On a warm summer day, reading by a pool, Gert kept hearing voices despite the being completely alone. He realised that these sounds were caused by the movement of the water causing the flaps of the filter to open and close. As a result of the water flowing in and out of the filters, the air inside was compressed and pushed out of the compartment through a small opening in the lid which caused it to vibrate. The housing of the lid started operating as a sound box. No sounds was discernible standing immediately next to these filters. However from a distance, and with the mind attuned to reading, these sounds turned once again into voices.

Inspired by Wolfgang von Kempelen’s research into speech synthesis as well as by his own experience of a whispering swimming pool, Gert Aertsen year-long research residency explored the mechanics of speech. Aertsens project attempted to create a mechanical speech synthesizer – a system copying the mechanisms of the human speech production organ. Not so much an object that speaks, but an object or installation with sounds reminiscent of talk, whisper, moan, groan.

Wolfgang Von Kempelen was an 18th century's Austrian inventor and probably most famous for his invention of the mechanical Turk, a chess playing automaton. His design can be divided in three main parts:

  • The bellow to invoke the function of the lungs
  • A reed, much like that in an organ pipe or other reed instruments to copy the function of the vocal chords
  • A resonator or acoustic cavity, to copy the function of the mouth.

Also Gert Aertsen is using these three parts: voicing, resonator and bellow. For visuals, check documentation.

The result was a site specific sound installation for the 11th floor of the Overtoon Kanal office building. Since this particular building stands high and free in the city center of Brussels, wind has a strong impact on it. By opening the windows on both sides of the space, a drafty situation occurs and the question becomes: can the energy created by the wind be used as a force or input to generate sound, which sounds remotely like speech?