top of page

Sound Systems

A Technical description of the Sound System by Elsa Garmire

Garmire Sound Sytsem Fig K2 sm.jpg

Sound System Diagram by Elsa Garmire



The speakers are located behind and broadcast right through the flexible mirror dome in a rhombic grid (see Figure Kl). The 37 speakers can be actuated separately to generate a myriad of sound experiences which are combinations of three basic types:


1.   The line sound;  a sound is switched at a rapid rate from speaker to speaker in any desired pattern.  Switching time from speaker to speaker is approximately a tenth of a second.  The amount of time a sound stays on any one speaker is programmable.


2.   Point sound;  a sound is heard from one speaker in the dome. Point sound can be shifted to any other speaker in the dome.


3.   Immersion, or environmental sound;  this effect is much like standing in a wood, or on a street, where sounds come from all directions.  This will require multiple recordings of sounds.


The acoustical properties of a dome are very similar to its optical properties since sound reflects from large smooth surfaces in a manner similar to light from a mirror.  Echoes and reverberation are really sound images occurring at a later time because of the slower travel time of sound compared to light.  This means that the voice of a person or speaker on one side of the Pavilion is heard distinctly as a sound image by those standing on the other side.  As with optical real images, the sound source appears to occur in the middle of space.  The position of sound images of voices is the same as the person's optical image.  This means that the real optical images appear to have voices.  However, because of the slower velocity of sound, the real image will look like it is speaking before the sound is heard.


The reason why sound and light images occur in the same part of the dome is that, just as with light, the angle of incidence equals the angle of sound reflection.  One difference is that a person can hear sounds coming from all directions at once  (unlike light), so that spherical aberration distorts the sound images.  However, since sound wave lengths are so much larger than those of light, accurate sound images would not be resolved as they are with light.


Special acoustic experiences occur in the center of the dome, which are analogous to the optical experience.  A person hears his voice in a very loud echo no matter which way he faces when he speaks.


The 37 speakers which comprise the sound system can be driven by as many as 32 inputs, which may be any signals in the audio-frequency range.  Available in the control room as sources are 16 quarter track monaural tape recorders and 16 microphone pre-amplifiers which can be connected to any combination of 25 microphone jacks.  Also available are a sine wave, square wave and pulse generator, and a record player. In addition, individual programmers may bring other audio inputs.


Groups of four signals from the 32 audio inputs are immediately summed in eight amplifiers.  (See figure K-2).  The 8 outputs from the summing amplifiers are in two groups of four.  Each of these groups of four is applied to an input switch.  Each input switch is a four by four matrix which can be controlled either manually from the control console or automatically via the input programmer which is a punched paper tape machine.  The input switch can place any of the four inputs on any of the four outputs.  The eight outputs from the two input switches are applied to the sound modification equipment.


The sound modification system allows each  output channel to pass through, or bypass, several stages of sound modification. The equipment consists of eight identical independent channels, one for each of the eight inputs.  Each channel consists of three sections, AM, FM, and filter (formant).  In Appendix E, the sound modification system is described in detail.


The 8 outputs from this system appear as audio inputs to the audio switch matrix.  In the switch matrix each of these eight channels is associated with a sequence program card and with 37 switches (one for each speaker).  A sequence card programs the order of switching speakers (which speaker or speakers receive the sound at any one time).  To provide flexibility, time of the speaker sequence is applied separately through the clock cards.  This controls the times at which the input sound switches from speaker to speaker.


Two sequence program cards are associated with one clock card. With 8 audio inputs, there are 8 x 37 = 296 switches, 8 program cards and four clock cards.  In addition, all sequence programs can be timed by a single master clock if desired.  Thirty-seven switches associated with the 37 audio outputs perform the actual switching of the input audio  to the power amplifiers.


The programming, (that is, speaker sequencing and channel selection) , cannot be manually operated but is performed by hard-wired programming cards.  A large number of these cards are made up with standard switching patterns.  However, individual programmers also make up their own cards.The speakers are 12" pioneer, dual axial type, with a frequency response from 20 to 20,000 cps.  Ten cubic feet of sound baffle is attached to the support structure behind each speaker.  The 50-watt driving amplifiers are located in bays behind the dome as close to the speakers as possible.

A Brief Introduction to the Sound Modifier Console
by Gordon Mumma

1.  Input-Output Programming


The Sound Modifier Console contains eight separate channels. The inputs and outputs of the Console are programmable by cards and tape.

The eight inputs (one to each modifier channel) are controlled by the Input Program.  Each of these eight inputs is the sum of up to four sound sources.  Each of the 32 sound sources has its own volume control.


2.  Basic Modifier Functions


On the Console itself the sound modifier functions for each of the eight channels are operated manually.  Each channel has three basic modifiers:  FM (frequency modulation), AM (amplitude modulation), and FILTER (a high-pass filter).  This configuration is indicated in Figure A, and detailed descriptions are given in Section 4 below.


The three basic modifiers are in series, so that a sound source can be modified first by FM, the result modified by AM, and that result can be modified by the FILTER.  This is the most complicated arrangement possible within a single channel.  Any or all of these three basic modifiers can be bypassed (by operation of an illuminated push-button).  Thus, it is easy to use the sound system with any amount of sound modification, or with no sound modification whatsoever.


3.  Examples of Programming Extremes


Since the sound sources are applied to the eight modifier channels by the Input Program, it is possible to program the type of modification to which the sources are subjected.


For example, if only a single sound source is to be used (such as a solo musician, or monaural tape), the Input Program can be arranged to route that source to any four of the eight modifier channels in any sequence.  If each modifier channel is preset to a different modification, then a sequence of up to four modification changes can be determined by the Input Program.


By two further procedures it is possible to extend the number of modifications for this example of a single sound source. First, the Input Program can be arranged to apply the sound source to two or more modifier channels at a time, allowing combinations of modifier functions.  By this procedure alone an enormous spectrum of modifications can be sequenced for a single sound source.

Second, the manual controls of the Sound Modifier Console can be operated independently of the input or output programs, allowing any re-arrangement of the functions within  any modifier channel.


At the other extreme, if all 32 sound sources are used (summed, of course, in groups of four sources to eight Console inputs) a programmed change of modification for any one input must be accomplished by exchanging modifier channels with another four of the eight modifier inputs.


4.  Sound Modifier Console Description


The FM function shifts the sound source upward or downward in pitch (or frequency) depending on the original pitch of the source.  The FM function has five controls:


REGISTER:  a switch with three positions (high, medium, low)


MODE:  a switch with four positions (fixed SHIFT, variable SHIFT, fixed VIBRATO, variable VIBRATO)


POS/WIDTH:  a continuously variable control which determines the pitch shift POSITION in the two SHIFT modes, and the vibrato WIDTH in the two VIBRATO modes


RATE:  a continuously variable control which determines the basic vibrato rate for the two VIBRATO modes


BYPASS:  an illuminated push-button which bypasses the entire FM function


Note:  the "variable" designation is a condition where the pitch shift (or VIBRATO, when in that mode) follows the amplitude envelope of the sound source.  This "envelope-variable' effect is direct.  In the variable SHIFT mode the pitch shift of the sound source increases with an increase in the loudness of the source. In the variable VIBRATO mode the rate-of-vibrato increases with an increase in the loudness of the sound source.  In all three basic modifiers (FM, AM, and FILTER) the envelope-variable circuitry always follows the amplitude envelope of the original (unmodified) sound source, regardless of the modifications which may be applied after the Console input.

The AM function imparts a periodically varying amplitude-envelope to the sound source.  The AM function has four controls:

RATE:  a continuously variable control which determines the basic period of the amplitude envelope.


WIDTH:  a continuously variable control which determines the on-to-off ratio of the AM.  The "on" part of the AM period is always the same loudness as the sound source itself.  The "off" part of the AM period can be adjusted from completely inaudible to the same audibility as the "on" part.


RATE MOD:  a continuously variable control to enable the AM RATE to follow the amplitude envelope of the original sound source. The maximum rate of change occurs at the maximum clockwise position of the RATE MOD control.

BYPASS:  an illuminated push-button which bypasses the entire AM function.


The FILTER function imparts a periodically varying high-pass formant to the sound source.  The FILTER has three controls:


RATE:  a continuously variable control which determines the basic period of filtering.


RATE MOD:  a continuously variable control to enable the FILTER RATE to follow the amplitude envelope of the original sound source. The maximum rate change occurs at the maximum clockwise position of the RATE MOD control.


BYPASS:  an illuminated push-button which bypasses the entire FILTER function.


5.  Further Notes on the Basic Modifier Functions


Of the three basic modifier functions, the FM function enables the most extreme modifications of a sound source.  Under some conditions a sound source can be modified beyond recognition.  With some sources certain FM control positions not only shift the pitch, but will compress or expand the harmonic spectrum of that source.


The SHIFT and VIBRATO positions which are "envelope-variable" will be of particular interest to musicians who are performing "live". By varying their dynamics they can have considerable "remote" control over the type and extreme of modification of their sounds.


Though of interest in itself, the AM function has special applications in conjunction with the FM and/or FILTER functions of the same modifier channel.  For instance, when the AM RATE is set to a different period than the FM VIBRATO or FILTER RATE, long-period changes of timbre or articulation are possible.  Further, the AM WIDTH control can be set in long-period conjunction with the FM VIBRATO control in such a way as to produce secondary articulation or modification of a single sound source.  One of these many effects is an inverted frequency-spectrum echo of the basic FM modification, a result which sounds as it two sources at different distances are being applied to the same modifier channel.


The FILTER function, which has the simplicity of only two controls (plus BYPASS), has particular relevance to the Pavilion. A considerable pc-.rl- of the sonic action of the Pavilion involves the rhombic grid of 37 loudspeakers which, in conjunction with the sound output programming (Switch Matrix and Sequence Programs), enables great diversity of sound movement in the Pavilion.  Thus, of the various kinds of filters possible, a high-pass filter was designed, in order to enhance the directionality of the sounds.  The higher frequencies of sound are inherently more directional than the lower frequencies.


This high-pass filter has a fixed cutoff-rate (approximately 1/2 db/octave) and a periodically varying cutoff-frequency. For reasons of increased diversity when all eight sound modifier channels are used simultaneously, the FILTER cutoff-frequency range is slightly different for each sound modifier channel.

bottom of page