Friday, 14 July 2006

Show / Release: Hijmans van den Bergh-building, University of Utrecht

Permanent installation of a neural network driven particle based sound piece in the
Hijmans van den Bergh-building, University of Utrecht, Netherlands.

A.A. Hijmans van den Berghgebouw
Universiteitsweg 98
3584 CG Utrecht

In a panel discussion at the International Computer Music Conference 1992 in Delphi, Iannis Xenakis stated:
"The point is that I don't need to try with computers to imitate this sound that exists already. You don't need that. What is interesting is to explore other paths or ways or sounds or even evolutions of sounds that have never been done or realized, and that is the interesting point."

The sound piece is based on a generative composition process running on a computer. The generative process is based on cellular automata driven neural networks which control several sound synthesis and processing algorithms. The neural networks are educated by the composer to produce desired output.

"These artificial neural networks are advantageous, especially in pattern recognition and classification tasks. Artificial neural networks are usually applied to tasks that are generally easy for humans to accomplish, but difficult for traditional computers. Traditional computers remain superior to neural networks in many situations, but neural networks are dominant in recognizing patterns in extremely large data sets. With the ability to learn from a set of examples and generalize this knowledge to new situations, neural networks are excellent for work requiring adaptive control systems."

The sound generation is created with a customized version of SuperCollider (see Software Links) and is running on GNU/Linux based operating system using a regular server computer (see Default Hardware). Just like clouds in the sky, sonic particles will accumulate and dissolve constantly to create an ever changing sound structure in the building. Oscillating between the constant light of the neon work and the pseudo endless light sequence of the chandelier, the generative sound piece is the third, almost holographic link between these different time-scales in light, and ultimately in sound.

Besides the range of deliberately determined structures of the sound piece, the neural network will be educated by my practice in 'doing' live electronic music, and installed at the university create music in 'the style of me'.

Software Development: Tommi R. Keränen