Professional Documents
Culture Documents
Auditorium Acoustics
and Architectural Design.
2nd edition. Spon Press.
(1) Introduction
Waves
The propagation of sound energy in an elastic medium, such as air, is
mainly in the form of longitudinal waves
http://en.wikipedia.org/wiki/Longitudinal_wave
http://en.wikipedia.org/wiki/Transverse_wave
More: http://www.physicsclassroom.com/mmedia/waves/lw.cfm
http://www.ndt-ed.org/EducationResources/CommunityCollege/
Ultrasonics/Physics/modepropagation.htm
Waves
Waves
Our hearing apparatus can handle sound intensity in a wide range, approximately
100 W
The hearing range in terms of intensity is fourteen powers of ten = 14 bel = 140 decibel.
Acoustic pressure is proportional to the square root of the intensity: real mean square, RMS.
In terms of pressure, the hearing range expressed above corresponds rather well to
200 N/m2
(20 Pa)
and the lower limit is by convention the reference level in acoustics (ANSI-1969/ISO-1963).
A sound level is indicated i.e. 60 dB re 20 Pa, or more succinctly, 60 dBSPL.
SPL = sound pressure level.
More: http://en.ossmanuals.net/csound/ch008_c-intensities/
Direct sound
Direct sound is the sound that travels from a source to a listener in a straight path (without reections).
It is the part of sound that arrives rst.
Direct sound in an enclosed space (e.g. a concert hall)
behaves much the same way as in a free eld (e.g.
outdoors). For every doubling in distance, the sound level
decreases by 6 decibel. This follows from the inverse
square law of propagation.
I = P/ 4r2
Intensity= Pressure/Area
Borb (2008). http://en.wikipedia.org/wiki/File:Inverse_square_law.svg
Seat-dip eect
It was discovered in 1962 that audience members receive a harmonically distorted direct sound, i.e. the
propagation does not have a at frequency response. Direct sound travelling at grazing incidence will be
subject to absorption over and above the inverse square law, but the eect is dierent at low and mid
frequencies.
At low frequencies, our hearing is less sensitive to interaural time dierences. Deciency in direct sound can thus
be compensated by late sound, e.g. a long enough
reverberation time at low frequencies.
The diagram also indicates that lateral reections are
unaected by the broad-band seating material absorption,
but the seat-dip eect is equally pronounced as for direct
sound.
The strong attenuation around 125 Hz is assumed to be due to resonance eects happening between rows of
seats. It is sometimes referred to as a seat-dip eect, and would be related to Helmholtz resonators.
Remarkably, The attenuation does not depend on distance but is uniformly distributed.
More: http://www.rexresearch.com/helmholtz/helmholtzresonators.htm
http://physics.kenyon.edu/EarlyApparatus/Rudolf_Koenig_Apparatus/Helmholtz_Resonator/Helmholtz_Resonator.html
The broad-band attenuation around 800 Hz would be due to absorbent seating materials.
Reection
Geometries for reection for light and sound are identical.
Pure reection (specular) of a sound wave on a nite-size reector happens for frequencies that are higher
(that is, with shorter wavelength) than a certain number. This cuto frequency depends on the apparent size of
the reector from the vantage points of both the source and the listener. (See Barron p. 457 for details).
Reection between a source and a listener can happen
when a reector is placed as a tangential to an ellips with
Fsource and Flistener as foci.
Good concert hall design maximises the density of the
subjectively important early reections. One way of achieving
this is to have shape and dimensions similar to an ellipse.
(Note: 50 ms is has been chosen based on psychoacoustic thresholds. Walls should generally
should not be curved, to avoid slap-back echoes at points near the ellipses two axes.)
Diraction
Diraction occurs whenever a propagating sound
wave encounters an obstacle. The bending or blocking
eects are most pronounced for waves where the
wavelength is roughly similar to the dimensions of the
diracting object.
Diraction of sound waves is normally a lowfrequency phenomenon.
At low frequencies, sound waves recombine behind a
small obstacle, while higher frequencies are reected
away (see page on reection).
Scattering
Scattering takes place when a reecting surface is irregular.
More: http://www.subwooferbuilder.com/qrdude.htm
http://www.digitalaudiorock.com/cgibin/qrd.cgi
http://www.pmerecords.com/
Diusor.cfm
Absorption
Absorption is the energy loss from the sound wave travelling in the propagation medium (e.g. the air) into
the reecting medium (e.g. a wall).
In practice, all building materials are scattering and absorptive to some degree.
The absorptivity of a material for construction or interior design is normally measured in a laboratory.
Standard tables are published and much in used in prediction of reverberation time.
More: http://www.sae.edu/reference_material/pages/Coecient%20Chart.htm
Background noise
Concert halls are often constructed on urban sites, where trac and other noise sources necessitate
substantial expenditure to achieve inaudibility inside the auditorium. (Barron p. 27)
The main contribution to background noise - in concert halls, laboratories and recording studios alike comes from ventilation systems. This is particularly the case in places where air conditioning is used a
lot (such as in Singapore).
As a measure of background noise, the NR (Noise
Rating) criterion is most common in Europe. (Other
measures include NC, PNC and RC).
NR is estimated by measuring SPL of ambient noise at
9 octave-band lters. The resulting curve is compared
with the template (see diagram) and moved upwards
or downwards so that no measured band exceeds the
templates at a corresponding frequency. The rating is
determined by the minimal distance between
measured curve and a NR template curve.
http://www.engineeringtoolbox.com/nr-noise-rating-d_60.html
NR15
NR20
NR25
Reverberation time
The most important acoustic measure for a room is its response. The sound that can be heard when a loud
source is turned o is call terminal reverberation. Subjectively, in situations of speech or music, the
running reverberation, i.e. the early part of the reverberation, has been shown to be more salient.
The RT60 measure is the time in seconds it takes for a
terminal reverberation to recede to a millionth of its energy,
i.e. by 60 decibel. In practice, it is calculated from the slope
between -5 and -35 dB below steady-state maximum. This
assumes a linear decay, which is not always the case. The
presence of coupled acoustic spaces with dierent
reverberation characteristics can give rise to non-linear
decay. In auditorium acoustics, such coupled spaces can be
found between the main hall and a reverberant ytower, or
an orchestra pit, or under large balconies.
Recommended RT60 values (in seconds) for auditoria depend on programme:
Drama theatre
Chamber music and opera
Classicist music
Romantic music
Organ
0.85 +- 0.15
1.55 +- 0.2
1.7 +- 0.1
2.0 +- 0.2
>2.5
Early decay time approaches RT60 for rooms that are highly diuse.
Reverberation time
Wallace Clement Sabine is the founder of modern acoustics. He discovered around 1900 that a rooms
reverberation time depends on two things only:
- room volume (V)
- total absorption (A)
The total absorption is the sum over all elements of the product of each elements size and absorptivity:
A = S*a
The equation for reverberation time (in metric units):
T = 0.161*V
A
Sabines formula for predicting total absorbtion was not correct, mainly because he was calculating seat
absorption on a per-seat basis rather than on per-square-meter basis. This caused him to grossly overestimate
the reverberation time in Boston Symphony Hall (Barron p. 89).
In auditoria, the major absorbing surface is people. Beranekt (1969) proposed a model where reverberation
time is estimated from volume and ST, the total oor area of audience, orchestra, and chorus (Kinsler & Frey p.
345):
The discussion of how to nd optimal measurement method and expression for a prediction is on-going. The
economic gains with a more exact prediction are obvious. For example, prediction methods currently in use
depend on precise values for the absorptivity of seating upholstery.
At higher frequencies, reverberation time also depends on air absorption, which in its turn depends on
temperature and humidity.
Reverberation time
Check predicted RT60, frequencies for standing waves, Schroeder
frequency etc with the RoomAcousticsCalculator! (Lindborg 2011)
Reverberation time
Calculate reverberation time from a soundle using the RT60 estimator!
(Lindborg 2011)
Subjective dimensions
Good acoustic design for good speech is easy speech must be intelligible!
With music, no direct assessment exists.
5D Cube
2. For each dimension, place a mark to indicate YOUR SUBJECTIVE IMPRESSION of the acoustics.
Muddy |---------|---------|---------|---------|---------|---------|---------|---------|---------|---------| Clear
CLARITY
REVERBERANCE
SOURCE WIDTH
ENVELOPMENT
INTIMACY
(do you feel acoustically involved and able to identify with the source?)
LOUDNESS
3. Indicate how you perceive the BALANCE between MID-range, and BASS and HIGH ranges:
TREBLE to MID
BASS to MID
4. BACKGROUND
NOISE
|---------|---------|---------|---------|---------|---------|---------|---------|---------|---------|
(inaudible)
(Acceptable)
5.
(Tolerable)
(Intolerable)
http://pcfarina.eng.unipr.it/Aurora/SAW/RoomSim.html
54.58
54.34
54.57
54
Intensity (dB)
42
36
30
24.01
24.12
24.92
0.806991
0.134193
0.12237
0.2
0.3
0.4
0.5
0.6
0.7
0.755128
0.763125
1.57667
Time (s)
http://students.uta.edu/MS/msb3676/THE%20CONTINUUM
%20OF%20SPEECH%20RHYTHM%20posterFinal.pdf
http://wwwu.uni-klu.ac.at/gfenk/posterLondon2.pdf
The perceived degree of source broadening also depends on direct sound level, and Barron writes:
[with 10 m as standard, L - L0
is in a range around 0 dB]
L - L0 = 10*log(d+er+l)
exponential function
of distance (r)
d = 100/r2
er = (31200 / V) e-0.04r/T * (1 - e-1.11/T)
l = (31200 T / V) e-0.04r/T * e-1.11/T)
Design recommendations
Barron discusses recommended values for the objective
characteristics that acoustic designers should aim for.
In Barrons study of British halls:
1.82.2 seconds
-22 dB
0.10.35
0 dB
Reverberation time
ISO 3382-1:2009
http://www.iso.org/iso/iso_catalogue/catalogue_tc/catalogue_detail.htm?csnumber=40979
- early reections
An orchestral shell is often advantageous. EDT onstage should be 70% of RT in the main hall.
- hearing damage is unfortunately common amongst orchestra musicians (around 25%, as quoted in the PDF above)
Historic overview
The width of a hall might be related to the maximum acceptable width for a stage, the length to visual
distance, and the height to reverberation time considerations (p. 75). While it seems that the overall
dimensions of an auditorium are set, (luckily enough for eager designers) the shape isnt.
Until around 1900, auditorium design was based on (serendipitous) precedence. With Sabine, enter science.
The four traditional forms, in
order of historical appearance:
Fan shape
Arena
Baroque theatre
Classical rectangular plan
Historic overview
In small halls, the room form matters relatively little for the acoustics.
The four most renowned classical concert halls (Beranek 1962), are
Grosser Musikvereinssall (Vienna)
Neues Gewandhaus (Leipzig)
Concertgebow (Amsterdam)
Symphony Hall (Boston)
The dimensions of these 4 halls are
similar (see p. 80 for details).
Averages are:
Construction around 1900
Volume
17000 m3
Seats
2000 pax
Length
48 m
Width
22 m
Height
17 m
1.85 s
H:W:L
1 : 1.22 : 2.70
Design factors
It is much easier to determine the reason for failure than for success (Barron p. 90)
The subjective acoustic experience is not due to any single factor. Risking absurd reductionism, which are the
design factors to keep in mind for good concert hall acoustics?
all seats are close to
a reecting surface
more envelopment
higher clarity
choice of material
more warmth
more envelopment
Which are the design factors that may cause bad concert hall acoustics?
too much reliance on
overhead reectors
colouration of higher
frequencies
bad sightlines
lack of intimacy
lack of power
Study of a failure
New York Philharmonic (1959) - the most publicized acoustic disaster of the twentieth century.
critics described sound as clear, a little dry, with not much
reverberation and a decided lack of bass steely hardness, the
ddles sounded harsh, and the orchestras sections failed to blend, as
if invisible walls stood between strings, woodwind and brasses
main oor rake is remarkably shallow
(bad sightlines + obstructed direct sound)
rear wall is concave
(slap echoes)
many seats have receive no lateral reection
stage too large = orchestra too widely spaced
(lack of intimacy and self-sound support for musicians)
double layer of small and regular suspended reectors
(some 30,000 small dowels, http://en.wikipedia.org/wiki/Avery_Fisher_Hall)
Study of a success
Berlin Philharmonie (1963) - people always gather around in circles when listening to music informally.
"product of the Expressionist movement and of organic
architecture, this concert hall in which the audience is seated around
the orchestra was worked out in accordance with the laws of
acoustics." (Stierlin http://www.GreatBuildings.com/buildings/Berlin_Philharmonic_Hall.html)
with the arena form there are two major acoustic concerns:
- the voice, and many instruments, are directional
- surfaces are needed to provide early lateral reections
solutions:
- not fully central stage, directionally biased
- vineyard terraced concert hall
tent-like prole of convex surfaces assist diusion
pyramidal diusers with slits acting as Helmholtz resonators
(limiting low-frequency reverberation: 1.9 s at mid, 2.1 s at low)
critic to balance: acoustic uniformity somewhat uneven
space acts as a positive force upon the life it contains
Music as the focal point no segregation between producers
and consumers, but rather a community of listeners the
construction follows the pattern of a landscape. with the
auditorium as a valley the ceiling, resembling a tent,
encounters this landscape like a skyscape (Scharoun ca. 1963)
absorbent
(5) British concert halls and conclusions for concert hall acoustics
Royal Albert Hall
Royal Festival Hall
Acoustical design of concert halls in
subjective and objective terms
More: http://peutz.fr/lacoustique/articles/salles/PaperIOA02.pdf
intimacy
loudness
warmth
level 0 dB
(6) Chamber music and recital halls (not part of course requirements)
Dierences between symphony halls and chamber music halls
Opera
Opera is an extravagant art form (Barron p. 333) and was at rst limited to wealthy courts. When brought to public
consumption the demands resulted in a theatre-building boom. In the 20th century, opera (as well as its various spinos: varit, cabaret, musical, rock show etc) came to fully exploit the stage, auditorium, electronics, screens, and all
the other media to extend its palette of expression.
Today, opera is a curious mixture of conservatism and renewal. Of the 1000+
composers programmed over the last ve seasons, over 500 are alive not bad for an
art-form that is sometimes thoughtlessly described as ready for the museum.
http://operabase.com/top.cgi?lang=en&
#opera productions globally, last 5 seasons
1638
SS Giovanni e Paolo
1993
Glyndebourne
Acoustic criteria
Opera house design has to resolve with many contending requirements, including:
- reverberation time (optimal speech intelligibility 1.0 s, optimal music 2.0 s)
- loudness balance between singers and orchestra
- source broadening (more important for orchestra than singers)
Barron argues that the performer-audience relationship is dierent in symphonic music and opera:
concert hall
opera house
auditorium
auditorium
stage
stage
Architectural design
Plan design in an opera has to originate from the proscenium opening.
To create early lateral reections (for both singers and orchestra), proscenium splays are helpful.
For visual reasons, the proscenium is typically more narrow than
the width of the hall, leading to at angle of the side splays. This
produces few early reections.
A solution is to place a retreated splay with a
more parallel angle behind a visually opaque but
acoustically transparent curtain. This will produce
a more dense set of early reections
More: http://peutz.fr/lacoustique/articles/salles/PaperIOA02.pdf
Recommendations:
RT = 1.31.8 seconds (+ slight bass rise; shorter for Mozart, longer for Verdi)
EDT RT (watch out for seats under balconies, with low EDT/RT ratio)