5 – Microtemporal listening dimensions: Timbre, Harmony and Pitch height
73
Risset and Wessel concluded that the attack transients constitute an impor-
tant part of instrument tones. If the attack segment of a tone is removed in
a tape recording, the instrument is no longer recognizable.
Many tones like those produced by the piano or percussion instruments
are characterized mainly by the complex temporal evolution of their tran-
sients, as they have no steady state at all. Transients are intrinsically
complex, and they are not reproducible from one tone to another. Houtsma
(1989) points out that high and low tones from an instrument normally
have different spectra; a low piano tone typically contains little energy at
the fundamental frequency and has most of its energy at higher partials,
while a high piano tone typically has a strong fundamental and weaker
higher partials.
The multidimensional nature of timbre has been investigated by Carol
Krumhansl (1989). With David Wessel, she conducted an experimental study
of the similarities and dissimilarities of 21 timbres synthesized by means a
frequency modulation technique. Most of the timbres were designed to
simulate traditional instruments such as horn, trombone, trumpet, oboe,
clarinet, bowed string, guitar, harpsichord and piano. A few others were
synthetic hybrid timbres such as "guitarnet", a hybrid of guitar and clarinet,
and "striano", a hybrid of strings and piano. A group of musically trained
listeners were asked to judge the relative similarities of these timbres, and the
obtained data were treated by a multidimensional scaling technique.
As a result of this study, three common dimensions of timbre were
found. The first dimension corresponds to the rapidity of attack, reflecting
differences for example between the sharp attack of plucked instruments
like harpsichord and guitar and the comparatively slow attack of horns or
bowed strings.
The second dimension corresponds to brightness, depending upon the
distribution of power in the sound spectrum. In relatively bright instru-
ments like the oboe and trumpet, energy is concentrated in the higher
components, while instruments such as horns and trombones are charac-
terized by energy concentration in the lower components.
The third dimension, named spectral flux, corresponds to the temporal
evolution of spectral components, reflecting differences for example
between woodwind and brass-like timbres, the latter characterized by
spectral contents changing with amplitude.
In addition, some timbres were found to possess specific qualities that
are not explained by these three dimensions, such as the clarinet timbre
which is unique in its absence of even harmonics.
The dimensions proposed by Krumhansl have been confirmed by later
acoustic analyses. Donnadieu et al. (1994) conclude that attack quality is
highly correlated with the logarithm of attack time, and brightness is highly
correlated with the spectral center of gravity. The third dimension, under-
stood as spectral fine structure, is well correlated with the ratio between the
amplitudes of even and odd harmonics.
ng.
5 – Microtemporal listening dimensions: Timbre, Harmony and Pitch height
75
When pitch height emerges in a timbral spectrum of a musical instrument
by a perceptual focusing at a certain level of the pitch continuum, the
sound is perceived in two simultaneous dimensions, the quality of the
timbre of a particular instrument and the quality of pitch height. Timbre
and pitch height are distinct qualities of the microtemporal continuum,
permitting the distinction of sound sources and the distinction between
higher and lower pitches. Between timbre and pitch height, a diffuse
quality can arise, the quality of harmonic color.
Harmony
Harmony arises as a specific color quality from the presence of several
simultaneous focal areas in a perceived timbral spectrum.
The simplest harmony is the musical interval, arising as a particular
sound color from the interaction of the focusing qualities of two simul-
taneous pitch heights. An interval is not an addition, but an interaction of
two components giving rise to a new emergent quality
Wright and Bregman provide this explanation;
Musicians are well acquainted with the idea that two tones sounding
simultaneously form a new whole exhibiting a quality which is more
than the sum of the qualities of the individual tones taken separately.
Such a quality might also be called an emergent quality. We depend
upon this quality to identify harmonic intervals, and to classify them
as consonant or dissonant according to their varying degrees of
qualitative roughness. Tonal simultaneities built up of one or more of
these intervals have been called "chords", their emergent qualities
can be called "chord color", and the process by which the indepen-
dent tones combine their effects to create this quality has been called
"tonal coalescence", or "chord fusion." (Wright and Bregman, 1987)
The musical interval has a double nature. It permits the identification of its
two constituting pitch heights, and simultaneously, it displays its specific
color quality, recognizable in different positions in the pitch
continuum.
If a third tone is placed between the two tones of an interval, it interacts
with the components of the interval, producing a new emergent color
quality. On the keyboard of a piano, a series of different colors can be
demonstrated as shown in Fig. 5.2.
ng.