CIPA 2003 XIX th International Symposium, 30 September - 04 October, 2003, Antalya, Turkey 
Figure 2. The relationship between the measured reverberation times 
and the volumes of mosques and churches:all pass values. 
Figure 4. The dependence of reverberation time with frequency for the 
byzantine churches. 
V [m3) 
Figure 3. The relationship between the measured reverberation times 
and the volumes of mosques and churches: averaged values of 500Hz, 
1kHz and 2kHz octave bands. Dashed lines indicate the occupied and 
unoccupied concert halls. 
In fact while the mosques show the maximum of reverberation 
in the lower frequency range the byzantine churches have 
generally the maximum in the medium frequrency range. This 
finding has an immediate link with the impression of a listener. 
The acosutic field in byzantine churches is extremely 
reverberant and the sounds overlap in time with little possibility 
of being separated. This applies also to spoken of sung phrases, 
whose intelligibility is greatly affected by the excess of 
reverberation. Such conditions of reverbeartion are also at the 
upper linit of acceptability for typical lyturgical music played 
with an organ. As regards the mosques the reverberation curves 
are easily grouped into two sub-groups of “large” spaces like 
Suleymanie and Selymie and the “small” ones like Sokullu and 
St. Sergius and St. Bacchus. For the former group the 
reverbeation time halves when passing from 500 Hz to 2 kHz 
causing a peculiar timbric effect in the spaces: the same 
phenomenon is observed also in the latter group but it is not so 
marked. In particular the frequency range of the singing 
formants (2 - 3 kHz) has a shorter reverberation time than the 
lower frequencies. This fact contributes in improving the oral 
unaided communication in the mosques since the masking due 
to the reverberation is controlled by the sound absorption in the 
higher frequency range. 
3.2 The sound level 
The sound level is investigated by means of the parameter 
called strentgth (G) reported in Fig. 6. The plot includes the 
overall values measured inside each space after averaging of all 
source and receiver positions. 
Lin 125 250 500 1K 2K 4K 
[Hz] 
Figure 5. The dependence of reverberation time with frequency for the 
mosques. 
Each point has the bars related to one standard deviation of the 
mean values. The first and clearest observation is that the sound 
level both in byzantine churches and in mosques has a 
decreasing trend with increasing of the volume. The logarithmic 
fit lines have quite a good regression coefficient. 
The values are good just for the smaller mosques and allow in 
general only a fair degree of communication over a small 
distance without the aid of electroacoustic devices. Then the 
sound level inside the byzantine churches is higher than that in 
the mosques with a comparable volume. This is true on average 
but the values in the different positions are not constant, and in 
both cases they decrease when moving away from the sound 
source. This means that there is a sort of overlap between the 
data as also shown in Fig. 6. Typically the lower levels for a 
smaller mosque (or church) can be compared with the higher 
levels measured inside a large one. In other words what happens 
is that, given a mosque and a byzantine church of comparable 
volume, the sound level measured at the same distance from the 
sound source will differ by a fairly constant amount depending 
on the volume. The churches show a higher sound level for 
most of the volumes tested. For larger volumes the difference 
tend to vanish since in both types of enclosures the shielding 
effects of architectural elements and also the air absorption 
become quite relevant. 
3.3 The clarity 
The attributes of the sound field concerning clarity and the 
definition of the message delivered in the spaces for worship 
have been measured according to suitable parameters like 
clarity (C50 and C80) and definition (D).