Assessing Acoustic Conditions in Hybrid Classrooms with COVID-19 Social Distancing at the University of Sharjah Hussein M. Elmehdi 1

Department of Applied Physics and Astronomy University of Sharjah Sharjah United Arab Emirates Ania Tato Iglesias 2 University of the Basque Country Bilbao, Spain

ABSTRACT In response to COVID-19 global pandemic, institutions are required to follow health protocols involving social distancing, which led to the need to redesign the teaching and learning spaces. The University of Sharjah (UOS), United Arab Emirates (UAE), has adopted a flexible hybrid teaching approach where the capacity of the classrooms is reduced to 50% and added the necessary audiovisual equipment to support the hybrid teaching model. As a result, the acoustic conditions are expected to be affected. The aim of this study is to assess the acoustic parameters in hybrid classrooms with a focus on reverberation time, sound clarity (C50), and strength (G). Measurements of the reverberation time, clarity, and strength were taken in six different classrooms were performed using a sound analyzer and ROOM EQ Wizard software in accordance with the guidelines set by ISO 3382. In addition, we compared the experimental data with the obtained from Sabine’s diffuse-field theory and from a model developed by E. Nilsson, which considers that the absorption in classrooms is mainly due to an absorbing ceiling. Reverberation times results show that Leq (dBA) varies from 43.0 – 50.9 dBA. Results obtained from the theoretical model showed the same overall behavior over the investigated frequency range. Sound clarity and strength measurements indicated that the acoustic conditions in hybrid classrooms do not satisfy the international standards with the exception of one classroom, which was acoustically treated.

1. INTRODUCTION

Institutions around the world had to switch to online teaching and learning early during the pandemic. However, institutions are gradually bringing back students to campus for face-to-face educational activities. Keeping in mind social distancing protocols, institutions are resorting to hybrid teaching

1 hmelmehdi@sharjah.ac.ae

2 Tato.ania20@gmail.com

models and approaches, which primarily rely on reducing the capacity of the classrooms by removing 40-60% of the seats. At the University of Sharjah (UOS), United Arab Emirates (UAE), the gradual return of students’ plans followed a Flexible Hybrid Model, which allows instructors to teach their classes on campus while broadcasting the lecture live (concurrent) via communication portals. This approach is highlighted by a number of attributes that have a direct impact on classroom acoustics. For instance, the reduction of seats in the classrooms affects the acoustic conditions and parameters like reverberation times [1,2,3]. In this study, we present a quantitative evaluation of the acoustic properties of flexible hybrid classrooms at the University of Sharjah (UoS). We will measure the reverberation time (T60) and clarity (C50) using an impulse response approach in selected classrooms. The results and calculations will be compared to the values obtained from two theoretical models: Sabine’s diffuse-field theory [4], applicable only to diffuse field conditions and a model developed by E. Nilsson et al. for Ecophon® Saint-Gobain, which takes into account that the absorption in classrooms is not uniform as it is mainly due to an absorbing ceiling [5]. Finally, we also calculated the sound strength, G, and compare it with the values obtained from Sabine’s and Nilsson’s model.

1.1. Sabine’s diffuse-field theory

Given the volume V, the total surface area S, and the average absorption coefficient a of a room, T 60 can be approximately calculated by Sabine’s formula [2],

0.163

𝑉

𝐴 (1) Where A is the total absorption coefficient of the room. In this paper, the reported values refer to the T 20 reverberation time at each octave center frequency. On the other hand, sound clarity (C50) is defined by ISO 3382-1 as the ratio of early-arriving energy to late-arriving energy [6]. Assuming a diffuse sound field and a perfect exponential decay, at distances well away from the sound source and for strength not much lower than 15 dB [5],

𝑇 60 =

𝑆.𝑎 = 0.163

−1−𝑒𝑥𝑝( −0.69

𝑇60 )

𝑇60 ) (𝑑𝐵) (2)

𝐶 50 = 10 log

log( −0.69

Finally, strength (G) represents the level at which sound is perceived. For a diffuse sound field [5],

𝑇 60

𝑉 (𝑑𝐵) (3)

𝐺= 45 + log

1.2. Nilsson’s model theory

In classrooms where all the absorption of the sound is concentrated in the suspended ceiling, diffuse conditions are not achieved [8]. The model proposed by E. Nilsson et al. [5] accounts for the non- uniformity of the sound field and thus provides better estimations of the acoustic parameters. According to these authors, the reverberation time, sound clarity, and strength are the following:

60

(4)

𝑇 60 =

(13.8 log( 1

𝑇𝑛𝑔 +𝐶.𝑒𝑥𝑝 ( 1

𝑇𝑛𝑔 )))

𝐶 50 = 10 log[(𝑑+ 𝑒 50 )/𝑙 50 ] (𝑑𝐵) (5)

𝐺= 10. log 𝑑+ 𝑒 50 + 𝑙 50 (𝑑𝐵) 6) T ng , C , and T g are intrinsic parameters that have to be estimated [5]. d , e 50 and l 50 refer to the energy of the direct sound, of the early reflected sound, and of the late reflected sound respectively. They depend on T ng , C , and T g . To calculate the parameters’ theoretical values we will use the Ecophon Acoustic Calculator [33]. To measure them, we will follow the methods described in ISO 3382. 2. MATERIALS AND METHODS

The classrooms assessed in this study were selected to represent the various classroom types used for lectures at the UOS and some of them are shown in Figure 1. Their characteristics are summarized in Table 1. The measurements were performed in empty occupancy conditions, except for an operator required for the measurement recording.

Table 1: Room characteristics

W10-007 W10-004 W10-110 W10-111 W10-106 W6-005 Surface area 507 m 2 322 m 2 326 m 2 232 m 2 268 m 2 315 m 2 Volume 526 m 3 352 m 3 303 m 3 200 m 3 240 m 3 335 m 3 No. of seats 47 40 41 30 48 32 Acoustic treatment

Absorbent

Absorbent

Absorbent

Absorbent

Absorbent ceiling, carpeted walls

Absorbent ceiling, one linoleum wall,

ceiling

ceiling

ceiling

ceiling

and floor

wooden

floor

The background noise was measured with a NOR140 Sound Analyzer (Norsonic AS, Norway) calibrated with the help of a Castle GA601 Single Level Class 2 Sound Meter Calibrator (Castle GROUP, UK) with an accuracy of ±0.1 dB for the equivalent sound pressure level. The measurements were carried out for three minutes in three different positions (front, middle, and back part of the classroom) and the results were averaged.

Reverberation time measurements were carried out following the guidance of ISO 3382-2. We used the integrated impulse response method with engineering precision. As a sound source, we used a non-omnidirectional speaker BX5 D3 5” Powered Studio Reference Monitor (M-audio, USA), which is consistent with ISO 3382-2. The microphone employed in these experiments was an omnidirectional Behringer UltraLinear Measurement Condenser Microphone ECM8000 (Behringer, Germany). The speaker and the microphone were connected to a portable computer via Steinberg UR12 (Steinberg Media Technologies GmbH, Germany) audio interface. Both speaker and microphone were positioned 1.5 m above the ground and the distance between them was always greater than 1.5 m, in compliance with ISO 3382-2 requirements. The experimental data was acquired using the software ROOM EQ Wizard (REW) and the theoretical values were calculated using Ecophon’s Acoustic Calculator which computes the numbers via Equations 1, 2, 3, 4, 5 and 6 [9].

Ssh. BRS CMR eas 7000 f(Hz) a)

59 (4B) (Hz) b)

Figure 1: Classrooms: (a) W10-007; (b) W10-004, similar to W10-110 and W10-111; (c) W10-

106; d) W6-005 3. RESULTS AND DISCUSSION

In Figures 2a and 2b we can observe that for the frequencies of 500 Hz, 1kHz, and 2kHz only the room W10-106 is consistent with the standards set for the reverberation time and sound clarity in classrooms and lecture halls [10, 11]. It should be noted that W10-106 is the classroom that has been acoustically treated with sound absorption panels and carpeted. This highlights the effectiveness of the acoustical treatment in this room and the poor acoustic conditions in the rest of the classrooms.

Figure 2: a) Reverberation, b) sound clarity and c) sound strength against frequency With respect to sound strength, we can infer from Figure 2c that the values obtained for G are well above the recommended 19 dB. In the classrooms with no special acoustic treatment, this is due to the rooms being reverberant and possessing highly reflective walls with nearly no absorbent materials on them. The classroom W10-106 presents a lower G than the rest, but it is still above the recommended values. This should be further investigated, as we expected a more effective reduction of the strength. Besides, we compared the experimental data with the two proposed models. We were expecting a good correlation of the experimental data with Nilsson’s model, but this is only achieved in classroom W10-106, as represented in Figures 2d, 2e and 2f. Due to the absorbent materials, in this classroom

the sound field is diffuse so Sabine’s theory and Nilsson’s model coincide for frequencies above a few hundred hertz.

GB) (Hz)

Figure 2: c) sound strength against frequency. Comparison of the theoretical and experimental data

of the d) reverberation time, e) sound clarity and f) strength in the classroom W10-106. Finally, it has to be mentioned that the background noise in all the classes ranges from 43.0 – 50.9 dBA, which complies with the standards set by some international organizations [12]. 4. CONCLUSIONS

It has been illustrated that the classrooms at the UOS do not comply with general international standards for acoustic comfort in learning and teaching centers. The only classroom that is proximate to fulfill these standards is W10-106, due to the absorbent materials placed on the walls and floor. Thus, this classroom should be taken as a reference for future renewals and improvements of other rooms at the University if the acoustic quality is to be enhanced. On the other hand, as diffuse field conditions are not generally achieved in the measured classrooms due to the absence of absorbent materials, we observed great discrepancies between Sabine’s diffuse field theory and the experimental data. We expected Nilsson’s model to describe the obtained data, as our measurements were performed in rooms where all the absorption is concentrated on the absorbent ceiling, but we could not find a good correlation between this model and the experimental data in most classrooms. The reason behind this is that we do not have access to the absorption coefficient of the ceiling, so we cannot perform the calculations correctly. Research on the determination of the absorption coefficient of the ceilings and the correct implementation of Nilsson’s model is left for future work. 5. ACKNOWLEDGEMENTS

Ty9(9) oberg W10-106 —— Sabine ~~~ Nilsson 1000 f(z) d)

The authors are very grateful to the Vice Chancellor Office for Research and Graduate Studies and the International Association for the Exchange of Students (IAESTE) for partially funding the research project. We are also gratefully thank the Career Advising and Student Training Office at the University of Sharjah for providing logistics for IAESTE Internship.

Cso (dB) W10-106 —— Sabine -- ~~ Nilsson 1000 f (Hz) e)

6. REFERENCES

1. Yang, Z., Becerik-Gerber, B., and Mino, L. A study on student perceptions of higher education

classrooms: Impact of classroom attributes on student satisfaction and performance. Building and Environment, 70:171–188, (2013). 2. Elmehdi, H.M., Alzoubi, H.A. and Lohani, S. Acoustic Quality of University Classrooms: a

Subjective Evaluation of the Acoustic Comfort and Conditions at the University of Sharjah Classrooms. Universitätsbibliothek der RWTH Aachen, (2019). 3. Elmehdi, H.M. and Sanjinez, M. Evaluation acoustic quality in lecture halls using onsite

measurements and survey questionnaire. In INTER-NOISE and NOISE-CON Congress and Conference Proceedings, volume 261, pages 1749–1755. Institute of Noise Control Engineering, (2020). 4. Everest, F. A. Master Handbook of Acoustics. McGraw-Hill Companies, (2001). 5. Nilsson, E., Marbjerg, G., Brunskog, J and Jeong, C.H.. Development and validation of a

combined phased acoustical radiosity and image source model for predicting sound fields in rooms. The Journal of the Acoustical Society of America, 138:1457–1468, 2015.6 6. ISO/TC 43/SC 2 Building acoustics. ISO 3382-2, 2008. URL: https://www.iso.org/obp/

ui/es/#iso:std:iso:3382:-2:ed-1:v1:en. (2008). 7. Harvie-Clark, J., Dobinson, N and Larrieu, F. Use of G and C 50 for classroom design. In Forum

Acousticum, (2014). 8. Harvie-Clark, J. and Dobinson, N. The practical application of G and C 50 in classrooms. In

INTER-NOISE and NOISE-CON Congress and Conference Proceedings, volume 247, pages 1510–1520. Institute of Noise Control Engineering, (2013). 9. Ecophon Saint-Gobain. Ecophon acoustic calculator, 2020. URL: https://www.ecophon.com/exp/e-tools/ecophon-acoustic-calculator/ ecophon-acoustic- calculator2/. 10. Berglund,B., Lindvall, T., and Schwela, D. H. World Health Organization. Occupational, and

Environmental Health Team. Guidelines for community noise, (1999). 11. American National Standard Institute. ANSI S12. 60.2010. Acoustical performance criteria,

design requirements, and guidelines for schools, part 1: Permanent schools, (2010). 12. NBR12179. Brazilian standard for acoustical treatment of closed rooms, (1992).