The
sounds in the following piece, Chréode, by Jean-Baptiste
Barrière, are based on something that all of us are familiar
with: the sound of the singing human voice.
All
of the sounds in this piece are synthesized by a computer: there
are no recordings of actual performers used. Instead, the computer
simulates the method of vocal production and provides the composer
with an array of parameters to control the synthesis. Barrière
manipulates these different parameters to create a sonic texture
that moves away and towards recognizable vocal sounds.
To
demonstrate the level of detail that Barrière worked at to
shape the sounds of the piece back in 1983, I will be using a program
that will represent graphically, in real-time, the frequency content
of sound across time. The time axis runs top to bottom, the most
recent events occurring towards the bottom, and the frequency axis
runs from left to right, higher pitches or frequencies towards the
right. Let's see what a series of vowels looks like in this representation:
(image 1, three vowels;
audio excerpt 1).
Notice
that each vowel has a different 'signature' that we can see visually,
made up of different 'peaks' of loudness in different parts of the
audible range. It is this arrangement of different elements that
combine into an end result that we can categorize as 'a vocal sound,'
and it is at this level of detail that Barri¸re is able to digitally
manipulate the synthesis.
All
of these sonic 'signatures' are grouped, transformed and developed
throughout the piece. In Chréode, Barrière
uses computer-aided processes to direct how the sounds develop from
one state to another. For example, one process might dictate the
rhythmic activity of a particular section while another process
governs the general trajectory of pitches. To demonstrate this I
will use another representation of the frequency content of sound
across time. This time the time axis scrolls to the left, while
the frequency axis runs bottom to top, the higher frequencies towards
the top. A very simple example of a process is an ascending glissando.
Let's see how this is shown on the screen (image 2,
audio excerpt 2).
Notice
that this representation is very similar to the previous one you've
seen, but now it allows us to get a better sense of how a particular
family of sounds change over time.
As
you listen to Chréode, feel free to look at--or ignore--these
two representations (spectral representations run in real-time during
performance).
|