
^5
700 2:3 702 2:3 702
^6 ^7 ^8
900 1100 1200 16:27 128:243 1:2
906 1110
1200 3:5
884
8:15
1088
1:2 1200 *
12ET
= 12-tone equally-temperament
Most intervals have 2 ratios
Would the real ratio please stand up?
interval note chr. pure/just Pythagorean P1 C 0 1:1 1:1 m2 C# M2 D 1 16:15 256:243 2 9:8 or 9:10 9:8 m3 D# M3 E 3 6:5 32:27 4 5:4 81:64 P4 F TT F# 5 4:3 4:3 6 45:32 729:512 P5 G 7 3:2 3:2 m6 G# 8 8:5 128:81 M6 A 9 5:3 27:16 m7 A# M7 10 9:5 or 7:4 16:9 B 11 15:8 243:128 P8 C 12 2:1 2:1
Strange ideas of ratio theorists
Pythagoreans
since 6 th Century BC
The universe is number and music reflects it
• •
Monochord mathematics
first four numbers (tetraktys) are special (1+2+3+4=10) all intervals by multiplying and dividing these numbers
Music of the spheres
Planets and stars move to these ratios a cosmic symphony!
Pythagoras could hear it! Did he have tinnitus? ;-)
Saint Bonaventure
Italian medieval theologian and philosopher, 1221 – 1274
God is number
“Since all things are beautiful and to some measure pleasing; and there is no beauty and pleasure without proportion, and proportion is found primarily in numbers; all things must have numerical proportion.
Consequently, number is the principal exemplar in the mind of the Creator
and as such it is the principal trace that, in things, leads to wisdom. Since this trace is extremely clear to all and is closest to God, it … causes us to know Him in all corporeal and sensible things” Itinerarium mentis in Deum, II, 7
Giovanni Battista Benedetti
Italian mathematician, 1530 –1590 • •
Consonance is all about waves
sound consists of air waves or vibrations in the more consonant intervals the shorter, more frequent waves concurred with the longer, more frequent waves at regular intervals
(letter to Cipriano de Rore dated around 1563)
Johannes Kepler
German mathematician, astronomer (1571-1630)
Music helps you understand the solar system
• Third law of planetary motion: The square of the orbital period of a planet is directly proportional to the cube of the semi major axis of its orbit.
• •
Aims: understand the music of the spheres express planetary motion in music notation
(Did he have tinnitus too?)
Gottfried Wilhelm Leibniz
German mathematician and philosopher (1646-1716)
Consonance is about subconscious counting
“Die Freude, die uns die Musik macht, beruht auf unbewusstem Zählen.” “Musik ist die versteckte mathematische Tätigkeit der Seele, die sich nicht dessen bewusst ist, dass sie rechnet.” (Letters)
Leonhard Euler
Swiss mathematician and physicist (1707-1783)
Consonance is based on numbers
“…the degree of softness of ratio 1:pq, if p and q are prime numbers … is p+q-1."
Tentamen novae theoriae musicae ex certissimis
harmoniae principiis dilucide expositae (1731) (A attempt at a new theory of music, exposed in
all clearness according to the most well-founded
principles of harmony)
Ross W. Duffin
Dept of Music, Case Western Reserve U, Cleveland OH
You can hear number ratios directly
“12ET major thirds are … the invisible elephant in our musical system today. Nobody notices how awful the major thirds are. (…) Asked about it, some people even claim to prefer the elephant. (…)
But I’m here to shake those people out of their cozy state of denial. It’s the acoustics, baby: Ya gotta feel the vibrations.
“
How equal temperament ruined harmony (and why you should
care). London: Norton, 2007 (pp. 28-29)
Kurt Haider
Institut für Musiktheorie und harmonikale Grundlagenforschung, Wien • • • •
Ratios can explain almost everything
harmonikale Grundlagenforschung: eine mathematische Strukturwissenschaft (Pythagoreer, Platon, Neuplatoniker) seit Kepler: auch eine empirische Wissenschaft
führt die Struktur der Naturgesetze auf ganzzahlige Proportionen zurück
durch die Intervallempfindung der ganzzahligen Proportionen werden nun qualitative Parameter wie Form, Gestalt oder Harmonie wieder Gegenstand der Wissenschaften kurthaider.megalo.at/node/49
Clarence Barlow
composer of electroacoustic music
Ratios help you compose
“Harmonicity” of an interval depends on
“digestibility” of the numbers in its ratio
(prime factors) Systematic enumeration of the most harmonic ratios within an octave 1:1, 15:16, 9:10, 8:9, 7:8, 6:7, 27:32, 5:6, 4:5, 64:81, 7:9, 3:4, 20:27, 2:3, 9:14, 5:8, 3:5, 16:27, 7:12, 4:7, 9:16, 5:9, 8:15, 1:2. Two essays on theory. Computer Music Journal, 11, 44-59 (1987)
Laurel Trainor
(Music) Psychologist, McMaster University
Infants process frequency ratios
“Effects of
frequency ratio simplicity
on infants' and adults' processing of simultaneous pitch intervals with component sine wave tones” (abstract) Effects of frequency ratio on infants' and adults' discrimination of simultaneous intervals.
Journal of Experimental Psychology: Human
Perception and Performance, 23 (5), 1427-1438 (1997)
Opposition to ratio theory
Aristoxenus “Harmonics”
(4 th Century BC; pupil of Aristotle)
There is more to music than number
“Mere knowledge of
magnitudes
does not enlighten one as to the functions of the tetrachords, or of the notes, or of the differences of the genera, or, briefly, the differences of simple and compound intervals, or the distinction between modulating and non modulating scales, or the modes of melodic construction, or indeed anything else of the kind.” “
we must not follow the harmonic theorists
in their dense diagrams which show as consecutive notes those which are separated by the smallest intervals [but] try to find what intervals
the voice is by nature able
to place in succession in a melody” Macran, H. S. (1902). The harmonics of Aristoxenus. London: Oxford UP.
Jean-Philippe Rameau
French composer and theorist (1683 -1764) • • • •
First tried to explain triads using ratios:
major triad 20:25:30 (4:5:6) Mm7 20:25:30:36 minor triad 20:24:30 (10:12:15) m7 25:30:36:45
Later referred to the corps sonore:
Foundation of harmony is the intervals between the harmonic partials of
complex tones in the human environment
Hermann von Helmholtz
German physiologist and physicist, 1821-1894 “Even Keppler (sic.), a man of the deepest scientific spirit, could not keep himself free from imaginations of this kind … Nay, even in the most recent times theorizing friends of music may be found who will rather
feast on arithmetical mysticism than endeavor to hear out partial tones
” (p. 229).
On the Sensations of Tone as a Physiological Basis for the
Theory of Music, 1863; 4 th ed. transl. A. J. Ellis (but Helmholtz theorized with ratios too…)
Ratios in Western music theory
1. Pythagoras (6 th C. BC) Boethius (6th C. AD) • Musical intervals
are
ratios • Based on prime numbers 2 & 3 • Spiritual, cosmic, religious
Ratios in Western music theory
2. Renaissance theorists
Ratios can include factors of 5
“just”
• Ramos de Pareja (1482) • Gioseffo Zarlino (1558) • Giovanni Battista Benedetti (1585) Can that explain the sonority of triads?
Ratios in Western music theory
3. Scientific revolution (18 th -19 th C.)
New concept of musical intervals
audible relationships between partials in harmonic complex tones • •
Consonance based on
harmonicity (Rameau, Stumpf) roughness (Helmholtz)
Shift of emphasis
from maths to physics, physiology, psychology
Ratios in Western music theory
4. 20 th -C. experiments on intonation in music • • • ≈12ET generally preferred Pythagorean preferred over just (e.g. rising leading tones) Just intonation: only for slow, steady tones with no vibrato
Many studies!
Ambrazevicius, Devaney, Duke, Fyk, Green, Hagerman & Sundberg, O’Keefe, Loosen, Karrick, Kopiez, Nickerson, Rakowski, Roberts & Matthews...
Just tuning: Impossible in practice
The fifth between ^2 (8:9) and ^6 (3:5) is not 2:3!
Must constantly shift scale steps to stay in tune If you don’t like it when your choir gradually goes flat or sharp, “just tuning” is not for you!
Renaissance choral polyphony
“Renaissance performers would have preferred solutions that favor just intonation wherever and whenever possible … deviations from it would have been
momentary adjustments to individual intervals,
rather than wholesale adoption of temperament schemes” Ross W. Duffin (2006). Just Intonation in Renaissance Theory and Practice. Music Theory Online
Johanna Devaney
with Ichiro Fujinaga, Jon Wild, Peter Schubert, Michael Mandel Participants: professional singers Task: sing an exercise by Benedetti (1585) to illustrate pitch drift in just • • •
Main results:
Intonation close to 12ET Standard deviation of pitch is typically 10 cents (!) Small drift in direction of Benedetti’s prediction
Limited precision of “Ideal tuning”
Just noticeable difference in middle register
for simultaneous or successive pitches under ideal conditions:
2 cents Uncertainty in f 0 of singing voice
vocal jitter of best non-vibrato voices:
3 cents Intervals in the audible harmonic series
all are stretched - physics & perception!
10 cents
M2 = 8:9 (204 cents) or 9:10 (182 cents):
20 cents Structure
Must tune all intervals between all scale steps!
Expression
Expressive intonation:
50 cents
So why do people sing in 12ET?
1. Familiarity with piano 2. Compromise between Pythagorean and Just
We don’t know which!
Point 1: since 18 th Century Point 2: for millennia!
Gregorian chant: Pythagorean? Or 12ET?
Renaissance polyphony: just? Or 12ET?
Thomas Kuhn’s “paradigm shift”
or scientific revolution • •
Paradigm
Entire landscape of knowledge and implications in a discipline Universally accepted • •
Long process of change
Gradual increase in number of anomalies Experimentation with new ideas crisis intellectual battles • •
Features of change
Old and new are incommensurable Shifts are more dramatic in previously stable disciplines
Examples
Physics: Classical mechanics Psychology: Behaviorism relativity and quantum mechanics cognitivism Music theory: Math & notation performance & perception
Carl Dahlhaus
German musicologist, 1928-1989 “Whereas in the ancient-medieval tradition
number ratios were considered to be the foundation or formal cause of consonance,
in modern acoustics and music theory they paled to an external measure that
says nothing about the essence of the matter
. … In the music theory of the 18 th and 19 th Centuries,
the overtone series is the natural archetype of the interval hierarchy
upon which rules of composition are founded. … The
surrender of the Platonic idea of number
meant nothing less than the collapse of the principle that had carried ancient and medieval music theory.” C. Dahlhaus (Ed.), Einführung in die Systematische Musikwissenschaft (1988)
Interval perception is not about ratios - it is about
Categorical perception
Color
e.g. range of wavelengths of the color red – “nature”: • physiology of rods and cones – “nurture”: • mapping between color words and light spectra
Speech sounds
e.g. range of formant frequencies of vowel /a/ – “nature”: • vocal tract resonances near 500 and1500 Hz – “nurture”: • learned formant frequencies of each vowel
Categorical perception of musical intervals
Burns & Campbell (1994)
Stimuli:
Melodic intervals of complex tones; all ¼ tones up to one octave
Participants:
Musicians
Task:
name the interval using regular interval names (semitones)
The ear acquires relative pitch categories
…from the distribution of pitches in performed music F F#/Db G In music, pitch varies on a continuous scale.
When some pitches are more common, categories crystalize.
These categories are the REAL ORIGINAL “musical intervals”.
In real performance, Just and Pythagorean
have no physical existence at all!
Just Pythagorean
Bimodal distribution
with tendency toward • pure (M3 = 386 cents) or • Pythagorean (M3 = 408 cents)
We never find this!
Normal distribution
sd ≈ 20 cents + 1 sd = acceptable tuning + 2 sd = pitch category
We generally find this!
Does the brain have a ratio-detection device?
If it did, we might expect:
1. bimodal interval performance and preference distributions 2. low tolerance to mistuning of harmonics in complex tones 3. an evolutionary basis for ratio detection
In fact:
1. distributions are unimodal 2. harmonics mistuned by a quartertone or semitone (!) are still perceived as part of the complex tone (Moore et al., 1985) 3. environmental interaction depends on identification of sound sources via synchrony, harmonicity… (Bregman, 1990)
For the psychoacousticians:
This is not a spectral approach!
Pitch is an experience of the listener – not a
physical or physiological measure
Pitch generally depends on both temporal and
spectral processing, which are inextricably
mixed and hidden in neural networks.
What influences intonation?
Real-time adjustment of frequency in performance
Perceptual effects (individual tones) – octave stretch (small intervals compressed) – beating of coinciding partials Cognitive effects (musical structure) – less stable tones are more variable in pitch – rising implication of leading tone; major-minor distinction
Effects of performance
– solo versus accompaniment (soloists tend to play sharp) – technical problems or limitations of instruments
Effects of interpretation
– intended emotion (e.g. tension-release) – intended timbre (e.g. deep = low)
“Authentic” Renaissance polyphony?
Some choristers practise just tuning with real-time computer feedback •
Pros:
improve intonation skills • • • •
Cons:
suppress expression construct fake authenticity just tuning produces pitch drift we cannot separate timbre & tuning
So what about quartertones?
• • Quartertones simply lie between half-tone steps Like half-tones, they are
pitch categories
- not ratios.
• •
Non-western music theories
Ratio theories exist in many music traditions All are problematic for the same reasons
Microtonal composition
• • Intervals are ALWAYS learned ANY microtonal scale can be learned, but: • • •
A new scale is easier to learn if
similar to existing scales roughly equal small intervals (JND) unequal larger intervals (asymmetry) • •
Relevance of ratios
Approximate: yes (familiar harmonic series; minimize roughness) Exact: no
Microtonal composition
ETs that most closely approximate simple ratios have 5, 7, 12, 19, 31, 53 tones per octave
0
C
Logical next step is 19ET:
1
C#
2
Db
3
D
4
D#
5
Eb
6
E
7
E# Fb
8
F
9
F#
10 11 12 13 14 15 16 17 18 19
Gb G G# Ab A A# Bb B B# Cb C
0
C
Cf. 12ET:
1
C# Db
2
D
3
D# Eb
4
E
5
F
6
F# Gb
7
G
8
G# Ab
9
A
10
A# Bb
11
B
12
B
In this music, 19ET is like 12ET!
• • • 12 pitch categories – not 19 exact pitches based on a 7-tone diatonic
subset
Tuning is more important for anchor tones which may be grouping, metrical, melodic, harmonic, durational accents
Conclusions
• • •
Musical intervals are:
cultural and psychological (not mathematical) approximate (categorical) learned from music (an
aural tradition
) • • • • •
Exact musical interval size depends on:
musical familiarity consonance: harmonicity, roughness physical and perceptual stretch structure and voice leading emotion and expression
Origin of Western intervals
• • •
Familiarity of
harmonic complex tones
in speech (audible harmonic series) Prehistoric emergence of
scales
(= sets of psychological pitch categories)
Consonance
of tone combinations in music
We don’t need ratios to explain…
•
Major and minor triads; harmonic cadences
harmonicity, fusion, smoothness •
Tuning of violin versus piano accompaniment
octave stretch, leading tones, expression •
Character of Renaissance choral music
pitch structure, rhythm, timbre, expression •
Ratio-based microtonality (e.g. Partch )
Form, development, timbre •
Music’s meaning, beauty, magic
chains of associations
Imagine: A music theory without ratios
We can explain the structure, beauty, power of music without ratios
But there is a paradox:
• • You have to understand ratios… to understand intervals to realise that intervals are not ratios