Frontiers of Architectural Research (2017) 6, 116

Available online at www.sciencedirect.com

Frontiers of Architectural Research

www.elsevier.com/locate/foar

RESEARCH ARTICLE

Acoustic improvement on two lecture

auditoria: Simulation and experiment
Soha Eldakdokya,n, Ahmed Elkhateebb

a
Department of Architecture, Faculty of Engineering, Helwan University, Egypt
b
Department of Architecture, Faculty of Engineering, Ain Shams University, Egypt

Received 19 April 2016; received in revised form 30 September 2016; accepted 7 November 2016

KEYWORDS Abstract
Acoustic performance; Based on the commonly used indicators for speech intelligibility, this work acoustically
Speech intelligibility; evaluates the two largest auditoria in the Faculty of Engineering, Helwan University, Cairo,
Reverberation time; Egypt, using experimental and digital simulation techniques. Design treatments were also
Background noise level; suggested to improve the acoustic performance of the auditoria, where the impact of these
Lecture auditoria
treatments was checked using the simulation as well. The models that were analysed using the
CATT software were rst validated utilizing the results of the eld work in the unoccupied
rooms. The results showed that the acoustic quality of the two auditoria are far from the
optimal conditions due to their improper acoustic characteristics and the high noise levels as
well. The results of improvement proposals showed that altering the ceiling shape and adding
efcient absorptive materials to the rear surfaces successfully reduced the excessive
reverberation time to the optimal values, increased the early reections and eliminated the
shadow zones. In addition, decreasing the noise levels by 20 dB due to improving the window
insulation noticeably improved the speech intelligibility at all receivers.
& 2017 The Authors. Production and hosting by Elsevier B.V. This is an open access article under
the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).

1. Introduction have been asked by the administration of the Faculty to

Due to frequent complaints about the low sound quality in
many lecture auditoria in the Faculty of Engineering at
Helwan University, particularly in large rooms, the authors
n
Corresponding author.
E-mail addresses: Scometd@yahoo.com (S. Eldakdoky),
ahmed_elkhateeb@eng.asu.edu.eg (A. Elkhateeb).
Peer review under responsibility of Southeast University.
investigate this problem and propose feasible and eco-
nomic solutions. The faculty was rst established in 1955
as an industrial institute of higher education. Later, in
1975, it joined Helwan University as the Faculty of
Engineering. It is located in a dense residential district
to the north of Cairo City and is surrounded on three sides
by narrow and congested streets. The faculty occupies
approximately 12.4 acres (52.000 m2), shown in Figure 1,
and contains six departments. The faculty contains more
than 70 auditoria of different areas and volumes; 54 of

http://dx.doi.org/10.1016/j.foar.2016.11.002
2095-2635/& 2017 The Authors. Production and hosting by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
2 S. Eldakdoky, A. Elkhateeb

them are small rooms (less than 265 m2), while the rest Room RAZ, the largest auditorium in the faculty, is the
are large rooms (more than 265 m2). main lecture room for the preparatory level, Building 4 in
Thus, the scope of this work includes two of the large Figure 1. It is a trapezoidal room, as shown in Figure 3. The
rooms in the faculty, namely, room C and the Abdolraseq slope of the audience area is approximately 9, which
room (RAZ). Room C, the second largest room in the faculty, provides a good sight line and grants a direct sound path
is one of four typical lecture rooms in the department of for every student. Similar to most auditoria in the faculty,
civil engineering, building 3 in Figure 1; hence it represents the two rooms utilize natural ventilation via open windows.
25% of the large auditoria in the faculty. It is a square room These windows directly overlook a crowded, noisy street,
in which the benches have been arranged diagonally, see parking areas and buildings 2 and 7 (see Figure 1). The main
Figure 2. The slope of the audience area is approximately architectural features of both auditoria are summarised in
7, which agrees with the minimum slope required for sight Tables 1 and 2. Speech intelligibility (SI) in lecture rooms is
lines (Elkhateeb, 2003). a major concern in educational spaces not only to support
the learning process (Reich and Bradley, 1998) but also for
lecturers comfort. Recent studies indicate that improper
1 Departments of
Architecture, Math acoustic conditions in lecture rooms cause severe vocal
and Physics problems for approximately one-tenth of teachers (Brunskog
2 Department of
automotive and et al., 2009). Speech intelligibility is a single measure for
locomotive, Library, both room acoustics and the speech-to-noise ratio (S/N)
Theatre and
administration.
(Reich and Bradley, 1998); it can be governed by different
3 Civil department indicators, such as the useful to detrimental sound ratio
auditoria (U80), speech transmission index (STI) or its common
4 Main auditorium
5 Civil Departments variant, the rapid speech transmission index (RASTI) (Reich
and Automotive& and Bradley, 1998; Shams and Rama krishnan, 2012). SI is
Locomotive
Workshops considered to be a function of reverberation time (T); early
6 Mechanical Power reection energy can increase the S/N by up to 9 dB
Department and
Laboratories.
(Bradley and Sato, 2003), while an S/N of + 15 dB can raise
7 New Building Under SI scores by up to 100% (Bradley, 2002). Thus, the optimal
construction reverberation time (Topt) is an important criterion support-
8 Mosque
ing SI (Shams and Rama krishnan, 2012). Background noise
Figure 1 Layout of the Faculty of Engineering. level (L) is another important acoustic indicator (Bradley

Figure 2 Typical lecture auditorium of the Civil department, room C.

Acoustic improvement on two lecture auditoria 3

Figure 3 Main lecture auditorium, room RAZ.

Table 1 Description of Room C.

Dimensions

L W H (avg.) Area m2 Volume m3 Perm.a D Max.b m

16.6 16.6 5.15 267.8 1390 66.9 19

Statistical Data

No. of Audience VP m3c Net Audience Area m2 APdm2 Topt@500 Hz Environmental Conditions
Net Gross RH % Temp. 1C
306 4.5 122.6 0.40 0.88 0.64 43% 26

a
perimeter.
b
maximum distance between the sound source and farthest receiver.
c
volume per person.
d
Area per person.

and Sato, 2008); because excessive noise distracts student
attention and reduces the S/N, the acceptable steady level
of noise in learning spaces should not exceed 35 dB
(Standard, ANSI/ASA S12.602010/Part1American National,
2010).
1.1. Related work
Many researchers have investigated the acoustics of learn-
ing spaces, either based on eld measurements or digital
simulation, in the design phase or after construction.
DiMarino et al. DiMarino et al. (2011) analysed the acoustic
performance of the proposed design for a preconstruction
lecture hall in the University of Windsor's Centre for
Engineering Innovation (CEI) using ODEON software. Four
acoustic parameters were considered in this study: T,
clarity (C50), early decay time and STI. The results
indicated that both T and early decay time were less than
the optimum values due to the excessive absorptive areas.
Thus, upgraded materials were suggested and explored
using ODEON software to enhance the acoustic quality of
the room. Hodgson (Hodgson, 2004) investigated the
acoustic performance of fourteen auditoria in the
4 S. Eldakdoky, A. Elkhateeb

Table 2 Description of Room RAZ.

Dimensions

L W H (avg.) Area m2 Volume m3 Perm. D Max. m

23.4 19.4 4.90 452.2 2266 95.3 21.20

Statistical Data

No. of Audience VPm3 Net Audience Area m2 AP m2 Topt@500 Hz Environmental Conditions

Net Gross RH % Temp. 1C
607 3.73 243 0.4 0.75 0.71 42% 26

University of British Columbia before and after renovation; 1. Field survey to specify the architectural
the values used in the evaluation of the room were features of the selected rooms.

estimated based on eld measurements using an MLSSA

system. The room verbal communication quality was 2. Analysis of rooms acoustic
quantied using the mean SI and STI. The results demon- performance
strated that the quality of verbal communication varied
2.1 Room shape analysis based
from good to poor due to a renewal effect that depended Validate the prediction
on the geometrical acoustics.
on the relationship between the variation in reverberation software by comparing
the TUcatt from the
and variation in the signal-to-noise level. Renovations simulation with the 2.2 Field measurements for
often rely on using sound absorption to manage reverbera- measured TMfor both TM& L in the unoccupied case.
unoccupied rooms.
tion, mainly at the rear of auditoria. To determine the
importance of early reection energy on SI inside class- 2.3 Prediction of TOcatt,
C50&STI for occupied rooms
rooms, Ellison and Germain (Ellison and Germain, 2013) using CATT software.
studied the effect of an active acoustics system on
controlling the early reection energy at the Sage Hill 2.4 Calculation of the optimal
School; this system consisted of microphones xed reverberation times (Topt)

throughout a room to capture the speaker's voice and send

the signals to processors that generate early energy
3. Analysis of the rooms performance by comparing the acoustic parameters
through loudspeakers. The results demonstrated that an resulting from the simulation with Topt and room shape analysis and the
active acoustic system could increase the strength and comparison of the measured L with ANSI recommended NC value
early reection benet while maintaining the reverbera-
tion times within acceptable limits. Elkhateeb (Elkhateeb,
4. Discussion of the results to specify the acoustic problems and to propose
2005a, 2005b) evaluated and analysed the acoustic envir- enhancement modifications for rooms interior design.
onment in nine lecture auditoria in Ain Shams University,
Cairo, Egypt. The work depended on eld measurements
for various acoustic indicatorsT, centre time (TS), C50, 5. Exploration of the proposed modifications using Catt software, and results

denition (D50)and Lin unoccupied rooms, except for one

room that was investigated twice, occupied and unoccu-
pied. The measured T in this room was used to validate the
methodology for calculating the occupied T based on the
unoccupied one. Then, the articulation loss of consonants
(Alcons) and occupied reverberation time were calculated
and ray diagram analysis was performed; these were
compared to their optimal values for all of the rooms
under consideration. The results showed that three of the
rooms had a good acoustic environment due to the suit-
ability of their nishing materials, two rooms needed
minor modications to be good lecture rooms and the
other four rooms had a poor acoustic environment due to
an excessive reverberation time; nally, some modica-
tions were suggested to mitigate those defects. Elkhateeb
(Elkhateeb, 2012) also acoustically designed the new
lecture room in the Faculty of Law at Ain Shams University,
using ODEON software. After construction, the unoccupied
room was evaluated by eld measurements for T, D50, STI,
and L. The measured indicators were compared to the
design objectives and to the values of the same indicators
Conclusion
Figure 4 The method.
previously predicted by ODEON during the design phase.
The results showed that the design successfully fullled
the recommended acoustic criteria, except for L, which
was relatively high due to the use of open windows for
natural ventilation.
1.2. Objective
The main purpose of this work is to assess the quality of
speech in two auditoria, room C and room RAZ, using eld
measurements and digital simulation. Based on the results
of the analysis, treatments to improve the acoustic quality
of the auditoria are suggested; the suitability of these
treatments are veried using digital simulation.
Acoustic improvement on two lecture auditoria
Figure 5 The existing nishing materials of room C and room RAZ.
2. Method
The method applied in this work to achieve the work
objective is summarised in Figure 4. For data collecting, a
comprehensive architectural survey for the rooms under
consideration was performed. The survey included the room
dimensions, capacity and nishing materials; see Figure 5
and Tables 1 and 2. The effect of the room shape on the
acoustic performance was rst investigated utilizing geo-
metrical acoustics, where the distribution pattern of the
acoustic rays was categorized into four zones, as shown in
Table 3 (Elkhateeb, 2005a, 2005b).
The quality of speech in the examined rooms was
evaluated by four acoustic indicators: T, L, C50 and STI. In
practice, eld measurements for the reverberation time
(TM) and L at the octave band centre frequencies were
carried out according to ISO 3382-2 (I. S. I. 3382-2, 2008)
and ISO 1996-2 (1996-2, International Standard ISO, 2007),
Figure 6. The measurement tools were Brel&Kjear instru-
ments, including, but not limited to, a sound analyser (type
2260), an Omni sound source (type 4296) and a power
amplier to drive the sound source. The height of the
microphone during the measurements was kept at 1.20 m
to simulate the height of a seated student's ears. A total of
24 reverberation times at a source height of 1.5 m were
recorded in unoccupied rooms utilizing alternating source
and receiver locations. The measurements of L were carried
out during class time, where all adjacent spaces were
5
occupied with normal activities. Three discrete measure-
ment positions were recorded in each room to cover the
signicant variation in noise emission. Since the windows
are designed to open for ventilation, the measurements of L
were performed with windows open. In addition, lighting
and electrical fans were operating normally. During the
measurements, the relative humidity and air temperature
were monitored and recorded using a thermo-hygrometer
device.
To estimate the effect of occupation and to calculate
both C50 and STI, CATT room acoustic software version 8.0b
was used. The absorption coefcients of the various sur-
faces used in the CATT calculations are shown in Table 4.
The properties and location of the sound source, as well as
the numbers and locations of the receivers, were simulated
based on those used in real measurements.
The different setups, calculation conditions and scatter-
ing coefcients used in the simulation are listed in
Tables 5 and 6. The results of the eld measurements were
used to validate the CATT model outputs. Figure 7 compares
the unoccupied reverberation time obtained using CATT
models (TUcatt) and that obtained from eld measurements
(TM). It can be seen that the difference between the TUcatt
and TM curves is less than 10%, which is the assumed error
due to differences in daily circumstances (Bradley, 2002).
After validation, the rooms, in simulation, were considered
to be occupied. The absorption coefcients () for the
model boundaries were kept constant with respect to the
6 S. Eldakdoky, A. Elkhateeb
Table 3 Zone patterns based on acoustical quality (Elkhateeb, 2005a, 2005b).
unoccupied case, except for the audience (students), for
which coefcients were taken according to Kuttruff
(Kuttruff, 2009); see Table 4. The block representing the
students was given the absorption coefcients of people
sitting on wooden chairs.
For comparison, optimal reverberation times (Topt) for both
auditoria at 500 Hz and above were calculated according to
the formula (2005a, , 2005b; William and Joseph, 1999):
V
T opt 0:3 log s 1
10
where V is the room volume in m3. The optimal reverberation
time at frequency bands less than 500 Hz (ToptOBCF) was
calculated by applying the formula (Elkhateeb, 2005a):
T optOBCF nTopt s 2 area not receiving any desirable early reections from the
where n is a ratio that can be calculated from (Elkhateeb,
2005a):
5:6716
n 3
F 0:2856
where F is the centre frequency of the band. The logarithmic
average of the background noise levels (Lave) for the mea-
sured L was calculated from (South, 2004):
1
Lave 10log L1 L2 dB 4 pied reverberation times calculated by CATT (TOcatt) and
x10 10 10Lx=10
10 10
where x is the number of measurements.
The results of the simulation were compared to the
optimal values, whereas Lave was compared to the ANSI
maximum acceptable level of noise (NC-35) for educational
spaces (Standard, ANSI/ASA S12.602010/Part1American
National, 2010). Based on the results, the acoustic defects
were specied. Accordingly, proposals to enhance and
improve the acoustic quality of the two rooms were
suggested. The effects of these proposals were explored
using CATT software.
3. Description of the acoustic performance in
the examined rooms
The acoustic performance of the rooms under consideration
as well as the architectural elements that negatively affect
their acoustic quality are detailed below.
3.1. Room C
Analysis of the room shape using geometrical acoustics, as
shown in Figure 8, showed an uneven distribution of early
and late reections due to the form of the walls and ceiling.
This irregularity results in the formation of many shadow
zones. No utter echoes were detected. Based on Table 3,
the room can be divided, generally, into two main zones:
medium and hard acoustic zones. The rst zone, medium,
encompasses the front audience area, which receives early
reections from the ceiling and uneven early and late
reections from the walls. In the second zone, hard, wide
strips of shadows appear due to the form of the ceiling in
addition to late reections from the rear walls. The shape of
the ceiling resulted in approximately 31% of the audience
ceiling.
As expected, due to the noisy environment surrounding
the faculty, the comparison between the measured spec-
trum of Lave and the maximum acceptable noise level for
learning rooms (Standard, ANSI/ASA S12.602010/
Part1American National, 2010), as shown in Figure 9(a),
illustrates that Lave (Nc-55) is higher than the acceptable
value (Nc-35) at all frequencies. The highest noise levels,
above 50 dB, occur at important frequencies for speech
intelligibility ( Z 1 kHz). The comparison between the occu-
Topt, as shown in Figure 9(b), indicates that there is a wide
gap between the two values at all frequencies. At 1 and
2 kHz, TOcatt is more than twice Topt and is almost doubled
at 500 Hz. The values of STI show that all of the receivers
lay in the poor zone, as shown in Figure 10(a), as expected
due to the long reverberation time and high noise level,
which reduce the S/N. The result agrees well with the
analysis of the room shape; the front receivers, Nos. 1 and
4, receive the strongest direct sound, 1storder early reec-
tions from the ceiling and 1st and 2ndorder late reections
from the rear walls. The others receive relatively abundant
1st and 2ndorder late reections from the rear walls in
addition to 1storder early reections, either from the side
walls (receivers Nos. 2, 3 and 5) or from the ceiling
(receiver 6). The long reverberation time and the excessive
reections result in a very low level of clarity, as shown in
Figure 10(b), where the average values of C50 are less than
0 dB at all frequencies except 500 Hz, which has a value of
0.3 dB, because of the total power absorption at this
frequency.
Acoustic improvement on two lecture auditoria 7

Figure 6 Omni source-microphone combinations in the two auditoria.

Table 4 The absorption Coefcients () for rooms surfaces and audience area (Kuttruff, 2009; William and Joseph, 1999).

Material Octave band centre frequency (Hz)

125 250 500 1000 2000 4000

Ceilings
Vinyl paint and cement plaster on expanded wire mesh ceiling 0.15 0.25 0.18 0.04 0.04 0.03
13 mm gypsum tiles over large air space 0.2 0.2 0.2 0.1 0.05 0.05

Walls and lining

Vinyl paint on cement plaster 0.02 0.02 0.02 0.02 0.02 0.02
10 mm Wood panelling 0.15 0.11 0.10 0.07 0.06 0.07
20 mm Marble cladding 0.01 0.01 0.01 0.01 0.02 0.02
3 mm Single pane of glass 0.08 0.04 0.03 0.03 0.02 0.02
Solid wooden door 0.14 0.10 0.06 0.08 0.10 0.10

Floors
Terrazzo tiles 0.02 0.02 0.02 0.04 0.05 0.05
Granite oor 0.05 0.05 0.05 0.05 0.05 0.05
Wooden oor on joists 0.15 0.11 0.10 0.07 0.06 0.07

audience area
Unoccupied wooden chairs 0.15 0.19 0.22 0.39 0.38 0.3
Audience on wooden chairs, 2/m2 0.24 0.4 0.78 0.98 0.96 0.87

Table 5 CATT setups and calculation conditions.

Parameter Value Parameter Value

Number of rays (recommended by 21514 for room C Background Logarithmic average of real measurements in each
Catt) 26772 for room noise room.
RAZ level
8 S. Eldakdoky, A. Elkhateeb

Table 6 Scattering coefcients (Elkhateeb, 2012; William and Joseph, 1999).

Audience Materials with rough surfaces Glass,

(e.g., gypsum boards, wood) marble

0.7 0.3 0.1

TM TUcatt TM TUcatt
2.5 2.5

Sound Pressure Level/dB

Sound Pressure Level/dB

2 2

1.5 1.5

1 1

0.5 0.5

0 0
125 250 500 1000 2000 4000 125 250 500 1000 2000 4000
Frequences/HZ Frequences/HZ
Room C Room RAZ
Figure 7 Comparison between TM and TUcatt for the unoccupied rooms.

Figure 8 Ray diagram analysis of room C.

90

80

70
Sound Pressure Level/dB

70
60 65 TOcatt Topt
60 2
50 55
Sound Pressure Level/dB

50
40 1.5
45
40
30
35 1
30
20
25
20 0.5
10
15
0 10 0
125 250 500 1000 2000 4000 125 250 500 1000 2000 4000
Frequency/Hz Frequences/HZ
Lave indicated on NC Comparison between TOcatt and Topt

Figure 9 Noise level and reverberation time in room C.

Acoustic improvement on two lecture auditoria 9

C50 ave.
0.8 3
0.7

Sound Pressure Level/dB

Good 2
0.6
Fair 1
0.5

STI
0.4 0
Poor
0.3 -1
0.2
0.1 -2
0 -3
R01 R02 R03 R04 R05 R06 125 250 500 1000 2000 4000
Audience Frequences/HZ
STI C50

Figure 10 STI and Clarity in room C.

Figure 11 Ray diagram analysis of room RAZ.

90

80

70
Sound Pressure Level/dB

70
60 65
60 TOcatt Topt
50 2
55
Sound Pressure Level/dB

50
40 45 1.5
40
30
35
1
30
20
25
20 0.5
10
15
0 10 0
125 250 500 1000 2000 4000 125 250 500 1000 2000 4000
Frequency/Hz Frequences/HZ
Lave indicated on NC Comparison between TOcatt and Topt

Figure 12 Noise level and reverberation time in room RAZ.

3.2. Room RAZ audience area and receives strong direct sound and early
sound reections from the ceiling and walls. In the second
The shape of room RAZ divides it into three main zones, as zone, medium, almost 19% of the audience receives early
shown in Figure 11: good, medium and hard acoustic zones. reections from the ceiling and late reections from distant
The rst zone, good, occupies approximately 25% of the walls. In the hard zone, 56% of the audience receives late
10 S. Eldakdoky, A. Elkhateeb

0.8 C50 ave.

3
0.7
Good

Sound Pressure Level/dB

0.6 2
Fair
0.5 1

STI
0.4 Poor 0
0.3
-1
0.2
0.1 -2
0 -3
R01 R02 R03 R04 R05 R06 125 250 500 1000 2000 4000
Audience Frequences/HZ
STI C50

Figure 13 STI and clarity in room RAZ.

Ocatt opt Ocatt opt

400 550
350 500
Absorption- Sabine

Absorption- Sabine
450
300 400
350
250
300
200 250
200
150
150
100 100
125 250 500 1000 2000 4000 125 250 500 1000 2000 4000
Frequences/HZ Frequences/HZ
room C room RAZ

Figure 14 Comparison between total room absorption for Topt and TOcatt.

reections from the rear walls and an uneven distribution of as expected due to the long T and abundant late reections
early reections and many shadows from the ceiling. Again, at the distant receivers.
the inappropriate shape of the ceiling resulted in approxi-
mately 22% of the audience area not receiving the desirable
1storder early reections from the ceiling. 3.3. Results and discussion
Similar to room C, the spectrum of Lave (Nc-55) is higher
than the maximum acceptable level (Nc-35) at all frequencies, The acoustic analysis of the two rooms showed that neither
as shown in Figure 12(a). The highest noise levels are at fulls the recommended criteria for learning spaces. Hence,
important frequencies for speech intelligibility. A comparison the two auditoria are noisy and suffer from poor speech
between TOcatt and Topt, as shown in Figure 12(b), showed that intelligibility. Although L is not more important than the
TOcatt exceeds Topt at all frequencies; TOcatt is higher than Topt acoustic design of a lecture room to achieve a satisfactory
by factors of 1.6 and 1.7 at 1 and 2 kHz, respectively, and by a S/N ratio (Bradley and Sato, 2003), the high noise levels in
factor of 1.4 at 125 Hz. STI values showed that other than the auditoria, 20 dB above the maximum acceptable levels,
receivers Nos. 1 and 4, the receivers nearest to the sound severely reduce the speech intelligibility. This noise mainly
source, all of the receivers lay in the poor zone, as shown in comes from the natural ventilation via the open windows
Figure 13(a). This is expected as T and L signicantly exceed that overlook a crowded, noisy street and yards for student
the acceptable limits. The result is also compatible with the assembly. The excessive TOcatt results from the absence of
room shape analysis, in which receivers Nos. 1 and 4 receive appropriate absorbing materials, although the volume per
the strongest direct eld, 1storder early reections from the person (VP) in both rooms is compatible with DIN-18041
ceiling and walls in addition to weak late reections from rear (3:6 m3/seat) (Cox and DAntonio, 2009), Tables 1 and 2.
surfaces. Receivers Nos. 2 and 5 receive 1storder early The volume per person in room C is higher than VP in room
reections, 1stand 2ndorder late reections from the rear RAZ by a factor of 1.2, which explains why room C is
walls and no early reections from the ceiling. Receivers Nos. acoustically worse. Moreover, the distribution and form
3 and 6 receive abundant late reections from the surrounding of the reective surfaces, especially the ceilings, are
walls in addition to 1storder early reections from the ceiling. inefcient, as 31% and 25% of the audience area in Room
The values of the C50 spectrum, as shown in Figure 13(b), C and RAZ, respectively, do not receive any early reections
varied between 4 dB and  2.3 dB. The negative values of from the ceiling (or walls, in room RAZ). These defects
C50 are at receivers in the hard acoustic zone; this result is resulted in negative values of clarity and poor STI levels.
Acoustic improvement on two lecture auditoria
Figure 15 Proposed treatments for room C.
Because early reections and direct elds are incorporated
together and perceived as one sound that is louder than the
direct sound alone by 10 dB or greater (Reich and Bradley,
1998), the internal surfaces of the auditoria must be
modied to minimize the late reections and excessive
reverberation time, maximize the desirable early reection
energy and eliminate the shadow zones. The total power of
absorption in room C should be almost doubled, Figure 14
(a). For room RAZ, the total power of absorption should be
increased over its recent values by a factor of 1.6 at low
frequencies and 1.7 at mid and high frequencies, Figure 14
(b). The existing ceilings in both auditoria must be modied
11
to regularly distribute the early reections throughout
the rooms.
4. Proposals for improvement
The results of the analysis show that reducing reverberation
by adding more absorptive materials and eliminating the
shadow zones by modifying the ceiling form will not be
sufcient to improve intelligibility as long as L is very high.
The classical approach to reducing L is to close the open
windows, which will improve the room insulation. Such a
12 S. Eldakdoky, A. Elkhateeb

Table 7 The absorption Coefcients () for proposed treatments (Kuttruff, 2009; William and Joseph, 1999).

Material Octave band centre frequency (Hz)

125 250 500 1000 2000 4000

Ceilings
Reective, painted steel tiles 0.1 0.2 0.2 0.1 0.1 0.2
mineral tiles 0.3 0.35 0.4 0.55 0.8 0.7

Walls and lining

100 mm of mineral wool and 3 mm of perforated plywood 0.7 0.8 0.75 0.6 0.5 0.4
50 mm thick wood-wool slabs on a solid backing 0.2 0.2 0.6 0.8 0.7 0.9
50 mm mineral wool (40 kg/m3), glued to wall, surface sprayed with thin 0.15 0.7 0.6 0.6 0.75 0.75
plastic solution
20 mm Marble cladding 0.01 0.01 0.01 0.01 0.02 0.02

Floors
wooden parquet 0.04 0.04 0.07 0.06 0.06 0.07

solution imposes the use of an air conditioning system
instead of natural ventilation, which is currently impracti-
cal. Therefore, different treatments were proposed and
CATT software was used to corroborate the selection of the
best proposal for each room as the following.
4.1. Room C
Three different proposals were suggested from which the
most feasible proposal was selected.
 The rst proposal includes changing the form of the
ceiling as well as the nishing materials for all of the
internal surfaces except the front wall and oor, see
Figure 15(a) and Table 7; the total absorption areas are
338 m2. The new materials include: i) covering all walls
with 100 mm of mineral wool and 3 mm of perforated
plywood (Kuttruff, 2009)and ii) using reective, painted
steel tiles for the front half of the new ceiling and
mineral tiles for the rear part (William and Joseph,
1999). The proposal also includes adding soundproof
windows over the existing single pane windows to raise
the sound transmission class (STC) from 28 to 54 to
reduce L by 20 dB (Safelex-Acoustical Guide, 2015).
 The second proposal is similar to the rst proposal but
without any changes to the existing ceiling, as shown in
Figure 15(b). The total absorption areas are 252 m2.
 The third proposal includes treating the back walls and
half of the side walls with 50 mm thick wood-wool slabs
on a solid backing, as shown in Figure 15(c) and Table 7,
as well as treating the plinths of other walls with 100 mm
of mineral wool covered by 3 mm of perforated plywood.
In addition, changes to the form of the ceiling and
treatment of the windows are the same as the rst
proposal. The total absorption areas are 269 m2.
The results of the three proposals are compared with
each other and with the results for the room parameters
before treatments, as shown in Figure 16. The comparison
illustrates that although the third proposal covers less area
by absorptive materials in comparison to the rst proposal,
it yields the best performance, as it is much closer to Topt at
all frequencies, Figure 16(a). This result is justied because
the total absorption for the new materials in the third
proposal is higher than the total absorption for the new
materials in the other proposals. The new shape of the
ceiling decreases the room volume by approximately 18%;
nevertheless, this does not reduce VP to less than the
recommended limits; it also enhances the acoustic quality
within the room by eliminating the shadows. For the above-
mentioned reason, the second proposal is the least effec-
tive. Due to the reduced L in the three proposals, a
noticeable improvement in the intelligibility of speech was
achieved at all receivers that are now located in the good
zone, as shown in Figure 16(b). Nevertheless, the third
proposal again achieves the highest STI due to its optimal
reverberation time. Clarity in the three proposals increased
by more than 2 dB at all frequencies, as shown in Figure 16
(c). Though the third proposal yields the highest clarity,
approximately 5 dB at mid and high frequencies, due to
proportional correlation between C50 and T.
4.2. Room RAZ
The following proposals were suggested to select the most
feasible proposal.
 The rst proposal includes installing 100 mm of mineral
wool covered with 3 mm of perforated plywood on the
back and rear two-thirds of the side walls, see Figure 17
(a) and Table 7; using marble cladding for the front wall
and the other third of the side walls; restructuring the
front third of the ceiling as reectors; and treating the
windows similarly to the proposals for room C. The total
absorption areas are 579 m2.
Acoustic improvement on two lecture auditoria 13

TOcatt Topt TP01 TP02 TP03

1.8
1.6

Sound Pressure Level/dB

1.4
1.2
1
0.8
0.6
0.4
0.2
0
125 250 500 1000 2000 4000
Frequences/HZ
Comparison between the values of T for the current situation and
different proposals

Ex P01 P02 P03

0.8
0.7 Good
0.6
Fair
0.5
STI

0.4 Poor
0.3
0.2
0.1
0
R01 R02 R03 R04 R05 R06
Receivers
Comparison between the values of STI for the current situation and
different proposals

C50 ave. P01 P02 P03

7
6
Sound Pressure Level/dB

5
4
3
2
1
0
-1
-2
-3
125 250 500 1000 2000 4000
Frequences/HZ
Comparison between the values of C50for the current situation and
different proposals

Figure 16 Effect of the different proposals on T, STI and clarity in room C.

 The second proposal includes xing wooden parquet to
the ooring, as shown in Figure 17(b) and Table 7.
Moreover, the windows, the ceiling shape and the back
wall are treated as in the rst proposal; the total
absorption areas are 641 m2.
The two proposals noticeably improve the acoustic qual-
ity of the room; the rst proposal yields better results
despite the fact that it contains a smaller absorptive area by
10% compared to the second proposal. This result is justied
for the same reasons previously mentioned in room C. In this
proposal, the resulting reverberation time is too close to
Topt at all frequencies, as shown in Figure 18(a). Due to the
reduction of both L and T, elimination of shadows by
reshaping the ceiling and avoidance of late reections from
the rear walls, all of the receivers are now located in a good
zone, as shown in Figure 18(b). The clarity rise above 3 dB
at mid and high frequencies, as shown in Figure 18(c),
reects the enhancement of the early reection energy and
S/N accordingly.
5. Conclusions
The results of the studies performed in this research clarify
that selecting the efcient absorptive materials for the rear
part of ceiling and distance walls achieves the required
total power of absorption, which successfully minimizes the
14
Figure 17 Proposed treatments for room RAZ.
late reections and excessive reverberation time to the
optimal value. Moreover, utilizing proper ceiling shape
eliminates the shadow zones and increases the desirable
early reection energy throughout the auditoria. Although
this corresponding to the optimal reverberation time and
even distribution of early reections, the negative values of
clarity and poor STI cannot be avoided without decreasing
the high background noise.
Based on an architectural survey, most surfaces in both
rooms are reective, which can cause an excessive rever-
berant sound. Shape analysis of the two rooms showed
that there are many shadow zones that cover more than
20% of the audience area as a result of the ceiling shape.
These shadows, in addition to the uneven distribution of
the early and late reections, divide the audience area in
both rooms into medium and hard acoustic zones. The
results show that TOcatt highly exceeds Topt at all frequen-
cies; for example, TOcatt at 1 and 2 kHz is higher than Topt
by more than a factor of 2 in room C and 1.5 in room RAZ.
It is obvious that they are highly reverberant rooms, and
room C is the worst because VP in room C is larger by 21%,
although the capacity of room C is smaller than that of
room RAZ by 40%. Furthermore, the measured spectra of
Lave (Nc-55) in the two rooms are higher than the
maximum acceptable noise level (Nc-35) by approximately
20 dB in the mid and high frequency ranges. This defect is
S. Eldakdoky, A. Elkhateeb
mainly due to the intrusion of outdoor noise through the
permanently open windows. Thus, good sound insulation
for these windows is vital for any acoustic enhancement.
In this case, natural ventilation in both rooms should be
accordingly switched to a mechanical ventilation system.
As a result of the excessive reverberation and the high
noise levels in both rooms, the values of STI and C50 are
dramatically low, excluding the receivers close to the
sound source. In conclusion, the two rooms suffer from low
speech intelligibility.
The analysis shows that room C requires approximately
1.9 times its current total power of absorption at low
frequencies or 2.2 times at mid and high frequencies,
whereas room RAZ requires approximately 1.6 times its
current total power of absorption at low frequencies and
1.7 times at mid and high frequencies. To improve the
acoustics within the two rooms, many scenarios have been
examined. Using large area of absorptive materials to achieve
the optimal reverberations time may reduce intelligibility due
to reduced benecial early reections. Thus, the main
approach is to achieve the required total power of absorption
using the less area of high absorptive materials at the
important frequencies for SI. The proposals also include
improving the STC of the windows and reforming the shape
of the ceiling, either partially or completely, to remove
shadow zones. The analysis of the suggested modications
Acoustic improvement on two lecture auditoria 15

TOcatt Topt TP01 TP02

1.6
1.4

Sound Pressure Level/dB

1.2
1
0.8
0.6
0.4
0.2
0
125 250 500 1000 2000 4000
Frequences/HZ
Comparison between the values of T for the current situation and
different proposals

Ex P01 P02
0.8

0.7
Good
0.6
Fair
0.5
STI

0.4
Poor
0.3

0.2

0.1

0
R01 R02 R03 R04 R05 R06
Receivers
Comparison between the values of STI for the current situation and
different proposals

C50 ave. P01 P02

7
6
5
Sound Pressure Level/dB

4
3
2
1
0
-1
-2
-3
125 250 500 1000 2000 4000
Frequences/HZ
Comparison between the values of C50for the current situation and
different proposals

Figure 18 Effect of the different proposals on T, STI and clarity in room RAZ.

using CATT software indicates that TOcatt in both rooms
becomes much closer to Topt when covering specic areas
with two different highly absorptive materials. The shape of
the new ceiling was designed to eliminate the shadow zones
and increase the early reections as well. In conclusion, the
speech intelligibility at all receivers, as can be concluded
from the new values of STI (good) and C50 (approximately
5 dB at mid and high frequency ranges).
Acknowledgments
The authors would like to thank Prof. Dr. Ahmed Abdin, Prof.
of Architecture and Environmental Studies, Faculty of
Engineering Cairo University, Prof. Dr. Randa Kamel, Prof.
of Architecture, Faculty of Engineering- Helwan University
and Prof. Dr. Mahmoud Taha Prof. of Urban Design, Faculty
of Engineering- Helwan University, for their collaboration.
References
Bradley, J., 2002. Optimizing Sound Quality for Classrooms.
SOBRAC, Rio de Janeiro.
Bradley, J.S., Sato, H., 2003. On the importance of early reections
for speech in rooms. J. Acoust. Soc. Am. 113 (6), 32333244.
16
Bradley, J.S., Sato, H., 2008. The intelligibility of speech in
elementary school classrooms. J. Acoust. Soc. Am. 123 (4),
20782086. http://dx.doi.org/10.1121/1.2839285.
Brunskog, J., Gade, A.C., Paya Bellester, G., Reig Calbo, L., 2009.
Speaker comfort and increase of voice level in lecture rooms. J.
Acoust. Soc. Am. 125 (4), 20722082.
Cox, T.J., DAntonio, P., 2009. Acoustic Absorbers and Diffusers
Theory Design and Application 2nd ed. Spon Press.
DiMarino, C., Fuerth, D., Gignac, D., Lunardi, A., Novak, C., Pikul, R.,
Simone, A., 2011. Acoustic enhancement of proposed grand
lecture hall using computer simulation. Can. Acoust. 39 (1), 4348.
Elkhateeb, A., 2003. Architectural Acoustics Theory and Application
1st ed. The Anglo Egyptian Library, Cairo.
Elkhateeb, A., 2005a. Sounds In Lecture Rooms In Ain Shams
University/ Acoustical Environment Assessment For The Main
Auditoria, Problem Analysis and Solution Discussion- Part One:
Faculty of Engineering. In: Proceedings of the 1st Ain Shams
University international Conference on Environmental Engineer-
ing, ASCEE-1, Cairo.
Elkhateeb, A., 2005b. Sounds In Lecture Rooms In Ain Shams
University/ Acoustical Environment Assessment For The Main
Auditoria/Problem Analysis and Solution Discussion- Part Two:
Faculties Of Arts, Law and Commerce. Ain Shams Eng. J. (ASEJ)
40 (4), 505541.
Elkhateeb, A., 2012. The acoustical design of the new lecture
auditorium, Faculty of Law/ Ain Shams University. Ain Shams
Eng. J. (ASEJ) 3 (2), 219235.
S. Eldakdoky, A. Elkhateeb
Ellison, S., Germain, P., 2013. Optimizing Acoustics for Spoken Word
Using Active Acoustics, 19. Acoustical Society of America, Canada.
Hodgson, M., 2004. Case-study evaluations of the acoustical designs
of renovated university classrooms. Appl. Acoust. 65 (1), 6989.
http://dx.doi.org/10.1016/S0003-682X(03)00108-7.
I. S. I. 3382-2, 2008. Measurement of room acoustic parameters Part
2: reverberation time in ordinary rooms. ISO, Switzerland.
Kuttruff, H., 2009. Room Acoustics 5th ed. Spon Press.
Reich, R., Bradley, J., 1998. Optimizing classroom acoustics using
computer model studies. Can. Acoust. 26 (4), 1521.
Safelex-Acoustical Guide, 2015. Saexs. [Online].
Shams, S., Rama krishnan, R., 2012. Acoustic metrics for classroom
performance - a literature review. Can. Acoust. 40 (3), 100102.
South, T., 2004. Managing Noise and Vibration at Work/A Practical
Guide to Assessment, Measurement and Control. Elsevier
Butterworth-Heinemann, UK.
Standard, ANSI/ASA S12.60-2010/Part1American National, 2010.
Acoustical Performance Criteria, Design Requirements and
Guidelines for Schools, Part1: Permanent Schools. Acoustical
Society of America, USA.
William, J.C., Joseph, A.W., 1999. Architectural Acoustics: Princi-
ples and Practice. John Wiley & Sons, Inc, USA.
1996-2, International Standard ISO, 2007. Acoustics-Description,
measurement and assessment of environmental noise, Part 2:
Determination of Environmental Noise Levels. ISO, Switzerland.
18041:2004-05, DIN, 2004. Acoustical Quality in Small to Medium-
sized Rooms. Beuth Verlag GmbH, Germany (05).