1876-6102 © 2015 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL
doi: 10.1016/j.egyp o.2015.11.767
Ene gy P ocedia 78 ( 2015 ) 3120 – 3125
ScienceDi ec
6 h In e na ional Building Physics Con e ence, IBPC 2015
Associa ions be ween class oom condi ions and eache ’s oice
p oduc ion
Leena Ran alaa*, Ee a Salab
aUni e si y o Tampe e, 33014 Tampe een yliopis o, Finland
bUni e si y o Tu ku, 20014 Tu un yliopis o, Finland
Abs ac
In his s udy, associa ions be ween class oom condi ions and eache s’ oce use we e esea ched (N=40). Acous ic
a iables we e e e be a ion ime (T60), ea ly decay ime (EDT), sound s eng h (G), he a io o ea ly and la e sound
ene gy a i ing o a lis ene (C50), he a io o ea ly a i ing sound ene gy o o e all sound ene gy (D50), speech
ansmission index (STI) and a e o spa ial decay o sound p essu e le els (DL2). Indoo clima e was also e alua ed.
Voice a iables measu ed we e sound p essu e le el (SPL), undamen al equency (F0) and spec um slope.
Acco ding o indings when he alues o T60, EDT and DL2 we e lowe o he alues o C50, D50 and G highe ,
eache s’ oice SPL was highe . When e e be a ion imes we e longe , F0 inc eased and when he alues o C50 and
D50 we e highe , men’s F0 was highe . Poo indoo ai inc eased occu ence o la yngi is. The esul s showed ha
indoo en i onmen a ec s oice use and hus, is associa ed wi h he heal h o oice o gans.
© 2015 The Au ho s. Published by Else ie L d.
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL.
Keywo ds: Class oom; acous ics; oice; eache s, indoo clima e
In oduc ion
Indoo en i onmen and ac i i y in class ooms a e an a ea whe e many ields o science mee , om hose
s udying human beha iou o he ones s udying physical cons uc ions and buildings. Class oom acous ics has been
esea ched om he pe spec i e o pupils’ lea ning, especially. Ano he poin o iew o oom acous ics is eaching,
whe e oice use comes o a ocus. Heal h oice is no only a ma e o impo ance o eache s, bu i a ec s also
pupils: i a speake ’s oice quali y is dysphonic, child en’s linguis ic-cogni i e pe o mance de e io a es [1]. Good
* Co esponding au ho . Leena Ran ala Tel.: +358 50 3812039.
E-mail add ess: leena.m. an [email p o ec ed]
A ailable online a www.sciencedi ec .com
© 2015 The Au ho s. Published by Else ie L d. This is an open access a icle unde he CC BY-NC-ND license
(h p://c ea i ecommons.o g/licenses/by-nc-nd/4.0/).
Pee - e iew unde esponsibili y o he CENTRO CONGRESSI INTERNAZIONALE SRL
Leena Ran ala and Ee a Sala / Ene gy P ocedia 78 ( 2015 ) 3120 – 3125 3121
oice e gonomics is also good o lis ening and lea ning. Voice e gonomics s udies ac o s ha may inc ease he isk
o oice diso de s [2].
A ac o known o h ea en he mos oice heal h is noise. Noise loads oice [3] and inc eases he numbe o
oice symp oms in eache s [2,3]. Noise also aises oice sound p essu e le el (SPL) and undamen al equency
(F0) [4,5] and changes sound spec um o oice (ene gy le el shi s o highe equency bands) [6].
Poo oom acous ics inc ease noise, especially excessi ely long e e be a ion (5]. In poo e acous ic condi ions,
speake s use loude oice, ha e longe speaking imes [7] and ha e mo e oice symp oms [8] han in be e condi ions.
In u n, an acous ically good oom suppo s oice p oduc ion [9].
The oice is also a ec ed by indoo clima e. Indoo ai is poo in many Eu opean schools [10] and isible mould
has been ound in 13.9% o 39.1% in buildings, o ins ance [11]. Exposu e o oxic subs ances o mold [12] o o
o ganic dus cause oice symp oms [13]. D y indoo ai – a p oblem o No dic coun ies du ing win e – is also a isk
o oice diso de s. I s i ens he co e o ocal olds and inc eases he iscosi y o mucosa [14], which in u n
de e io a es he ib a ion o ocal olds [15].
Al hough s udies ha e shown he e ec o indoo en i onmen on human beha iou he e a e only ew oice
e gonomic esea ch on he opic. Hence, he aim o his esea ch was o s udy connec ions be ween class oom
condi ions – acous ics and indoo clima e − and acous ic pa ame e s o oice. Class ooms wi h highe and lowe
noise we e also s udied sepa a ely. The s udy was ound o be e hically sa e y o pa icipan s.
2.
Ma e ials and me hods
2.1.
Teache s and schools
Fo y eache s and hei class ooms we e s udied in he esea ch. O he eache s, 32 we e women (mean age 45
yea s, ange 27–57) and 8 men (mean age 39 yea s, ange 31–45). The mean numbe o pupils was 20 pe
class oom. Nine eache s had wo ked less han i e yea s, o he s o e ha . Fou een pa icipan s (36%) epo ed
hoa se and 21 (54%) i ed oice a leas weekly.
The classes we e om 14 schools in 5 di e en ci ies. The sizes o he ooms we e qui e simila (Table 1).
Because noise le el has been shown o be he mos impo an ac o a ec ing oice use in eache s [7], on he basis
o L90 he class ooms we e di ided in wo g oups: noisy (n=23) and quie ones (n=17). The pa ame e s o ooms
acous ics analysed in his s udy did no di e be ween he wo class oom g oups (S uden ’s - es ; see Table 4 o
acous ic a iables). Table 1. Class oom (N=40) dimensions
Magni ude
Mean
SD
Min
Max
Room leng h (m)
8.4
1.0
6.8
10.20
Room b ead h (m)
7.5
0.9
6.1
10.00
Room heigh (m)
3.2
0.24
3.0
3.80
2.2.
Va iables o class ooms
Acous ic a iables (Table 2) and noise le els we e measu ed wi h he p ecision sound le el me e o 01dB-S ell
Symphonie and 01dB-S ell Ha monie placed in he cen e o he class oom. Re e be a ion T60(s) was measu ed in
acco dance o s anda ds ISO 354 and ISO 3382 and STI in acco dance o he S anda d IEC 60268-16:2003. Noise
caused by pupils (L90), also called ac i i y noise in his s udy, was measu ed du ing lessons (measu ed ime 2−6
hou s).
2.3.
Indoo clima e and espi a o y ac diseases
Indoo clima e assessmen was pe o med in compliance wi h he Voice E gonomic Assessmen in Wo k
En i onmen - Handbook and Checklis [2]. The o al numbe o isk ac o s was 16 (Table 3). The inding o a isk
3122 Leena Ran ala and Ee a Sala / Ene gy P ocedia 78 ( 2015 ) 3120 – 3125
ac o was assessed dicho omicly (0 = no isk, 1 = isk o oice diso de ). Respi a o y ac diseases we e s udied by
asking i pa icipan s we e gi en he diagnosis o la yngi is in he pas 12 mon hs (answe s ‘‘yes/no’’).
Table 2. Acous ic a iables measu ed in class ooms
Room acous ic
pa ame e
Symbol
Uni
De ini ion
Re e be a ion ime
T60
second
[s]
The ime du ing which he sound le el dec eases by 60 dB a e he
sound sou ce has ceased o ope a e.
Ea ly decay ime
EDT
second
[s]
The ime du ing which he sound le el dec eases by 10 dB a e he
sound sou ce has ceased o ope a e.
Sound s eng h
G
decibel
[dB]
Desc ibes he sound olume o sound sou ce a he dis ance o 10
me e s compa ed o ha a he same dis ance in an anechoic oom.
Cla i y o speech
C50
decibel
[dB]
The a io o ea ly and la e sound ene gy a i ing o he lis ene , ha
is, he ene gy ha a i es du ing he i s 50 milliseconds.
Desc ibes cla i y o pe cei ed speech sound.
Deu lichkei
D50
pe cen
[%]
Ra io o ea ly a i ing sound ene gy o o e all sound ene gy.
Desc ibes pe cei ed cla i y o speech.
Speech ansmission
index
STI
-
Speech in elligibili y be ween a speake and a lis ene . Desc ibes
he quali y o speech ansmission om he poin o iew o
syllable dis inc ion.
Ra e o spa ial decay
o sound p essu e
le els
DL2
decibel
[dB]
Ra e o spa ial decay o sound p essu e le els pe dis ance
doubling. Desc ibes sound a enua ion when dis ance om he
sound sou ce doubles.
Table 3. Risk ac o s o indoo clima e. Sco ing: a ac o causes isk o a oice diso de =1; no isk = 0.
Risk ac o
Numbe o
ac o s
Room empe a u e ou side he ange o 20–23° C
1
Dus in ai /on su aces, deco a ion ma e ial in oom, a icles di icul o clean
4
Unpleasan odou s, signs o wa e leaks, ea lie wa e leaks
3
S u y ai , hea y eeling in head, d augh , aul y in ake /ou le o ai en
5
D y indoo ai , d y eyes, i i a ed eyes
3
To al numbe o isk ac o s in indoo ai quali y
16
2.4.
Voice samples and acous ic analyses
The pa icipan s ead a 102-wo d-long ex be o e and a e a wo king day. The eco ding was made wi h a
Zoom-H2 and a headse mic ophone (AKG CL-555) loca ed a he dis ance o 3 cm om he co ne o he mou h.
The eco dings we e calib a ed (gene a o BOSS TU-120; a sound le el me e B üel & Kjæ 2206).
F om he oice samples, F0, SPL and he il o a sound spec um slope (alpha a io) we e analyzed. Alpha a io
is he ela ionship o oice ene gy le els be ween he le el o 50 Hz–1 kHz and he le el o 1–5 kHz and exp esses
oice quali y (hypo/hype unc ional quali y; highe alue exp esses mo e hype unc ional oice quali y). F0 was
analysed sepa a ely o men and women. In addi ion, he change o he alue o a oice pa ame e du ing wo k was
also aken as one a iable. The acous ic analyses we e made wi h P aa -p og am ( e sion 5.3.18).
2.5.
S a is ical analysis
The associa ions we e s udied wi h ei he Spea man co ela ion ank coe icien o Pea son p oduc -momen
co ela ion coe icien . The e ec o indoo clima e was calcula ed wi h Wilcoxon signed ank es .
Leena Ran ala and Ee a Sala / Ene gy P ocedia 78 ( 2015 ) 3120 – 3125 3123
3.
Resul s
3.1.
Values o a iables o oice, oom acous ics and ac i i y noise
Pa icipan s’ mean oice SPL was 67 dB (SD 5.4 dB) be o e and 69 dB (SD 4.3 dB) a e wo k. Mean F0 o
emales was 182 Hz (SD 15.7 Hz) be o e wo k and 185 Hz (SD 17.2 Hz) a e i ; o males he alues we e 97 Hz
(SD 11.6 Hz) and 102 Hz (SD 14.9 Hz), espec i ely. Alpha a io alues we e be o e wo k -15.7 dB (SD 1.28 dB)
and a e i -15.2 dB (SD 2.06 dB). Table 4 p esen s he alues o acous ic a iables and noise (L90) in he
class ooms.
Table 4. The alues o acous ic a iables and noise in he class ooms (N=40).
T60 (s)
EDT (s)
G (dB)
C50 (dB)
D50 (%)
STI
DL2 (dB)
L90 (dB)
Mean
.55
.53
15
4.5
73
.74
2.8
42
SD
.114
.118
2.8
1.56
6.7
.036
.86
4.1
3.2.
Associa ions be ween oom acous ics and oice pa ame e s in all class ooms
The longe he e e be a ion ime (T60) and he highe he DL2 alue, he lowe he oice SPL (Table 5). F0
co ela ed nega i ely wi h T60 bu posi i ely wi h he a iable exp essing speech cla i y (C50). Howe e , in
class ooms wi h longe T60, eache s’ F0 has inc eased mo e du ing a wo king day. Ac i i y noise (L90) co ela ed
wi h no oice a iables.
Table 5. Associa ion be ween oom acous ic a iables and pa ame e s o eache s’ oices be o e and a e
wo k, and he change du ing wo k, M=males, ns= no signi ican (N=39−40).
Room
acous ic
pa ame e
Voice SPL
F0
Be o e
A e
Be o e
A e
Change
T60
ns
-.29 (.037)
ns
-.62 (.05) M
.29 (.036)
C50
ns
ns
ns
.64 (.045) M
ns
D50
ns
ns
ns
.66 (.038) M
ns
DL2
-.28 (.042)
-.37 (.011)
.65 (.039)M
ns
ns
3.3.
Associa ions be ween oom acous ics and oice pa ame e s in noisy and quie class ooms
Noisy class ooms The mean ac i i y noise le el (L90) was 45 dB (SD 2.5 dB) o he noisy class ooms. The a iables
o oom acous ics we e associa ed only wi h oice SPL and mainly a e wo k (Table 6). In he class ooms whe e
e e be a ion imes (T60, EDT) we e longe , he eache s’ oice SPL was lowe a e wo k. In u n, i lis ening
condi ions we e be e (inc ease in he alues o C50, D50, STI), eache s used highe SPL ei he be o e o a e wo k.
Table 6. Associa ions be ween acous ic a iables o loud class ooms and
eache s’ oice SPL be o e and a e wo k, ns= no signi ican , (n= 21−22).
Room acous ic
pa ame e
Be o e
A e
T60
ns
-.6 (.002)
EDT
ns
-.55 (.004)
G
.35 (.05)
ns
C50
ns
.54 (.004)
D50
ns
.55 (.004)
STI
ns
.51 (.007)
3124 Leena Ran ala and Ee a Sala / Ene gy P ocedia 78 ( 2015 ) 3120 – 3125
Quie class ooms The mean ac i i y noise le el (L90) was 38 dB (SD 3 dB) o he quie class ooms. In hese
class ooms, oom acous ic pa ame e s co ela ed wi h mo e oice pa ame e s han in he noisy class ooms. The
s onges connec ion was be ween T60 and F0 change du ing a wo king day: he longe he e e be a ion ime, he
mo e he eache s’ oice pi ches inc eased du ing wo k (Table 7). The eache s used loude oice a e wo k i oom
sound s eng h (G) had highe alues and he a e o spa ial decay o SPL (DL2) lowe ones. In addi ion, he lowe
he alues o he cla i y o speech (C50), he smalle was he spec um slope il meaning ha oice quali y changed
in o mo e hype unc ional di ec ion du ing a wo king day.
Table 7. Associa ion be ween oom acous ic a iables o quie class ooms and pa ame e s o eache s’
oices be o e and a e wo k, and he change du ing wo k, ns= no signi ican (n= 17–18).
Room acous ic
pa ame e
Voice SPL
F0
Alpha a io
A e
Change
Change
T60
ns
.51 (.018)
ns
G
.42 (.047)
ns
ns
C50
ns
ns
-.47 (.029)
DL2
-.49 (.024)
ns
ns
3.4.
Indoo clima e
Median numbe o isk ac o s o indoo clima e was 8 ( ange 0−12). The isk ac o s mos o en ound in
class ooms we e ma e ials collec ing dus o di icul o clean (90 % o he class ooms) and dus on su aces (85 %).
Ai was o en also s u y (72.5 %) and d y (70 %). Teache s who wo ked unde poo e indoo clima e, ha e had
mo e o en la yngi is du ing pas yea han hose wi h be e clima e (p=0.051; z =1.994).
Discussion
Ou esul s showed ha indoo en i onmen is associa ed wi h eache s’ oice p oduc ion and good indoo
clima e inc eases ocal heal h. Especially oice SPL seems o be a ea u e ha eac s o oom acous ics and indoo
clima e. In addi ion, ac i i y noise a ec s how he in luences o oom acous ics eme ge: in he noisy class ooms, he
connec ions we e ound only wi h oice SPL, bu in he quie ones, mo e oice pa ame e s we e a ec ed by oom
acous ics. Inc easing oice SPL may be ela i ely unconscious eac ion o a speake o demanding speaking
condi ions, such as noise igge s so called Lomba d e ec [16].
O he acous ic a iables T60, EDT, C50, D50, STI a ec ed oice simila ly: i hey indica ed poo e oom
acous ics, he eache s’ oice SPL was lowe . This sugges s muscle a igue in oice o gans. Fu he mo e, i
e e be a ion imes we e long, he eache s had a endency o aise oice pi ch ha may be a eac ion o inc eased
noise le els due o poo a enua ion [4,5]. Re e be a ion ime is ypically he only pa ame e ha is desc ibed in
egula ions o acous ics in class ooms. Howe e , e e be a ion ime is nei he a di ec p edic o o speech
in elligibili y, no o ac i i y noise le els du ing classes. In his s udy, ne e heless, i was connec ed o he eache s’
oice p oduc ion.
The indings ega ding T60 and DL2 and oice SPL we e somewha con adic o y. Namely, he esul s showed
ha when he ooms we e mo e e e be an (T60) and sound a enua ion (DL2) was bigge , he eache s used lowe
oice SPL. I he space is e e be an , speech oice ca ies easily a , and i he a e o spa ial decay o sound
p essu e le els pe dis ance doubling is high, he speech sound is a enua ed. On he o he hand, T60 and he a e o
DL2 measu e di e en en i ies. This ge s suppo om ou pos hoc analysis acco ding o which T60 and DL2 did no
co ela e wi h each o he in his da a.
F0 ( om he samples be o e and a e wo k) was analysed o he gende s sepa a ely. Only men’s oice pi ch
eac ed o oom acous ics p ima ily. The poo e acous ics on he basis o he pa ame e s o T60, C50, D50, DL2, he
lowe he men’s F0. This may o igina e om issue damage o ocal olds (inc eased oedema) o om he eache s
a emp o inc ease au ho i y in noisy en i onmen .

Leena Ran ala and Ee a Sala / Ene gy P ocedia 78 ( 2015 ) 3120 – 3125 3125
Highe sound s eng h (high G alue) was connec ed wi h highe oice SPL. In e es ingly, hose wo king in he
noisie class ooms al eady used loude oice be o e wo k bu hose wo ked in he quie e ones had inc eased oice
SPL du ing a wo king day ( ound in oice samples a e wo k). The i s men ioned eache s’ oice may e eal a
long- e m e ec o noise because o highe sound s eng h ( ha in ensi ies bo h noise and speech oice), possibly
due o swollen ocal olds ha leads o inc eased phona ion h eshold p essu e and u he , o ces o use high
subglo ic p essu e and consequen ly, highe SPL. The la e men ioned eache s’ oices, in u n, p obably showed a
mo e di ec (acu e) loading e ec .
The a iable indica ing oice quali y (alpha a io) was associa ed wi h he pa ame e desc ip ing pe cei ed
cla i y o speech (C50). The be e he cla i y, he mo e hypo unc ional was he oice p oduc ion, which means less
loading o oice o gans.
The e a e some me hodological conside a ions ega ding he indings. The esul s e ealed long- e m
consequences due o indoo en i onmen a he han di ec ones because mos o he pa icipan s had wo ked as
eache s yea s be o e his esea ch. The e a e also se e al isk ac o s ha a ec oice p oduc ion o he han ound in
wo king en i onmen . Fu he mo e, ou indings canno no be gene alized comple ely due o he small numbe o
class ooms and eache s.
Re e ences
[1] Lybe g-Åhlande V, Haake M, B änns öm J, Schö z S, Sahlén B. Does he speake 's oice quali y in luence
child en's pe o mance on a language comp ehension es ? In J Speech Lang Pa hol 2015;17:63-72.
[2] Ran ala LM, Hakala SJ, Holmq is S, Sala E. Connec ions be ween oice e gonomic isk ac o s and oice
symp oms, oice handicap, and espi a o y ac diseases. J Voice 2012 11;26:819.e13-819.e20.
[3] K is iansen J, Lund SP, Pe sson R, Shibuya H, Nielsen PM, Scholz, M. A s udy o class oom acous ics and
school eache s' noise exposu e, oice load and speaking ime du ing eaching, and he e ec s on ocal and men al
a igue de elopmen . In A ch Occup En i on Heal h 2014; 87:851-60.
[4] A onsson C, Bohman M, Te ns öm S, Söde s en M. Loud oice du ing en i onmen al noise exposu e in pa ien s
wi h ocal nodules. Logoped Phonia Vocol 2007;32:60-70.
[5] Pa el R, Schell K. The In luence o Linguis ic Con en on he Lomba d E ec . J Speech Lang Hea Res
2008;51:209-20.
[6] Ga nie M, Hen ich N, Dubois D. In luence o sound imme sion and communica i e in e ac ion on he Lomba d
e ec . J Speech Lang Hea Res 2010;53:588-608.
[7] As ol i A, Puglisi GE, Pa ese L, Ca ullo A. Long- e m ocal pa ame e s o p ima y school eache s and
class oom acous ics wi h and wi hou an acous ical ea men . Fo um Acus icum 2014;7.-12.9. K akow.
[8] Cu i a LC, Bu do A. E ec s o noise and acous ics in schools on ocal heal h in eache s. Noise Heal h
2015;17:17-22.
[9] Peleg ín-Ga cía D, B unskog J. Speake s' com o and oice le el a ia ion in class ooms: labo a o y esea ch. J
Acous Soc Am 2012;132:249-60.
[10] Simoni M, Annesi–Maesano I, Sigsgaa d T, No back D, Wieslande G, Nys ad W, e al. School ai quali y
ela ed o d y cough, hini is and nasal pa ency in child en. Eu Respi J 2010;35:742-49.
[11] An o a T, Pa enden S, B unek ee B, Hein ich J, Rudnai P, Fo as ie e F, e al. Exposu e o indoo mould and
child en’s espi a o y heal h in he PATY s udy. J Epidemiol Communi y Heal h 2008;62:708-14.
[12] Pa o i a R, Meklin T, Ne alainen A, Husman T. E ec s o mould emedia ion on school eache s' heal h. In J
En i on Heal h Res 2004;14:415-27.
[13] Geneid A, Rönkkö M, Ai aksinen L, Vou ilainen R, Toskala E, Alku P, e al. Pilo s udy on acu e oice and
h oa symp oms ela ed o exposu e o o ganic dus : p elimina y indings om a p o oca ion es . Logoped Phonia
Vocol 2009;34:67-72.
[14] Hemle R, Wieneke G, Lebacq J, Dejoncke e P. La yngeal mucosa elas ici y and iscosi y in high and low
ela i e ai humidi y. Eu A ch O o hinola yngol 2001;258:125-29.
[15] Wi R, Taylo L, Regne M, Jiang J. E ec s o su ace dehyd a ion on mucosal wa e ampli ude and equency
in excised canine la ynges. O ola yngol Head Neck Su g 2011;144:108-13.
[16] an Heusden E, Plomp R, Pols LCW. E ec o ambien noise on he ocal ou pu and he p e e ed lis ening
le el o con e sa ional speech. Applied Acous ics 1979;12:31-43.