Skip to main content
NCBI home page
Elsevier - PMC COVID-19 Collection logoLink to Elsevier - PMC COVID-19 Collection
. 2022 Dec 8;202:109149. doi: 10.1016/j.apacoust.2022.109149

Effects of face masks and acoustical environments on speech recognition by preschool children in an auralised classroom

Miji Kwon a, Wonyoung Yang b,⁎
PMCID: PMC9729249  PMID: 36510558

Abstract

The potential impact of mask-wearing specifically on early-childhood speech and language development in classrooms has not been widely reported yet, although face masks are compulsory even in educational settings during the COVID-19 pandemic. This study investigated the combined effects of face-mask usage (no mask, surgical and KF94 masks) and room acoustics (RT 0.6 s and 1.2 s, SNR 12 dB and 22 dB) on speech recognition (KS-MWL-P) in preschool children (N = 67) in realistic classroom-acoustic settings using the auralisation technique.

The face mask and reverberation time affected pre-schoolers’ speech recognition scores. Reducing RT in the classroom improved the pre-schoolers’ speech recognition that was reduced by face masks. Children aged 4 and 5 years were affected by face masks and RT more significantly than children aged 6 years.

Appropriate room acoustics for classrooms and clear speech of teachers are recommended for better speech recognition in preschool, where pre-schoolers’ language and speech development usually occur.

Keywords: Face masks, Speech recognition, Preschool children, Monosyllabic word list, WIPI test, Reverberation time, Noise, Classroom acoustics, Auralisation, COVID-19

1. Introduction

Owing to the recent COVID-19 pandemic, many countries are mandating the use of face masks in public [1], [2], [3] as they are effective in mitigating virus spread. [4], [5] Face masks are compulsory even in educational institutions. [6] Recent studies on the pandemic’s potential impact on child development have reported evidence of learning loss [7], negative mental health outcomes [8], and lowering of language skills. [9], [10] Nevertheless, how mask-wearing affects early-childhood speech and language development in educational institutions specifically, has not yet been widely reported. Sfakianaki et al. [11] reported a tendency of lower performance in word recognition among children (aged 6 years and 8 months to 7 years and 6 months) owing to wearing masks. Although the effects of face masks on speech recognition have been investigated in adults, to the best of our knowledge, no studies exist regarding pre-schoolers. Language-and-speech development usually occurs during early childhood as part of the gradual acquisition of receptive and expressive skills. [10] It is widely known that changes in language development are rapid up to the age of 3.5 years. Subsequently, the developmental process continues at a slower rate for the next few years. [12].

In adults, speech recognition was significantly lower with a mask, than without a mask. A surgical mask significantly reduced speech recognition in both, quiet and noisy environments. [11] Note that such a mask was reported to cause the least acoustic attenuation compared to other types of masks. [13], [14] Face masks reduced the sound levels by approximately 6–7 dB for frequencies between 100 Hz and 1.6 kHz and 7–13 dB for frequencies between 2 kHz and 5 kHz. [15] The speech transmission index (STI) at a distance of 2 m was decreased upon face-mask type. Speech levels in the octave band at 2 kHz or higher were also decreased when wearing face masks. [16] In acoustic measures, the mask tissue reduced amplitudes up to 8 dB at frequencies above 1 kHz, whereas no reduction was observed below 1 kHz. [17] Transparent masks could facilitate the ability to understand target sentences by providing visual information. [18], [19].

In addition to the transparent mask, the classroom-acoustic environment is considered helpful for speech recognition. In previous room-acoustic studies, it was identified that poor classroom acoustics, such as long reverberation time, high background-noise level, or low signal levels negatively affected speech perception and listening comprehension. This effect was more pronounced in younger children than in older children or adults. [20], [21] In one-talker masker, speech reception-threshold performance was estimated to be adult-like by 10–12.9 years of age. In two-talker masker, the performance was not projected to be adult-like until 16.1–16.8 years of age. [20] Thus, children need more favourable listening conditions than adults for decoding and processing oral information. [22] In addition, younger children (11–13 years), compared to older ones (15–17 years) and adults, were less able to use contextual cues to reconstruct noise-masked words presented in a sentential context. [23] Young children (5–7 years), unlike older children (10–12 years) and adults, were unable to recognise spectrally degraded speech. [24] The parameters for classroom acoustics were level of speech, level and characteristics of background noise and/or competing talkers, and reverberation time (RT) of the room. [25] Maximum background-noise levels of 35 dBA and a maximum RT of 0.6 s for typical medium-sized classrooms have been recommended. [26] In real classrooms, which are rarely noiseless, reverberance must be optimised rather than minimised. [27], [28], [29] Early studies reported that speech intelligibility decreased with increased RT, which indicated zero optimal reverberation for children and adults. [30], [31], [32], [33] However, these early experimental studies were unrealistic because they effectively assumed a diffused sound field, by involving exponential sound decays, without accounting for the interaction between reverberation and sound levels. [28] In contrast, theoretical studies reported non-zero optimum RTs for speech intelligibility. [34], [35], [36] In a relatively realistic study, noise was incorporated into a theoretical model, and non-zero optimum RTs for speech intelligibility were validated with children and adults. [27], [28], [29] Children in elementary schools were asked to participate in most studies. However, younger children (2–5 years), for whom language development is the most important, rarely participated in speech-recognition testing in classroom-acoustics studies. Young children (4–6 years), at high risk for language impairment, spend most of their day in classrooms. Here, acoustics have a more important role, particularly when face masks have been unavoidable in preschool.

This study’s purpose was to investigate the combined effect of face-mask usage and room acoustics on speech recognition in preschool children. We used the auralisation technique in realistic classroom-acoustic settings. The effects of RT and signal-to-noise ratio (SNR) were investigated in a classroom environment with children aged between 4 and 6 years, using face masks.

2. Methods

2.1. Participants

Sixty-seven children (4–6 years old) were recruited from two preschools in Gwangju, Korea, with parental consent. The children received toys and books to take home as compensation for their participation. The Institutional Review Board of Gwangju University approved the informed consent procedure. With no hearing tests performed at this stage, hearing-impaired students were excluded, as reported by their teachers and parents.

The Receptive and Expressive Vocabulary Test (REVT) [37], [38] for Koreans was used to evaluate the children’s language development. As the REVT factor was not a research topic in this study, the children identified as ‘delayed’ did not need to be excluded from data analysis for statistical validity. Further analysis was not performed on the REVT of this study. Table 1 lists the participant numbers according to age.

Table 1.

Description of the participants.

Age 4 yr
 5 yr
 6 yr
 Sum
 Total
Gender Female Male Female Male Female Male Female Male
Sum 10 9 17 11 10 10 37 30 67
REVT Normal 9 9 14 8 7 8 30 25 55
Delayed 1 – 3 3 3 2 7 4 12
Open in a new tab

2.2. Anechoic speech recording wearing face masks

The Korean-standard monosyllabic-word list for pre-schoolers (KS-MWL-P) [39], developed based on the international standard for speech audiometry [40] and word intelligibility by picture identification (WIPI) test [41], was used for the speech-recognition test. The KS-MWL-P consists of four lists, each of 25 words, and has 26 colour plates (one for practice) with six pictures per page. Depending on the situation, only ten words can be used. [42].

The lists were recorded by a professional voice-actress in a flat-walled fully anechoic chamber [43] (8.2 m × 7.0 m × 7.5 m,  = 50 Hz), using a class-1 sound level metre (RION NL-52). The microphone was 1 m away from the seated speaker. Three sets of 100 KS-MWL-P words were recorded based on mask types: no mask as the control, surgical mask (PURO GUARD Daily Mask, 175 mm × 90–155 mm, 3 layers of non-woven fabric, 50 pieces in a box, PURO Corp.), and KF94 mask (ECOLTER KF94 fine dust mask, 210 mm × 48 + 80 + 48 mm, 4 layers of non-woven fabric, individually packaged in plastic, EcoDream Corp.); as shown in Fig. 1 .

Fig. 1.

Fig. 1

Open in a new tab

Two different face masks used in the study (left: surgical, right: KF94).

A surgical mask is a loose-fitting disposable device that creates a physical barrier between the wearer’s mouth-and-nose and potential contaminants in the immediate environment. [44] Although a surgical mask with non-woven fabric fails to provide complete protection, many people prefer it for its breathability. [45] A KF94 mask is the “Korea filter” standard, equivalent to the N95 mask; 94 (%) refers to its filtration efficiency. [46].

2.3. Room acoustics in an auralised classroom

A typical preschool classroom (6.80 m × 8.00 m × 2.64 m) was chosen for this study, as shown in Fig. 2 . Two RTs (T30500Hz, 1kHz, 0.6 s and 1.2 s) were fitted at the listener’s position using ODEON 15.16 to change the surface materials. Table 2 presents the absorption coefficients and the scattering coefficients of the materials fitted for each RT. Table 3 lists the two fitted RTs throughout the octave frequency bands above the Shroeder frequencies (129 Hz and 183 Hz).

Fig. 2.

Fig. 2

Open in a new tab

Odeon room model with positions for the speaker (red) and listener (blue) in classroom.

Table 2.

Absorption coefficients and scattering coefficients of the surface materials in T30s.

T30 Material Area (m2) Absorption Coefficient
 Scattering Coefficient
707 Hz
250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz
0.6 s Floor 54.40 0.15 0.10 0.10 0.05 0.10 0.10 0.1
Ceiling 54.40 0.25 0.40 0.55 0.60 0.60 0.60
Wall1 + 3 32.77 0.30 0.20 0.17 0.15 0.10 0.10
Wall2 21.14 0.21 0.10 0.08 0.06 0.06 0.06
Walls 9.83 0.22 0.17 0.09 0.10 0.11 0.11
W1 9.72 0.06 0.04 0.03 0.02 0.02 0.02
W2 1.16 0.25 0.18 0.12 0.07 0.04 0.04
Doors 3.60 0.10 0.06 0.08 0.10 0.10 0.10
1.2 s Floor 54.40 0.01 0.015 0.02 0.02 0.02 0.02 0.1
Ceiling 54.40 0.30 0.2 0.17 0.15 0.10 0.10
Wall2 21.14 0.20 0.15 0.13 0.10 0.08 0.08
Wall3 11.63 0.02 0.02 0.02 0.02 0.02 0.02
Walls 30.97 0.14 0.09 0.06 0.05 0.05 0.05
Windows 10.88 0.05 0.03 0.03 0.02 0.02 0.02
Doors 3.60 0.10 0.06 0.08 0.10 0.10 0.10
Open in a new tab

Table 3.

Octave band RTs in T30s.

T30 250 Hz 500 Hz 1 kHz 2 kHz 4 kHz 8 kHz T30500Hz, 1kHz
0.6 s 0.56 0.60 0.58 0.58 0.57 0.45 0.59
1.2 s 0.79 1.14 1.30 1.41 1.47 0.88 1.22
Open in a new tab

After the reverberation fitting of the simulation models, the anechoic speech recordings were auralised in the two simulated classrooms, with RTs of 0.6 s and 1.2 s, respectively. The speech source had the directivity of a talking human provided by ODEON 15.16 (BB93. RAISED NATURAL.SO8). [47] Headphone transfer functions (Sennheiser HD600) and head-related transfer functions (KEMAR) were applied to the auralisation. The input parameters that were used include impulse response length of 1000 ms, number of late rays of 1000, and transition order of 2. Fig. 3 shows the impulse responses of the simulated classrooms.

Fig. 3.

Fig. 3

Open in a new tab

Impulse responses from ODEON simulations (a) RT = 0.6 s, (b) RT = 1.2 s.

2.4. Speech recognition test design and procedure

12 experimental configurations (3 masks × 2 RTs × 2 noise conditions) were designed for this study. Each auralised speech source was presented at 62 dBA through a headphone. A babbling sound of 50 dBA, which is natural in an actual classroom, was used as the background-noise source. As a control, no-noise conditions were tested in this study. The background-noise levels were monitored to be<40 dBA during the tests. Therefore, the two signal-to-noise ratios of the tests were 12 dB and>22 dB, respectively.

Data were collected between January and February 2022. Fig. 4 shows a photograph of the preschool-classroom test. Speech-recognition tests were conducted in a quiet preschool classroom, specifically allocated for testing. It was administered on a tablet pad and Sennheiser HD 600 headphones. Ten words were randomly selected for each experimental configuration using a program, specifically developed for this study. The 6 colour pictures (3 columns × 2 rows) were displayed on a tablet screen for children to point a correct word picture of six. No text was displayed on a tablet for children. Their responses were automatically saved in a database. A total of 12 sessions (120 words) were performed with each child. They performed a full test, divided into 2–4 sessions, based on their level of concentration. In general, each session lasted less than five minutes.

Fig. 4.

Fig. 4

Open in a new tab

Speech recognition testing.

Descriptive statistics were used to summarise and describe variables for the speech-recognition score. Based on the Anderson-Darling normality test, the data were not normally distributed. The data were analysed by applying a non-parametric statistical methodology. The Mann-Whitney U and Kruskal-Wallis tests were used to verify the effects of face masks on speech-recognition.

3. Results

3.1. Overall data analyses for the test score

Table 4 lists the results of the descriptive statistics. Speech-recognition scores varied according to the face-mask type, RT, and noise. Fig. 5 illustrates the average speech-recognition scores according to the face-mask type. Age, gender, and REVT, which are well-known factors affecting children’s language development, [48] were also observed.

Table 4.

Descriptive statistics results for the test scores.

Factor N Mean StDev Skewness Kurtosis
Mask No Mask 261 8.3 1.35 −1.85 5.07
Surgical 258 7.9 1.59 −1.34 2.41
KF94 256 7.7 1.66 −1.08 1.35
RT 0.6 s 390 8.2 1.40 −1.47 3.30
1.2 s 385 7.8 1.68 −1.26 1.77
Noise SNR > 22 dB 395 8.1 1.52 −1.45 2.69
SNR = 12 dB 380 7.9 1.59 −1.32 2.31
Age 4 yr 226 7.6 1.53 −0.82 0.32
5 yr 330 7.7 1.68 −1.56 3.00
6 yr 219 8.8 1.13 −1.25 1.15
Gender Girl 429 8.3 1.33 −1.09 1.16
Boy 346 7.7 1.76 −1.36 2.08
REVT Typical 632 8.6 1.35 −1.09 1.09
Delayed 143 7.4 2.10 −1.14 0.97
Open in a new tab

Fig. 5.

Fig. 5

Open in a new tab

Effects of wearing masks on speech recognition according to (a) RT and (b) age (with 95% Confidence Intervals).

The average children's speech-recognition scores varied according to the RT. The younger the children, the more they were affected. Children aged 4–5 years were affected, but those aged 6 years showed no difference in the average scores for RTs of 0.6 and 1.2 s. Compared with the RT of 1.2 s, RT of 0.6 s caused higher speech-recognition in children up to the age of 5. However, descriptive statistics could not verify the statistical significance of variables. Negative skew was observed in each data set; it is presumably due to the ceiling effect of the test material. Normality was only observed at severe conditions at the RT of 1.2 s (AD = 0.384 and P = 0.394) and age of 4 (AD = 0.576 and P = 0.133), which had relatively lower absolute skewness values.

The Mann-Whitney U test was performed to compare the pairwise data sets. For overall data, mask-wearing significantly affected the speech-recognition scores, but the type of mask was not critical. The speech-recognition score was affected by RT with statistical significance. However, it was not influenced by noise herein. Age, gender, and REVT, which are observed factors affecting average speech-recognition scores in descriptive statistics, affected the speech-recognition scores with statistical significance.

3.2. Combined effects of the face Mask, Noise, and reverberation time

Four acoustic environments were tested in this study, as listed in Table 6 . When the SNR was greater than 22 dB, no combined effect of the face mask and RT was observed. However, in the severe acoustic environment, combined with the SNR of 12 dB and RT of 1.2 s, the children’s speech-recognition test scores were significantly affected by the face mask, as listed in Table 7 . The average scores of children wearing the KF94 mask were significantly lower than the scores of those without the mask.

Table 6.

Kruskal-Wallis test results of the speech recognition scores.

Factor RT0.6 sNo Noise
(SNR > 22 dB)
 RT0.6 sBabble
(SNR = 12 dB)
 RT1.2 sNo Noise
(SNR > 22 dB)
 RT1.2 sBabble
(SNR = 12 dB)
Median H P Median H P Median H P Median H P
Mask No Mask
Surgical
KF94		 9.0 1.59 0.101 9.0 2.19 0.075 9.0 3.20 0.202 9.0 10.76 0.005
9.0 9.0 9.0 9.0
9.0 9.0 8.0 8.0
Overall 9.0 9.0 9.0 8.0
Open in a new tab

Table 7.

Mann-Whitney U test results of the speech recognition scores for babble noise conditions.

Factor Variabl1 Variable2 RT0.6 sBabble
(SNR = 12 dB)
 RT1.2 sBabble
(SNR = 12 dB)
W P W P
Mask No Mask Surgical 5,221 0.053 4,597.5 0.103
Surgical KF94 4,407 0.811 4,566.5 0.102
KF94 No Mask 5,196.5 0.047 4,906 0.001
Open in a new tab

3.3. Combined effects of the face mask and Children’s age

In children aged 4–5 years, the face mask significantly affected speech recognition, as shown in Table 8 . The youngest group, children aged 4 years, was affected by the surgical mask, whereas those aged 5–6 years were not. Children aged 4–5 years were affected by the KF94 mask. However, children aged 6 years were not affected by any face mask. As age decreased, the effects of wearing masks on speech recognition increased.

Table 8.

Mann-Whitney U test results of the speech recognition scores according to children’s age.

Factor Variabl1 Variable2 Age 4 yr
 Age 5 yr
 Age 6 yr
W P W P W P
Mask No Mask Surgical 6,743.5 P < 0.0005 13,123.5 0.168 5,288.5 0.751
Surgical KF94 5,889 0.364 13,003 0.035 5,141 0.345
KF94 No Mask 6,998.5 P < 0.0005 14,089.5 P < 0.0005 5,069 0.215
RT 0.6 s 1.2 s 14,454.5 0.007 30,453 P < 0.0005 11,790.5 0.839
Noise No Noise
(SNR > 22 dB)		 Babble
(SNR = 12 dB)		 11,900.5 0.054 27,914.5 0.896 11,686 0.052
Gender Girl Boy 14,543.5 0.091 36,417.5 P < 0.0005 12,717 0.249
Open in a new tab

The children's speech-recognition scores were affected by RT. Children aged 4–5 years were affected, but those aged 6 years showed no difference in the mean scores for RTs of 0.6 s and 1.2 s. Compared to RT of 1.2 s, RT of 0.6 s caused higher speech recognition in children up to age 5.

4. Discussion

4.1. Effects of face mask on speech recognition for preschool children

In this study, we found that face masks had a significantly negative impact on children’s speech recognition. This is consistent with the findings of Sfakianaki et al. [11], the only reference, to the best of our knowledge, regarding the effects of face masks on speech recognition in children with normal hearing. They observed that word identification produced with a surgical mask was significantly lower in children (N = 10). Note that it was reported to cause the least acoustic attenuation compared to other types of masks. [14] The negative effects were more pronounced in children than in adults. Lipps et al. [49] tested the impact of face masks on audio-visual word recognition in young children with hearing loss (3–7 years, N = 13). Word recognition was significantly poorer for surgical and transparent-apron masks than without any mask.

Experimental studies on how face masks affect speech recognition have been conducted more frequently on adults than on children. Generally, the masks affect speech recognition in adults with normal hearing. Thicker the opaque mask-material, lower the speech intelligibility. In addition, more vulnerable the group, lower the speech intelligibility. Yi et al. [50] reported that in the presence of noise, listeners (N = 26) performed poorly when speakers wore a disposable-paper mask. Choi [15] found that the intelligibility scores for 48 university students, obtained in N95-mask conditions at an SNR of 5 dB, decreased by a maximum of 10 %, compared to no-mask conditions. Toscano and Toscano [14] reported that face-mask effects appeared at an SNR of 3 dB (77 observations per parameter). Bottalico et al. [13] found that the use of fabric masks yielded a significantly greater reduction in speech intelligibility than surgical or N95 masks did for college students (N = 40). Thibodeau et al. [19] compared transparent and opaque masks in a noisy environment for speech recognition in an online study with 154 participants. They concluded that the use of transparent masks could significantly facilitate speech recognition in noisy environments. Thus, such masks might reduce stressful communication challenges experienced in medical, employment, and educational settings during the global pandemic. Brown et al. [51] investigated the degree to which different types of face masks and noise levels affected speech intelligibility in young (N = 180) and older (N = 180) adults. They concluded that they were similarly affected by face masks and noise in terms of intelligibility. Older adults showed poorer overall intelligibility than younger adults. Smiljanic et al. [52] found that masks affected non-native speech processing, even at easier noise levels; and clear speech accuracy improved significantly despite wearing a mask. Yi et al. [18] also tested 26 adults and observed that clear speech alleviated challenging communication situations, including lack of visual cues and “masked” acoustic signals. Homans and Vroegop [53] showed that even for speech perception in no-noise environments, surgical face masks, and face shields to a smaller extent, had a negative effect for patients with moderate to severe hearing loss.

Younger the children, lower the score in this study. This reflected the degree of language development, along with the test scores. As the mask became thicker, the children seemed to choose a completely different word rather than worrying about which word to listen to. To the best of our knowledge, no information on response time has been reported regarding the effects of masks on speech recognition. Ease of listening or listening difficulty was investigated using auditory response times. [54], [55] The response time was not used to further develop a proper unambiguous subjective rating. There were probably inherent difficulties in managing the measurement process in the field, or with multiple subjects at once. [54] Information on the response process is now more easily available due to computer-based assessments, such as in this study. In the future, a more in-depth analysis of speech recognition with a mask could be possible using the response time.

Previous studies on adults are generally consistent with our findings. In this study, the KF94 mask showed a significant negative effect on speech recognition, specifically in children aged 4 and 5 years. Children aged 6 years were not affected by the face mask under the conditions tested. The effects of a surgical mask on speech recognition was observed more in children aged 4, the youngest group in this study. Authors previous study [56] found that children aged between four and five seemed to perceive the mask as a physical self, while children aged six did not. Thus, chidren aged four and five are vulernable to mask in perception and speech recognition as well. In summary, the thicker the opaque material of the face mask, the lower the speech-recognition score, and the younger the pre-schoolers, the lower the scores because of face masks. However, surgical masks are less effective in filtering viral particles with coronavirus (SARS-CoV-2) than KF94 masks. [46], [57] Thus, individual speech effort [18], [50] and better room-acoustic conditions [11], [13] are recommended for the situation under question.

4.2. Effect of classroom acoustics with face masks on speech recognition

Four virtual classrooms with combined acoustical conditions, two RTs (0.6 and 1.2 s), and two SNRs (12 and>22 dB) were used for the speech-recognition test. RT showed a significant negative effect on the test in both, no-mask and mask-wearing conditions. The longer the RTs, the lower were the children’s speech-recognition scores. This is consistent with Bottalico et al. [13], who found that a longer RT reduced speech recognition, both with and without face masks. Reverberation effects on face-masked speech have not been fully investigated. Till date, Bottalico et al. [13] was the only publication regarding the effect of masks on speech in realistic listening environments with reverberance. For 6-year-olds, RT did not affect speech recognition with statistical significance in the conditions tested in this study. It seemed that children aged 6 years showed a ceiling effect for the testing conditions (KS-MWL-P and high SNRs) due to their language development. Table 5 shows that degradation of speech recognition by the RT of 1.2 s was approximately similar to the reduction by face masks at the RT of 0.6 s. Thus, better classroom acoustics could ameliorate the negative effects of face mask on speech recognition for pre-schoolers.

Table 5.

Mann-Whitney U test results of the speech recognition scores for all participants.

Variable1 N Median Variable2 N Median W-value P-value
Mask No Mask 261 9.0 Surgical 258 9.0 72,492.5 0.005
Surgical 258 9.0 KF94 256 8.0 68,885.5 0.136
KF94 256 8.0 No Mask 261 9.0 74,718 < 0.0005
RT 0.6 s 390 9.0 1.2 s 385 9.0 161,982.5 < 0.0005
Noise SNR > 22 dB 380 9.0 SNR = 12 dB 395 9.0 148,007.5 0.083
Age 4 yr 226 8.0 5 yr 330 9.0 61,584 0.454
5 yr 330 9.0 6 yr 219 9.0 77,634 < 0.0005
6 yr 219 9.0 4 yr 226 8.0 41,021.5 < 0.0005
Gender Girl 429 9.0 Boy 346 8.9 177,992 < 0.0005
REVT Normal 632 9.0 Delayed 143 8.0 44,016 < 0.0005
Open in a new tab

Noise did not affect speech recognition with statistical significance overall. This is probably due to the high SNR used in this study. However, noise nearly thwarted speech recognition for the group aged 4 years. Thus, the vulnerable group was affected by noise even at an SNR of 12 dB in face-masked conditions. Previous studies on adults have shown that noise affects speech recognition. In such studies, the SNRs were relatively lower than those used in the present study: SNRs of −9 and −5 dB [51], −8.3 to 25.4 dB [15], −5 dB [18], [19], [50], −5 to 5 dB [52], 0 and 5 dB [11], 3 dB [13], 3 dB, and 13 dB [14] were used.

In summary, when face masks are mandatory in the classroom, a shorter RT (approximately 0.6 s and higher SNR) can improve speech recognition among pre-schoolers.

4.3. Limitations and future works

First, this study was conducted in two randomly chosen preschools. Socioeconomic status or parenting, which could affect children’s early language development [58], [59] was not considered in the choice of participants. Accessibility to children was the priority of this study during the COVID-19 pandemic.

Second, the children’s language-development stage was not balanced in the recruitment process. For children with delayed development in the REVT group, few samples did not show statistically significant results. However, language development was not one of our research objectives; we only accepted and confirmed their language development.

Third, identical test materials and conditions were used for all children, although their social and cognitive development varied throughout the age range of 4–6 years. Children’s proficiency in using a touch pad and their ability to interpret pictures showed differences by age group. Furthermore, the two noise conditions used in this study also affect the negative skewness of the speech-recognition score distributions. The SNR of 12 dB was good enough for high scores for children aged 6. However, young children aged 4–5 were very sensitive to the noise environment of the SNR 12 dB. The test materials and conditions were set for children aged 4, and the results for children aged 6 showed ceiling effects.

Fourth, we examined only one female speaker with two types of face masks: a surgical mask and a KF94 mask. Male speech while wearing a face mask has been less investigated than female speech. The effect of the opacity of face masks has been investigated, but the effect of their shape is not yet clearly known. Future studies should consider the differences between male and female speakers and among the various types of face masks.

Additionally, in the room acoustic simulation process, the scattering coefficient for the materials was set as a single value of 0.1, which corresponds to 707 Hz. A potential problem is that with a so low scattering coefficient, multiple reflections and flutter echo might occur between parallel reflective surfaces, especially in the conditions with the RT of 1.2 s. In this study, no severe reflections or flutter echoes were observed at the listener's location, despite not being a typical sound field in a classroom. Realistic scattering coefficients for all materials or typical preschool classrooms are difficult to measure accurately; some generalization is required. However, it is necessary to gradually develop the virtual classroom model in a more realistic way in future studies.

Finally, the vocal effort of the speakers owing to the face mask was not considered in this study. The speech level of each mask type was adjusted to 62 dBA. Recently, the impact of face masks on vocal effort has been reported. At this stage, it is difficult to properly implement an objective experimental procedure to determine the effect on vocal effort, because it is a self-reported perception. To further investigate the effects of face masks on language, future studies should consider the vocal efforts of speakers due to face masks.

5. Conclusions

The face mask negatively affected pre-schoolers’ speech recognition in a realistic classroom environment, which included reverberance and noise. Reducing the RT and noise level in the classroom improved the recognition. Children aged 4 and 5 years were affected by face masks and RT more significantly than those aged 6 years.

Appropriate acoustics for classrooms and clear speech by teachers are recommended for better speech recognition in preschools, which usually facilitate children’s language-and-speech development.

CRediT authorship contribution statement

Miji Kwon: Data curation, Visualization, Validation, Investigation, Writing – original draft. Wonyoung Yang: Conceptualization, Methodology, Software, Writing – review & editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

Acknowledgement

We gratefully acknowledge the contributions of all the children who participated in our experiments, and the teachers and parents of the children who supported the experiments. We would like to express our deepest gratitude to the directors of Hanam Green and Forest Love Kindergarten, Jeong Suk-ja, and Choi Myeong-soon for understanding the importance of this study, convincing parents, and providing permission for the experiment.

This study was supported by the Basic Science Research Program of the National Research Foundation (NRF) [grant no. 2018R1D1A1B07048157], funded by the Ministry of Education, Republic of Korea. This study was also conducted with research funds provided by Gwangju University in 2022.

Appendix A. Korean Standard-Monosyllabic Word List for Pre-schoolers (KS-MWL-P)

graphic file with name fx1_lrg.jpg

Open in a new tab

Data availability

The data that has been used is confidential.

References

  • 1.Raina MacIntyre C., Jay H.S. Community universal face mask use during the COVID 19 pandemic—from households to travellers and public spaces. J Travel Med. 2020:27. doi: 10.1093/jtm/taaa056. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Brooks J.T., Butler J.C., Redfield R.R. Universal masking to prevent SARS-CoV-2 transmission—the time is now. JAMA. 2020;324:635–637. doi: 10.1001/jama.2020.13107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Mitze T, Kosfeld R, Rode J, et al. Face masks considerably reduce COVID-19 cases in Germany. Proc Natl Acad Sci 2020; 117: 32293-32301. [DOI] [PMC free article] [PubMed]
  • 4.Wang X., Ferro E.G., Zhou G., et al. Association between universal masking in a health care system and SARS-CoV-2 positivity among health care workers. JAMA. 2020;324:703–704. doi: 10.1001/jama.2020.12897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Harrison S, Duval D, Walters B, et al. The effectiveness of face coverings to reduce transmission of COVID-19 in community settings; A rapid revew (update 2). In: Care DoHaS, (ed.). UK Health Security Agency, 2021.
  • 6.Education UDf. Evidence Summary, Coronavirus (COVID-19) and the use of face coverings in education settings. In: Education Df, (ed.). London, UK2022.
  • 7.Suroyo S, Putra BM, Ibrahim B, et al. the effect of learning loss on student’s cognitive development during pandemic of COVID-19. In: Proceeding of International Conference on Islamic Education (ICIED) 2021, pp.169-176.
  • 8.Araújo L.A., Veloso C.F., Souza M.C., et al. The potential impact of the COVID-19 pandemic on child growth and development: a systematic review. J Pediatr (Rio J) 2021;97:369–377. doi: 10.1016/j.jped.2020.08.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Charney S.A., Camarata S.M., Chern A. Potential impact of the COVID-19 pandemic on communication and language skills in children. Otolaryngol Head Neck Surg. 2021;165:1–2. doi: 10.1177/0194599820978247. [DOI] [PubMed] [Google Scholar]
  • 10.Rocha P.M.B. The Covid-19 pandemic and its possible consequences to language/speech development and delay in children: an urgent issue. Audiology-Communication Research. 2021:26. [Google Scholar]
  • 11.Sfakianaki A., Kafentzis G.P., Kiagiadaki D., et al. Effect of face mask and noise on word recognition by children and adults. ExLing. 2021;2021:217. [Google Scholar]
  • 12.McCarthy D. Language development in the preschool child. Child behavior and development: A course of representative studies. New York, NY, US: McGraw-Hill, 1943, pp.107-128.
  • 13.Bottalico P., Murgia S., Puglisi G.E., et al. Effect of masks on speech intelligibility in auralized classrooms. J Acoust Soc Am. 2020;148:2878–2884. doi: 10.1121/10.0002450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Toscano J.C., Toscano C.M. Effects of face masks on speech recognition in multi-talker babble noise. PLoS One. 2021;16:e0246842. doi: 10.1371/journal.pone.0246842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Choi Y.-J. Acoustical measurements of masks and the effects on the speech intelligibility in university classrooms. Appl Acoust. 2021;180 [Google Scholar]
  • 16.Jeong J., Kim M., Kim Y. Changes on speech transmission characteristics by types of mask. Audiology and Speech Research. 2020;16:295–304. [Google Scholar]
  • 17.Rahne T., Fröhlich L., Plontke S., et al. Influence of surgical and N95 face masks on speech perception and listening effort in noise. PLoS One. 2021;16:e0253874. doi: 10.1371/journal.pone.0253874. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Yi H., Pingsterhaus A., Song W. Effects of wearing face masks while using different speaking styles in noise on speech intelligibility during the COVID-19 pandemic. Front Psychol. 2021:12. doi: 10.3389/fpsyg.2021.682677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Thibodeau L.M., Thibodeau-Nielsen R.B., Tran C.M.Q., et al. Communicating during COVID-19: the effect of transparent masks for speech recognition in noise. Ear Hear. 2021;42:772–781. doi: 10.1097/AUD.0000000000001065. [DOI] [PubMed] [Google Scholar]
  • 20.Buss E., Leibold L.J., Porter H.L., et al. Speech recognition in one- and two-talker maskers in school-age children and adults: Development of perceptual masking and glimpsing. J Acoust Soc Am. 2017;141:2650–2660. doi: 10.1121/1.4979936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Valente D.L., Plevinsky H.M., Franco J.M., et al. Experimental investigation of the effects of the acoustical conditions in a simulated classroom on speech recognition and learning in children. J Acoust Soc Am. 2012;131:232–246. doi: 10.1121/1.3662059. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Klatte M., Bergström K., Lachmann T. Does noise affect learning? A short review on noise effects on cognitive performance in children. Front Psychol. 2013;4:578. doi: 10.3389/fpsyg.2013.00578. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Elliott L.L. Performance of children aged 9 to 17 years on a test of speech intelligibility in noise using sentence material with controlled word predictability. J Acoust Soc Am. 1979;66:651–653. doi: 10.1121/1.383691. [DOI] [PubMed] [Google Scholar]
  • 24.Eisenberg L.S., Shannon R.V., Schaefer Martinez A., et al. Speech recognition with reduced spectral cues as a function of age. J Acoust Soc Am. 2000;107:2704–2710. doi: 10.1121/1.428656. [DOI] [PubMed] [Google Scholar]
  • 25.Bradley J.S. Predictors of speech intelligibility in rooms. J Acoust Soc Am. 1986;80:837–845. doi: 10.1121/1.393907. [DOI] [PubMed] [Google Scholar]
  • 26.ANSI/ASA. S12.60-2010/Part 1 (R2015) Acoustical performance criteria, design requirements, and guidelines for schools, Part 1: Permenent Schools. Washington American National Standard Institute, 2010.
  • 27.Hodgson M., Nosal E.-M. Effect of noise and occupancy on optimal reverberation times for speech intelligibility in classrooms. J Acoust Soc Am. 2002;111:931–939. doi: 10.1121/1.1428264. [DOI] [PubMed] [Google Scholar]
  • 28.Yang W., Hodgson M. Auralization study of optimum reverberation times for speech intelligibility for normal and hearing-impaired listeners in classrooms with diffuse sound fields. J Acoust Soc Am. 2006;120:801–807. doi: 10.1121/1.2216768. [DOI] [PubMed] [Google Scholar]
  • 29.Yang W and Bradley JS. Effects of room acoustics on the intelligibility of speech in classrooms for young children. J Acoust Soc Am 2009; 125: 922-933. 2009/02/12. doi: 10.1121/1.3058900. [DOI] [PubMed]
  • 30.Nábělek A.K., Pickett J. Reception of consonants in a classroom as affected by monaural and binaural listening, noise, reverberation, and hearing aids. J Acoust Soc Am. 1974;56:628–639. doi: 10.1121/1.1903301. [DOI] [PubMed] [Google Scholar]
  • 31.Nabelek A.K., Pickett J.M. Monaural and binaural speech perception through hearing aids under noise and reverberation with normal and hearing-impaired listeners. J Speech Hear Res. 1974;17:724–739. doi: 10.1044/jshr.1704.724. [DOI] [PubMed] [Google Scholar]
  • 32.Finitzo-Hieber T., Tillman T.W. Room acoustics effects on monosyllabic word discrimination ability for normal and hearing-impaired children. J Speech Hear Res. 1978;21:440–458. doi: 10.1044/jshr.2103.440. [DOI] [PubMed] [Google Scholar]
  • 33.Nábělek A.K., Robinson P.K. Monaural and binaural speech perception in reverberation for listeners of various ages. J Acoust Soc Am. 1982;71:1242–1248. doi: 10.1121/1.387773. [DOI] [PubMed] [Google Scholar]
  • 34.Bradley J.S. Speech intelligibility studies in classrooms. J Acoust Soc Am. 1986;80:846–854. doi: 10.1121/1.393908. [DOI] [PubMed] [Google Scholar]
  • 35.Plomp R., Steeneken H., Houtgast T. Predicting speech intelligibility in rooms from the modulation transfer function. II. Mirror image computer model applied to rectangular rooms. Acta Acust Acust. 1980;46:73–81. [Google Scholar]
  • 36.Bistafa S.R., Bradley J.S. Reverberation time and maximum background-noise level for classrooms from a comparative study of speech intelligibility metrics. J Acoust Soc Am. 2000;107:861–875. doi: 10.1121/1.428268. [DOI] [PubMed] [Google Scholar]
  • 37.Kim YT, Hong GH, Kim K, et al. Receptive and Expressive Vocabulary Test (REVT). Seoul Community Rehabilitation Center, 2009.
  • 38.Hong G.H., Kim Y.T., Kim K. Content and reliability analyses of the receptive and expressive vocabulary test (REVT) Commun Sci Disorders. 2009;14:34–45. [Google Scholar]
  • 39.Kim J., Lee J., Lee K.W., et al. Test-retest reliability of word recognition score using Korean standard monosyllabic word lists for adults as a function of the number of test words. J Audiol Otol. 2015;19:68. doi: 10.7874/jao.2015.19.2.68. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.ISO 8253-3:2022 Acoustics. Audiometric test methods-Part 3: Speech audiometry.
  • 41.Ross M., Lerman J. A picture identification test for hearing-impaired children. J Speech Hear Res. 1970;13:44–53. doi: 10.1044/jshr.1301.44. [DOI] [PubMed] [Google Scholar]
  • 42.Kim J.-S., Lim D., Hong H.-N., et al. Development of Korean Standard Monosyllabic Word Lists for School Aged Children (KS-MWL-S) and preschoolers (KS-MWL-P) Audiology. 2008;4:141–160. [Google Scholar]
  • 43.Doo S, Oh S, Brandstätt P, et al. Anechoic Chamber Design using Broadband Compact Absorber. In: 2003 Spring Meeting Mokpo, 2003, pp.393-396. The Korean Society for Noise and Vibration Engineering.
  • 44.USFDA. N95 Respirators, Surgical Masks, Face Masks, and Barrier Face Coverings, https://www.fda.gov/medical-devices/personal-protective-equipment-infection-control/n95-respirators-surgical-masks-face-masks-and-barrier-face-coverings (2022).
  • 45.Ma L., Kim M.-S. A study of the purchasing tendency of healthcare masks based on the user-centered design concept-centered on the form and color of the mask. J Korea Convergence Society. 2020;11:143–154. [Google Scholar]
  • 46.Kim M.-C., Bae S., Kim J.Y., et al. Effectiveness of surgical, KF94, and N95 respirator masks in blocking SARS-CoV-2: a controlled comparison in 7 patients. Infect Dis. 2020;52:908–912. doi: 10.1080/23744235.2020.1810858. [DOI] [PubMed] [Google Scholar]
  • 47.Rindel JH. Auralisation and how to calibrate the sound level for presentations. 2015.
  • 48.Owens R.E. Merrill Columbus; OH: 1988. Language development. [Google Scholar]
  • 49.Lipps E., Caldwell-Kurtzman J., Motlagh Zadeh L., et al. Impact of face masks on audiovisual word recognition in young children with hearing loss during the Covid-19 pandemic. J Early Hearing Detection Intervention. 2021;6:8. [Google Scholar]
  • 50.Yi H, Pingsterhaus A, Song W. The adverse effect of wearing a face mask during the COVID-19 pandemic and benefits of wearing transparent face masks and using clear speech on speech intelligibility. 2021. [DOI] [PMC free article] [PubMed]
  • 51.Brown V.A., Van Engen K.J., Peelle J.E. Face mask type affects audiovisual speech intelligibility and subjective listening effort in young and older adults. Cognitive Res: Principles Implications. 2021;6:1–12. doi: 10.1186/s41235-021-00314-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 52.Smiljanic R., Keerstock S., Meemann K., et al. Face masks and speaking style affect audio-visual word recognition and memory of native and non-native speech. J Acoust Soc Am. 2021;149:4013–4023. doi: 10.1121/10.0005191. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.Homans N.C., Vroegop J.L. The impact of face masks on the communication of adults with hearing loss during COVID-19 in a clinical setting. Int J Audiol. 2021:1–6. doi: 10.1080/14992027.2021.1952490. [DOI] [PubMed] [Google Scholar]
  • 54.Prodi N., Visentin C., Farnetani A. Intelligibility, listening difficulty and listening efficiency in auralized classrooms. J Acoust Soc Am. 2010;128:172–181. doi: 10.1121/1.3436563. [DOI] [PubMed] [Google Scholar]
  • 55.Gatehouse S., Gordon J. Response times to speech stimuli as measures of benefit from amplification. Br J Audiol. 1990;24:63–68. doi: 10.3109/03005369009077843. [DOI] [PubMed] [Google Scholar]
  • 56.Kwon Miji, Yang Wonyoung. Mask-Wearing Perception of Preschool Children in Korea during the COVID-19 Pandemic: A Cross-Sectional Study. International Journal of Environmental Research and Public Health. 2022;19(18):11443. doi: 10.3390/ijerph191811443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Wilkinson C, Low F, Rindelaub J, et al. Improving mask use to reduce COVID-19 transmission. 17 September 2021 2021. Auckland: University of Auckland.
  • 58.Pungello E.P., Iruka I.U., Dotterer A.M., et al. The effects of socioeconomic status, race, and parenting on language development in early childhood. Dev Psychol. 2009;45:544. doi: 10.1037/a0013917. [DOI] [PubMed] [Google Scholar]
  • 59.Schwab J.F., Lew-Williams C. Language learning, socioeconomic status, and child-directed speech. Wiley Interdiscip Rev Cogn Sci. 2016;7:264–275. doi: 10.1002/wcs.1393. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that has been used is confidential.

Articles from Applied Acoustics. Acoustique Applique. Angewandte Akustik are provided here courtesy of Elsevier

RESOURCES