Architectural Acoustics II
Indoor Acoustical Phenomena
Prof S K Tang
Preface
Concepts on direct and reverberant sound
fields and reverberation time have been
introduced in past years
Direct sound follows inverse square law
Reverberant field is uniform (ideal case)
Enclosure has acoustic response
Human responses to sound is subjective –
psychology & physics
Acoustic Responses of Enclosure
Similar to a mechanical & electrical system, an
enclosure responses to acoustic excitation
Impulse decay g(t) – impulse response function
g includes information of direct sound,
reflections and the amount of room absorption
These responses then modify the direct acoustical
signal and contribute directly to human
responses
Modulated acoustic signal is heard
Psycho-acoustics Consideration
Reflections cannot be avoided
A mixture of direct sound and reflections
reaches human ears
How these reflections are perceived?
Two questions to deal with :
Psycho-acoustics Consideration
Under what condition is a reflection
perceivable at all, without regard to the
way in which its presence is manifested,
and under what condition is it masked by
the direct sound?
Under what condition does the presence
of a reflection rate as a disturbance of
the listening impression?
Absolute Perceptibility
function of the time delay of the reflection
The threshold also depends on type of
sound
For a sound of 70dBA and frontal
incidence :
L  0.6to  8
Human is more sensitive to reflections
arriving from lateral directions
Threshold of absolute perceptibility for
speech
Threshold of absolute perceptibility for
music
Our hearing is less sensitive to reflection
when music is concern
Echoes & Colouration
a reflection may not reach the consciousness
of the listener even it is perceived
At lower levels, it manifests itself by an
increase of the overall loudness of the total
sound
At higher levels, a reflection can be heard as
a separate event – ‘Echo’
Not welcomed in indoor environment. It may
distract listener’s attention, reduce enjoyment
of music and impair speech intelligibility
Perception of an Echo
The superposition of a strong isolated
reflection or multiple significant regular
reflections onto the direct sound results in
a characteristic change in timbre (spectral
characteristics) – ‘Colouration’.
This is important when music is concerned
Acoustical Parameters
Echoes are not the only factor affecting
the intelligibility.
The duration of a sound decay, which is
brought about by multiple reflections in the
presence of boundary absorption, is of
equal importance
Commonly used parameter : RT
Haas Effect
Early reflection,
which is the sound
energy reaches the
listener within 50
ms after the arrival
of the direct sound,
enforces
intelligibility as it
integrates nicely
with the direct
sound
Early Decay Time (EDT)
time it takes the sound energy to decay by
10dB multiplied by 6 = EDT
RT equals the time taken for the sound
energy to decay by 60dB
EDT includes a few isolated early
reflections
A short EDT enhances clarity
Longer RT gives the feeling of ‘liveness’
Clarity
Clarity is defined as the difference (in dB) of the
sound energy received at a listener in the first
80ms minus the (late) reverberant energy, which
is the entire energy arriving at the listener after
80ms
 80 ms

2
  g (t ) dt 
 0

C80  10 log 10  

 g (t )2 dt 
 80ms

Clarity
characterize the transparency of music in
a concert hall and describe the fullness of
tone
A C80 of -3dB is still acceptable.
The typical range of this parameter is from
-3dB to +5dB
Definition
describes the fullness of tone but it is
related to speech intelligibility instead of
music.
50 ms
 g (t ) dt
2
D
0



g
(
t
)
dt

2
0
100%
Definition
Centre Time

 g (t ) tdt
2
ts 
0

 g (t ) dt
2
0
Gravity / Strength Factor
C80 and D will become meaningless if the
sound is weak to be heard at a
comfortable loudness
assumption of uniform reverberation field
is not valid in halls or long enclosures


2
  g (t ) tdt 
0

RT


G  10 log 10  
 G  10 log 10 
  45
 Volume 
 g (t )2 dt 
 0 A

Initial Time Delay Gap, ITDG
This parameter
describes intimacy subjective impression
of the size of a hall
ITDG - interval in
milliseconds between
the arrival of the
direct sound and the
first reflection at the
listener
Early Lateral Energy Fraction, LEF
correlate with the spaciousness of an
enclosure
independent of other reflections and of the
presence or absence of reverberation
80 ms
 g (t ) cos  dt
2
LEF 
5 ms
80 ms
 g (t ) dt
2
0
Specular Reflector
Diffuse Reflection
Convex Surfaces
Quadratic Residue Diffuser
diffusers made from surface modulation
have two major limitations :
protrusions and recesses have to be large
to provide good diffusion at low
frequencies.
no objective method to determine the
extent of scattering
Consists of an array of linear slits (or wells)
of constant width.
Wells are separated by thin rigid walls.
Number of wells inside a period of the
diffuser is a prime number
Design for Speech
Speech should be intelligible without an undue
strain on the listener.
to retain the natural character of the speaker’s
voice.
Providing optimum RT
Eliminating acoustical defects such as echoes
and flutter echoes
Maximizing loudness in the audience
Minimizing the ambient noise level
Providing a speech reinforcement system where
needed
Speech Transmission Index STI
STI is derived from 7 octave band
modulation transfer functions m(F)
F represents a modulation frequency :14
Fs (1/3 octave bands from 0.63Hz to
12.5Hz).
m(F) is related to RT and S/N :
 
RTi 
m( F )  1   2F

13.8 
 
2



0.5
1  10

 S / N i 10 1
STI
STI i  (S / N app  15) 30
S / N app, F
 m( F ) 
 10 log 10 

1  m( F ) 
7
STI   wi STI i
i 1
STI
Quality
Speech Intelligibility Loss of Consonants
0 – 0.2
Bad
< 50% intelligible
> 50%
0.2 – 0.4
Poor
50 – 60% intelligible
40 – 50%
0.4 – 0.6
Fair
60 – 78% intelligible
22 – 40%
0.6 – 0.8
Good
78 – 98% intelligible
2 – 22%
0.8 – 1
Excellent
> 98% intelligible
< 2%
Articulation Index (AI)
AI is a parameter obtained from the weighted
average of S/Ns in the octave bands from 250Hz
to 4kHz
AI
< 0.3
0.3 – 0.5
0.5 – 0.7
 0.7
Speech
intelligibility
Poor
Marginal
Good
Very
Good
Percentage Articulation Loss of
Consonants (%ALcons)
% ALcons  100(10  2[( A  BC )  ABC ]  0.015)
A = -0.32 log[(Lr+Ln)/(10Ld+Lr+Ln)]
B = -0.32 log[Ln/(10Lr+Ln)]
C = -0.5 log (RT/12)
Ld = Direct level (energy within 35 ms)
Lr = Reverberant level
Ln = Background noise level
Office Design
For an office, it is the speech privacy which is of the
major concern.
Certainly, the noise level and NC or RC are also in the
design specification.
Speech privacy is basically the opposite of speech
intelligibility.
STI and AI and even %ALcons can also be used to
describe it
AI
Speech privacy
0 – 0.05
0.05 – 0.15
0.15 – 0.2
> 0.2
Very Good
(confidential)
Good
(normal)
Marginal
(little)
Poor
Privacy between Rooms
An index called privacy P is derived from
the isolation property of the partition wall
(STC), receiver room characteristics (RR),
background noise level in receiver room in
dBA (N), voice level (VL) and a source
room factor (SR) :
P  STC  RR  N  VL  SR
Privacy in Open-plan Office
Acoustic privacy in open-plan offices is achieved
by using absorption partitions (of sufficient
height), absorptive ceiling/floor and in some
cases by artificial sound masking.
Common absorptive ceiling usually has a NRC >
0.75 (NRC is the average of the sound
absorption coefficients of the octave bands from
250Hz to 2kHz).
HVAC noise provides some masking effects
L  VL  P  R  STCeq
Masking
masking noise level
cannot be too high
and must be
acceptable to the
occupants.
In general, a masking
noise level of 48dBA
is the practical upper
limit for open-plan
offices
Considerations for Music
The acoustical phenomena related to
musical attributes may be divided into the
following two categories :
(a) to early sound : reverbance (EDT),
Clarity, Intimacy, Spaciousness (laterality)
(b) to reverberant sound : warmth
(liveness), brilliance, loudness (G)
Bass Ratio BR
Warmth feeling is associated with low
frequency reverberation
BR is between 1.1 to 1.25 for RT  2.2s
Between 1.1 to 1.45 for RT < 1.8s
Larger BR than the recommended values
should be avoided
RT125  RT250
BR 
RT500  RT1000
Brilliance
related to the early decay of high
frequency sound
EDT2000
 0.9
EDT500  EDT1000
EDT4000
 0.8
EDT500  EDT1000
Computational Approach