A case study of acoustic intervention in classrooms

Abstract

High reverberation times (RTs) have always been an acoustic barrier to effective learning in classrooms. Acoustic corrections to reduce RT involve complex acoustic treatment. Previous studies have indicated that classrooms in most schools do not meet the established acoustic criteria, as the school authorities refrain from such acoustic treatment. Aim of the study was to optimize the RT within classrooms through easily-implementable acoustic corrections. Different combinations of acoustic corrections have been experimented in eight classrooms, through a step-by-step approach to optimize RT. After each acoustic modification, the RT was measured and the speech clarity parameter C₅₀, was estimated. At the final step, RT of the classrooms was diminished to a mean value of 0.74 s (standard deviation = 0.04) from the initial mean value of 4.37 s (standard deviation = 0.42). C₅₀ values corresponding to the final acoustic correction were found to fall within good speech intelligibility scale.

Keywords

Classroom acoustics easily-implementable acoustic corrections reverberation time speech clarity

Introduction

Numerous previous studies have investigated and documented the difficulties experienced by students in correctly perceiving speech amidst acoustic barriers within a classroom.^1–4 One example of a major acoustic barrier that alters speech perception within a classroom is reverberation.⁵ The reflection of sound from the walls, floor, and ceiling of closed rooms, including classrooms, lead to the prolongation of sound, known as reverberation. These reflected sounds reach the listener a few seconds following the direct sound and result in its persistence. These delayed reflections mask direct sounds, thereby affecting the perception of speech.⁶ For example, the perception of consonant sounds will be affected due to the prolongation of preceding vowel sounds. Controlling these reverberations is essential to meet the expected acoustic quality in a classroom, since these interferences will have an adverse impact on a student’s academic achievement; hence, they are a very important consideration in defining the acoustical environment of a classroom.

The reverberation within a room is usually expressed by Reverberation Time (RT₆₀), which is defined as the time taken (expressed in seconds) for the sound pressure level (SPL) of a sound originating from a source to reduce by 60 dB after the source has abruptly stopped.⁷ After reviewing several previous research investigations, Hodgson and Nosal⁸ commented that speech intelligibility is inversely related to the RT. Nabelek and Robinson⁹ found that, irrespective of age, the percentage word recognition score reduces with an increase in the RT. The effect of RT on speech intelligibility is addressed in the standard document American National Standards Institute (ANSI) S.12.60,² which specifies that the optimal RT is 0.6 s for classrooms whose volumes are less than 283 m³ and 0.7 s for classrooms whose volumes are between 283 m³ and 566 m³. Optimal values of the RT according to Indian standards¹⁰ are 0.75 s for an occupied classroom and 1.25 s for an empty classroom. Meanwhile, permissible values of the RT in primary classrooms according to the respective standards of countries such as France, the Netherlands, Sweden, Norway, Portugal, the USA, Great Britain, Australia, New Zealand, and Finland are given in Mikulski and Radosz.¹¹

Mikulski and Radosz¹¹ assessed RT values and the speech transmission index for 110 classrooms in five typical primary schools in Warsaw and discovered that many of those classrooms did not meet the established acoustic criteria defined by building bulletin¹² 93. Acoustic criteria considered in their study, were speech transmission index (STI) and RT. The optimal value of RT was 0.6 s for primary schools and 0.8 s for secondary schools. The minimum permissible value for STI was 0.6 s. Losso and Viveiros¹³ conducted an acoustical measurement evaluation of schools in Southern Brazil and observed RT values ranging from 1.15 to 1.68 s, which well exceeded the recommended limits. Seidel et al.¹⁴ examined the RT within 65 classrooms of 23 different schools in Germany and found that in many classrooms, the observed values surpassed the limits specified by the German Institute for Standardization (DIN) 18041. Iannace et al.¹⁵ measured the RT of seven classrooms located within historical buildings in Italy, whereupon the RT was observed to be substantially high. RT values estimated by Sundaravadhanan et al.¹⁶ for 23 classrooms within four governmental primary schools in southern India were found to be greater than 2.6 s in all of the classrooms, which were again found to be higher than the optimal values specified by the Bureau of Indian Standards.¹⁰ All of these previous studies have demonstrated that the RTs in the majority of investigated classrooms were higher than the optimal levels specified by the respective national or international standards. As a consequence, the RT of those classrooms necessitates a reduction to enhance the quality of speech reception and academic performance.

Sabine¹⁷ devised the following formula to estimate the RT value:

{RT}_{60} = \frac{0.161 V}{\sum S \propto}

(1)

where “V” is the volume of the room (in cubic meters), “S” is the surface area (in square meters) of the various materials that form the surfaces of the room, and “α” is the absorption coefficient (at a given frequency) of each of those materials. According to this formula, there are two possible options for reducing the RT: either through reducing the room volume or by increasing the product value of “Sα.” The latter can be accomplished either by increasing the surface area by adding materials into the room or by treating the reflecting surfaces with acoustic materials, which possess higher absorption coefficients. Treating walls and ceilings with acoustic dampening materials necessitates enormous additional costs, and as a consequence, it is difficult to obtain the appropriate funding from school management. Sundaravadhanan et al.¹⁶ reported that classrooms in India were scarcely equipped with acoustical treatment. Iannace et al.¹⁵ attempted to reduce the RT within classrooms by adding sound-absorbing panels composed of shredded flakes (40 mm × 10 mm × 3 mm) of giant reeds. However, this solution is not easily implementable, as the material used is tailor-made and thus not widely available. John et al.¹⁸ attempted to diminish the RT within 12 unoccupied classrooms in a government-operated higher secondary school situated in the Kerala State of southern India by making use of mats composed of coir, screw-pine, and bamboo. While these efforts were successful in reducing the values of the RT, they were not sufficiently successful at confining the RT within the optimal values. Moreover, the corrections were not easily implementable by the school authorities. Choi¹⁹ made an attempt to arrive at an optimum combination of acoustic treatment with different diffusing and absorbing materials on the walls and ceiling for achieving the optimal RT in a 1/10 scale model classroom. RT was brought down to 0.67 s (T₃₀ averaged over 4 octave bands from 500 Hz to 4 kHz) by treating the entire ceiling with absorptive material and by treating the front wall with diffusing material. The materials used for this treatment were not widely available. Labia et al.²⁰ attempted to arrive at an optimal placement of acoustic panels on the walls and ceiling in two meeting rooms, one room of volume 520 m³ and another room of volume 298 m³. Among the various combinations experimented, the optimal configuration was the one with absorptive material on the ceiling and upper part of one of the side wall and rear wall. However, the experiment was based on simulations and hence the solution was not easily implementable. It is therefore apparent from these previous studies that, to increase the “Sα” of a classroom, widely available and easily implementable materials must be employed, the implementation of which needs to be investigated and validated. Consequently, the aim of the present study was to find such appropriately effective acoustic modifications for classrooms that fit these criteria to optimize a decline in the RT value and thereby improve the acoustic quality.

Quoting previous studies^21,22 Choi¹⁹ stresses the significance of early reflections in achieving the required speech intelligibility in classrooms. Bradley²³ stated that the clarity measure C₅₀ – representing the ratio of early reflections to late reflections at an early time interval of 50 ms – is the most preferred measure of clarity for speech sounds. The basis for C₅₀ is the fact that early reflections (within the time interval of 50 ms) contribute positively to the intelligibility whereas late reflections may not. C₅₀ is influenced by the location of the speaker, the amount of absorption at the listener, distance between the speaker and the listener and the presence of reflecting surfaces close to the speaker or sound source.²⁴ Clark and Dobinson²⁵ opined that the RT alone may not be sufficient to measure the acoustic quality within classrooms and that speech intelligibility and loudness are equally important in determining the acoustic response. They also suggested that the speech clarity parameter C₅₀ correlates better with loudness and intelligibility and thus proposed the parameter C₅₀ as a more comprehensive measure of the acoustic quality in classrooms. Campbell et al.²⁶ observed that an equivalent reverberation time in two identical rooms does not necessarily indicate equivalent levels of speech clarity and that a high speech clarity is as equally important as a low RT value toward achieving acoustic comfort in classrooms. Labia et al.²⁰ in their study focussed on improving the acoustic quality of two meeting rooms of medium size, observed RT and C₅₀ to be more sensitive compared to STI. Ansay and Zannin²⁷ suggested that an evaluation of the acoustic quality of a classroom should utilize the Definition parameter (D₅₀) in addition to the RT. D₅₀ as defined by Fasold and Veres²⁸ is the ratio of the early received sound energy (0–50 ms after direct sound arrival) to the total received energy. Since C₅₀ is related to D₅₀, C₅₀ is assessed during the present study to validate and quantify the resulting acoustic comfort values obtained from experiments that are conducted with the ambition of lowering the RT value within classrooms.

The objectives of this study were to (1) Implement acoustic corrections in eight empty classrooms, each of size 11 m × 8.28 m × 3.6 m, using widely available materials through a step-by-step approach, (2) Measure the RT value according to the standard procedure in each of these eight classrooms following each individual step during the acoustic correction, (3) Estimate the speech clarity parameter C₅₀ at each of these steps to quantify the improvement in acoustic quality and (4) Analyse the results and derive a easily-implementable solution, which will optimize the RT within classrooms to achieve acceptable speech clarity.

Method

Location

Eight classrooms located in the Academic Block of a higher education institute at Mysuru, India, each 11 m × 8.28 m × 3.6 m in size, were selected for the study. The classrooms did not possess false ceilings; the walls were painted with an emulsion paint and were plain and smooth. The floors were paved with mat-finish vitrified tiles, each of size 0.38 m × 0.38 m. The classrooms selected for the study did not possess any furniture or sound-treated walls, ceilings or floors. The layout of the investigated classrooms is shown in Figure 1.

Figure 1.

Layout of classroom.

Materials

Two types of window curtains were considered for this study, one of which (0.33 kg/m²) was composed of polyester mixed with cotton at a ratio of 60:40. The width of the stitched curtain was sustained at twice the width of the window (Window 1:1.8 m wide, Windows 2 and 3:1.2 m wide) so that the curtain possessed a greater number of folds (Figure 2(a)).

Figure 2.

(a) Curtains composed of polyester-cotton fabric blend. (b) Curtains composed of handloom fabric with calico lining.

The second type consisted of curtains (0.50 kg/m²) composed of a handloom fabric lined with calico cloth (Figure 2(b)), which were also maintained at twice the window width. Calico is a plain-woven textile made from cotton that is not fully processed and is unbleached. To increase the surface area, length of the second type of curtains was extended beyond the bottom of the window till the skirting level. Thirty-nine peacock chairs with seat and back rests upholstered with cloth cover (Figure 3) were set out throughout the classroom. In addition, a 10 mm thick soft acrylic carpet was spread out over the empty spaces of the classroom floor (Figure 4) for this study. The mean absorption co-efficient shown in Table 1 is calculated by taking the average of values of sound absorption coefficients at 250 Hz, 500 Hz, 1000 Hz, and 2000 Hz for each material, extracted from Vorländer.²⁹

Figure 3.

Layout of chairs and tables within the classroom.

Figure 4.

Carpet spread out over the floor to cover empty spaces.

Table 1.

Mean absorption coefficient of materials used for acoustic modifications.

Material	Mean sound absorption coefficient -α-	Surface area per classroom
Curtains (0.33 kg/m²) composed of polyester-cotton blend (covering windows)	0.23	5.4 m² for window 1 3.6 m² for windows 2 & 3
Curtains (0.50 kg/m²) composed of handloom fabric with calico cloth lining (covering windows till skirting level)	0.56	7.2 m² for window 1 4.8 m² for windows 2 & 3
10 mm thick soft acrylic carpet	0.21	91.08 m²
Empty chair upholstered with cloth cover	0.77	0.3125 m² per chair
Table	0.10	0.54 m² per table
Student on chair	0.58	1 student per m²

Procedure

The RT was measured within each of the classrooms through the interrupted noise method, using a standard RT measurement setup as shown in Figure 5. To utilize this method,³⁰ white noise, which was switched on for a few seconds and then abruptly switched off, was employed as a sound source. The response of the classroom to this white noise was recorded by a Bruel and Kjaer (B & K) 4189 microphone coupled to a B & K 2250 Sound Level Meter. The source, a white noise generator, was embedded within a B & K 2734B power amplifier. B & K 4292L omni-power sound source was used as the transducer, and the Sound Level Meter provided the noise source trigger to synchronize the sound level measurements. The subsequent RT values were measured using B & K BZ-7228 building acoustic software.

Figure 5.

RT measurement setup for this study.

Figure 6 illustrates the layout of noise sources as well as microphone positions, at which measurements were collected. These positions were fixed according to the guidelines of International Organization for Standardization (ISO) 3382-2.⁷ The engineering method⁷ was employed for deciding the source microphone positions, since the objective of the data acquisition was to verify the acoustic performance. The engineering method recommends measurements of reverberation time with at least two source positions and at least six independent source microphone combinations. Two source positions (S1, S2) and three microphone positions (P1, P2, P3) were chosen accordingly (Figure 6), and the RT was measured for six independent source-microphone combinations. Source positions S1 and S2 are located on the platform where the teacher would be positioned while teaching. The subsequent arithmetic mean of the RT acquired for frequencies of 250 Hz, 500 Hz, 1000 Hz, and 2000 Hz was calculated¹¹ for each source-microphone combination (Table 2), following which the arithmetic mean of RT of the source-microphone combinations in the classroom were taken for analysis.

Figure 6.

Source (S1 and S2) and microphone (P1, P2, and P3) positions for RT measurement.

Table 2.

Mean values of the RT and of the speech clarity index (C₅₀) for each acoustic condition (N = 8).

Condition/acoustic modification	Parameter	Mean	Standard deviation	Median
Empty classroom without curtains, chairs or carpet	RT (s)	4.37	0.42	4.31
Empty classroom without curtains, chairs or carpet	C₅₀ (dB)	–7.4	0.4	–7.4
Empty classroom with polyester-cotton blend curtains	RT (s)	3.11	0.15	3.17
Empty classroom with polyester-cotton blend curtains	C₅₀ (dB)	−5.6	0.2	−5.6
Empty classroom with calico cloth-lined curtains	RT (s)	2.49	0.39	2.36
Empty classroom with calico cloth-lined curtains	C₅₀ (dB)	–4.5	0.2	–4.5
Empty classroom with chairs, tables and calico cloth-lined curtains	RT (s)	1.66	0.36	1.50
	C₅₀ (dB)	–2.2	0.5	–2.1
Empty classroom with chairs, tables, 10 mm soft carpet on floor, and calico cloth-lined curtains	RT (s)	1.10	0.04	1.08
	C₅₀ (dB)	0.3	0.3	0.3
Classroom with chairs occupied by students, tables, 10 mm soft carpet on floor, and calico cloth-lined curtains	RT (s)	0.74	0.04	0.72
	C₅₀ (dB)	2.0	0.5	2.3

To check for test–retest reliability, the RT was again measured in two of the eight classrooms 1 week after the initial experiment was conducted. The differences between the measurements were compared to determine whether the mean difference was within the tolerance limits of measurement error.

Mikulski and Radosz¹¹ opined that the RT does not always reflect a subjective evaluation of the acoustic properties of a room. Hence, the speech clarity parameter (C₅₀) was included in the conducted research. The value of C₅₀ was estimated by the following formula, which was developed by Ahnert and Schmidt³¹:

C_{50, E} = 10 . \log_{10} \frac{γ_{s} {(\frac{r_{H}}{r_{x}})}^{2} + 1 - e^{- \frac{0.69}{T}}}{e^{- \frac{0.69}{T}}} d B,

(2)

where r_x represents the distance between the source and a listener’s seat in meters, γ_s is the front to random factor of the source, which is equal to 1 in this study, and r_H denotes the half-room diffuse-field distance, which is defined as follows:

r_{H} = 0.057 \sqrt{\frac{V}{T}}

(3)

where V is the volume of the room in cubic meters, and T is the value of the RT in seconds. Based on a diffuse statistical sound-field structure, the known room volume V and the reverberation time RT₆₀, C50 was computed as a function of the distance between the sound source – listener seat (r_x). As C₅₀ is a source-to-receiver distance-dependent parameter, it should be measured in each position or in regular areas of the room. Accordingly the present study computed C₅₀ at three positions in the room (P1, P2, P3) with source at positions S1, S2 as indicated in Figure 6.

Results

The mean RT for each of the eight classrooms considered within this study was observed to be substantially high (mean = 4.37, SD = 0.42) when the classrooms were empty, without any acoustic modification (Table 2).

Efforts were made to reduce the RT initially by adding curtains to the three windows of the classroom. When polyester-cotton blend curtains were added, the mean RT was lessened to 3.11 s [decreased by more than 5 Just Noticeable Difference (JND)]. JND for RT is ±5% as per ISO 3382-1:2009(E).³² The RT value was diminished to a mean value of 2.49 s (decreased by more than 3 JND) when handloom curtains with a calico cloth lining were used in place of the polyester-cotton blend curtains.

Furthermore, 39 chairs upholstered with cloth covering the seats and backrests, which were complemented by wooden tables (with sizes of 0.9 m × 0.6 m), were added to the classrooms, whereupon a further improvement in the RT (mean = 1.66, decreased by more than 6 JND) was observed. In addition, soft acrylic rugs with thicknesses of 10 mm were spread throughout empty spaces in the classroom, following which a further improvement of the RT (Mean = 1.10, decreased by more than 6 JND) was observed. After these modifications, students were allowed to sit in each of the classrooms, whereupon they occupied all of the 39 chairs. The measured RT for the occupied classrooms demonstrated further improvement, bringing down the RT to a mean value of 0.74 s (decreased by more than 6 JND), which is very close to the optimal value of 0.7 s as per ANSI Standard² and well within 0.75 s according to Indian Standards.¹⁰ Mean value of the measured RT for the occupied classrooms without acoustic corrections was 3.72 s (SD = 0.36).

Normality was tested using Shapiro-Wilk’s test,³³ which demonstrated that the five conditions characterized by an empty classroom were defined by normally distributed RT values (p > 0.05). Consequently, the first five conditions were compared using a parametric test, namely, the Repeated Measures Analysis of Variance (ANOVA). Significant differences in the RT values were observed between the acoustic modifications [F(4,28) = 613.99, p < 0.001, partial η² = 0.989]. F ratio is the ratio of two mean square values. The p value is the probability of getting a result at least as extreme as the one that was actually observed. Partial η² is a measure of effect size and reflects the percentage of the variance in the dependent variable explained by the independent variables in a sample. Pair wise comparisons (i.e. between the five conditions) were administered using an adjusted approach of Bonferroni’s multiple comparison test, which confirmed that the five conditions were significantly different from one another (p < 0.05). Next, since the last condition (i.e. for occupied classrooms) was not normally distributed, it was compared (via a pair wise comparison) with the other conditions using a nonparametric test, namely, Wilcoxon’s Signed Rank test.³⁴ The application of this test showed a significant difference between the final condition and the other five conditions (p < 0.05).

To check the test–retest reliability, the initially measured RT values were compared with measurements taken a week following the original experiment within two of the eight classrooms. The errors were observed not to exceed ±0.05 s.

The corresponding values of the clarity parameter (C₅₀) were also estimated for each of the eight classrooms considered for this study. The mean C₅₀ values for each acoustical modification are shown in Table 2. JND for C₅₀ is ±1 dB as per ISO 3382-1:2009(E).³² The first intervention (polyester-cotton blend curtains) resulted in increase of C₅₀ by more than 1 JND, the second one (calico cloth-lined curtains) by more than 1 JND³¹ and the third (chairs and tables) and fourth (10 mm acrylic carpets) interventions each resulted in increase of C₅₀ by more than 2 JND. An increase of C₅₀ by more than 1 JND was observed when the modified classroom was occupied by students. Normality was tested using Shapiro-Wilk’s test, which similarly demonstrated that the first five conditions (i.e. for unoccupied classrooms) possessed characteristic C₅₀ values that were distributed normally (p > 0.05). The fifth condition had a high standard deviation, and the sixth was not normally distributed. Therefore, these two conditions were compared with the first four conditions (as a pair wise comparison) using the nonparametric Wilcoxon’s Signed Rank test, while the first four conditions were compared with each other using the parametric Repeated Measures ANOVA test. The ANOVA test showed a significant difference between the four conditions [F (3,21) = 475.82, p < 0.001, partial η² = 0.986]. Pair wise comparisons, which were administered using an adjusted approach of Bonferroni’s multiple comparison test, between the same four conditions also demonstrated that the four conditions were significantly different from one another (p < 0.05). The fifth and sixth conditions were compared with the other four conditions using Wilcoxon’s Signed Rank test, which confirmed a substantial difference between all of the pairs (p < 0.05).

Discussion

Acoustic corrections in classrooms

This study implemented a series of acoustic corrections within eight empty classrooms, each characterized by a size of 11 m × 8.28 m × 3.6 m, which were located in the Academic Block of a higher education institute at Mysuru, Southern India. The volume of each of the classrooms considered in the present study is 364.32 m³; hence, the optimal reverberation time considering the ANSI S 12.6-2010 standards is 0.7 s. An observed mean RT of 4.37 s for the empty classrooms was lowered to 0.74 s through acoustic corrections accomplished with easily-available materials. The acoustic corrections were experimented in one classroom. After achieving satisfactory results, the same was implemented in the other seven classrooms, following a step by step approach.

The acoustic corrections by Iannace et al.¹⁵ could achieve an RT (averaged over 125 Hz–4000 Hz) of 2.13 s only. Choi¹⁹ experimented with 13 different configurations of absorbing and diffusing treatments on the walls and ceiling of a model classroom. Even though they could achieve an RT of 0.67 s (averaged over 500 Hz–4000 Hz), they had to add absorptive material for the entire ceiling and diffusive material for lower front wall. Moreover, their investigation was in a scaled down (10:1) model classroom. The present study could bring down the RT to 0.74 s without any treatment for the walls and ceiling in an actual classroom.

RT values obtained after each successive acoustic correction step

The present study measured RT values for each of the eight classrooms using the measurement setup (Figure 2) and procedure prescribed by relevant standard,⁷ unlike a similar previous study¹⁶ wherein the RT values were calculated for classrooms using Sabine’s formula. As a consequence of the experimental methodology, the corresponding improvements in RT after each acoustic correction were able to be accurately recorded. A mean RT value of 4.37 s observed for the unoccupied classrooms is similar to that obtained from previous studies.^{11,13,14,16,18} The corrections were implemented sequentially using widely available materials through a step-by-step approach. The first step in the correction process, wherein polyester-cotton blend curtains were added to the three windows, helped to reduce the RT to 3.11 s. The curtains were sustained at twice the width of the windows and hence, the absorption surface area was increased by 12.60 m². The second acoustic correction, wherein the polyester-cotton blend curtains were replaced with handloom fabric curtains with a calico cloth lining, lessened the RT to 2.49 s. As these curtains were up to the skirting level, the absorption surface area was increased from 12.60 m² as in the first step to 16.80 m². Shoshani and Rosenhouse³⁵ found that the microstructure of the fabric substantially influences the sound absorption coefficient. The higher mean absorption coefficient (0.56) for the handloom fabric curtains, compared with that of the polyester-cotton blend curtains (0.23), as shown in Table 1 along with the increase in absorption surface area has resulted in a proportionate increase in sound absorption and respective reduction in RT.

An addition of 39 chairs, each of which was outfitted with cotton fabric on the seat (0.4 m × 0.5 m) as well as on the back rest (0.45 m × 0.25 m), further increased the surface area of absorbing materials by 12.19 m². The addition of 39 tables (0.9 m × 0.6 m), which were equipped with a plywood top overlaid with a veneer, also enlarged the surface area of available absorbing materials by 21.06 m². This increase in absorption due to the combined effect of tables and chairs (Table 1) resulted in further diminishing of the RT to 1.66 s, which is likely the consequence of the tables and chairs preventing the sound from reflecting off of the floor. Similarly, to further curtail the RT, the empty spaces within the floor (91.08 m²) were covered with absorbing carpets, from which a further reduction of the RT to a mean value of 1.10 s was detected. When each of the 39 chairs was occupied with a student, a significant reduction in the RT to a mean value of 0.74 s was observed, due to the sound absorption (Table 1) by the students.

While investigating the guidelines to optimize classrooms for learning enhancement, SanSouci³⁶ discovered that excessive reverberation within classrooms can inhibit learning. Castro-Martíne³⁷ observed that specific changes, aimed at decreasing reverberation values in classrooms affect positively the levels of attention and student performance. According to Astolfi et al.³⁸ optimal reverberation in the learning environments has a crucial role, as it contributes to preserve the amplitude modulation of the speech produced by the teacher which will prevent the student from potential reading disorders. Puglisi et al.³⁹ observed that, inadequate acoustics in classrooms negatively influences the reading development in children. The acoustic corrections investigated for the present study, which resulted in bringing down the initially excessive reverberations to an optimal level, are thus expected to ensure an enhancement of learning within classrooms.

Speech clarity parameter C₅₀ after each successive acoustic correction step

The study also investigated the corresponding changes in the clarity measure (C₅₀) after each step in the acoustic correction process, as suggested by Ansay and Zannin.²⁷ Many of the previous studies^{25,26,40–43} indicated that the RT alone cannot describe the acoustic quality within a closed space. Accordingly, Marshall⁴⁴ proposed a subjective rating scale that linked C₅₀ values to speech intelligibility, as shown in Table 3. C₅₀ is measured with a source with directivity that approximates the one of the human mouth and torso, and should be measured in fixed points in the room. In the present study C₅₀ was estimated at three positions P1, P2, P3 with source at positions S1, S2 as indicated in Figure 6. Mean values of the clarity measure C₅₀ after each step of the acoustic correction procedure demonstrated that the reductions in the RT resulted in corresponding improvements of the C₅₀ value.

Table 3.

Range of potential C₅₀ values and a subjective rating scale for speech intelligibility.⁴⁴

Range of C₅₀ values (dB)	Subjective rating scale
−12 to −7	Bad
−7 to −2	Poor
−2 to +2	Fair
+2 to +7	Good
Above 7	Excellent

Figure 7 illustrates the distribution of C₅₀ values corresponding to each modification in each of the classrooms plotted against the subjective rating scale for speech intelligibility proposed by Marshall.⁴⁴ It is evident that the C₅₀ values estimated for the sixth condition in most of the classrooms fall within the good speech intelligibility region, whereas the intelligibility was poor prior to the acoustic corrections. This improvement in speech intelligibility could be explained by the inference of Choi¹⁹ that, reduction in sound absorption and enhancement of early reflections (as indicated by higher C₅₀ values in the present study) would result in high speech intelligibility. Speech clarity (C₅₀) was found to be significantly correlated with all the investigated reading tasks, in an investigation done on the influence of classroom acoustics on the reading speed, in classrooms of second grade students.³⁹

Figure 7.

Speech clarity plotted against the subjective rating scale.

Ease of implementation of the acoustic corrections

During an investigation of the acoustic quality of school buildings, Losso and Viveiros¹³ found that most of the necessary acoustic modifications were not implemented primarily as a consequence of the cost and time involved. The modifications investigated within the present study can be implemented easily and at a nominal cost and should thus encourage school authorities to apply the corrections in a timely fashion. The expenditure incurred for the acoustic modifications is shown in Table 4, which demonstrates that the additional cost involved in implementing the acoustic modifications is only USD 1110.10 for a classroom that can be occupied by 39 students. Only the difference in cost between regular chairs and upholstered chairs was considered, as regular chairs would any way be a part of the regular classroom. The expenses represented by the tables were not included in the cost estimation because they are essential elements in a classroom. The intervention in the present study is easily implementable compared to the interventions in similar studies by Iannace et al.¹⁵ and Choi.¹⁹ The materials employed during the study were curtains, carpets and upholstered chairs with cloth covering the seat and back rests. These materials are readily available compared to the tailor-made materials utilized by Iannace et al.¹⁵ and John et al.¹⁸

Table 4.

Cost of the acoustic modifications.

Material	Quantity per classroom	Cost (USD) per classroom
Handloom fabric curtains with calico cloth lining	16.80 m²	165.45
Soft acrylic carpet of 10 mm thickness	91.08 m²	424.65
Difference in cost between regular chairs and upholstered chairs	39 numbers	520.00
Total		1110.10

Use of heavy curtains may shade the sunlight. However, the classrooms under study are fitted with multimedia projectors which demands reduced ambient light. Thus the curtains are closed when the classrooms are used for teaching and learning. Likewise, using upholstered chairs and carpets may raise questions regarding the cleanliness of the classroom. The acoustic solutions identified in the study were implemented 3 years back and are in regular use since then. No issues regarding cleanliness have been reported so far.

Limitations of the study

The present study focused on the RT, which is only one of the acoustic barriers in a classroom. Background noise inside of a classroom represents another barrier, especially for primary school children. For example, a teacher’s voice will be affected if the background noise level is high. This study did not investigate this particular acoustic barrier. However, the level and impact of background noise within 23 classrooms of four government primary schools in southern India were recently investigated by Sundaravadhanan,¹⁶ who recently suggested several methods to reduce background noise and thereby enhance the speech to noise ratio.

Conclusion

The acoustic conditions of the classrooms in this study prior to acoustic corrections were unable to meet either national or international standards for the reverberation time. The present study succeeded in discovering a acoustic solution to overcome the acoustic barrier represented by the RT and therein optimize the RT as specified according to respective standards. By adding handloom curtains with a calico cloth lining to the classrooms, by using wooden chairs upholstered with cloth cover, and by providing carpets on the floor, the RT within occupied classrooms was diminished very closely to the optimal value of 0.7 s specified in ANSI S 12.60-2010.² The observed RT values after acoustic corrections were also well within the optimal value of 0.75 s set by Indian standards¹⁰ for an occupied classroom. Corresponding speech clarity index (C₅₀) estimates showed that by adopting these modifications, the intelligibility in the classrooms could be raised to a level required for the good intelligibility of speech, thereby removing the adverse impact on the academic achievement of the students. The solution is both easily and quickly implementable, as it involves only stitching and installing curtains as well as laying of carpet. Thus, the present study has successfully achieved its objective of discovering a easily implementable acoustic correction to optimize the reverberation time in classrooms.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Ajish K Abraham

References

Acoustical Society of America. Classroom acoustics. Technical Committee on Architectural Acoustics, http://asa.aip.org. 2000.

ANSI S 12.60/Part 1, American national standard acoustical performance criteria, design requirements, and guidelines for schools, Part 1: Permanent Schools. 2010.

Bradley

JS.

Speech intelligibility studies in classrooms. J Acoust Soc Am 1986; 80: 846–854.

Rabelo

Santos

Oliveira

de Magalhães

MC.

Effect of classroom acoustics on the speech intelligibility of students. Codas 2014; 26: 360–366.

Crandell

Smaldino

JJ.

Classroom acoustics for children with normal hearing and with hearing impairment. Lang Speech Hear Serv Sch 2000; 31: 362–370.

Nabelek

Room acoustics and speech perception. In: Katz

(ed.) Handbook of clinical audiology. 4th ed. Baltimore, MD: Williams & Wilkins, 2011, pp.624–637.

International Organization for Standardization ISO 3382-2:2008. Acoustics measurement of room acoustic parameters. Part 2: reverberation time in ordinary rooms, 2008.

Hodgson

Nosal

Effect of noise and occupancy on optimal reverberation times for speech intelligibility in classrooms. J Acoust Soc Am 2002; 111: 931–939.

Nabelek

Robinson

PK.

Monaural and binaural speech perception in reverberation for listeners of various ages. J Acoust Soc Am 1982; 71: 1242–1248.

10.

Bureau of Indian Standards. National Buildings Code of India. https://bis.gov.in/index.php/standards/technical-department/national-building-code. 2016, p.7.

11.

Mikulski

Radosz

Acoustics of classrooms in primary schools results of the reverberation time and the speech transmission index assessments in selected buildings. Arch Acoust 2011; 36: 777–793.

12.

Building Bulletin 93. School Acoustics. 2003.

13.

Losso

Viveiros

Acoustical quality in educational buildings: measurements in Brazilian public schools. In: Proceedings of 11th international congress on sound and vibration (ICSV), St. Petersburg, Russia, 2004, pp.599–560.

14.

Seidel

Weber

Leistner

, et al. Acoustic properties in German classrooms and their effect on the cognitive performance of primary school pupils. In: Proceedings of Forum Acusticum 2005, Budapest, pp.2065–2069 Budapest: OPAKFI..

15.

Iannace

Trematerra

Acoustic correction using green material in classrooms located in historical buildings. Acoust Aust 2013; 41: 213–218.

16.

Sundaravadhanan

Selvarajan

McPherson

Classroom listening conditions in Indian primary schools: a survey of four schools. Noise Health 2017; 19: 31–40.

17.

Sabine

WC.

Collected papers on acoustics. Cambridge, MA: Harvard University Press, 1922.

18.

John

Premlet

Thampuran

Improvement of acoustic comfort in school classrooms of Kerala. Int J Adv Res Eng Technol 2015; 3: 21–26.

19.

Choi

Y-J.

An Optimum combination of absorptive and diffusing treatments for classroom acosutic design. Build Acoust 2014, 2, 175–180.

20.

Labia

Shtrepi

Astolfi

Improved room acoustics quality in meeting rooms: investigation on the optimal configurations of sound-absorptive and sound-diffusive panels. Acoustics 2020, 2, 451–473.

21.

Bradley

Sato

Picard

On the importance of early reflections for speech in rooms. J Acoust Soc Am 2003;113(6): 3233–3244.

22.

Bradley

Reich

Norcross

On the combined effects of signal-to-noise ratio and room acoustics on speech intelligibility. J Acoust Soc Am 1999; 106(4): 1820–1828.

23.

Bradley

JS.

Review of objective room acoustics measures and future needs. Appl Acoust 2011; 72: 713–720.

24.

Room acoustic descriptors – RT, C50 and Strength/Gain. Acoustic bulletin, https://www.acousticbulletin.com/room-acoustic-descriptors-rt-c50-and-gain. 2018.

25.

Clark

Dobinson

. The practical application of G and C50 in classrooms. In: Proceedings of Inter-noise, 42nd international congress and exposition on noise control engineering, Austria, 2013, p.207.

26.

Campbell

Svensson

Nilsson

. The same reverberation time in two identical rooms does not necessarily mean the same levels of speech clarity and sound levels when we look at impact of different ceiling and wall absorbers. Energy Procedia 2015; 78: 1635–1640.

27.

Ansay

Zannin

PHT

. Using the parameters of definition, D50, and reverberation time, RT, to investigate the acoustic quality of classrooms. Can Acoust 2016; 44: 6–11.

28.

Fasold

Veres

Schallschutz + Raumakustik in der Praxix (noise control and room acoustics). Berlin, Germany: Huss Medien Gmbh. Verlag, 2008 (in German).

29.

Vorländer

Auralization: fundamentals of acoustics, modelling, simulation, algorithms and acoustic virtual reality. 1st ed. Berlin, Germany: Springer-Verlag, 2008, pp.305–307.

30.

Horvat

Jambrosic

Domitrovic

. Methods of measuring the reverberation time. In: Proceedings of 3rd congress of the alps adria acoustics association, 2007. Alps Adria acoustics association.

31.

Ahnert

Schmidt

Fundamentals to perform acoustical measurements, http://renkusheinz-sound.ru/easera/EASERAAppendixUSPV.pdf. 2005.

32.

International organization for standardization ISO 3382-1:2009(E). Acoustics measurement of room acoustic parameters. Part 1: Performance Spaces, 2009.

33.

Shapiro

Wilk

MB.

An analysis of variance test for normality (complete samples). Biometrika 1965; 52(3–4): 591–611.

34.

Wilcoxon

Individual comparisons by ranking methods. Biom Bull 1945; 1(6): 80–83.

35.

Shoshani

Rosenhouse

Noise absorption by woven fabrics. Appl Acoust 1990; 30: 321–333.

36.

Sansouci

. Classroom acoustics: new principles for improved scholastic achievement. In: Acoustical Society of America 159th Meeting, 2010.

37.

Castro-Martínez

Effects of classroom-acoustic change on the attention level of university students. Interdisciplinaria. 2016; 33(2): 201–214.

38.

Astolfi

Puglisi

Minelli

, et al. Effects of classroom acoustics on potential struggling first-grade readers. In Proceedings of euronoise, Heraklion (Crete—GR). 2018.

39.

Puglisi

Prato

Sacco

, et al. Influence of classroom acoustics on the reading speed: a case study on Italian second-graders. J Acoust Soc Am 2018; 144: 2.

40.

Iannace

Ciaburro

Trematerra

. The acoustics of the holy family church in Salerno. Can Acoust 2020; 48: 25–31.

41.

Jianxin

Peng

The effects of the noise and reverberation on the working memory span of children. Arch Acoust 2017; 43: 123–128.

42.

Astolfi

Castellana

Puglisi

, et al. Speech level parameters in very low and excessive reverberation measured with a contact-sensor-based device and a head worn microphone. J Acoust Soc Am 2019; 145: 2540–2551.

43.

Subramaniam

Ramamurthy

Effect of mode of delivery and background noise on speech characteristics of talkers in a classroom environment. Build Acoust 2020; 27: 113–135.

44.

Marshall

LG.

Speech intelligibility prediction from calculated C₅₀ values. J Acoust Soc Am 1995; 98: 2845–2847.

A case study of acoustic intervention in classrooms

Abstract

Keywords

Introduction

Method

Location

Materials

Procedure

Results

Discussion

Acoustic corrections in classrooms

RT values obtained after each successive acoustic correction step

Speech clarity parameter C50 after each successive acoustic correction step

Ease of implementation of the acoustic corrections

Limitations of the study

Conclusion

Footnotes

Declaration of conflicting interests

Funding

ORCID iD

References

Speech clarity parameter C₅₀ after each successive acoustic correction step