BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 - pages 29 – 47 29
Chu ch Acous ics and he In luence o Occupancy
Vic o Desa naulds1, An ónio P.O. Ca alho2 and Gilbe Monay3
1Ecole Poly echnique Fédé ale de Lausanne (EPFL), CH-1015 Lausanne,
Swi ze land ([email p o ec ed]);
2Labo a o y o Acous ics, Facul y o Enginee ing, Uni e si y o Po o, R. D . Robe o
F ias, P-4200-465 Po o, Po ugal ([email p o ec ed]);
3Bu eau G. Monay, 25 Vine , CH-1004 Lausanne, Swi ze land ([email p o ec ed])
(Recei ed 28 May 2001 and accep ed in e ised o m in 10 Oc obe 2001)
ABSTRACT
Speech in elligibili y is usually e alua ed, by he use o objec i e pa ame e s, in unoccupied
ooms due o p ac ical conside a ions. Howe e , unde no mal condi ions, he oom
occupancy can inc ease o dec ease he alues o speech in elligibili y by he e ec o he
addi ional sound abso p ion p esen in he oom o by he change in S/N a io. Measu emen s
we e ca ied ou in six chu ches wi h and wi hou occupancy. The esul s show ha
occupancy induces a mean inc ease o he speech in elligibili y o ∆STI (STIoccup-
STIunoccup) o 0.050 wi h he use o a public add ess sys em (PA) and 0.035 wi hou a PA.
This inc ease is caused mainly by he educ ion o he oom e e be a ion ime, due o he
inc eased sound abso p ion and by he educ ion o he addi ional RT induced by he PA
(using loudspeake s wi h s ong di ec i i y o ien ed owa ds he cong ega ion a ea). The
a ia ions in he speech in elligibili y alues wi h occupancy can be p edic ed in chu ches by
employing new empi ical o mula ions. The dec ease in S/N a io (due o he backg ound
sound le el augmen a ion by he p esence o people and by sound le el dec ease wi h dis ance
in he cong ega ion a ea), o he poo di ec i i y and o ien a ion o he sound sou ces, could
educe he speech in elligibili y gain in pa icula cases. O he e ec s induced by he
cong ega ion, such as he a ia ions o humidi y and empe a u e, gene ally ha e negligible
e ec on speech in elligibili y.
1 INTRODUCTION
Fo easons o con enience, he measu emen o speech in elligibili y using objec i e
pa ame e s akes place gene ally in unoccupied ooms. Unde no mal condi ions o
use, he oom occupa ion can inc ease o dec ease he speech in elligibili y. The
a ailable bibliog aphy conce ning he acous ic e ec o occupancy, ela es p ima ily
o he sound abso p ion o he audience ( o calcula e he e e be a ion ime in
occupied oom see e e ences1-3) o on he a enua ion wi hin he public4,5 (in
pa icula he 125-350 Hz "sea dip a enua ion" 6-8).
This pape conce ns a s udy ha aims a de e mining he e ec o occupa ion on
30 Chu ch Acous ics and he In luence o Occupancy
speech in elligibili y based on compa isons made wi h one speech in elligibili y
objec i e pa ame e measu ed in unoccupied and occupied chu ches. The analysis o
he a ious e ec s induced by he p esence o he cong ega ion on he speech
in elligibili y will esul in a p oposed a model o de e mine he a ia ions o
objec i e speech in elligibili y acco ding o he chu ch occupancy.
2 THE VARIOUS EFFECTS OF THE OCCUPANCY ON SPEECH
INTELLIGIBILITY
Speech in elligibili y is in luenced by many pa ame e s easonably well accoun able
in e ms o he objec i e desc ip o s9. Among hese pa ame e s, he mos signi ican
a e he signal- o-noise a io (S/N), ha ep esen s he eme gence o he speech le el
o e he backg ound noise, he e e be a ion ime (RT) and he p esence o dis u bing
la e e lec ions.
Ce ain e ec s ela ed o he oom occupancy can inc ease he speech
in elligibili y alues, such as:
1. Reduc ion in he e e be a ion ime induced by he addi ional sound abso p ion
b ough by he audience;
2. Reduc ion in he e e be a ion induced by he sound sys em. In he case o an
adap ed public add ess sys em (loudspeake s wi h high di ec i i y di ec ed
owa ds he cong ega ion a ea), he public abso bs a la ge pa o he ene gy
di used by he loudspeake s, whe eas his is sen o all he emp y oom olume
when he chu ch is unoccupied;
3. Reduc ion in he phenomena o ocusing. The public by i s sound abso p ion
can emo e a possible lu e echo be ween he loo and he ceiling (especially
i his is a ched);
4. In win e , he hea ing o he chu ch du ing chu ch se ices (changing he
empe a u e and humidi y) induces a educ ion in he e e be a ion ime alues
(by inc easing he speed o sound p opaga ion and he ai abso p ion, a low
and medium equencies) as well as a bending downwa d o he sound ays
(posi i e a ia ion in empe a u e wi h heigh ).
O he e ec s, on he con a y, can dec ease he speech in elligibili y, such as:
1. Inc ease in he backg ound noise le el ela ed o he p esence o he
cong ega ion (sea ing noise, alks, coughing, e c.);
2. The weakening o he sound le el wi h dis ance wi hin he cong ega ion due o
he sound abso p ion and he e ec o shielding.
3 MEASUREMENT
3.1 Me hod and da a desc ip ion
In o de o cause he leas possible dis u bance o he cong ega ion du ing se ices
and o allow he collec ion o a maximum o in o ma ion, a measu ing sys em o ype
MLS (MLSSA) was used wi h asynch onous da a p ocessing (acquisi ion wi h wo
eco de s DAT Casio DA-7), excep o Opo o's S John Chu ch (measu emen
wi h a B&K 3361 sys em10). The da a p esen ed in his s udy ela e mainly o he
BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 31
speech in elligibili y exp essed by he STI (Speech T ansmission Index9) as well as
he e e be a ion ime (RT). Reco ding o he se ices acili a ed analysis o he
sound le els (in dBA, dBLin and oc a e bands om 125 o 4000 Hz), he backg ound
noise and he a ious spoken in e en ions du ing he p oceedings.
The ield measu emen s ook place in six chu ches (unoccupied and occupied)
whose main cha ac e is ics a e gi en in Table 1.
Table 1. Main cha ac e is ics o he chu ches sampled
Chu ch (symbol) Place, coun y Denomina ion Volume
(m3)
Occupa ion
(pe sons)
Sac é-Cœu (LC) La-Chaux-de-Fonds, Swi ze land Ca holic 9137 183
Pasqua (P) Bienne, Swi ze land P o es an 4472 114
Fille-Dieu (FD) Romon , Swi ze land Ca holic 5600 124
Cheseaux (C) Cheseaux, Swi ze land P o es an 575 53
Romanel (R) Romanel, Swi ze land P o es an 477 28
S . John (J) Po o, Po ugal Ca holic 6048 200
Mean 4385 117
S anda d de . 3365 68
3.2 Resul s
Tables 2 and 3 p esen he esul s o he measu ed speech in elligibili y index and
he e e be a ion ime alues (RT).
Table 2. STI esul s o he chu ches sampled
(PA = Public Add ess sys em, DI = Di ec i i y Index)
Chu ch Condi ion STI occup. STI unocc. ∆STI
= STIoccup - STIunoccup
S anda d
de ia ion
Sac é-cœu Wi hou PA 0.278 0.242 0.036 0.035
Wi h PA 1 (high DI) 0.472 0.410 0.062 0.043
Wi h PA 2 (low DI) 0.405 0.332 0.073 0.049
Pasqua Wi hou PA 0.523 0.495 0.028 0.013
Wi h PA 0.576 0.558 0.018 0.029
Fille-Dieu Wi hou PA 0.313 0.290 0.023 0.015
Cheseaux Wi hou PA 0.515 0.449 0.066 0.019
Romanel Wi hou PA 0.615 0.577 0.038 0.012
S . John Wi hou PA 0.435 0.415 0.020 0.045
Wi h PA 0.555 0.508 0.048 0.029
Mean alues Wi hou PA 0.447 0.411 0.035 0.014
Wi h PA 0.534 0.492 0.050 0.024
32 Chu ch Acous ics and he In luence o Occupancy
Table 3. Re e be a ion ime esul s o he chu ches sampled
F equency (Hz)
Chu ch Si ua ion 125 250 500 1000 2000 4000
Sac é-cœu Occupied - 3.93 4.66 4.13 3.54 2.75
Unoccupied
3.26 5.88 6.83 6.47 5.91 4.19
di e ence - 1.95 2.17 2.33 2.37 1.43
Pasqua Occupied 1.58 2.14 2.32 2.25 2.24 1.71
Unoccupied
1.74 2.43 2.74 2.77 2.53 1.93
di e ence 0.16 0.29 0.42 0.52 0.29 0.22
Fille-Dieu Occupied 3.00 3.80 3.50 2.80 2.60 2.10
Unoccupied
3.15 4.40 4.65 4.20 2.80 2.60
di e ence 0.15 0.60 1.15 1.40 0.20 0.50
Cheseaux Occupied 1.04 1.05 1.16 1.13 1.08 1.05
Unoccupied
1.17 1.51 1.75 1.84 1.86 1.60
di e ence 0.13 0.46 0.59 0.71 0.78 0.55
Romanel Occupied 0.69 0.83 0.89 0.87 0.86 0.75
Unoccupied
1.17 1.51 1.75 1.84 1.86 1.60
di e ence 0.48 0.68 0.86 0.97 1.00 0.85
S . John Occupied 2.86 2.45 2.43 2.18 1.92 1.70
Unoccupied
3.09 3.32 3.31 3.22 2.73 2.42
di e ence 0.23 0.87 0.88 1.04 0.81 0.72
Mean alues Occupied 2.1 2.3 2.5 2.2 2.0 1.7
Unoccupied
2.2 3.1 3.4 3.3 2.9 2.3
di e ence 0.2 0.8 1.0 1.2 0.9 0.7
Figu e 1. Va ia ions in he STI mean alues in chu ches (wi h and wi hou a
public add ess sys em, PA).
dSTI (wi h PA) = 0.276Ln(RTunoc/RTocc) - 0.052
R2 = 0.95
dSTI (no PA) = 0.102Ln(RTunoc/RTocc)
R2 = 0.55
0,00
0,02
0,04
0,06
0,08
0,10
0,12
1,2 1,3 1,4 1,5 1,6 1,7 1,8
RTunoccupied (2 kHz)/ R occupied(2 kHz)
dSTI = STI occ. - STI unocc.
wi h PA
no PA
wi h PA
no PA
BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 33
4 ANALYSIS
4.1 Speech T ansmission Index and Re e be a ion Time
The
∆
STI esul s shown in Table 2 and Figu e 1 highligh an inc ease in he speech
in elligibili y alues wi h he occupancy. This imp o emen is mo e signi ican when
he sound sys em is used:
∆
STI = STIoccup - STIunoccup = 0.035 ± 0.014, wi hou a sound sys em;
= 0.050 ± 0.024, wi h a sound sys em.
The a ia ion in he speech in elligibili y alues (∆STI) co ela es well wi h he
educ ion in he e e be a ion ime alues o 2 kHz (o mo e p ecisely wi h he
ac o : ln(RTunocc/RTocc)), and his pa icula ly when he sound sys em is used (see
Figu e 1). Wi hou he use o a sound sys em, he imp o emen o he speech
in elligibili y inc eases sligh ly wi h he educ ion in he e e be a ion ime.
The e e be a ion ime measu emen s show a educ ion in he RT alues wi h
occupancy om 0.13 o 2.37 seconds, s ongly dependen on he equency band
(a e age o 0.8 ± 0.6 seconds o a 95% con idence in e al) (Table 3). Chu ch
occupancy induces a signi ican dec ease in he e e be a ion ime alues,
pa icula ly in he 1 kHz and 2 kHz equency bands, which a e some o he mos
signi ican in e ms o speech in elligibili y quali y (see Table 3 and Figu e 2). This
dec ease in he RT alues can be explained by se e al e ec s as desc ibed below.
Figu e 2. Va ia ions in he e e be a ion ime alues in chu ches (occupied and
unoccupied condi ions).
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
5,5
6,0
6,5
7,0
125 250 500 1000 2000 4000
F equency (Hz)
RT (s)
La Chaux de Fonds occ.
La Chaux de Fonds non occ.
Pasqua occ.
Pasqua non occ.
Fille-Dieu occ.
Fille-Dieu non occ.
Romanel occ.
Romanel non occ.
Cheseaux non occ.
Cheseaux occ.
S . John non occ.
S . John occ.

34 Chu ch Acous ics and he In luence o Occupancy
4.1.1 Inc eased sound abso p ion due o he audience
The educ ion in he e e be a ion ime alues comes p ima ily om he inc eased
sound abso p ion b ough by he audience. The alue o ha educ ion in he
e e be a ion ime alues in a chu ch depends mainly on i s olume and on he
e e be a ion ime o he unoccupied building, as well as on he numbe and
dispe sion o he pe sons in i .
The calcula ion o he addi ional a e age equi alen sound abso p ion by pe son
(Ape s in m2) can be done s a ing om he Sabine equa ion, ha is:
whe e N, numbe o pe sons, V olume (m3), RTocc and RTunocc, he a e age
e e be a ion ime alues in he occupied and unoccupied chu ch. Fo he chu ches
sampled (excep ed o Pasqua ha has sound abso ben sea ing), he a e age
equi alen sound abso p ion alues Ape s (calcula ed using he Sabine equa ion o all
he sampled chu ches) and he op imised Ape s' alues (calcula ed using he leas
squa e me hod, by minimiza ion o he
χ
2 calcula ion-measu emen o he chu ches
sampled) a e gi en in Table 4 and Figu e 3.
Table 4. Equi alen sound abso p ion by pe son (Ape s) acco ding o equency
F equency (Hz) 125 250 500 1k 2k 4k a g. 500-1k
Mean calcula ed alues Ape s (m2) 0.24 0.50 0.56 0.73 0.76 0.77 0.64
S anda d de ia ion (m2) 0.18 0.22 0.06 0.17 0.34 0.25 0.09
Op imised calcula ion Ape s' (m2) 0.13 0.53 0.54 0.73 0.79 0.87 0.63
Figu e 3. Sound abso p ion (pe pe son) o he chu ch audience (La-Chaux
chu ch has dispe sed sea ing and he o he s ha e dense sea ing;
Pasqua chu ch has abso ben chai s, o he s ha e sound e lec i e
pews).





−= RTunoccRToccN
V
Ape s
11
*
*16.0
0,0
0,1
0,2
0,3
0,4
0,5
0,6
0,7
0,8
0,9
1,0
125 250 500 1000 2000 4000 a g 500-1k
F equency (Hz)
Sound abso p ion pe pe son (m2)
La-Chaux
Fille-Dieu
Romanel
Cheseaux
Pasqua
Calcul op
(wi hou P)
Mean
(wi hou P)
BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 35
The mean sound abso p ion by pe son (Ape s) calcula ed o he a e age
e e be a ion ime (500-1000 Hz) and o all he chu ches, equals o 0.64 ± 0.09 m2.
When he pe sons a e ela i ely g ouped in a chu ch (case o he chu ches Fille-
Dieu, Cheseaux and Romanel) he sound abso p ion a he highes equencies ( > 2
kHz) is abou 0.7 m2. When he ai h ul a e highly dispe sed (case o La-Chaux-de-
Fonds chu ch) he sound abso p ion pe pe son in he highes equencies inc eases o
app oxima ely 0.9 m2. The s anda d de ia ion be ween he sound abso p ion alues
(Ape s) calcula ed o each chu ch is high. This can be explained by he s ong
a ia ion o he occupa ion densi y om one chu ch o ano he and maybe e en by
he ype o clo hing wo n du ing measu emen s ha ook place bo h in summe and in
win e .
The smalle Ape s alues ound o he Pasqua chu ch (Figu e 3) we e caused by
he sound abso p i e cha ac e is ics o i s sea s. The a e age sound abso p ion pe
pe son calcula ed in his case co esponds o he di e ence be ween he a e age
sound abso p ion pe pe son on he pews (s anda d si ua ion as on he o he chu ches)
and he sound abso p ion o he s u ed sea s.
The minimiza ion o he di e ences be ween measu ed and calcula ed absolu e
alues makes i possible o de e mine an op imised sound abso p ion pe pe son
which allows a sa is ac o y es ima e o he RT in all cases ( he a e age e o be ween
he compu ed and he measu ed occupied e e be a ion ime alues is lowe han 0.1
second, see Figu e 4). The Sabine equa ion seems hen su icien o calcula e he
educ ion in he e e be a ion ime alues gi en by he addi ional sound abso p ion
caused by he p esence o he audience.
The alues ob ained o he equi alen sound abso p ion pe pe son a e gene ally
highe han hose a ailable in he li e a u e o a sea ed pe son, because o he low
Figu e 4. Re e be a ion Time alues - measu ed ( ull lines) and calcula ed wi h
op imised Ape s (do ed lines) o occupied chu ches.
0,5
1,0
1,5
2,0
2,5
3,0
3,5
4,0
4,5
5,0
125 250 500 1000 2000 4000 a g
F equency (Hz)
RT occupied chu ches (s)
Sac é-coeu
Fille-Dieu
Pasqua
Romanel
Cheseaux
36 Chu ch Acous ics and he In luence o Occupancy
occupa ion densi y in he occupied chu ches compa ed o alues calcula ed using a
e e be a ion chambe o a conce hall. The edges e ec being mo e signi ican o a
dispe sed assembly, one could expec an inc ease in he equi alen sound abso p ion
pe pe son.
Ano he app oach, sugges ed by se e al au ho s11-12, in pa icula o conce and
ope a halls, consis s in de ining he audience sound abso p ion no by pe son (Ape s in
m2) bu pe uni o loo a ea o he building (αpe s) by conside ing as an addi ional
pa ame e he a io o pe ime e /a ea o he cong ega ion sea ing a ea1,2,13.
Conside ing he di icul y in clea ly de ining hese pa ame e s in he case o chu ch
buildings, o en wi h a e y sca e ed sea ing a angemen s, his la e app oach does
no seem well adap ed o hese si ua ions bu seems o emain a he speci ic o an
audience in dense sea ing a eas (which is gene ally he case o hea es, conce o
ope a halls).
The p edic ion me hod sugges ed by Be anek11 o conce halls whe e he
abso p ion cha ac e is ics o he sea s a e no known, was es ed in his s udy, ha is,
RTocc = c - d*exp(-RTunocc)
In spi e o op imisa ion o he pa ame e s c and d o each equency band by he
me hod o leas squa es, his me hod does no make i possible o calcula e, in a
sa is ac o y way, he e e be a ion ime in chu ches sligh ly occupied (de ia ion o ±
0.9 s in pa icula in he 500 and 1000 Hz equency bands).
4.1.2 The e ec o hea ing
In summe , he empe a u e and humidi y wi hin a chu ch p ac ically do no change
wi h he occupa ion (app oxima ely 20 o 23 °C and 60 o 70% ela i e humidi y ( h)
in Swi ze land). In con as , in win e , hea ing he chu ch be o e and du ing he
se ice modi ies i s hyg o he mic condi ions. In a unhea ed chu ch, he empe a u e is
abou 8 o 10 °C wi h a ela i e humidi y o abou 60 o 70%, whe eas in a hea ed
chu ch ( o he pe iods o occupa ion), he empe a u e goes up o 18-20°C and he
humidi y goes down o 40-50%14. The a ia ion in ela i e humidi y is mainly due o
he inc ease in empe a u e. Wi h cons an empe a u e, humidi ica ion o he ai by
he p esence o a cong ega ion emains weak ( he inc ease o he ela i e humidi y
due o he audience is lowe han 10%).
In win e , he chu ch hea ing sys em du ing ce emonies causes se e al e ec s
a ou able o he speech in elligibili y alues, such as:
1. A small inc ease in ai abso p ion be ween 500 and 2000 Hz ha in ol es a sligh
educ ion in e e be a ion ime alues a hese equencies. A high equencies ( > 2
kHz), he opposi e phenomenon, wi h mo e signi ican alues is obse ed (Table 5).
The abso p ion o he ai is gi en by he ollowing o mula15:
Aai = 4m*V
whe e V is olume, (in 103 m
3) and he 4m (m-1) he ai equi alen abso p ion
coe icien .
BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 37
Fo he win e clima ic condi ions, e e ed o abo e, he ai abso p ion alues
p esen ed in Table 5 a e ob ained9. The alues show an inc ease in ai abso p ion in
he low and medium equency bands ( he alues emain howe e low) and a
educ ion in he abso p ion a highe equencies ( > 2 kHz). As an example,
conside ing a chu ch o olume 6000 m3 wi h a e e be a ion ime (unoccupied) o 4
s, he educ ion in he a e age e e be a ion ime by he modi ica ion o he ai
abso p ion is abou 0.1 s.
Table 5. Ai abso p ion (4m) o win e clima ic condi ions (m-1) 9
F equency (Hz) 125 250 500 1000 2000 4000 8000
Non-hea ed chu ch (10 °C, 60% h) 0.4 0.9 1.7 3.1 6.9 23.1 82.3
Hea ed chu ch (20 °C, 40% h) 0.5 1.3 2.4 3.9 7.4 22.0 77.5
2. Hea ing o he chu ch du ing se ices also induces an inc ease in he speed o
sound p opaga ion ha causes a educ ion in he building e e be a ion ime.
Re e be a ion ime is gi en by he ollowing o mula8:
whe e c( ) is he eloci y o he sound p opaga ion in he ai a he empe a u e (in
°C) gi en by:
The a io o he e e be a ion ime (RT) o a oom a wo di e en empe a u es ( )
is hen equal o he in e se a io o he eloci ies o he sound p opaga ion (c) a hese
wo empe a u es:
Conside ing ypical empe a u es a lues o 1 = 20 ºC and 2 = 10 °C, yields:
RT(20°C) = c(10°C) / c(20°C) * RT(10°C) = 337 / 343 * RT(10°C) = 0.98 * RT(10°C)
ha is, a dec ease o app oxima ely 2% in he e e be a ion ime alues as a esul o
he empe a u e inc ease due o no mal win e hea ing condi ions in a ypical chu ch.
Conside ing a chu ch wi h a e e be a ion ime o 4 s, he educ ion in
e e be a ion ime as a esul o inc easing he eloci y o he sound p opaga ion in
he ai , by hea ing is abou 0.1 s (Table 6).
Fo a chu ch o olume 6000 m3 wi h a e e be a ion ime (unoccupied) o 4 s, he
c += 16.273*04.20)(
)1(
)2(
)2(
)1(
c
c
RT
RT =
A c
V
RT *)(
*3.55
=
44 Chu ch Acous ics and he In luence o Occupancy
In one case howe e (La-Chaux chu ch), backg ound noise inc eases su icien ly wi h
he numbe in he cong ega ion, o dec ease he S/N a io o less han 15 dB. In his
case (Soccup=14 dB), he con ibu ion o backg ound noise o he a ia ion o speech
in elligibili y is, acco ding o he heo e ical model (Equa ion 3)
∆STI = -0.001.
This negligible alue is s ongly a ec ed by he e e be a ion ime alue used ha is
a he high (RT = 4 s) and ha is signi ican o speech in elligibili y. In chu ches a
small dec ease in he S/N has only a e y small e ec i he e e be a ion ime is
high. The esul s o he heo e ical model (wi h his co ec ion) a e p esen ed in
Figu e 7.
4.2.2 - Empi ical model
F om he esul s ob ained o he ield measu emen s epo ed in his s udy, an
empi ical app oach can be s a ed ha allows a simple o mula ion han he abo e
heo e ical models. The ield esul s showed ha he p incipal ac o in luencing
speech in elligibili y (STI) was he a ia ion o e e be a ion ime ( he dis ance ac o
is no e y signi ican ). As he clima ic pa ame e s ( empe a u e and ela i e
humidi y) we e no measu ed in si u, he p oposed loga i hmic model is simply o he
ype:
∆
STI = a*ln(RTocc / RTunocc) + b
whe e he pa ame e a akes in o accoun he con ibu ion o a ia ion o
e e be a ion ime, and he pa ame e b ga he s he o he di e se con ibu ions.
The op imisa ion o hese wo pa ame e s ( o medium size chu ches) gi es he
ollowing o mula:
∆
STI =0.276*ln(RTunoccup(2kHz)/RTocc(2kHz))–0.052 wi h a sound sys em (R2 = 0.96),
and
∆
STI =0.102*ln(RTunoccup(2kHz)/RTocc(2kHz)) wi hou PA (R2 = 0.55).
These o mulae we e no es ed o e y la ge chu ches and may be no alid
beyond he ange o he da a. The alues ob ained by hese o mulae a e p esen ed in
Figu e 7.
De e mina ion o he alue o he pa ame e a, o si ua ions wi h sound sys ems,
gi es a highe alue han o a heo e ical condi ion wi hou backg ound noise (a =
0.190 acco ding o 20). The educ ion o he addi ional RT induced by he PA sys em
(using loudspeake s wi h s ong di ec i i y o ien ed o he cong ega ion a ea) could
howe e explain he highe slope ob ained empi ically wi h PA.
I one conside s Ea ly Decay Time (EDT) ins ead o he Re e be a ion Time, he
heo e ical alue o a is 0.216 wha would esul in an inc ease in STI o 0.15, by

BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 45
educ ion o e e be a ion ime by a ac o o 2. (This las alue was highligh ed in a
s udy unde aken in uni e si y class ooms21). In hese pa icula cases (low speci ic
olume and a he high e e be a ion ime), occupancy can a ec he STI
conside ably, om 0.47 o 0.76 in unoccupied class ooms, 0.61 o 0.77 wi h hal
occupancy and 0.65 o 0.82 in ully occupied.
5 - SUMMARY
Measu emen s we e ca ied ou in six chu ches wi h and wi hou occupancy (mean
olume o 4385 m3 and mean occupancy o 117 pe sons) o es he e ec o
occupancy on STI alues. The esul s show ha occupancy causes a mean inc ease in
speech in elligibili y o
∆
STI (STIoccup - STIunoccup) o 0.050 wi h he use o a public
add ess sys em (PA) and 0.035 wi hou a PA. This inc ease is caused mainly by he
educ ion in oom e e be a ion ime, due o he addi ional sound abso p ion caused
by he p esence o people. The a ia ions in he speech in elligibili y alues wi h
occupancy can be p edic ed in occupied and unoccupied chu ches (mean e o
∆
STI <
0.007) by empi ical o mulae based on RT
∆
STI = 0.276*ln(RT unoccup(2kHz) /RT occ(2kHz) ) – 0.052 wi h a PA (R2 = 0.96), and
∆
STI = 0.102*ln(RT unoccup(2kHz) /RT occ(2kHz) ) wi hou a PA (R2 = 0.55).
The educ ion o RT caused by he PA sys em (using loudspeake s wi h s ong
di ec i i y o ien ed owa ds he cong ega ion a ea) could explain he high slope
ob ained in he abo e empi ical model wi h PA (highe han he alues expec ed
heo e ically).
The educ ion in e e be a ion ime in an occupied chu ch depends on i s olume
and unoccupied e e be a ion ime, as well as on he numbe and dispe sion o he
people on i . The mean sound abso p ion pe pe son (Ape s) calcula ed o he a e age
e e be a ion ime (500-1000 Hz) o each chu ch, is equal o 0.64 ± 0.09 m2.
The dec ease in he S/N a io (due o he backg ound sound le el inc ease by he
p esence o people, by inc eased sound le el dec ease wi h dis ance in he
cong ega ion a ea, o by poo di ec i i y and o ien a ion o he sound sou ces) could
educe he STI gain in pa icula cases. Howe e , hese e ec s a e gene ally
negligible.
The a ia ions o humidi y and empe a u e caused by he cong ega ion and by
hea ing cause a small change in RT alues (less han 0.2 s) and gene ally a negligible
change in STI alues.
REFERENCES
1. B adley, J. "The sound abso p ion o occupied audi o ium sea ing", J. Acous .
Soc. Am., 1996, 99(2), 990-995.
2. B adley, J. "Measu ing and p edic ing he e ec o an audience", 124 h
Acous . Soc. Am. mee ing 1992, New O leans.
46 Chu ch Acous ics and he In luence o Occupancy
3. A au, H. "Va ia ion o he e e be a ion ime o places o public assembly wi h
audience size", Building Acous ics, 1997, 4(2), 87-97.
4. Schul z, T. and Wa e s, B. "P opaga ion o sound ac oss audience sea ing", J.
Acous . Soc. Am., 1964, 35(5), 885-896.
5. Momme z, E. "Einige Messungen zu s ei enden Schallausb ei ung übe
Publikum und Ges ühl", Acus ica, 1993, 79, 42-52.
6. Da ies, W. and Lam, Y. "New a ibu es o sea dip a enua ion", Applied
Acous ics, 1994, 41, 1-23.
7. LoVe i, J., Ma dane, D. and Soulod e, G. "Modelling o he sea dip e ec
using he ini e-di e ence ime domain me hod", J. Acous . Soc. Am., 1996,
100(4), 2204-2212.
8. B adley, J. "Some u he in es iga ions o he sea dip e ec ", J. Acous . Soc.
Am., 1991, 90(1), 324-333.
9. EN 60268-16, "E alua ion objec i e de l'in elligibili é de la pa ole au moyen
de l'indice de ansmission de la pa ole", Commission Elec o echnique
In e na ionale, 1998.
10. Ca alho, A. "The signi icance o he Chu ch Occupancy in he RASTI alues
in a Ca holic Chu ch", Cong ès F ançais d'Acous ique CFA2000, 2000,
Lausanne, 477-480.
11. Be anek, L. "Conce and Ope a Halls. How hey sound." Acous . Soc. Am.,
1996, New Yo k.
12. Ba on, M. "Audi o ium Acous ics and A chi ec u al Design", E&FN Spon,
London, 1993.
13. B adley, J. "P edic ing he abso p ion o pew cushions", 124 h Acous . Soc.
Am. mee ing 1992, New O leans.
14. Hunge bühle , E. "Dimensionie ung, Sanie ung und Be ieb on
Elek oheizungen in Ki chen", I. Ra el, Be n, 1994.
15. P EN 12354-6 (4 h d a ), "Es ima ion o acous ic pe o mance o buildings
om he pe o mance o elemen s. Pa 6: Sound abso p ion in enclosed
spaces", B ussels, 2000.
16. Desa naulds, V., Bossoney, S. and Eggenschwile K. "S udie zu Raumakus ik
on Schweize Ki chen", DAGA 1998, Zü ich, 710-711.
17. Desa naulds, V. "Mesu ages des b ui s de en ila ion dans la ca héd ale de
Lausanne pou di é en s égimes de onc ionnemen ", Rappo No 1704 du
Bu eau ing. G. Monay, 1995, Lausanne.
18. Lo e mose , W. "Nachhalzei en in Ba ockki chen", Acus ica, 1952, 2, 109-
111.
19. Ahne , W. and S e en, F. "Beschallung echnik. G undlagen und P axis",
1993.
20. Hou gas , M. and S eeneken, H. "The modula ion ans e unc ion in oom
acous ics", B üel&Kjae Technical Re iew, 1985, 3, 3-12.
21. Hodgson, M. "Expe imen al in es iga ion o he acous ical cha ac e is ics o
uni e si y class ooms", J. Acous . Soc. Am., 1999, 106(4), 1810-1819.
BUILDING ACOUSTICS . Volume 9 . Numbe 1 .2002 47
22. Be anek, L. "Rela ion o acous ical pa ame e s wi h and wi hou audience in
conce halls and a simple me hod o simula ing occupied s a e", J. Acous .
Soc. Am., 2001, 109(3), 1028-1042.