Acoustics in architectural fabric

structures: the case of ETFE 9

pillows
S. Chiu, D. Noble, E. Valmont
University of Southern California, Los Angeles, CA, USA

9.1 Introduction
Architectural acoustical engineering usually prompts one to imagine music concert
halls for which sound quality is a high priority in design development. Acoustics is
important for more than just specialized performance spaces. One unique condition
faced by designers is the acoustics of fabric membrane structures. For fabric structures,
acoustics are particularly challenging due to the lightweight physical property of the
membrane materials.
A principal building acoustic concern is noise. In general, noise is unwanted
sound, whether created deliberately or unintentionally. Urban noise sources might
include automobiles, trains, or aircraft, as well as building activities, industrial
machinery, and equipment operations. Other potential noise sources include human
speech, music performances, and impacts of objects. Natural sources can be included
among other noise sources to be considered, including noise caused by wind and pre-
cipitation. Excessive noise levels in urban settings is typically described as noise
pollution.
Excessive noise can negatively impact human health and behavior. Humans
exposed to high noise levels for excessive periods of time can exhibit stress,
changes in behavior, loss of sleep, cardiovascular illnesses, a rise in blood pressure,
and temporary or permanent hearing loss. In addition, noise pollution can cause
annoyance and aggression, hypertension, and other harmful effects. Negative
effects from noise pollution can lead to other serious illnesses and cause significant
health issues.

9.2 Principles of sound

Sound is generated by vibration transmitted through air, liquid, and solids. The period
of vibration is the number of cycles per second defined as frequency, measured in
Hertz (Hz) (1 Hz equals one cycle per second).
Not all sound can be detected by the human ear. Audible sound is limited to the fre-
quencies between 20 and 20,000 Hz, with particularly sensitivity at 3000e4000 Hz

Fabric Structures in Architecture. http://dx.doi.org/10.1016/B978-1-78242-233-4.00009-7

Copyright © 2015 Elsevier Ltd. All rights reserved.
242 Fabric Structures in Architecture

Figure 9.1 Examples of sound frequency, measured in Hertz (Hz).

Courtesy of E. Valmont/ARUP.

(Figure 9.1). Sounds above the audible frequencies are called ultrasonic, and sounds
below audible frequencies are called infrasonic.
Higher frequency produces a higher pitch sound. For example, the sound generated
by a grand piano is at 27.50 Hz from the left-most number 1 key, and 4186.01 Hz from
the right-most number 88 key.

9.2.1 Sound level

The measurement of sound is called sound level, or sound pressure level. It is a log-
arithmic measure of the effective sound pressure of a sound relative to a reference
value. Sound level is measured in decibels (dB) above a reference level. At the
threshold of hearing, sound level is at 0 dB. Normal speech level occurs around
60 dB. Audible sound levels can be as high as 120 dB near a jet aircraft engine. Sound
levels over 140 dB can cause pain to a human ear (Figure 9.2).

9.2.2 Reflection
When sound energy travels and hits a surface, the energy will be distributed in different
directions. Part of the energy will be bounced off the surface in the form of sound
reflections. The level of reflections will depend on the material properties and the
form and shape of the surface. Unwanted reflected sound can negatively impact
the acoustics in the space. A concave surface will generally concentrate the
reflected sound, whereas a convex surface will generally disperse the reflected sound.
A material with a higher absorptive surface will absorb more sound energy than a less
absorptive material.
Acoustics in architectural fabric structures: the case of ETFE pillows 243

Figure 9.2 Example of sound pressure level source.

Referenced from Cavanaugh, 2009.

9.2.3 Reverberation time

The prolongation of the sound from a sound source once it stops emitting energy is
defined as reverberation. The prolongation is caused by the sound reflections. Rever-
beration time is particularly important to sound quality. In a concert hall, the reverber-
ation time influences the quality of music. The sound dies away from the source over
time as the sound energy is absorbed by the material surfaces in the hall. If the surfaces
of the room are highly absorptive materials, the sound may appear to be solid and
dead, as the reverberation time is cut short. If the surfaces of the room are made of
reflective materials, the sound will appear to be more alive or bright. However,
if the reverberation time is too high, the noise level in the space will build up to
undesirable levels.
The desirable level for reverberation time varies depends upon the use of the room.
For an orchestra concert hall, the ideal reverberation time for sound is 1.8e2.6 s over
the range of high to low frequency. For opera, the ideal value is 1.4e1.8 s. For
classrooms or conference rooms, the reverberation time shall be between 0.6
and 1.0 s.
The reverberation time also varies over the range of sound frequency. Lower
frequency is desirable at longer reverberation time than higher frequency.

9.2.4 Sound intensity

While preference characteristics of the sound quality in a concert hall is somewhat sub-
jective, the desirable reverberation time depends upon the sound intensity. Professional
organizations and standards groups set the target sound intensity levels for various
244 Fabric Structures in Architecture

Figure 9.3 Comparison of sound absorption coefficient among different construction materials.
Courtesy of E. Valmont/ARUP.

types of space. For example, the target sound intensity is approximately 60 dB for
defining the standard reverberation time for a music concert hall. It is derived from
the difference between the loudest crescendo of typical orchestral sound of 100 dB,
and the desire background level of 40 dB for good music.

9.2.4.1 Transmission loss

Sound absorption: When sound energy travels and hits a surface, some of the energy
will convert into a different form of energy, such as heat energy. As a result of the
transformed energy, there is a loss of sound energy due to sound absorption by the
material in contact (Figure 9.3). The energy transformed into another form of energy
is considered acoustic transmission loss or sound absorption.

9.2.5 Sound isolation

Sound isolation occurs when a sound source passes through a material, and as a conse-
quence, sound level is reduced. The reduction is a result of sound reflection and
absorption by the material. Sound reduction is measured in decibels (dB).
However, sound absorptiveness of building materials is expressed as the noise
reduction coefficient (NRC). It is a single-number rating index resulted from standard-
ized testing, by organizations such as the American Society for Testing and Materials
(ASTM). An NRC of zero means the material reflects 100% of sound and absorbs no
sound energy (Figure 9.4). An NRC of one means the material reflects 0% of sound
and absorbs 100% of the sound energy. The NRC value is the average of the mid-
frequency sound absorption coefficients (250, 500, 1000, and 2000 Hz rounded to
the nearest 5%). However, NRC does not represent how absorptive a material is in
the low and high frequencies. It also does not represent the material’s sound isolation
effect.
Acoustics in architectural fabric structures: the case of ETFE pillows 245

Figure 9.4 Comparison of sound reduction between a 600 concrete wall and gypsum board with
batted wall construction.
Courtesy of E. Valmont/ARUP.

9.2.6 Sound transmission class

Sound transmission class (STC) is a rating of sound isolation of a building wall assem-
bly. The higher the STC rating, the better sound isolation the wall assembly is to
achieve. STC is widely used to rate interior partitions, ceilings/floors, doors, and win-
dows. The STC rating figure represents approximately the decibel reduction in noise
that the assembly can provide.
For an interior wall that is made of 3 5/800 metal studs with one layer of 5/800 gyp-
sum board on both sides, the STC rating is about 38e40. However, if the space be-
tween the gypsum boards and the metal studs is filled with batt insulation, the STC
can be improved to 43e45. For an interior wall that is made of 200  400 wood studs
at 1600 apart with one layer of 5/800 gypsum board on each side, the STC rating is 37.

9.2.7 Acoustics in membrane structures

Architects can overlook acoustics in designing fabric membrane structure due to a
mistaken perception that fabric membrane is acoustically transparent. Acoustics are
particularly challenging in fabric membrane architecture due to the physical property
of the materials.

9.2.7.1 Acoustic principles of lightweight, stressed ETFE pillows

Acoustics is one of many design considerations required to assess if a lightweight, ten-
sile fabric is appropriate for a building enclosure. Materials such as ETFE and polyte-
trafluoroethylene (PTFE) are widely used worldwide, but thus far they are used less
often in the US. Although there are many advantages from the material properties
(i.e., flexibility, waterproofing, UV resistance, high reflectance, light transmittance,
246 Fabric Structures in Architecture

Figure 9.5 Multidisciplinary design considerations.

Image Courtesy of E. Valmont/ARUP.

life expectancy), further investigation must be taken into the suitability of the material
based on the space use and occupancy.
Multidisciplinary Design Considerations (Figure 9.5)
• Acoustics
• Fire
• Lighting
• Maintenance
• Thermal
• Architectural
Lightweight fabric membranes such as ETFE pillows have interesting acoustic prop-
erties due to the inherent thickness, weight, and airspace:
• At low frequencies (31.5e250 Hz), the stretched material is almost 100% acoustically trans-
parent, and sound at low frequency will pass directly through.
• At middle to high frequencies (500e8000 Hz), the material absorbs a maximum of 30% of
the sound, and the majority is reflected back into a space.
Basic Properties of ETFE
• Thickness: 12e250 m (Figure 9.6)
• Weight: 2e3.5 kg/m2
• Layered: typical three layer 250 m/100 m/150 m
• Airspace
Acoustics in architectural fabric structures: the case of ETFE pillows 247

Figure 9.6 Typical properties of an ETFE pillow with only two foils.
Image by Courtesy of E. Valmomt/ARUP.

The advantage of this acoustic transparency is that the reflected sound can enhance
acoustic energy back into the space (Figures 9.7 and 9.8). An example of this effect
can be seen in stadia design where noise from a crowd can enhance the sense of
excitement.
A disadvantage is that the reverberation and noise build-up from the middle to high
frequencies is difficult to control for speech intelligibility, that is, for emergency and
public address sound systems. Additionally, noise generated from inside a space
enclosed with a membrane skin façade may disturb nearby buildings and people. Envi-
ronmental noise (e.g., aircraft, cars, railways) may easily be heard inside the space and
may be distracting to the occupants.
Rain noise is another important consideration of lightweight fabric membranes.
Impact from rain falling on the lightweight fabric system can cause a drumming noise
inside the space. Often, noise mitigation is required to minimize the effect of this
impact noise. Manufacturers of ETFE and PTFE materials are aware of this issue
and offer integral acoustic solutions to reduce rain noise.

20

16
Sound reduction, dB

12

0
0 5
10 12
0

00
50
0
0

00
0

00
5
0
0

00

50
00

00
0
80
16
20
25
31
40
50
63

16
12

25

50
20

31
10

40

1/3 Octave band center frequency, Hz

Figure 9.7 Sound isolation performance of vector Foiltec TexlonÒ, one-third octave band
center frequencies.
Data referenced from Vector Foiltec, 2008.
248 Fabric Structures in Architecture

0.5

0.4
Sound absorption, α

0.3

0.2

0.1

0.0

25 0
00

00
00
0

10 0
12 0
50
00

50
0
0
5
0
0
0
5

0
40

63
80

0
50
10
12
16
20
25
31

20

40
50
16

31
1/3 Octave band center frequency, Hz
Figure 9.8 Sound absorption performance of Vector Foiltec TexlonÒ, one-third octave band
center frequencies.
Data referenced from Vector Foiltec (2008).

9.2.7.2 Tensile fabric manufacturers

• Vector Foiltec, United Kingdom
Vector Foiltec is headquartered in the UK, with offices in the USA, Germany, Australia,
France, Sweden, and China. They manufacture pneumatically stressed ETFE foil cushions,
called the Texlon Cladding System. They offer rain noise suppressers, an ETFE rain shield
mesh, comprised of a very fine porous net weave. It is usually installed a few inches (4e600 )
from the outermost layer of the main ETFE façade and reduces the drumming impact of the
rain by making the spectrum more broadband in the mid to high frequencies (Figures
9.9e9.11).

80
70
Sound pressure level, dB

60
50
40
30
20
10
0
125 250 500 1k 2k 4k 8k
Octave band center frequency, Hz

Ambient noise Simulated rain Rain suppresser

50 dB(A) 77 dB(A) 58 dB(A)
Figure 9.9 Rain noise reduction with Vector Foiltec Rain Suppresor.
Reference from Vector Foiltec TexlonÒ RSÒ performance data (Vector Foiltec, 2008).
Acoustics in architectural fabric structures: the case of ETFE pillows 249

Figure 9.10 ETFE TexlonÒ rain suppressor. Image Courtesy of Arup.

• Convertex, Germany
Convertex is headquartered in Germany, with offices in the UK and China. They manufac-
ture pneumatically pre-stressed ETFE foil cushions and mechanically pre-stressed ETFE
sheeting; they also engineer cladding systems based on the alternative fluoropolymer fabrics
PTFE, TFA/PFA, THV, FEP, and PVDF.
• Hightex Group
The Hightex Group manufacturer ETFE roof solutions. They have innovative methods of
achieving improved sound insulation and sound absorption with a three-layer EFTE pillow
assembly. It is comprised with a layer of Plexiglas to improve the sound insulation and an
inner pillow that is perforated for sound absorption.

9.2.7.3 Importance of rain noise

Rain impact on a lightweight membrane can generate excessive reverberant noise
levels within the enclosed space. In design work, it is the sound level in the room
due to rainfall that is of primary interest to the affected users. A laboratory test study
conducted by the British Research Establishment (BRE) for schools investigated and
confirmed the significance of rain noise on various roof membranes (Figure 9.11).
250 Fabric Structures in Architecture

Polycarbonate
68
66 64.6 dBA
64
62 ETFE
Intensity level, L1 (dB re 10-12 W/m2)

60 68.7 dBA
58
56
54
52 ETFE with rain
50 suppressor
48
46 type 1
44 62.9 dBA
42
40
38 ETFE with rain
36
34 suppressor
32 type 2
30 55.5 dBA
28
26
24
22 6–12–6.4
20 glazing
18 52.0 dBA
100 160 250 400 630 1000 1600 2500 4000

Third octave frequency band (Hz)

Figure 9.11 Rain noise levels from different lightweight roof materials, test conducted by the
British Research Establishment. BRE Report Number 220312 BRE: Rain noise from roof
glazing, polycarbonate roofing, and ETFE roofing.
The measured test data compares levels caused by artificial heavy rainfall on glass,
polycarbonate, and ETFE membranes (with and without rain suppressors). Results
from the test data are shown in the figure above.
Measured sound intensity levels for different roof elements
• ETFE 69 dB
• Polycarbonate 65 dB
• ETFE (Type 1) 63 dB
• ETFE (Type 2) 56 dB
• IGU (6/12/6.4) 52 dB
For further information see reference from BRE.
9.2.7.4 Acoustically transparent
Tensile fabric membranes weigh about 1% the weight of an equivalent glass enclosure
and membranes are often described as acoustically transparent. The term acoustic
transparency can be misused in describing the intent of the acoustic environment
that the tensile fabric provides. Generally, if a material is acoustically transparent,
it does not isolate or absorb significant amounts of sound energy. The following
(anonymous) phrases have been used to describe ETFE or PTFE in practice, and
they demonstrate some of the semantic misconceptions:
• “Known to be ‘acoustically transparent’ compared to other materials such as glass, so
undesirable background noise (is reduced) .”—anonymous industry professional
Acoustics in architectural fabric structures: the case of ETFE pillows 251

Acoustic transparency is applied in reference to reducing background noise levels.

Background noise levels are measured from mechanical services noise, and ambient noise
levels are measured from activity noise over a certain period of time. The intent of the reference
is that sound energy is not trapped within the space and is transmitted through the material.
This is true for low-frequency noise, as the large sound waves react with the membrane as a
large drum, and the membrane dissipates the acoustic energy through the process of vibration.
High-frequency wavelengths will reflect back into the space.

• “A foil roof is acoustically relatively ‘transparent.’”—anonymous industry professional

The reference is in the context of describing the foil as a sound absorbing material, which
may be used to optimize the room acoustic environment. The material is frequency depen-
dent and will, therefore, influence the acoustic environment mostly at mid to high
frequencies, making the space sound tinny.
• “The ETFE cushions are ‘transparent’ to sound, meaning that sound travels through and
very little noise is reverberated back”—anonymous industry professional
In this case, noise generated within a space is beneficial as the ambient noise will transmit
through the façade enclosure and dissipate into the atmosphere. This also means that noise
generated outside the membrane enclosure (i.e., mechanical equipment noise, aircraft) will
more easily transmit into the space.
Acoustically transparent materials are highly porous, similar to a fabric membrane that
covers a loudspeaker. An ETFE foil is not porous and, therefore, not invisible to all
sound frequencies. An ETFE foil can have a mass as small as 1 kg/m2 (0.2 psf) and
at some frequencies is indeed acoustically transparent, but the material should be eval-
uated for the acoustic impact on the space use and environment at the early design
phases.
Table 9.1 shows examples of existing venues with tensile fabric enclosures and an
associated acoustic consideration that may have been evaluated. Many acoustic

Table 9.1 Project typologies with ETFE or PTFE building envelopes

Potential
Project Year Capacity Material Architect consideration

Eden project, 1997 e ETFE Grimshaw & Noise ingress

Cornwall Partners
The O2 1999 20,000 PTFE Richard Room acoustics
(Millennium Rogers for
Dome), performances
Greenwich
Allianz Arena, 2005 71,137 ETFE Herzog & de Enhancing crowd
Munich Meuron excitement
National 2006 17,000 ETFE CSCEC Evaluating rain
swimming PTW noise impact
center, architects
Beijing
252 Fabric Structures in Architecture

parameters are evaluated and targeted for project design on a case by case basis; only
one acoustic concept is listed.

9.3 Case study: National Aquatic Center Beijing—the

Water Cube
The Water Cube erected for the 2008 Beijing Olympics was intended as a 17,000 seat
swimming center to be used after the Olympics as a multipurpose leisure and elite
swimming center (Figure 9.12). The design and construction is a result of the collab-
oration between PTW architects, Arup, and China Construction Design Institute.
The façade design of ETFE pillows is based on a concept related to the arrangement
of organic cells and the natural formation of soap bubbles. Additional drivers of this
form included interdisciplinary requirements for structural, thermal, and acoustical
properties. It was important to provide an appropriate internal acoustic environment,
to reduce structural weight due to the high seismicity of the location, and to reduce
the consumption of resources. The final design of the aquatic center integrated
100,000 m2 of pressurized ETFE pillows striking a balance between aesthetic and
technical purposes.
Generally, swim centers are noisy due to the acoustic energy generated by the
patrons (e.g., shouting and screaming) and exacerbated by the sound-reflecting surface
finishes within the space. ETFE pillows are largely transparent to sound and provide an
acoustic response more similar to an open air venue.

9.3.1 Acoustic criteria

The criteria for the aquatic center followed Chinese Standard Code for “Acoustic
Design and Measurement of Gymnasiums” (JGJ/T131-2000/J42-2000). Acoustic

Figure 9.12 Beijing National Aquatics Center (the Water Cube) © Ben McMillan.
Image by Courtesy of ARUP.
Acoustics in architectural fabric structures: the case of ETFE pillows 253

targets were established for reverberation time, sound isolation, and speech intelligi-
bility. This case study is focused on how rain noise can influence the speech intelligi-
bility requirements for public address and life safety announcements.
Noise levels created by the impact of rain on the lightweight façade can be exces-
sive and interfere with pool activities, particularly during heavy rainfall. This will
significantly reduce the intelligibility of announcements made by the public address
systems.

9.3.2 Arup acoustic studies

Rain noise suppression formed part of the Arup design process. The following factors
influencing noise impact were studied (Field, 2003):
• Rain flow rate (liters/minute)
• Rain droplet size
• Surface area over which rain impacts the ETFE
• Velocity of droplets: simulated rain (simulated rain at 1 m above ETFE) was compared to
terminal velocity of natural rain (expected to be higher than simulated rain velocity).

At the very early concept phase, three alternative principle approaches were considered
to reduce rain impact noise on the lightweight roof (Field, 2003):
1. Adding an external surface above the ETFE structure to interrupt rainfall and reduce the
impact when falling on the pillows (Figure 9.13).

2. Adding a sound isolating surface inboard of the ETFE structure to attenuate noise generated
by the direct impact of rainfall on the ETFE pillows (Figure 9.14).
3. Increasing the damping within the roof structure to reduce the noise generated by the rain
droplets on impact. This is achieved by reducing tension (i.e., air pressure) at the outermost
surface. To achieve any benefit from this approach, a multilayer ETFE roofing system would
be required (Figure 9.15).
At the early concept phase, these considerations needed to be evaluated for aesthetic
and multidisciplinary performance impacts and cost implications.

Figure 9.13 Additional external surface above the ETFE structure.

Image courtesy of Field, 2003.
254 Fabric Structures in Architecture

Figure 9.14 Internal acoustic insulation.

Image courtesy of Field, 2003.

Lower pressure of air in outermost

layer—reduces noise generated by
impact of rainfall. Impact noise
further attenuated by inner laver(s)

Figure 9.15 Additional mass layer as acoustic treatment option.

Image courtesy of Field, 2003, ARUP.

In subsequent design phases, other acoustic treatment options were recommended

by adding a mass layer, sealed airtight, covering the entire surface area of the roof.
Diagrams of these options are given below (Figure 9.16).
The sound source data used to predict internal rain noise levels was measured dur-
ing test simulations—rain droplets over a 1 m2 area of ETFE at a controlled flow rate.
Acoustic predictions based on this data were applied to the surface areas of ETFE in
the aquatic venue. A further 8e10 dB reduction in rain noise levels were predicted
by the use of a rain mesh suppressor (per the manufacturer), to be used externally
above the ETFE cushions (Field, 2003).
The acoustic studies demonstrated that it would require significant architectural
intervention to control the indoor levels caused by rain noise.
Acoustics in architectural fabric structures: the case of ETFE pillows 255

Roof construction Schematic Expected

indoor
noise level
at nearest
spectator,
dB(A)

An internal mass layer (12 kg/m2)

sealed airtight within the void
between the ETFE cushions. The
mass layer will be formed according
60
to the shape of the space frame
structure between the ETFE
cushions

ETFE cushion (no rain mesh), 6 mm

polycarbonate plate (12 kg/m2)
58
within cushion (based on published
lab test data)

An external mass layer (12 kg/m2)

backed with 100 mm mineral wool
insulation (mininum density 48
kg/m3) covering 80% of roof area, +
with remaining 20% covered with 53
ETFE pillow incorporating 6 mm
polycarbonate plate (12 kg/m2) 80% 20%
inside`

Figure 9.16 Predicted noise levels of acoustic treatment of the ETFE roof based on options for
added mass.
Image Courtesy of C. Field and E. Valmont (Field, 2003)/ARUP.

9.4 Conclusions
Acoustic design investigations should be conducted at early design stages to decide if a
lightweight tensile fabric system is appropriate.
Some early considerations are the following:
• Noise survey: sound measurements of external ambient noise, including an appraisal of noise
impact from transit systems, for example, aircraft, watercraft, vehicular, or rail.
• Mechanical and electrical noise assessment: assess preliminary noise data to predict noise
break-in through the ETFE roof.
• Feasibility of speech intelligibility requirements per code or life safety.
The implications of the acoustic transparency of a lightweight mass need
to be assessed. Certain frequencies of sound may adversely influence the use of the
space.
256 Fabric Structures in Architecture

Designers should be aware that various façade membrane products will have
different acoustic performances based on the physical attributes related to mass and
material.

References
BRE. (November 2004). Measurement of rain noise on roof glazing, polycarbonate roofing and
ETFE roofing. Report 220312. Available at website: http://www.bre.co.uk/pdf/Rain
NoiseGuidance.pdf.
Cavanaugh, W. J., Tocci, G. C., & Wilkes, J. A. (2009). Architectural Acoustics: Principles and
Practice. John Wiley and Sons.
Field, Chris, & Arup. (October 27, 2003). Preliminary Design 100% Report e National
Swimming Centre. Beijing. Project Design. Arup.
Foiltec, V. (January 25, 2008). Texlon: Outline Design Guidance. London UK: © Vector Special
Projects Studio 4 The Cooperage. [email protected].