Sound design team -> Activities -> Installations -> Reflections


What - where

 

"Reflections" was part of the Resonances festival at Ircam in October 2002. It is an audio only installation - no visuals.

It uses experimental speakers based on new technologies, such as the Audio Spotlight and the Cube.
Not only these speakers can produce interesting effects by themselves, but their simultaneous use enables a whole range of unheard of virtual spaces.

"Reflections" was located at Ircam, research space, level -3, in a way that can be seen on this general layout.

 


The music

"Reflections" is based on a 15 minute long original music composed by Emmanuel Deruty & Pierre-Yves Macé.
This is a 6 part piece, with occasional transition elements derived from the first part.

 

* Structure

Each part features very different sources and timbres.

Part I : modified piano and high pitched metal percussions, with imaginary radio speaker
Part II : modified marimbas and magnetic type noise
Part III : electronic oscillators following very distinct rythmic structures
Part IV : low pitched metallic sounds and modified organic instruments, in a very large acoustic space
Part V : tonal string ostinato and filtered drum programmings, with imaginary radio speaker
Part VI : natural clarinet on a piano ostinato, with modified vibraphones

 

* Writing

Mostly, the piece is based on what can be referred to as "weak counterpoint" technique.
This is a polyphonic writing technique in which the links between different "horizontal" parts are approximative.
The trick is to find the optimum point where two simultaneous parts are still related, but not too strictly, not too precisely.

This technique aims at the maximum perceptive wealth. The result takes place at an interesting limit between strict order and unrecognizable disorder.
It was inspired by the observation of certain real acoustic situations, which are full of informations to a point in which it is quite difficult, but still possible, to describe every element which takes place in the audio scene.

The counterpoint, instead of relying on strict and hard relations between parts, shows what can be called "weak forms".

A "strong form", like a square for example, is immediately recognizable, categorizable, and can easily be put in memory.
A "very weak form" would be a lot of little dots placed randomly on a 2D surface, seen from afar.
In classical counterpoint, relations between parts can be seen as "strong forms". In "weak counterpoint", the forms are "weaker".

Two typical "mild counterpoint" examples : one taken from a piece from Emmanuel Deruty, the other from a piece from Pierre-Yves Macé.
( - Examples in Real Audio 8+ format, streaming 176kbs - the whole pieces can be found on 1-1-1-1.net )

 


Spatialization techniques

On this layout ( view from top ), one can basically see the principle of use for each source.
Different spatialization techniques have been used in relation to musical writing - here are the most interesting.

 

* Cube techniques


Mostly, the Cube was used taking profit of the volume approach, in order to create a wide sound halo around it.

In part III for instance, there was a space polyphony in which the electronic oscillators, which are very precise and thin sounds, were directed to the listening zone 1, while a shapeless pad was projected in volume approach. This contrast was enhanced by the use of a "large hall" reverb on the shapeless pad, while the oscillators were strictly dry.

Another example of space polyphony, in part V.
The imaginary radio speaker voice, treated in terms of reverb with another large hall, is projected on the wall, thus creating a seemingly distant space behind the Cube.
In the meanwhile, the strings are projected laterally using a slightly vibrating dipole ( figure of eight ), thus creating a nearer, but very diffuse and unlocalizable space.

A more extreme example : in part VI, the pianos and marimbas are projected using an horizontal figure of eight pattern - but this pattern is constantly and widely vibrating. This spatial vibrato creates a strange kind of vintage Leslie effect, which goes very well with the actual piano timbre, modified in order to sound like an old piano.

 

* Timbre and space split principles


Let's take an example : part I. In this part, we can find, amongst other sounds, hi pitched metallic percussions.

These metallic percussions are projected using the three speakers simultaneously.
- The standard speaker does its job all right and the percussions sound like themselves
- The Cube projects the sound using a slowly rotating figure of eight. This leads, in terms of spectrum, to a low cut filter, 6dB per octave.
In term of space, this leads to a virtual source which is very diffuse and difficult to locate.
- The ASL projects the sound on the reflective panel - the sound is cut below 2kHz.

Cut because 2kHz, because the AudioSpotlight is still some kind of a prototype - when using it, one can encounter certain problems, one problem being the fact that the ASL cannot playback frequencies under somewhere like 2kHz....

Anyway, this is a very interesting example - it gives an illustration to two of the main principles used during this installation :

- The timbre split principle
Speaking in terms of spectrum, one single sound is "spread" on different speakers.
This does not mean that the whole spectrum is strictly divided in three bands, and that these three bands are then projected on three different speakers - which would lead to moving sounds in case of gliding sounds, and split sounds otherwise.
Here we want to create living sounds : the sound's spectral diversity is translated into inner movement - the sound gets very large and the placement of the sound's "inner components" never stop moving inside the global shape.

- The space split principle
It is obvious that the three speakers radiate in very different ways.
This difference can be called "non uniform projection" - and this kind of projection leads to a great perceptive wealth : depending on the listener's position, the music's aspect in terms of timbre, perspective and texture is very different.
Whereas the listening zones emphasize on certain space phenomena - in a static listening position ; but when the listener moves, the colors never stop changing.

It is possible to find more details concerning these spatialization techniques and other ones, on the multi-speaker diffusion dedicated page.