THE ALGORITHMIC AUDITORIUM

A computational model for auditorium design

GANAPATHY MAHALINGAM
Department of Architecture and Landscape Architecture
North Dakota State University
Fargo, North Dakota
USA

Abstract. Auditorium design is a complex task. Various programmatic, functional and

acoustical parameters have to be resolved in the spatial design of an auditorium. This
ongoing research project deals with the development of a computer-aided design system
for the preliminary spatial design of proscenium type auditoriums. The concept of
“acoustic sculpting” is used to generate the spatial form of the auditorium from
programmatic, functional and acoustical parameters. These parameters are incorporated
using a combination of mathematical, empirical and statistical methods. The generation
of the spatial form of the auditorium is implemented as an algorithm that is executed on
the computer. The spatial form of the auditorium generated by the system is exported as
a computer model for design development and acoustical analysis.

1. Introduction

Auditorium design is a complex task. Various programmatic, functional and

acoustical parameters have to be resolved in the spatial design of the
auditorium. The emergence of sophisticated computational modeling tools has
now enabled the creation of design systems that treat the design of auditoriums
as an algorithmic process. In this paper, the design of proscenium-type
auditoriums is presented as an algorithmic process. This process is implemented
in a design system where the generator of the spatial form of the auditorium is
modeled as a “virtual computer.”

2. Auditorium Design Parameters

The complexity of auditorium design arises from the need to resolve many
interacting parameters. Some of the programmatic design parameters of the
auditorium include the type of performance that is to be presented in the
2 GANAPATHY MAHALINGAM

auditorium and the capacity of the auditorium. Programmatic parameters help

decide the dimensions of stage enclosures and seating areas. Functional
parameters include anthropometric constraints such as the area per seat, visual
constraints such as sight lines, and conditions for visual clarity. However, the
key parameters that influence the generation of the spatial form of the
auditorium are the acoustical parameters. Acoustical parameters are integrated
in the auditorium design system using the concept of acoustical sculpting.

3. Acoustic Sculpting

Acoustic sculpting is the creation of architectural shapes and forms based

primarily on acoustical parameters. It can be likened to sculpting, not with a
chisel, but with abstract entities such as acoustical parameters. Acoustical
parameters become special abstract tools that shape the environment in their
own characteristic way, hence the term acoustic sculpting.
In this context, it will be interesting to introduce the concept of a locus. In
planar geometry, loci are lines traced by points according to certain rules or
conditions. A circle is the locus of a point that is always equidistant from a
given point. An ellipse is the locus of a point whose sum of distances from two
given points is always equal. From these examples, it can be seen that a
particular rule or condition can trace a particular locus. The scope of application
of the concept of a locus can be dramatically widened by realizing that the word
locus in Latin means place. Architecture involves the creation of places and
spaces. A question can be posed - What is the locus of an acoustical parameter?
In answering that question, architecture based on acoustical parameters can be
created. Acoustics can become a form-giver for architecture. Figure 1 shows
how the time delay gap, an acoustical parameter, is used to generate a semi-
elliptical spatial field using the concept of the locus.
Acoustical parameters are often measured to assess the acoustical quality of
a space or a scaled architectural model. They are indicators of the acoustical
quality of the space in which they are measured. However, it is important to
realize certain facts about acoustical parameters. Acoustical parameters are
location specific. For a given sound source in a room, acoustical parameters
vary systematically at different locations in the room. Acoustical parameters
also vary when the sound source is varied. Hence, a set of acoustical parameters
at a given location, for a specific sound source, can be used only to generate the
general features of the architectural space around that location. Figure 2 shows
the source-receiver locations used in the design system. This, to stay within the
metaphor of sculpting, will result only in a first cut. Different sets of acoustical
parameters from different locations can further refine the definition of the
architectural space encompassing those locations. It has been found by
researchers that at least 10 to 12 sets of acoustical parameters are required to
 THE ALGORITHMIC AUDITORIUM 3

derive the mean values of acoustical parameters in an auditorium (Bradley &

Halliwell, 1989). If architectural shapes and forms can be created from
acoustical parameters, then a rational basis can be established for the creation of
acoustical environments.

Figure 1. Concept of locus used to derive a spatial field from an acoustical parameter.

Currently, the creation of acoustical environments is a trial-and-error

process that tries to match the acoustical parameters of the space being created,
probably in the form of a physical model, with acoustical parameters that have
been observed in other well-liked spaces. The manipulations of the space's
shape and form to achieve the match, are done in an arbitrary fashion, with no
explicit understanding of the relationships between the shape and form of the
space and the corresponding acoustical parameters. There has been extensive
research conducted in the 1960s, 1970s and 1980s by Ando (1985), Barron
(1988), Barron & Lee (1988), Beranek (1962), Bradley (1986, 1990), Cremer
(1978), Hawkes (1971) and Sabine (1964) to establish those aspects of the
auditory experience that are important in the perception of the acoustical quality
of a space, and how they relate to objectively measured acoustical parameters in
that space. There has not been much research conducted except by Gade (1986,
1989) and Chiang (1994) regarding the relationships between acoustical
parameters and the shapes and forms of the spaces in which they are generated.
Acoustic sculpting attempts to define the latter relationships and uses them
to create a system that generates spatial forms of auditoriums based on
acoustical parameters. This generative system is used as a tool for creating
preliminary designs of proscenium-type auditoriums.
4 GANAPATHY MAHALINGAM

Figure 2. Spatial form of the auditorium showing the source-receiver pair for acoustical
parameters.

3.1. METHODS OF ACOUSTIC SCULPTING

The process of generation of the spatial form of the auditoriums is related to the
set of acoustical parameters both statistically and theoretically. The acoustical
parameters for the generative system are drawn from, but are not limited to, the
set presented in the following section. This set of parameters is used by
acousticians to study concert hall and lecture room acoustics. These parameters
are derived from response graphs of sound intensity variations at the receiving
location. Figure 3 shows a response graph. Though the set is extensive, not all
of the parameters are used in the spatial form generation stage.

3.1.1. Acoustical Parameters

The acoustical parameters include Reverberation Time, Early Decay Time,
Room Constant, Overall Loudness or Strength of Sound Source, Initial Time
Delay Gap, Temporal Energy Ratios: Early/Total Energy Ratio (Deutlichkeit),
Early/Late Energy Ratio (Clarity), Center Time, Lateral Energy Fraction,
Spatial Impression, Bass Ratio, Bass Level Balance, Early Decay Time Ratio
 THE ALGORITHMIC AUDITORIUM 5

and Center Time Ratio, Useful/Detrimental Ratio, Speech Transmission Index

and the Rapid Speech Transmission Index.
The different acoustical parameters cited above resolve into related groups
that have corresponding subjective perception characteristics. These subjective
perception characteristics are classified as Reverberance, Loudness, Clarity,
Balance and Envelopment.
A limited set of acoustical parameters related to these subjective perceptions
are incorporated in the system (using both statistical and theoretical methods)
that derives architectural parameters from the acoustical parameters. It must be
remembered that, in the spatial form generation stage, acoustical parameters are
not the only factors determining the shapes and forms of the auditoriums. Other
factors like seating requirements, visual constraints and other programmatic
requirements, along with the acoustical parameters, determine the spatial forms
of the auditoriums. The values of the acoustical parameters for use in the
generative system are drawn from a database of objectively measured readings
in different architectural settings that have been subjectively evaluated as
desirable. Based on studies done so far, a generative system based on
macrostatic statistical relationships and some analytical theory has been
developed by the author. Details of this system are to be found in another paper
by the author (Mahalingam, 1992).

Figure 3. Response graph showing sound intensity variation over time at receiver
location.
6 GANAPATHY MAHALINGAM

4. Spatial Model Of The Auditorium

The spatial form of the auditorium is modeled as a parametric object. Each

vertex that makes up the topology of the auditorium is spatially located by a
function of multiple parameters. These parameters may be directly input by the
user of the design system or derived from the user input using calculations. The
various parameters are linked in a spatial form generating algorithm using a
structure that resembles an ASIC (application specific integrated circuit). Figure
4 shows this relationship of the various parameters. This structure can also be
reconfigured as a network or semi-lattice. The connectivity of the vertices that
establishes the topology of the auditorium is derived from the spatial type of the
proscenium auditorium. The whole design system is a “virtual computer” that
outputs spatial designs of proscenium-type auditoriums.

Figure 4. The relationship of the various parameters of the spatial form generator as an
ASIC (application specific integrated circuit)
 THE ALGORITHMIC AUDITORIUM 7

5. What Is A Virtual Computer?

In object-oriented computing, entities are modeled as encapsulations of data,

and operations that can be performed on that data. Encapsulation is a computer
abstraction. A collection of data and operations normally performed on the data
are closely related, so, they are treated as a single entity (rather than separate)
for purposes of abstraction. Each encapsulation can be thought of as a virtual
computer that is mapped onto a physical computer (see Figure 5) with its own
private memory (its data) and instruction set (its operations). The reference to
objects as computers was made by Alan Kay (1977). He envisaged a host
computer being broken down into thousands of computers (virtual?), each
having the capabilities of the whole, and exhibiting a certain behavior when
sent a message which is a part of its instruction set. He called these (virtual?)
computers "activities." According to him, object-oriented systems should be
nothing but dynamically communicating "activities." As such they form an
interesting model with which to simulate architectural design. Mitchell’s recent
call (1994) for a “society of design” with a “collection of agents of different
kinds interacting over a network” echoes the ideas of Alan Kay. In another
interesting perspective, encapsulations have been likened to integrated circuits
rather than virtual computers by Ledbetter & Cox (1985) (see Figure 4).

Figure 5. The concept of a virtual computer (or computational object) being mapped
onto a physical computer.
8 GANAPATHY MAHALINGAM

6. The Auditorium Design System

The design system used to generate the preliminary spatial designs of

proscenium type auditoriums is based on acoustical, functional and
programmatic parameters. The computational model of the auditorium is
parametric. The various acoustical, functional and programmatic parameters are
its data. Procedures that compute the spatial parameters of the auditorium and
create its graphic representation are its operations. These data and operations,
when encapsulated, act as a virtual computer that is mapped onto the physical
computer (see Figure 5). The function of this virtual computer is to output
auditorium designs.
The generative system involves an algorithmic procedure for the design of
the auditoriums based on constants, user input of independent variables and
derived variables. These constants and variables are used to calculate the spatial
location of sets of vertices in 3D space that are linked to form wire-frame and
shaded plane images of the auditoriums. The topology of the auditorium is
based on the proscenium-type auditorium typology. It is a variant topology with
the introduction of balconies only when necessary (see Figure 6). The vertices
are parametrically controlled and change with changing parametric inputs.

Figure 6. Topological model of the auditorium showing the variant topology for the
balconies.
 THE ALGORITHMIC AUDITORIUM 9

The algorithmic procedure is implemented in the SmalltalkTM object-

oriented programming language. The software has a user-friendly menu and
graphic interface with which to input acoustical, functional and programmatic
parameters. When any aspect of the model is changed, the spatial form is
updated. The system provides a dynamic design environment. In the system, the
spatial form changes in real time with changing input of the parameters. The
auditorium is depicted in true perspective. Once the spatial form is generated, it
can be viewed from any angle and from any distance. The systems can be used
to rapidly generate alternate designs based on the various parameters.
To limit the scope of the software design to manageable limits, the initial
version of the generative system has a limited set of 21 independent variables.
However, the total number of variables (both independent and derived) in the
system is large, indicating a complex system. An interface has been developed
that can transfer the computer model generated by this system in a format
readily acceptable by commercial CAD packages (the DXF format) for design
development. An interface has also been developed to link this system to
acoustical simulation software (EASE and EARS) to predict what the
auditorium will sound like if it is built. The view of the computer screen when
running the software is shown in Figure 7.

Figure 7. View of the screen of the design system software for auditoriums.
10 GANAPATHY MAHALINGAM

Acknowledgements

Much of this research work was completed as part of the author’s doctoral
dissertation under the supervision of Professors John Alexander and Gary
Siebein at the University of Florida, Gainesville, USA.

References

Ando, Y.: 1985, Concert Hall Acoustics, Springer-Verlag, Berlin.

Barron, M.: 1988, Subjective Study of British Symphony Concert Halls, Acustica,
Vol.66, No.1, pp. 1-14.
Barron, M. and Lee J.: 1988, Energy Relations in Concert Auditoriums, Journal of the
Acoustical Society of America, Vol. 84, No. 2, pp. 618-628.
Beranek, L. L.: 1962, Music, Acoustics and Architecture, John Wiley & Sons, New
York.
Bradley, J. S.: 1990, The Evolution Of Newer Auditorium Acoustics Measures,
Canadian Acoustics, Vol.18, No.4, pp. 13-23.
Bradley, J. S.: 1986, Auditorium acoustics measures from pistol shots, Journal of the
Acoustical Society of America, Vol.80, No.1, pp. 199-205.
Bradley, J. S. and Halliwell R. E.: 1989, Making Auditorium Acoustics More
Quantitative, Sound and Vibration, February, pp. 16-23.
Chiang, Wei-Hwa: 1994, Effects of Various Architectural Parameters on Six Room
Acoustical Measures in Auditoria, Ph.D. Dissertation, University of Florida,
Gainesville.
Cremer, L.: 1978, Principles and Applications of Room Acoustics, Vol. 1, (Translated
by T. Schultz), Applied Science Publishers, London, England.
Gade, A. C.: 1989, Acoustical Survey of Eleven European Concert Halls, Report No.
44, The Acoustics Laboratory, Technical University of Denmark, Lyngby.
Gade, A. C.: 1986, Relationships Between Objective Room Acoustic Parameters And
Concert Hall Design, Proceedings of the 12th International Congress on Acoustics,
Vol./Band D-G, E4-8, Toronto.
Hawkes, R. J.: 1971, Experience in Concert Auditoria, Acustica, Vol. 24, pp. 236-250.
Kay, A. C.: 1977, Microelectronics and the Personal Computer, Scientific American,
September, pp. 230-244.
Ledbetter, L. and Cox B.: 1985, Software-ICs, in Byte, June, pp. 307-316.
Mahalingam, G.: 1992, Designing the Sound Environment: Acoustic Sculpting,
Proceedings of the ACSA Technology Conference, San Diego, February.
Mitchell, W. J.: 1994, “Three paradigms for computer-aided design,” in Automation in
Construction, Vol. 3, Numbers 2-3.
Sabine, W. C.: 1964, Collected Papers on Acoustics, Harvard University Press,
Cambridge, Massachusetts, Reprinted by Dover Publications.