CIPA 2003 XIX th International Symposium, 30 September-04 October, 2003, Antalya, Turkey 
Figure 6. The measured values of the parameter strength G. 
The deviation bars of one standard deviation are included. 
600 
550 
500 
450 
400 
350 
300 
250 
200 
150 
100 
/ 
/ 
. 
* / 
/ 
X/ 
V 
, Mosques 
i Churches 
Mosques: 
y = 118.27Ln(x) - 943.27 
R 2 = 0.9224 
Churches: 
y = 113.63Ln(x) - 832.49 
R 2 = 0.9452 
10000 100000 
V [m3] 
600 
550 
500 
450 
400 
£ 350 
- 300 
w 250 
200 
150 
100 
50 
/ 
/ 
/ 
> Mosques 
i Churches 
Mosques: 
y = 75.947Ln(x) - 510.72 
R 2 = 0.9435 
Churches: 
y = 114.51 Ln(x) - 835.44 
R 2 = 0.904 
10000 100000 
V [m3] 
Figure 7. The centre of gravity of energy. Top: all pass values; Bottom: 
averaged values of 500Hz, 1kHz and 2kHz octave bands. 
Also the so-called center of gravity of energy was measured 
(Ts) and this parameter is selected for being reported in this 
section. Actually Ts is highly correlated with the former indexes 
C or D but it is not as scattered from position to position as C 
and D are. 
In Fig. 7 one can see that there is a sort of bias in the parameter 
between mosques and byzantine churches. The slope of the two 
lines is quite similar so that the values of a church are 
reproduced by a mosque of larger volume. In general the Ts 
values are higher for greater volumes and inappropiate for the 
comprehension of speech. In any case the sound field inside 
mosques is more suited for speech showing lower values. The 
graph can also be interpreted in terms of useful listening areas. 
That is, both in mosques and churches the positions close to the 
sound source are remarkably better than the others, where also 
the parameter Ts is out of range for both speech and music. In 
the case of mosques the good listening area is usually wider that 
in a byzantine church of comparable volume. 
4. THE ACOUSTICAL PERFORMANCE OF THE 
FLOOR 
Many features of the sound field inside mosques are governed 
by the sound absorbing characteristics of the floor. This surface 
can be found in different arrangements from mosque to mosque. 
In particular it can be set by simply laying carpets on the floor 
or by building a wooden structure of some centimeters height, 
where the carpets are rested. Moreover this wooden structure 
defines an air-backing of a few centimeters with respect to the 
floor, which can be either left void or filled with material for 
thermal insulation. 
In order to study the acoustical properties of mosque floor 
layouts a simple project was prepared. Then the mosque floor 
was built as a modular structure summing up to a sample of 
12m 2 . The wooden structure consists of 25mm thick panels and, 
in order to model the air-gap underneath the structure, two 
heights were chosen: 40mm and 150mm. 
The mosque floor model was then taken to a reverberation 
chamber and measurements of acoustical absorption were made 
(Prodi, Marsilio, Pompoli, 2001). Four configurations were 
tested, namely: carpet alone, structure (40mm) alone, structure 
(40mm) with carpet, structure (150mm) alone. The 
measurements consisted in the placing of the dodecaedric sound 
source in three positions in the chamber and sampling the sound 
field in six positions each time with omnidirectional 
microphones. Impulse responses were calculated and later 
parameters regarding reverberation time were extracted. The 
measurements followed the specific standard and the alfa 
absorption coefficients in the diffuse sound field were obtained. 
From the results of these measurements reported in Fig. 8 it can 
be seen that the behaviour of the floor is a combination of the 
absorption of the carpet and of that of the wooden panel 
structure. 
100 160 250 400 630 1000 1600 2500 4000 
Freq. [Hz] 
Figure 8. The absorption coefficient for each of the four tested 
configurations. 
In particular the carpet works in the medium and higher 
frequency range whereas in the lower range one can see the 
typical absorption of a panel. As a confirmation of this, the 
effect of the air backing disappears when the carpet is placed 
alone. Moreover, the height of the structure sets the frequency 
showing the maximum sound absorption in the lower range so 
that for the higher structure the resonant frequency is shifted 
downwards.