The concept of re-synthesizing instrument spectra has been thoroughly explored

by spectral composers throughout the past three decades. The idea of re-
synthesis is based on sonographic analyses of the spectra of (acoustic)
instruments. A composer studies the characteristics of specific aspects of an
instrument's spectrum, analyzing the overtone structure of a specific pitch. This
includes determining the existing partials and their respective intensities, and
existing formants and their amplitude(s). Since these elements define the timbre
of an instrument, this technique has great potential - apart fromthe organization
of pitches - in theparameter of orchestration and the composition of timbre. For
example, an orchestra can "imitate" the sound of acello pizzicato, apedal-tone
from a horn, or a trumpet tone played with a mute, thereby becoming a "new
instrument" itself.
The technique of re-synthesizing (instrument) spectra has often been used
by French spectral composer Tristan Murail (b. 1947) to formthe basis of pitch
organization for a whole work as well as for certain sections of a piece. The
most striking characteristic of Murail' s musical aesthetic is his concept of a
continuum of 'harmony-timbre' 14. For Murail, both parameters are inseparably
linked with each other and thus become hybrid structures that combine the
characteristics and qualities of the two initial concepts. It is for this reason that
Murail makes frequent use of re-synthesis, since such operations provide the
composer with exactly those hybrid structures. Here, the timbral characteristics
of the re-synthesized spectra are mutually conditional to the construction of
I harmony.
Claes J. Biehl:
Spectral Analysis and
Re-synthesis`
(excerpt)
A second maj or aspect of Murail' s individual spectral aesthetic is his conviction,
that also harmony and form should be strongly interlinked. In other words,
harmony generates the formal course of a work resulting in further hybrid
structures defined by harmony and form. This idea illustrates Murail' s
connection to the tradition of tonal western art-music, which also shows strong
links between harmony and form, but at the same time it also points out the
composer's disapproving attitude towards atonal composition philosophies. For
Murail, harmony is the reference point for virtually all other musical parameters,
which means that a logic, meaningful and perceivable harmonic language is a
necessity for asuccessful and valuable composition.
A third aesthetic concept of great importance to the composer is the idea of
continuity. The "permanent" state of flux is a typical characteristic of Murail's
music, since most of his works are predominantly made of processes, which the
composer describes as 'sculpting sonic material'.15 This does not mean that the
music is entirely free of interruptions or irregularities, but the works as awhole
are strongly directional which is owed to the consequent and meaningful
interlinking of the aforementioned musical parameters.
These general observations about Murail' s composition philosophy can be
verified by analyzing his work Desintegrations. Written in 1982/83, the piece is
scored for seventeen instruments (2fl, lob, 2el, 1bsn / Ihm, ltr, ltromb / 2perc,
pno / 2vl, 1vIa, 1vel, 1db) and tape. The title mainly refers to the predominant
musical processes in the work, which are the (mostly gradual) transitions from
harmonious spectra to inharmonious ones and vice versa. The transfer of certain
partials to another octave within a spectrum could therefore be described as the
"disintegration" of pitches. Other techniques of disintegration are the
'exaggeration,16 of particular partials through the means of orchestration, or the
use of other filtering techniques. In addition, the title could also refer to the
15 op.cit, 123
160p.cit,141
processes of extracting melodies or melodic fragments from a spectrum, when
overtones arepresented horizontally and in different (not "natural") orders.
The work Desintegrations is subdivided into eleven sections, each of them
describing the transition from a harmonious spectrum to an inharmonious or
vice versa. The organization of harmonic progression, however, is based on a
variety of different techniques, as the following analysis will show.
Re-synthesis
Ring-modulation
Frequency shifting (linear and non-linear)
Frequency modulation (FM)
Harmonic distortion
The following paragraphs will give detailed analyses of representative passages
to illustrate how and to what effect these techniques are used by Murail in this
particular work.
The harmonic material fromthe extensive opening section of Des integrations is
entirely based on the results of the spectral analysis of aCl fromthe piano. This
sonographic analysis revealed - among other aspects - the formants of the
analyzed sound. In other words, Murail obtained the information on the intensity
of the resonance of each partial within the analyzed spectrum. The composer
was particularly interested in the spectrum's formants (loudest areas of
resonance), since the individual timbre of the low piano sound is defined by the
order and intensity of the formants.
Murail's spectral analysis
1
? showed the following results: there are six fields of
formants among the first forty partials of the spectrum based on aC1of apiano.
Partials 2-7, with astrong emphasis on 2and 7
Partials 9-13, with an emphasis on 11
Partial 15
Partials 17-23, with astrong emphasis on 21
Partials 27-30, with an emphasis on 29
Partials 35-38
~ 1
III
a:; 0,8
. . .
l::
~ 0,6
>
~ 0,4
a:;
0:: 0,2
o
The composer used this information to re-synthesize the spectrum and timbre of
a low piano sound as the underlying concept for the opening section of
Des integrations .
[]
13
12
Example 6 shows the first twelve aggregates
18
of section 1. The results of the
sonographic analysis have been applied to two different spectra, one with the
fundamental BflatO and the other with the fundamental Csharp2 through the
means of transposition. This way, the concept of re-synthesis is not only used as
a technique of imitating the instrumental model; in addition it also produced a
unique variant - the timbre of a particular spectrum eCl) transferred to two
different fundamentals at the outset of the piece.
Murail made use of this material by introducing acontinuous shuttling back and
forth between the two spectra, a process, which dominates the whole first
section of the piece. An analysis of the overtone structure of each aggregate
shows, that the composer exclusively used partials fromthe zones of formants of
the analyzed piano spectrum. Each partial located outside the six zones of
formants (partials 8, 14, 16,24-26,31-34, 39,40) is therefore left out. The
example also illustrates, that all twelve aggregates areharmonious.
Over the course of the section, Murail intuitively explores the two different
spectra. Each aggregate reappears in a slightly different form, e.g., partial 11
from aggregate I is left out in aggregate III whilst being reintroduced in
18 The term 'aggregate' describes an excerpt from aspectrum. Such an excerpt is ahybrid structure since it can
be perceived both as achord (harmony) and as asonority or atimbre.
aggregate V, where the thirty-sixth partial is left out instead. Notes are
constantly added or removed while the overall structure of each aggregate,
however, remains almost unchanged, which explains the subtlety of change that
the composer applied to theharmonic and timbral progressions.
The fundamentals of the two spectra (BflatO and Csharp2) are introduced with
the first four aggregates, but are then consequently left out only to be imagined
(added) by the listener instead. This phenomenon hints at Murail' s subtle and
sophisticated concepts of perception and his reflection onpsycho-acoustics.
Even though the fundamentals of each respective spectrum are left out, the
relationship and the distance between the two spectra remain clearly audible.
This is due to the fact, that the predefined range of each spectrum used for the
aggregates is almost the same for both spectra (Bflat: partials 1to 36; Csharp:
partials 1to 28). This means, that the abstract distance from the interval of the
tenth separating the two fundamentals, is constantly apparent, although the
fundamentals are actually not being played.
It is to a great extent the sudden disappearance of the two fundamentals which
strikingly points out the potential of re-synthesizing instrumental sounds: the
tape "extracts" and reconstructs the piano's timbre, which is usually hidden
behind the initial piano note (fundamental), and then heightens the listener's
awareness of it. In other words, the "subsidiary" parameter timbre becomes
harmony without losing its former timbral identity. This phenomenon can serve
as an example for Murail's concept of 'harmony-timbre'.
There are two further layers of differentiation applied to the re-synthesis of the
piano spectrum in section 1. One of them is achieved through the means of
orchestration. The sounds ontape, which contains the complete pitch material of
all aggregates in section I, are supported by the acoustic instruments. These
instruments filter or exaggerate certain partials from each aggregate. In
aggregate I, for instance, partials 7, 20, 29, and 36 are doubled by flute,
clarinets, and percussion, whereas in aggregate III partials 7, 13, 29, and 36 are
doubled. Murail enhances and reinforces certain formants within each aggregate
thereby adding another dimension to the subtle changes of harmony-timbre. A
further heightening of the complexity of sounds is achieved through the
harmonic overlapping of the acoustic instruments. In bar 5, for instance, partials
7and 11 from aggregate II are still sustained by the flutes, although the next
aggregate based on BflatO has already been introduced by tape, clarinets, and
percussIOn.
Another (inevitable) phenomenon of sound-enrichment is created through the
addition of the inherent spectra of each acoustic instrument. These spectra
increase the complexity of the composed sounds, but at the same time they
thwart the concept of an undisturbed instrumental re-synthesis as executed by
the tape, which makes the audible harmonic results ambiguous to a certain
degree. It should also be remarked in this context that due to practical reasons
the microtonal notation for the acoustic instruments has been limited to quarter
tones.
19
On the one hand the results obtained contradict and potentially disturb
the originally calculated harmonic structures played by the tape. On the other
hand, however, this compromised notation also generates a variety of complex
sounds that can be of great interest to the composer.
The second layer of differentiating the actual execution of the re-synthesis is the
dynamic distinction between the partials within each aggregate performed by the
tape. The dynamic hierarchy follows the hierarchy of the formants: the strongest
formants (partials 7, 11) are playedfortissimo, the second-strongest (12, 13,21)
forte, the third-strongest (5, 10, 19, 29) mezzo forte etc.
20
Here, the electro-
acoustic sounds actually imitate the timbre of the low piano spectrum.
19 Murai1 even abandoned all microtones in section VII due to the fast tempo.
20 op.cit, 213
As was mentioned before, there are further techniques used by Murail to
organize harmony in the first section of Des integrations . One moment of great
importance is the superposition of the two spectra inbar 30.
<l l>
13
7
# # = S :
#8~b__ J 1
8~b_J
[]
Bar 38
13
<l
2
#u-
The first part of Example 7 shows the merging of two aggregates from the
progression described before. The resulting aggregate incorporates excerpts
from the two different spectra whose respective characteristic qualities become
assimilated by the unified outcome. The sound heard in bar 30is ambiguous in
terms of its structure: an analysis is required to filter out the two components
that formthe final aggregate. Therefore, the spectrum cannot be considered to be
a harmonious one any more, so that at this point the previously explored
progression of harmonious aggregates is interrupted by the composer. The final
bars of section I eventually present a quick evolution from harmonicity to
inharmonicity.
The second part of Example 7 shows the final aggregate of section 1. Again, the
sound as a whole is ambiguous, but the analysis shows, that both subordinate
aggregates already contain inharmonious structures. Partials 17and 21 fromthe
BflatO spectrum as well as partial 13 from the Csharp2 spectrum have been
transposed down by an octave, while partial 19 from the latter spectrum has
been transposed down by even two octaves.
In summary it can be said, that Murail applied three essential methods to
organize harmony inthe first section of Desintegrations:
Partial re-synthesis of an instrumental spectrum and its transpositions
Creation of a double-spectrum through the superposition of
aggregates
Transforming harmonious aggregates into inharmonious ones
There are many examples from Desintegrations or other works by Murail that
could be chosen to discuss this last aspect. The exploration and interaction of
harmonious and inharmonious structures is probably the most frequently used
technique in spectral music. This is not surprising since this concept is strongly
related to that of "consonance and dissonance" in non-spectral music: it is to a
great extent its equivalent.
The second half of section III(bars 169to 192) introduces afurther technique to
generate harmony, which Murail frequently uses in his works: Frequency
shifting. This technique resembles that of traditional transposition since both
concepts describe a shifting of notes into a previously specified direction. The
major difference, however, lies in the fact that transposition is a linear process
whereas frequency shifting is non-linear (transposition refers to the shifting of
intervals, whereas frequency shifting deals with abstract frequencies).
Transposition byasemitone Frequency shift bv33Hz
~ ~I/~~~==- _ ===- - ====='~ :Z
The example illustrates the difference between transposition and frequency
shifting; the semitone shift from C5 to H4 (33 Hz) was chosen as the reference
point for the frequency shifting.
This phenomenon of "non-linear transposition" can be observed in section IIIof
Des integrations .
II III IV V
Ex. 8 Frequency shiftinginSectionIII (bars 169-192)
Example 8 gives an overview of the basic harmonic progression in bars 169 to
192. At the outset Murail introduces aharmonious five-tone aggregate based on
the fundamental Fl. The composer intended to compose a gradual frequency
shift that would take the course of the bass line as the point of reference. This
shift was supposed to cover the distance between the C3 of the first and the
Dflat2 of the final aggregate. Due to considerations of duration and form, this
process had to be divided up into ten steps. The distance between the
frequencies of C3 and Dflat2 amounts to 61.5Hz, so for each step of the
frequency shifting all frequencies of an entire aggregate had to be lowered by
6.8Hz. The results, however, did not please the composer (there was, for
instance, an F-major triad among the resulting aggregates), so Murail decided to
recalculate the whole process.
21
This time, 25 steps were calculated (each
aggregate lowered by 2.56Hz per step). Murail then intuitively chose the ten
aggregates from the collection of 25, which would eventually be used in the
This variant of constructing aharmonic progression based on frequency shifting
explains the irregularities of the harmonic progression shown in Example 8.
Aggregates separated by very small frequency shifts (e.g., I to II) alternate with
those separated by much larger shifts (e.g., VIII to IX), which means that afew
aggregates between VIII and IX have been left out. This approach, which could
be described as a merging of system and intuition, is very typical for Murail' s
way of composing, where the composer's ear always controls the results
obtained fromtechnical and mathematical operations.
In addition, this excerpt is agood example for Murail' s conviction, that neither a
'totally predictable nor a totally unpredictable discourse,22 are desirable for a
composition, and that 'structural predictability needs to be contradicted
constantly by some type of unpredictability within the discourse' .23 By
"filtering" the results he gained from his calculations through his own musical
intuition, Murail created subtle moments of irregularity within the continuity of
the macro-process (here: the frequency shifting), which precisely generate the
effect the composer was interested in.
Other than in section I, where the transformations from harmonious to
inharmonious spectra were deliberately composed, in section III these
transformations are literally inevitable. This is due to the fact, that frequency
shifting is furthermost a mathematical operation (each partial is shifted
individually!), which does not consider the construction of the various spectra.
In other words, the detailed structure of the resulting aggregates is of an entirely
theoretical nature.
Example 9 shows the overtone structure of the first and the last aggregate from
Example 8:
22 op.cit, 217
23 op.cit, 216
It becomes obvious, that the non-linear transpositions of the five pitches of each
aggregate have nearly entirely destroyed the previously harmonious structure of
the spectrum, which led to a final aggregate whose structure is almost
completely inharmonious.
A closer look at the score discloses further elements that take place
simultaneously to the frequency shifting described above. The processing of the
harmonic material has become much more complex than in section I, where
there were mainly homophonic structures rather plainly presented by the electro-
acoustic sounds. In this example from section III the material is distributed to
both the whole ensemble and the tape. The partials of each aggregate are
permanently interchanged by the means of orchestration (e.g., the "bass notes"
are shared by bass clarinet, bassoon and trombone), and there is a great amount
of figuration and inner motion (e.g., trills and melodic gestures). Furthermore,
there is an extensive amount of harmonic overlapping between the aggregates, a
concept which Murail had already introduced in section 1. In addition, the
textures are sometimes "filled up" with additional (distorting) pitches that are
not part of the underlying harmonic progression of frequency shifting. This
increasing amount of complexity is mainly due to the conceptions on the macro-
level (the third section serves as part of alarger build-up to the climax in section
VIII). It is on the one hand for this reason that the music gains additional layers,
but on the other hand this is also another way to avoid harmonic and structural
monotony and predictability.