Conference PaperPDF Available

From Pattern Making to Acoustic Panel Making Utilizing Shape Grammars

Authors:

Abstract and Figures

This paper presents the application of shape grammars in a real case design problem. The design problem is stated as developing computational acoustic panel solution for classroom with various acoustic problems by modifying 2D pattern, which basically utilizes shape grammars. The study demonstrates interdisciplinary environment of design education and discussions of shape grammars in acoustic panel design and making. It includes different methods to design, like intuitive tendencies, computational thinking, computational tools, and computer simulations. The rule sets of the 2D (pencil-paper-based) pattern are intuitively created by the designer with simultaneous studies of understanding shape grammars. The study consists of three stages. The first stage illustrates 2D pattern generation utilizing computational thinking via shape grammar methodology, second stage illustrates computer generation of 2D pattern with the help of computational tools, and the third stage utilization and modification of this 2D pattern into 3D acoustic panel with feedbacks of computer simulations.
Content may be subject to copyright.
From Pattern Making to Acoustic Panel Making Utilizing
Shape Grammars
Ecenur Yavuz1, Birgül Çolakoğlu2, Begüm Aktaş3
1Istanbul Technical University, Institute of Informatics, Graduate Program of Ar-
chitectural Design Computing 2,3Istanbul Technical University
1,3{ecenuryavuz|aktasbegum}@gmail.com 2bigi@alum.mit.edu
This paper presents the application of shape grammars in a real case design
problem. The design problem is stated as developing computational acoustic
panel solution for classroom with various acoustic problems by modifying 2D
pattern, which basically utilizes shape grammars. The study demonstrates
interdisciplinary environment of design education and discussions of shape
grammars in acoustic panel design and making. It includes different methods to
design, like intuitive tendencies, computational thinking, computational tools,
and computer simulations. The rule sets of the 2D (pencil-paper-based) pattern
are intuitively created by the designer with simultaneous studies of understanding
shape grammars. The study consists of three stages. The first stage illustrates 2D
pattern generation utilizing computational thinking via shape grammar
methodology, second stage illustrates computer generation of 2D pattern with the
help of computational tools, and the third stage utilization and modification of
this 2D pattern into 3D acoustic panel with feedbacks of computer simulations.
Keywords: computational design, computer-generated geometrical design,
shape grammar, acoustic, odeon
INTRODUCTION
Shape grammar has been used for generating ge-
ometries and forms as a generative design tool, sys-
tem, and language. According to Stiny, a shape
grammar defines a set called language (Stiny, 1980).
From the introduction of the shape grammar to de-
sign methodology onwards, shape grammar, that is
widely used in design education and practice has
been evolved as a generative design tool by recur-
sively applying shape rules to the initial shape. As
Knight stated that, “A shape grammaris a set of shape
rules that apply in a step by step way to generate
a set, or language, of design. Shape grammars are
both descriptive and generative. The rules of a shape
grammar generate or compute designs, and the rules
themselves are descriptions of the forms the gener-
ated designs.” (Knight, 2000). Novel use of the shape
grammar in design education in practice enrich the
abilities of designers as well as providing design al-
ternatives to designers.
In the early 1970s, George Stiny and James
Gips first presented shape grammar theory to de-
sign methodology. As Stiny and Gips point out,
shape grammars are similar to phrase structure gram-
SHAPE, FORM & GEOMETRY | Applications - Volume 2 - eCAADe 36 |477
mars, which were introduced by Chomsky in lin-
guistics. Where phrase structure grammars are de-
fined over an alphabet of symbols and generate one-
dimensional strings of symbols, shape grammars are
defined over an alphabet of shapes and generate n-
dimensional shapes (Stiny & Gips, 1971). Through
development and application in shape grammar ap-
proaches, the generation of complex forms and pat-
terns are defined by basic formal rule sets. Within this
context, a designer has a chance to formalize or de-
sign the generation of any shape. Thus, shape gram-
mar is not only capable of generation of complex de-
signs, but also a powerful tool for computational pat-
tern generation.
In the design education and practice, shape
grammars have been used as a tool for teaching com-
putational thinking and design to architects and de-
signers in various graduate level courses and design
practice. “The first two decades of shape grammar
applications focused almost exclusively on analysis.
(Knight, 1999). In the earlier stages of the shape
grammar approach, shape grammar is utilized as de-
veloping new design languages based on existing
ones as well as capturing the inferring grammar of
the design, transferring the defined rule sets to gen-
erate designs as an infinite number of new designs
such as new design language or style. Knight (1999)
states that shape grammars are used by students to
understand the design language and manipulate the
existing design with its rules to generate a new de-
sign. Within this context, initially, students analyse
the existing patterns which could be architectural
styles or design languages to formulate (extract) their
own rules for their designs. Therefore, analysing pro-
cess has potential in design education and practice.
“There is no better way to learn about styles or lan-
guages of designs (at least compositionally) than by
either studying shape grammars already written for
languages or by writing grammars oneself.”(Knight,
1999). Shape grammars, on the other hand, have a
basis of cognition-oriented design theories as com-
putational theory and making. Cognitive computa-
tion in the shape grammar is conducted with the em-
bedded integration of seeing and doing as is hand-
eye coordination in the design and making process
in the creative design process like as Stiny concludes
that; “What you see is what you get”(Stiny, 2006). Un-
dergone a relationship between seeing and doing,
shape grammars is defined as computation design
and thinking as cognitive computation.
As a way of understanding the trajectory of
Stiny’s ideas of shape grammar as are the importance
of the integration of the seeing and doing, and a ba-
sis of the mathematics behind the whole process is
maintained as follows; “I am going to use my formula
to get the idea that calculating is visual if it can deal
with shapes like this. I want to use rules to determine
what parts I see and what I can say about them, and to
allow for what I see and what I say to change freely as
I calculate. And I want this to happen every time I try
a rule.”(Stiny, 2006) Therefore, the pattern analysing
process is driven recursively depending on the cog-
nitive computation, while seeing the sub-shapes on
the pattern and intuitively decoding the sub-shapes
of the pattern to determine the geometric definition
of the pattern. Computer application of shape gram-
mar make geometric logic embedded in pattern easy
to understand, enabling better control, modification,
and construction. This study differs from the earlies
studies where rules for pattern making are extracted
from analyses of existing pattern. Here, the rules of
the pattern are created by the designer as a cogni-
tive computation based pattern. Then, they are trans-
formed to mathematical rules for automatic genera-
tion of the patterns.
The study consists of three stages. The first stage
illustrates 2d pattern generation through the shape
grammar theory with developing a computational
thinking, second stage illustrates computer genera-
tion of 2d pattern with the help of digital tools, and
the third stage utilization and modification of this 2d
pattern into a 3d acoustic panel with the feedbacks
of computer simulation results.
478 |eCAADe 36 - SHAPE, FORM & GEOMETRY | Applications - Volume 2
FIRST STAGE - Paper Based Pattern Genera-
tion
Shape grammar is the computational design
methodology, derived based on mathematics and
algorithmic thinking. The shape grammar based de-
sign process, uses geometric elements, shapes as ba-
sic elements of the design like applying mathemati-
cal operations to compute calculations with shapes.
The rules of the shape grammar generate designs
by computing directly with shapes made of basic
spatial elements (points, lines, planes, and solids),
rather than with symbols, words, numbers, or other
abstract structures that represents visual shapes indi-
rectly (Knight, 2018). Within the context of the shape
grammar definitions, the hexagon is accepted as the
initial shape of the pattern. The designer defined the
rule sets of the 2D pattern with simultaneous lectures
and shape grammars.
In the first stage of the study, main objective was
to create paper-based 2D geometric pattern, basic
(initial) pattern making steps are defined for creating
the hexagonal grid to generate the hexagon-based
pattern. The smallest unit, the hexagon shape is cho-
sen intuitively by the designer. Hexagon is a six-sided
polygon and the total of the internal angles of any
simple (non-self-intersecting) hexagon is 720°. Fol-
lowing this step, the smallest repetitive unit is cre-
ated based on defined rule sets. This hexagon’s pe-
ripheral length is conceived as a single total length
and it is divided into 12 and all the division points are
numerated from 0 to 11 counters to clockwise. Then
4 lines: from 0 to 4 L1, from 0 to 5 L2, from 10 to 4
L3, and from 10 to 6 L4 are drawn on the hexagon,
and then all these lines are mirrored vertically (on
the y axe). 8 lines are acquired at the end of the de-
sign process (Figure 1). As a consequence of applying
rule sets pattern generation process, pattern genera-
tion process could easily be controlled by changing
rules. While generating pattern, making changes on
the rule sets can be clearly seen through the whole
pattern.
SECOND STAGE - Computerization of 2D
Pattern
In the second stage of the study, the embedded al-
gorithm of the pattern was coded in the Grasshop-
per as enabling designer to do changes in the rule
sets which are defined according to the shape gram-
mar operations (Figure 2). In the digital environment,
hexagonal grid system is defined to create alterna-
tives and variations in the pattern.
In the coding process, defined rule sets trans-
posed to digital environment, while the number of
division of the peripheral length of the hexagon pro-
vide alterable values to designers. While the num-
ber of the division changes from 12 to 36 on the
hexagon the location of the points on the hexagon
change, resulting changing the location of the lines
on the hexagon. As consequence the entire pattern
changes (Figure 3). In grasshopper coding, changing
the value of one component or altering the number
of the division of the hexagon changes the entire pat-
tern easily.
To get dynamic complex patterns, an attraction
point was defined in the Grasshopper code. While
coding the entire pattern, the construction point was
defined to control the starting point of the pattern in
case of the need for the change of the starting point
of the pattern. In addition to construction point, at-
traction point is defined which led the changes in the
number of the division on the hexagon. Attraction
point was defined as the distance between the attrac-
tion point and grids which were defined as the cen-
tre of the hexagons. Varied values of the distance be-
tween the attraction point and the entire grid gener-
ates irregular gradient pattern alternatives (Figure 4).
Mathematically defined algorithms in grasshop-
per generate hundreds of complex pattern alterna-
tives that can be used in different designs such as
ceramic design, fabric design, facade panel designs,
etc.. The use of these designs as ceramic tiles and di-
vider panel is illustrated in (Figure 5).
SHAPE, FORM & GEOMETRY | Applications - Volume 2 - eCAADe 36 |479
Figure 1
Rule set of the
pattern making the
process of the
study.
Figure 2
Grasshopper code
of the pattern.
480 |eCAADe 36 - SHAPE, FORM & GEOMETRY | Applications - Volume 2
THIRD STAGE - Acoustic Panel Design
Acoustics is important performance criteria which
determines the quality of an architectural space. Ac-
cording to Peters (2009), as light illuminates our vi-
sual environment, sound sources illuminate our sonic
environment. ... Human activities produce sound,
and our architecture constantly interacts with us
through its modification of the sounds we create (Pe-
ters, 2009). Since then, computer aided tools be-
come apparent in the architectural acoustic designs
as well as light design in architecture. Altering the
acoustic performance of a space by changing the ma-
terial properties, surface geometries, and etc is an
important factor that stated the acoustic quality of
the space. Since the widespread use of the com-
putational tools, firstly in architectural design, and
then, acoustical practices, provided acoustically well-
balanced space designs.
Third stage of the study investigated how pat-
tern generated in stage two can be utilized in acous-
tic cloud panel making. Here, real classroom acous-
tic problem in the Architectural School of Istanbul
Technical University Classroom 214 used as a case
(Figure 6). In this stage, first design students make
3D model of the classroom, then graduate students
of the building physic analyse acoustics of the class-
room, using 3D model in Odeon simulation soft-
ware. As the consequences of the analysis and sim-
ulations of Classroom 214 has acoustical problems
which are mainly related with speech are identified
as the persistence of the sound in the classroom and
the speech intelligibility across the listeners.
Reverberation time is one of the primary impor-
tance of acoustic parameters that gives the first idea
Figure 3
Varied pattern
designs as a result
of diverse dividing
options.
SHAPE, FORM & GEOMETRY | Applications - Volume 2 - eCAADe 36 |481
Figure 4
Dynamic pattern
design depending
on the location of
the attraction point.
Figure 5
Ceramic art wall
design alternatives.
Figure 6
Architectural
School of Istanbul
Technical University
Classroom 214.
about an acoustical behaviour of a room. (Figure
7). Wallace Sabine is the founder of the science of
the Architectural Acoustic, discovered the relation-
ship and effects between sound, space, and material
of acoustic quality of the space. By knowing the sur-
face area of the constituent materials, their absorp-
tion coefficients, and the volume of the space, the
reverberation time can be predicted (Peters, 2010).
Reverberation is an objective measurable indicator
of the persistence of a sound after its source turned
off and it is mainly depended on the volume and ab-
sorption amount of the room. As the absorption de-
482 |eCAADe 36 - SHAPE, FORM & GEOMETRY | Applications - Volume 2
Figure 7
Recommended
reverberation times
(Long, 2005).
Figure 8
Reverberation times
of the classroom for
different frequency
bands. creases and reflection increases, reverberation time
gets longer and the room sound dies in a long time
in the room. A larger volume also affects the rever-
beration time to be longer. Reverberation time de-
fines the appropriate acoustical conditions for differ-
ent functions. A longer reverberation time causes the
speech to be less understandable which it is not a
preferred situation for a room.
While site visit to classroom 241, very long rever-
beration time that affects speech intelligibility is ex-
perienced as is also observed in Odeon software sim-
ulation. The simulation results for different frequency
bands are listed below (Figure 8). As it can be seen
from the results the large volume of the class and a
large number of reflective surfaces, cause reverbera-
tion time to be much higher than the criteria. How-
ever, reverberation time does not change too much
around a room as a constant average level is perva-
sive that gives an idea of the speech intelligibility in a
room while it is not the only parameter that effects it.
The understanding of the speech depends on many
different things while the main factors are the char-
acter of the source, the distance between the source
and receiver, and the background noise.
There are different ways to measure the speech
intelligibility. One of most common and standard-
ized objective method is the speech transmission in-
dex Speech Transmission Index (STI). It is an index
gives a result between 0 and 1, which indicates worst
and best speech intelligibility respectively (Figure 9).
Odeon Software was used to measure the STI for
different positions in the simulated classroom. Under
the assumptions about the speaker, surface materials
and background noise, STI values as grid response of
the whole classroom can be seen in the figure 10. As
expected from the experience the results are very low
and even at the front seats conditions are poor to un-
derstand the speaker.
SHAPE, FORM & GEOMETRY | Applications - Volume 2 - eCAADe 36 |483
Acoustic Proposal and Pattern Modification
While proposing an acoustical element to enhance
the indoor quality, it is considered that the building’s
being listed as a first-class historic building. To avoid
closing the ribs on the ceiling and making a perma-
nent change, the absorptive sound panel in other
saying an acoustical ceiling cloud was proposed. It
is thought that they have an advantage in terms of
barely attaching to the ceiling and does not entirely
cover and close the ribs. Acoustical ceiling clouds
are widely used for noise reduction and reverberation
control in auditoriums, restaurants, atriums and mul-
tipurpose rooms.
During the studies of the acoustic ceiling cloud
design, the entire pattern as structure and substruc-
tures were analysed to design fabrication steps and
system for the acoustic ceiling cloud design pattern.
To enable design for this, a mathematical connec-
tion between acoustics and architecture is required
(Reinhardt, Cabrera, & Hunter). In this process, the
pattern was optimized according to construction re-
quirements as defining continuous and discontinu-
ous axes on all lines in the acoustic cloud pattern (Fig-
ure 11). The line from 0 to 5 and the line from 0 to
7 were discontinuous lines in the pattern which were
hard for implementation of the acoustic ceiling cloud
pattern. Thus, these lines were removed from the
entire pattern. As it can be seen in the new single
hexagon module, the red axe was added to the pat-
tern structure to make the pattern more rigid (Figure
12).
To construct this 3D composition, the 2D pat-
tern was transformed in 3D cloud by transforming 2D
lines into 3D surfaces. Afterwards, the pattern struc-
ture was treated like a Lego consisting of 5 layers to
increase both the number of reflective surfaces as an
acoustic solution and to prevent the cutting of the
each and every element of the pattern. In the layer
system of the pattern for construction, pattern struc-
ture was named as from the top to bottom as the
first layer is the top layer, the third layer is the mid-
dle layer, and the fifth layer is the bottom layer that
the structures were designed without slits between
layers which provide nesting to all the layers easily in
the construction process (Figure 13). And the metal
pipe was designed to hang the acoustic cloud design
to the ceiling from the four corners of the rectangu-
lar acoustic cloud. Therefore, acoustic cloud panel is
designed as a self-supporting system to provide the
acoustic solutions to the classroom 214.
Figure 9
Distribution of STI
in three different
cases.
Figure 10
Acoustic properties
of the classroom as
ODEON outputs.
Figure 11
Analysis of structure
and substructures.
Figure 12
Rules of the new
structure.
Figure 13
Layer system of the
acoustic panel unit.
Figure 14
Acoustic panel
design renders for
the classroom 214.
484 |eCAADe 36 - SHAPE, FORM & GEOMETRY | Applications - Volume 2
Figure 15
Absorption
coefficients of
Basotect 50mm
Figure 16
Frequency bands in
which human
speech is most
effective for three
different cases.
Figure 17
Distribution of STI
in three different
cases.
SHAPE, FORM & GEOMETRY | Applications - Volume 2 - eCAADe 36 |485
It is aimed to reduce the reverberation time in ac-
cordance with the speech function by increasing the
absorptive surface in the classroom, and to provide
a more homogeneous distribution of sound in the
volume. The coffered slab at the ceiling is a part of
the architectural character of the building, so a sus-
pended ceiling which would conceal the slab and
lower the volume was not possible. So, a transpar-
ent absorptive sound baffle was suggested by the
designer to overcome the problem. The desired ab-
sorption was aimed to be provided with the material,
depth, and complexity of the suspended structure as
well as acoustic panel design (Figure 14).
50mm Basotect® melamine foam is used as the
absorbing material because its absorption efficiency
is 50mm (Figure 15). Sound baffles positioned at a
distance of 2m from the ceiling which is expected
to enhance its effect. The room is modelled with-
out and with the ceiling baffles and the reverbera-
tion time results are listed below for before and after
the acoustical treatment (Figure 16). It is seen that
baffles would be mostly effective. Especially for the
mid-frequencies 250Hz, 500Hz, 1000Hz in which hu-
man speech is most effective the reverberation times
are lowered to the recommended levels as 0,70-1,00
s. The simulated STI results for the speech intelligibil-
ity are listed below for before and after the acoustical
treatment (Figure 17).
According to the simulation outputs, the rever-
beration time is decrease in frequency bands where
human speech is most effective with the final design.
CONCLUSION
This study explored computational thinking and
making from paper-based 2D geometric pattern gen-
eration to its evolution into 3D acoustic panel and its
realization. It also includes various methods as intu-
itive tendencies, collaborative multi-disciplinary re-
search conducted with building physics students by
utilizing integrative thinking and making.
Consequently, the 3D pattern not only have a
graphical value that could be described as an art, but
also alter the acoustic conditions of the classroom.
However, when the geometry was structurally anal-
ysed and materialized, it has become an architectural
element at the end. As a result of the analysis made
for the acoustic panel design, the absorptive acous-
tic panels to be designed at the ceiling to reduce the
reverberation time and create more suitable environ-
ments for lectures as well as other functions. The
complexity of the pattern with the accurate material
would provide most of the solutions for the long re-
verberation.
REFERENCES
Knight, Terry 1999 ’Applications in Architectural Design
and Education and Practice., Report for the NSF/ MIT
Workshop on Shape Computation, Massachusetts In-
stitute of Technology, Cambridge
Knight, Terry 2000, ’Shape Grammars in Education and
Practice: History and Prospects’, International Jour-
nal of Design Computing, 2, p. -
Knight, Terry 2018, ’Craft, Performance, and Grammars’,
in Lee, Ji-Hyun (eds) 2018, Computational Studies on
Cultural Variation and Heredity, Springer, Singapore,
pp. 205-224
Long, Marshall 2006, Architectural Acoustics, Elsevier,
USA
Peters, Brady 2009 ’Parametric Acoustic Surfaces’, ACA-
DIA 2009, Chicago, Illinois, USA, pp. 174-181
Peters, Brady 2010, ’Acoustic Performance as a Design
Driver: Sound Simulation and Parametric Modeling
Using Smart Geometry’, International Journal of Ar-
chitectural Computing, 2010, Vol 8, Issue 3, pp. 337-
358
Reinhardt, Dagmar, Cabrera, Densil and Hunter,
Matthew 2017 ’A Mathematical Model Linking Form
and Material for Sound Scattering Design, Robotic
Fabrication and Evaluation of Sound Scattering
Discs: Relating Surface Form to Acoustic Perfor-
mance’, CAADFutures 17, Istanbul, pp. 150-163
Stiny, George 1980, ’Introduction to Shape and Shape
Grammars’, Environment and planning B: Planning
and Design, Vol 7, Issue 3, pp. 343-351
Stiny, George 2006, Shape: Talking about Seeing and Do-
ing, The MIT Press, Cambridge, Massachusetts
Stiny, George and Gips, James 1971 ’Shape Grammars
and the Generative Specification of Painting and
Sculpture’, IFIP Congress 71, Ljubljana, pp. 125-135
486 |eCAADe 36 - SHAPE, FORM & GEOMETRY | Applications - Volume 2
... The optimal placement of BIPV for the optimized building is determined. Yavuz et al. (2018) proposed a novel shape grammar to guide the generation of acoustic panels in order to create an optimal indoor acoustic environment, from the generation of 2D geometric to the evolution of 3D acoustic panels. Rodrigues et al. (2018) proposed a step-bystep method to generate and evaluate schemes. ...
Chapter
Full-text available
In recent years, generative design methods are widely used to guide urban or architectural design. Some performance-based generative design methods also combine simulation and optimization algorithms to obtain optimal solutions. In this paper, a performance-based automatic generative design method was proposed to incorporate deep reinforcement learning (DRL) and computer vision for urban planning through a case study to generate an urban block based on its direct sunlight hours, solar heat gains as well as the aesthetics of the layout. The method was tested on the redesign of an old industrial district located in Shenyang, Liaoning Province, China. A DRL agent - deep deterministic policy gradient (DDPG) agent - was trained to guide the generation of the schemes. The agent arranges one building in the site at one time in a training episode according to the observation. Rhino/Grasshopper and a computer vision algorithm, Hough Transform, were used to evaluate the performance and aesthetics, respectively. After about 150 h of training, the proposed method generated 2179 satisfactory design solutions. Episode 1936 which had the highest reward has been chosen as the final solution after manual adjustment. The test results have proven that the method is a potentially effective way for assisting urban design.
... General recommended reverberation time for speech and orchestral music (Barron, 2010) The best speech intelligibility is 1, while the worst speech intelligibility value is 0 (Yavuz et. al. 2018). ...
Conference Paper
Full-text available
The complexity of the wall form affects the acoustics of the space. In this study, the effect of the complex form walls produced by nCloth dynamic simulation on the acoustics of an office space was investigated. In this research, reverberation time and Speech Transmission Index (STI) values of the pilot office space with one wall having complex form and the office space with all of the walls as flat were measured by acoustic simulation. As a result of the comparison, it has been found that, within speech intelligibility and reverberation time, the acoustics of the space with one wall having complex form is better than the acoustics of the space with all the walls as flat.
Thesis
Full-text available
Architecture arouses feelings in its experiencers. Anything, colors, lights, scales, and shapes we perceive by our senses, causes a cognitive process and results in a feeling. The aesthetic experience arises from our perception, and it plays an important role in the design fields such as product design, interior design, and architecture. The general design goal in these applications is to choose the shape and material for the object in such a way that the product is desirable by its perceivers/users. Here, we can distinguish two kinds of desirable properties of a product. One kind is the fulfillment of a utilitarian purpose. Another kind is the aesthetical pleasure, which arises solely from the act of visually perceiving the object. Defining the visual pleasantness of a shape is a ‘soft’ issue. Pleasantness is a linguistic concept with associated imprecision and uncertainty. This is because it is a form of qualitative evaluation. As preferences related to shape, and in particular those that are due to aesthetical inclinations, are merely loosely related to rationality, the reasoning underlying an evaluation is problematic to pinpoint. For instance, commonly designers are having difficulty expressing the rationale both behind their own aesthetical preferences and those of the user. The study aims to devise a computational method to get to know the rationale that underlies certain shape preferences. The method consists of three steps: abstracting the detailed shape attributes (1); aggregating several attributes to characterize the shape in more general terms (2); and tuning the representation of the shape character based on preference data. The suitability of the method is verified by applying it to a shell type of shape which has got a triangular plan and has got three support locations. This shape has been selected to have simplicity in the explanation, as it involves only a few parameters. As a first step, the shape is analyzed with respect to its physical attributes in the vertical and horizontal sense. The basic shape elements are the boundary curves, which give the salient character to the overall shape. The exemplary shell shape is defined with six support location points, two-point pairs of which determine the support locations as well as triangularity of the shape; additionally, the control point locations of three NURBS curves determine the symmetry, verticality, and height of the shape. Based on the detailed attributes of the shape, four conditions of symmetry are distinguished. These are that all three boundary curves are symmetric (1), two of the curves are symmetric (2), two of the curves are asymmetric (3), and all three curves are asymmetric (4). Six conditions as to verticality are distinguished. These are that all three curves are oriented such that each forms an overhang (1); all three curves are oriented vertically so that there is no overhang (2); all three curves are recessed with respect to the vertical orientation (3); two of the curves yield overhangs (4); two of the curves are oriented vertically (5); two of the curves are recessed (6). Three conditions of height relations among the boundary curves are distinguished. These are that the three curves have similar heights (1); their heights are gradually changing (2); they have heights that are different in a diverse manner (3). Four conditions as to the triangularity of the plan are distinguished. These are that the plan is an equilateral triangle (1); it is an isosceles triangle with one pointed corner (2); it is a right-angled triangle (3); it is a skewed obtuse triangle (4). When the physical attributes of a shape are exactly matching to the prototypical example of one of the shape conditions above, then the linguistic label expressing the condition undoubtedly applies to the shape. However, when there is some discrepancy between the triangle at hand and its characteristic prototype, then the applicability of the label is dependent on the magnitude of the discrepancy. To express the association strength of linguistic shape attributes, in this thesis, fuzzy sets are used. However, the precise shape of fuzzy membership functions is problematic to specify due to the imprecision associated with the concepts the functions are to represent. In particular, the imprecision refers that people differ to some extent with regards to the meaning they associate with the concepts. As understanding implies some general validity of the knowledge, the identification of the membership functions that characterize a shape is a matter of obtaining and modeling data from multiple people. This is accomplished in this study by means of surveys. The data obtained in this way is used to identify the shape parameters of the membership functions through curve fitting. In this way an ontology of the basic shape features is established. While the meaning of the basic shape descriptors is subject to agreement among different designers, when we come to more general shape characteristics, different designers can have different aesthetic preferences. Therefore, the rationale underlying individual aesthetical preference requires obtaining multiple manifestations instances of the individual’s preference. In this study, this takes the following form. The architect selects among a set of random shapes the subset of shapes that he/she deems as preferable over the others. The next step toward understanding the architect’s shape preference is to identify common characteristics among the shapes in the preferred subset. The method to represent the range of distinct shape characters is a neural computing method known as fuzzy neural tree (FNT). Having distinguished several possible character types respectively for a shape’s symmetry, verticality, height topology, and plan triangularity, and noting that the relative importance among these four characteristics is not to assume for an individual designer, it is clear that the number of possible explanations for a designer’s preference is large. Each possible explanation is represented in the work by one particular FNT. The task in this modeling step is to identify among the large set of understandings of the designer’s preference, the understanding that is most likely to be correct. This is accomplished by searching the space of possible FNT models using an evolutionary algorithm. The evolutionary algorithm is run multiple times until as many as possible among the preferred designs are represented. Multiple runs are necessary because different patterns of preferences have been applied by the architect so that a single FNT model does not represent all of the preferred solutions. With different possible combinations of property conditions and with a different possible degree of memberships, there appear a variety of shape characters. The identification process reveals that there are four patterns of preferences applied by the architect. Thus, there are four different resulting FNT models explaining the preferences. Among four different trees, the first tree can be considered as the most important one as it represents four of the ten preferred samples whereas the other three respectively represent one solution each. It can be considered as the model of the aesthetic preference of the architect. And finally, this FNT model can be used to understand the preference for any shape. When we put any shape to this model, the magnitudes at the FNT outputs predict the likelihood the architect will prefer the shape. Due to the transparency of FNT, the reason why he/she will like the shape is also given. Due to the interpretability of the reasoning, it is appropriate to refer to the result from this modeling effort as ‘understanding’ the designer’s preference, beyond merely representing it in some computational form of black-box type. The benefit of attaining the understanding of preferences by computation is that then the preferences become subject to satisfaction in extreme form by a systematic search. A second benefit is that aesthetics related objectives become compatible with non- aesthetic related ones so that the best compromise is found taking the aesthetic preference duly into account. The subject matter dealt with in the thesis is of generic relevance in architectural design research, since the precise description of soft objectives has been a long-standing problematic issue. The study puts forward an innovative approach to tackle the problem, going beyond using computation as a mere mathematical tool for representing a non-linear relationship in a data; but yielding insight into a cognitive process. The particularity of the approach is the synergetic treatment by several computational intelligence methods in such a way that the result is subject to interpretation in the verbal form the designers are familiar with.
Article
Acoustic performance is an inevitable part of architectural design. Our sonic experience is modified by the geometry and material choices of the designer. Acoustic performance must be understood both on the level of material performance and also at the level of the entire composition. With new parametric and scripting tools performance driven design is possible. Parametric design and scripting tools can be used to explore not only singular objectives, but gradient conditions. Acoustic performance is often thought of in terms of singular performance criteria. This research suggested acoustic design can be understood in terms of gradients and multiple performance parameters. Simulation and modeling techniques for computational acoustic prediction now allow architects to more fully engage with the phenomenon of sound and digital models can be studied to produce data, visualizations, animations, and auralizations of acoustic performance. SmartGeometry has promoted design methods and educational potentials of a performance-driven approach to architectural design through parametric modeling and scripting. The SmartGeometry workshops have provided links between engineering and architecture, analysis and design; they have provided parametric and scripting tools that can provide both a common platform, links between platforms, but importantly an intellectual platform where these ideas can mix. These workshops and conferences have inspired two projects that both used acoustic performance as a design driver. The Smithsonian Institution Courtyard Enclosure and the Manufacturing Parametric Acoustic Surfaces (MPAS) installation at SmartGeometry 2010 are presented as examples of projects that used sound simulation parametric modeling to create acoustically performance driven architecture.
Article
Acoustics are important performance criteria for architecture; however, architects rarely consider them, except, perhaps, when designing concert halls. Architectural spaces can be said to perform well or poorly in terms of their acoustic qualities. By altering the geometry or material characteristics of the surfaces within a room in specific ways, the acoustics can be controlled. Once the geometric rules governing these acoustic alterations are understood, these rules can be encoded into a CAD system through parametric modeling or the use of computer programming. The architectural designer can then generate acoustically regulating surfaces according to desired performance criteria. In this way, acoustic engineering links to architectural design, and allows architectural design to become acoustically performance-driven. This paper considers three primary types of acoustic surfaces: absorbers, resonators, and diffusers. complex surfaces that combine these three performance characteristics in different ways are proposed.
Article
The definitions pertaining to the shape grammar formalism are developed in detail.
Applications in Architectural Design and Education and Practice
  • Terry Knight
Knight, Terry 1999 'Applications in Architectural Design and Education and Practice. ', Report for the NSF/ MIT Workshop on Shape Computation, Massachusetts Institute of Technology, Cambridge
Shape Grammars in Education and Practice: History and Prospects' , International Journal of Design Computing
  • Terry Knight
Knight, Terry 2000, 'Shape Grammars in Education and Practice: History and Prospects', International Journal of Design Computing, 2, p. -Knight, Terry 2018, 'Craft, Performance, and Grammars', in Lee, Ji-Hyun (eds) 2018, Computational Studies on Cultural Variation and Heredity, Springer, Singapore, pp. 205-224
  • Marshall Long
Long, Marshall 2006, Architectural Acoustics, Elsevier, USA
A Mathematical Model Linking Form and Material for Sound Scattering Design, Robotic Fabrication and Evaluation of Sound Scattering Discs: Relating Surface Form to Acoustic Performance
  • Dagmar Reinhardt
  • Cabrera
  • Matthew Hunter
Reinhardt, Dagmar, Cabrera, Densil and Hunter, Matthew 2017 'A Mathematical Model Linking Form and Material for Sound Scattering Design, Robotic Fabrication and Evaluation of Sound Scattering Discs: Relating Surface Form to Acoustic Performance', CAADFutures 17, Istanbul, pp. 150-163
George and Gips, James 1971 'Shape Grammars and the Generative Specification of Painting and Sculpture
  • George Stiny
Stiny, George 2006, Shape: Talking about Seeing and Doing, The MIT Press, Cambridge, Massachusetts Stiny, George and Gips, James 1971 'Shape Grammars and the Generative Specification of Painting and Sculpture', IFIP Congress 71, Ljubljana, pp. 125-135