A study of classroom acoustics, hearing and learning difficulties as experienced in four classes in a Middle School

Ruth Tolland, Educational Audiologist, CSSS, Norfolk

Back to index of articles

Hearing is affected by teacher voice level, reverberation time, distance from the teacher, background noise and class noise levels. Improved speech-to-noise ratios and low reverberation times within classrooms not only improves audibility but also increases intelligibility of individual phonemes, fundamental to speech perception. Ease of hearing encourages good listening, an activity that takes at least 60% of pupil time and is fundamental to education regardless of hearing status or learning ability (Crandell et al 1995). It follows, therefore, that the classroom environment should facilitate both listening and learning

Sound field FM systems are widely accepted as a means to facilitate improved listening in problematic classrooms and with this in mind a study was undertaken in a local middle school with an above average number of pupils with special needs - 60%. The school was on 'special measures' and located in a Government Action Zone. Concern was expressed regarding the large size and high levels of reverberation within a number of classrooms. There was concern also for the teachers competing in such conditions on the grounds of health and safety, particularly as one had been diagnosed with noise-induced deafness.

The main intention of the study was to look in detail at the conditions within the chosen classrooms before installing trial sound field systems and assessing their respective benefits. The selected classrooms had large volumes (Table 5) and varying acoustic modifications (Table 1). The Year 6 classes had part-brick, part-ceramic-tile walls. The Year 7 rooms had wooden floors, high vaulted ceilings incorporating a glass panel and part-brick, part-wooden-panelling for walls. 7B had one pupil with known moderate, bilateral, sensori-neural hearing loss who refused to wear hearing aids and a second pupil with Aspergers Syndrome.

Table 1: Current acoustic treatments available in the 4 classrooms
	6A	6B	7A	7B
Suspended ceiling	yes	yes
Acoustic tiles on ceiling	yes
Carpet	yes	yes
Acoustic wall panels	-	-
Curtains	-	-

Measurements and assessments were made for each classroom and class as follows with all sound measurements in dBA:

1. ambient noise in the unoccupied room
2. teacher voice level at 5 points around the room during reading to a very quiet class and at a point 1m in front of the teacher
3. speech to noise ratios during reading to a quiet class
4. class working noise level
5. physical measurement of each room and calculation of volume
6. calculated reverberation time (RT) in the unoccupied classroom using the RT = .05 V/A formula and sound absorption coefficients as detailed in Crandell et al (1995)
7. screen of hearing at 20 dB HL for all pupils present on the day (with parental permission) with further hearing assessment and tympanometry for those who did not pass
8. completion of a screen for targeting educational risk (SIFTER) by class teachers for selected pupils with listening and attention deficits (Anderson 1989)

A two-week sound field FM trial was arranged next for each class, starting with 6A and 7B. This was followed by:

1. measurement of teacher voice level at points A-E (see below) around the room during reading to a quiet class.
2. questionnaires completed by staff and pupils in 6A and 7B.

Fixed positions measured in each room:



Results

Table 2: Noise levels (dBA)

	6A	6B	7A	7B	DfEE recommended levels (1998)
Ambient noise in unoccupied classes	36	37	42	44*	40 max
Working noise	60	60-65	65	70	-
Chatter	65-70	75	70	75	-
Laughter	-	-	-	90	-

* Increased to 47dB with windows open to unusually quiet traffic noise.

Background noise is within DfEE guidelines for 6A and 6B only. However, hearing-impaired children require a maximum 30dB of background noise (DfEE 1998).

Table 3: Teacher voice level (tv) during reading to a quiet class compared with very quiet class noise (cn) at 6 fixed positions A - F (dBA)

Position	6A	6B	7A	7B
	tv/cn	tv/cn	tv/cn	tv/cn
A	70/35	65/40	65/40	62/45
B	62/35	62/35	65/40	60/42
C	60/36	63/36	61/43	65/42
D	70/40	65/36	63/45	62/45
E	63/35	65/36	61/40	65/40
F - Teacher voice level at 1m	70	68	77	70

Table 4: The average speech-to-noise ratio (SNR) during teacher reading to a quiet class

	6A	6B	7A	7B
Average SNR (dBA)	+29	+27	+21	+20

The SNR reflect the acoustic treatments present in each classroom (see Table 1). An acceptable SNR is in the range of +15dB to +20dB for pupils with normal hearing and without additional learning difficulties. However, if listening and speech perception are to be positively enhanced for children with listening deficits then a SNR of +30dB is required during instruction. To be able to hear all the consonant sounds in the acoustic signal the teacher's voice needs to be 33dB above the ambient noise level. To receive all vowel sounds a SNR of 22dB is required (Ross, Maxon & Brackett 1982). Consonants give words their intelligibility, vowels give volume.

The four teachers in the study used a louder voice during their normal teaching to compensate for increased class noise and activity compared with voice levels used when reading to a quiet class. For example, the 7A teacher increased her voice level by 9 dB and this in turn, arrived at points A-E at an average level of 72 dB. The SNR, however, dropped to a range of +7 to +12 dB. At 1m, the same teacher's voice was measured at 81 dB. This must put great strain on the individual's voice.

Table 5: Reverberation times (RT) as calculated for 4 classrooms

	6A	6B	7A	7B
Calculated RT	0.5	1.1	2.8	2.8
Classroom volume	270m3	228m3	316.7m3	316.7m3
DfEE recommended RT per classroom volume	0.7	0.65	0.7	0.7

The RT scores in Table 5 reflect the acoustical treatment within the classrooms (see Table 1) and the larger volumes in 7A and 7B result from their high arched ceilings. Reverberation in a classroom influences speech perception as the direct speech signal is 'smeared' or masked by reflected energy. Reflected signals arrive at the ear later than the direct signals causing degradation of the speech signal (Crandell et al 1995). Speech recognition decreases as RT increases. Children with learning and listening deficits are affected most and research indicates that RT in their classrooms should be no greater than 0.4 seconds (Finitzo-Hieber & Tillman 1978). All classrooms in this study had RT poorer than 0.4 seconds but one (6A) was within DfEE guidelines.

When an average-sized classroom (150 m³) has a RT of 0.6 seconds, pupils sat within 3m of teachers hear speech directly, uncontaminated by the effects of reverberation (Crandell et al 1995). This could apply in 6A but unfortunately, most pupils sit beyond this distance. Interestingly, the 6A teacher does most whole class teaching with pupils sat close together on the carpet. In 7A and 7B, the distance for direct speech reception would be much less, a situation especially disastrous for the reluctant hearing aid wearer in 7B.

The distance a pupil sits from the teacher also affects speech perception. In 6A, 6B and 7A, pupils may be up to 7.4m from the teacher. In 7B the greatest teacher-pupil distance was reduced to approximately 6m by arranging the tables in a double horseshoe, thereby limiting the seating area used within the classroom. Unfortunately the teacher was criticised by OFSTED inspectors as the arrangement restricted group work.

Table 6: A summary of SIFTER and hearing test results for each class

	6A	6B	7A	7B
Number in class	34	33	32	32
Number hearing tests (screen only) and %	27 (79%)	24 (73%)	26 (81%)	27 (84%)
Total number of pupils with hearing loss	11	4	11	16
Unilateral mild loss - conductive	6	2	4	8
Bilateral mild loss - conductive	2	1	6	6
Unilateral moderate loss - 1 conductive, 2 sensori-neural	3	-	-	-
Bilateral moderate loss - sensori-neural (hearing aids)	-	-	-	1
Unilateral high frequency loss - sensori-neural	-	-	1	1
Bilateral high frequency loss - sensori-neural	-	1	-	-
Number of SIFTERs not passed	4	5	2	10
Number of SIFTERs with hearing loss	3	0 (1 not tested for hearing)	1	7

NB: Pupils not screened for hearing loss were either absent or lacked parental consent.

83% of all pupils found to have a hearing loss in the 4 classes proved to have a mild, conductive hearing loss in at least one ear. This was unexpected as middle ear problems tend to be resolved by the age of 11 and 12. Some pupils reported fluctuating hearing due to ear infections, colds and hay fever. Results and subsequent discussions suggested that many of these pupils had a history of conductive hearing loss resulting in listening and language deficits that had not been overcome. In 6A and 7B the majority of pupils reported to experience learning and listening difficulty in class were found to have a hearing loss. In addition, the teachers were found to be highly reliable in their selection of pupils for SIFTER as only one of the selected pupils passed the screen. It is worth noting, however, that fewer pupils were assessed with SIFTER in 7A although 11 were found to have some hearing loss. Perhaps the particularly loud teacher voice and charismatic teaching style helped maintain an adequate SNR most of the time and especially when the class was quiet.

Table 7: Percentage of pupils in each class with additional difficulties as identified by SIFTER and hearing tests

	6A	6B	7A	7B
No. in class	34	33	32	32
% pupils with additional difficulties	35%	27%	38%	59%

A detailed report was produced for the school to help secure funding for acoustic improvements. Some of the conclusions and recommendations of that report are detailed below.

1. Acoustical Treatment
Covering hard reflective surfaces with absorptive material such as acoustic panelling and cork boards can reduce RT. Absorptive materials also reduce class noise in the room by 5-8 dB. It is recommended that the four classrooms receive the following treatments:

6A	curtains; acoustic panels on the brick walls.
6B	as above; also acoustic tiles on the ceiling.
7A	suspended ceiling with acoustic tiling or a translucent panel in the centre of the new ceiling to maintain current levels of light; thick carpet, heavy curtains, acoustic panels on the brick walls.
7B	as for 7A. In addition, the windows on the back wall could be covered with acoustic panelling. Light is more restricted on this side due to the high wall behind. A reflective surface would be eliminated and a much needed display area would be available.

2. Background Noise
Ambient noise in the four rooms is within DfEE guidelines for unoccupied rooms. However, with the high numbers of pupils with learning and listening deficits in each class, background noise should be nearer 30 dB than 40 dB. This may be difficult to achieve, however, the following measures will help:

• The doors of the room between 7A and 7B should be kept shut during lessons to avoid noise overspill. Noise and reverberation from corridors should be reduced outside 6A and 6B.
• Arrange moveable noise absorbent partitions or screens at varying angles within the classrooms. This is an effective way to reduce reflection and reverberation.

3. Teacher Voice Levels
Teacher voice strain and fatigue would be reduced if class noise was kept to a maximum of 50-55 dB. It is currently between 60-75 dB. In general, teachers should avoid using a voice level in excess of 70 dB at 2m to avoid strain. Three teachers were able to achieve this whilst reading to a quiet class, however, normal teaching voices were about 9 dB louder and SNR were much reduced (+7 to +12 dB). It is of concern that one teacher was using a voice of over 80 dB routinely during teaching and found great difficulty modifying their voice when speaking to colleagues in the staff room. A SNR of +20 dB is generally recommended, however, a SNR of +30 dB is better for children with listening and learning difficulties. High numbers of such pupils (27 - 59%) were found in the 4 classes as identified by SIFTER and the hearing screen (Table 7). It is worth noting that pupils also suffer listening fatigue especially when listening in noise. Additional energy is expended on extracting speech from noise that would be better spent in processing and comprehending the message.

4. Teacher-pupil distance
All four classrooms are large and the greatest distance between teacher and pupils in three of the rooms is about 7.4m. As a result, speech signals arrive at the ear more quietly and listening is more difficult due to degradation of elements of the speech sounds. Pupils with learning and listening deficits experience greater difficulty filtering speech from noise and teachers have to work harder to maintain a louder voice level. It is advised that teachers consider means by which to reduce the space used by the class during direct teaching sessions. The current seating arrangement in 7B has reduced distances and decreased the overall classroom space in use. Preferential seating is necessary for pupils with poor SIFTER and/or hearing results as these pupils experience greater difficulty attending to speech.

5. Sound Field FM Systems (SF FM)
SF FM systems help overcome the problem of distance between teacher and pupil and excessive teacher voice levels. The system provides up to 10 dB of teacher voice amplification and effectively brings most pupils within a 2m range of the teacher. Loss of critical speech elements is thus overcome as the distance travelled by the speech signals is reduced. Masking of the speech signals by ambient noise is minimised and overall audibility is increased. This allows pupils to focus on speech signals and ignore the relatively less intense class noise. With increased ease of listening, classes tend to become quieter and less time is expended on control and repetition.

A SF FM trial will be carried out in the four classes and teacher and pupil opinion will be sought as to the perceived benefits. It is anticipated most benefit will be experienced in 6A as acoustic modifications and acceptable RT levels are in place already, with remaining problems being teacher/pupil distance, class noise and teacher voice strain. Teachers will be required to undertake routine checks and maintenance of the systems.

Results of the SF FM trial

Table 8: Questionnaire results

	6A	6B	7A	7B
Teacher reaction	very positive	very negative	mostly positive	positive
Pupil reaction	93% positive	very negative	positive (reported)	74% positive

The trial was highly successful in class 6A. The teacher admitted to initial reluctance towards wearing the system (mainly cosmetic) but this was quickly overcome as the benefits became apparent. Voice strain and stress levels were reduced to such an extent after the two weeks that the teacher refused to part with the system. Given the nature of the children in the class, the following comments by the teacher were particularly telling:

• Particularly useful when calling the class to order when they are working on a task, or speaking to a child across the room. Without amplification I can only do this by shouting, which can be interpreted as aggression by the children.
• I can reprimand a child from a distance quietly repeating a request until he complies. This is then completely non-confrontational and impersonal - a disembodied voice from the wall - and he can't answer back! This is a big, unexpected advantage.
• I am considerably more relaxed, much less tired and ragged.
• I do not find it very comfortable but I enjoy using it!
• I still have to repeat everything several times because although they can hear, they are not listening. I wonder if this would improve over a period of time as listening is easier and their habits change. I suspect it would.

The pupils were equally appreciative although aware that the temporary fitting of the system caused uneven amplification in some areas of the room. Some pupils wanted more microphones and speakers. They enjoyed remaining at their desks for whole class teaching rather than sitting on the carpet but did not like the occasional squeaks from the speakers. Two pupils mentioned headaches. During week three, visiting inspectors commented on the quieter environment within the classroom compared with all other classes. It was explained that the difference stemmed from improved listening conditions and a more relaxed teacher rather than any radical change of teaching method. The pupils were benefiting from direct speech signal reception and increased clarity of consonants.

The benefits of SF FM in 6A were clear. They were possible because of the good acoustical treatments already in the room. As predicted, the remaining problems of teacher-pupil distance, poor speech recognition and voice strain were overcome by the SF FM system. In 7A and 7B - two rooms with large volumes and no acoustical treatment - the situation was rather different because although the teacher and the majority of pupils were positive about the easier listening afforded by the amplification, one third of 7B disliked the sound quality and complained of headaches. It is thought that the combined effect of class noise and too few loudspeakers i.e. four instead of 6, resulted in a more distorted sound field compared with 6A. Over amplification of the teacher's voice may also have produced some of the distortion. Despite this the teacher commented, 'went home less stressed, voice not strained, ears not ringing.' One pupil wanted two loudspeakers on each wall and several wanted more mics to pass around as pupils could not hear each other.

The SF FM trial in 6B was not successful. Both pupils and teacher appear to have rejected it from the outset. Unfortunately the teacher was absent for some of the measurements and explanations and this may have influenced the outcome. The 7A teacher although positive about the benefits of the system, commented that it affected her teaching style making it somewhat flat. She felt some time was needed to adjust to a quieter style that could still be dynamic. However, her conversational voice had returned to acceptable levels in the staff room.

Table 9: Average teacher voice levels and SNRs during teacher reading to a quiet class +/- SF FM

	6A	6B	7A	7B
Average teacher voice level + SF FM at points A-E (dBA)	65	NT	66	65
Average teacher voice level - SF FM at points A-E	65	64	63	63
Average SNR +SF FM	+29	NT	+23	+20
Average SNR -SF FM	+29	+27	+21	+19

Table 9 shows that teachers used the SF FM system to moderate their voices and reduce strain whilst keeping the SNR much as before. The Year 7 teachers, however, could afford to raise their voice by 5 dB when using SF FM for quiet reading to enhance further SNRs for pupils with listening deficits and more closely approximate the +30 dB SNR advocated by Ross et al (1989).

The purpose of this study was to examine the acoustic conditions in each of four classrooms and to trial a SF FM system. The greatest satisfaction with the SF FM system for both teacher and pupils was found in 6A - the classroom with the best acoustic treatments and RT within DfEE guidelines. SF FM was rejected in 6B, a classroom with a RT of 1.1 that also had the largest floor area (but not the largest volume) of the four rooms. Although this is a small sample, it suggests that poor classroom acoustics should be improved as far as possible before the introduction of SF amplification to ensure good long term use.

Following the production of the report for this study, the school approached the LEA and managed to obtain funding for all recommended classroom modifications except the suspended ceilings for 7A and 7B. It seems this falls between two departments. A SF FM system with 6 speakers was purchased for 6A and it is hoped that more systems will be obtained for the other classes as soon as possible.

References

Anderson, K.(1989). Screening instrument for targeting educational risk (SIFTER). Austin, TX: PRO-ED
Crandell, C., Smaldino J. & Flexer, C. (1995). Sound-field FM Amplification. Pub. Singular Publishing Group
DfEE Building Bulletin 87. Guidelines for environmental design in schools, Section A, Acoustics.
Finitzo-Hieber & Tillman, (1978). Room acoustics effects on monosyllabic word discrimination ability for normal and hearing-impaired children. Journal of Speech & Hearing Research, 21, 440-458. and in Crandell et al (1995)
Ross, Maxon & Brackett (1982) in Crandell et al (1995)