The installation uses the following equipment:

  • Silicon Graphics O2 with a sound card providing 8 separate outputs in ADAT format. The O2 is also fitted with converters to enable it to read and send MIDI signals.
  • Converter for an 8-channel ADAT digital audio optical output to 8 balanced analog audio outputs.
  • A (minimum) 12 course lighting console that must be MIDI-controllable and send DMX512 out.
  • One (minimum) 12 course dimmer that can be controlled by DMX512.
  • 9 infrared transmitters and receivers
  • 1 interface box for photocells to communicate w/ O2 (MIDI or serial)
  • 8 speakers and 1 subwoofer
  • 6 small PAR16 spotlights
  • 6 larger theatre lights (either PAR or Fresnels)
  • 3 pairs of infrared glasses, mounted on tripods

The O2 monitors the photocells via an interface box that sends MIDI signals. Depending on the state of the beams, it then plays appropriate sounds to the relevant sound outputs. It also sends MIDI data to the lighting console to control the lights.

The sound card for the O2 provides 8 channels of 24 bit digital audio output through an ADAT port. ADAT is an optical port which interleaves 8 channels of sound in one cable. This then has to be converted through a D/A converter that will take an ADAT input and give eight analog outputs. One such converter is the Korg 880 D/A (not to be confused with the 880 A/D which takes eight analog inputs and gives one ADAT output).

All sound is stored in digital format on disk and indexed according to its content. All characters have their text divided into soundbytes (a few words, a phrase, a speech), as is the music. The computer keeps track of which soundbytes have been played and makes sure no repetition takes place until the character or music has exhausted its vocabulary. All sounds are in CD quality (44kHz, 16 bit). At any given time, 6 tracks are devoted to playing the characters and 2 tracks to playing music.

The music consists of layers of sound, so that there will be a constant bottom layer of bass that is only rarely silenced and at least one layer for playing shorter pieces of music on top of this bottom layer. In addition, music may be mixed into the voices as they speak.