Child-Centred Room Acoustic Parameters of Public Preschools in Sweden

Abstract

Preschool should promote children’s well-being and development, but the indoor sound environment is commonly problematic. The aim of our research project Supportive Preschool ACoustic Environment (SPACE) is to identify acoustic quality factors resulting in a supportive sound environment for children. This paper presents the first phase of the project where acoustic conditions were measured in unoccupied preschool rooms and analysed in terms of reverberation time, early decay time, sound strength, speech clarity, unoccupied sound pressure levels, and several room features. The results were compared with current target values, building year, and socioeconomic status of the preschool. A child perspective on room acoustics was, in addition, applied and it was revealed that children may be exposed to a lower sound strength than adults, and that adults may have better speech intelligibility conditions than children. Rooms in newer buildings had a longer reverberation time in the 125 Hz band, lower unoccupied levels, and lower sound strength. These differences could be explained by the trend towards larger rooms and porous acoustic ceilings in newer buildings. We found no significant correlations with the socioeconomic status. Ongoing work will facilitate an analysis of the correlation between the room acoustic parameters, the sound environment and children’s perception.

1. Introduction

A majority of Swedish children attend preschool where they spend an average of 31 h each week [1,2]. The curriculum of the Swedish preschool was revised in 2018 with the intention to improve quality and equity. However, the sound environment in many Swedish preschools is inconsistent with these objectives, as children are being subjected to high sound pressure levels over extensive time periods [3,4]. In addition, children are exposed to 4–8 dB higher levels than the staff [5,6].

How the sound environment at preschool influences the children, their well-being, and their development is rarely considered in the literature. However, a growing number of studies over the past years indicate that a child perspective may be needed when assessing supportive acoustic environments for younger children. Preschool-aged children require a better signal-to-noise ratio to be able to understand speech signals than normal hearing adults do, which is particularly important as pre-schoolers are in an intensive phase of acquiring language [7,8,9,10]. In addition, children with a non-native background are even more vulnerable as they have a poorer listening comprehension compared to native-speaking children [7]. Children’s behaviour causes them to be exposed to higher sound pressure levels, and, conversely, their behaviour may be affected by high levels [11,12].

According to Serpanos and Gravel, a young child’s hearing function is not comparable to that of an adult [13]. The diffraction and reflection properties of the head, pinna, and torso mean that the Head Related Transfer Functions (HRTF) will differ between children and adults. As the HRTF plays an important role in localising sounds, this may be of great relevance for the understanding of speech in a room under noisy conditions. The HRTF of a child amplifies the frequencies around 6 kHz, and the child’s ear canal further adds to this frequency amplification; in comparison, the HRTF and ear canal of an adult lead to amplification around 3 kHz [14].

It is not fully known how the structural anatomical development of the outer and middle ear affects sound transmission into the inner ear, and, thus, the hearing experience for younger children in comparison to adults. However, there is some evidence that children between the ages of 4 and 12 experience sounds louder than they would be for an adult exposed to the same stimuli [13]. It can also be hypothesised that the amplifications of higher frequency sounds may be perceived as more painful and distressing by children. Initial support for such reactions was obtained in a qualitative study of 36 preschool children aged four to five [15]. Uncontrollable sounds and distressing sounds (i.e., angry, loud, and scraping sounds) were experienced as both physically and emotionally painful. The importance of a child perspective for acoustic interventions at preschools was further shown by Persson Waye and Karlberg [11]. They found that a reduction in sound pressure levels was significantly associated with a 30 per cent (OR = 0.69) reduction in children’s perceptions of scraping and screeching sounds. The decrease in scraping and screeching sounds was associated with a 60 per cent (OR = 0.40) reduction in negative emotional reactions and an 80 per cent (OR = 0.20) reduction in stress-related symptoms.

Children’s awareness of their acoustic environment also encompass decision-making related to goal achievement as was shown in a recent work by [16]. However, the amount of freedom for children to choose an appropriate acoustic environment for a specific task is, in reality, limited in preschool. In addition, we do not know to what extent the present acoustic target values for preschools are met and if they result in facilities that enhance a supportive sound environment for children. What we do know, however, is that non-native language speaking children and children with language disorder and hearing impairments are at particular risk for language acquisition in rooms with poor acoustics [7,17].

In order to evaluate typical preschool room designs in relation to room acoustic parameters and children’s specific needs, the following acoustic parameters were measured and analysed in the present study. They are among the most commonly suggested in the literature, and may be of importance for a future child perspective on the acoustics of preschool rooms:

-

Reverberation time is the most used room acoustic parameter. The T₂₀ version is robust with respect to repeatability and background noise, and is the only room acoustic parameter with a target value in the Swedish standard [18]. Reverberation time measurements may be difficult to interpret in a non-diffuse field, such as a room where all available sound absorption is concentrated at the ceiling. Another critique is that T₂₀ is insufficient in itself, as it does not include the early reflections that are vital for intelligibility [19]. Early decay time (EDT) is considered to be better adapted to the perception of reverberance [20], but it is a less “global” measure. It may vary with source–receiver distance depending on the main reflections in the room, and repeatability problems have been reported for measurement positions close to the sound source [21]. The just-noticeable difference (JND) for reverberation time metrics in the literature ranges from 5 to 24 per cent [20,22,23,24].

-

Sound strength (G) accounts for the complete energy decay curve and correlates with the loudness of sounds in a certain room and is thus an important parameter for room acoustic characterisation. It is a measure of the sound energy emitted from a source in a room compared with what it would be in a free field at 10 m distance [20]. Sound strength has been suggested to be used to measure the acoustic quality of classrooms [25,26], and may also be appropriate for describing the acoustic quality of preschool rooms. The JND for the mid frequency sound strength is expected to be of the order of 1 dB, as reported for performance spaces [20].

-

Speech clarity (C₅₀) includes the effects of early reflections and has been found to correlate with the perceived clarity of speech. Puglisi et al. found a correlation between children’s reading speed and the C₅₀ of their classroom, but not the reverberation time [27]. In a recent study by Astolfi et al., speech clarity was found to be the most suitable parameter to classify the room acoustic condition of a school classroom [28]. The JND for speech clarity parameters is expected to be of the order of 1 dB in rooms for speech [29].

-

Unoccupied sound pressure levels in preschool rooms are regulated nationally in Sweden [30]. It is not explicitly known how low frequency noise affects children, but effects may include masking of speech and reduced performance [31,32]. Shield et al. found a small but significant correlation between unoccupied noise level in the classrooms of English secondary schools, and the overall lesson noise level and the time lost to disruptive activities (e.g., students talking or shouting during the lesson) [33].

In addition to the room acoustic parameters describing general properties of the room, some parameters characterise properties related to certain transmission paths in the room.

The speech transmission index, STI, is an objective measure of the intelligibility of speech [34]. It includes effects of the room, especially the share of early-to-late acoustic energy, and the effects of signal-to-noise level. Correlations between STI with C₅₀ and U₅₀ (useful-to-detrimental sound ratio) have been reported [29,35]. With a stationary transmission path and knowledge of the source strength and effective background level, the STI has been found to correlate well with subjective speech intelligibility, and this is also the case for schoolchildren [36]. However, since STI depends on the source and receiver positions, it may not be representative of other paths or the general situation in a room.

Studies on the connection between room acoustic properties and perception and health outcomes are sparse for the preschool setting, especially from a child perspective. In one of the rare intervention studies [11], the acoustic conditions of preschool rooms were improved with respect to perceived well-being by increasing the absorption area and introducing measures to reduce friction noise. The results showed moderate effects on objective parameters such as occupied sound level and reverberation time, but significant effects on children’s perception of their sound environment. Another study examined the effects of a multi-measure intervention program on preschool staff [37]. Improvements to the acoustic environment (more absorption, lower ventilation noise, fewer and quieter noise sources) was found to affect the personnel’s subjective ratings of the sound environment even though the measured sound level did not change significantly.

Care must be taken when applying results from studies on school classroom acoustics to preschool rooms; not just because preschool children are younger, but also because the organisation, the activities, and the conditions and prerequisites may differ substantially. For many children, preschool is the first environment where they need to actively use the Swedish language. According to the National Agency for Education, 25 percent of the children in Swedish preschools have a foreign background, meaning they were born abroad or have two parents who were born abroad [1]. In a traditional school classroom, there is a need for reverberation to support the teacher’s voice during full class lecturing [38]. This is not the most common type of activity in a typical Swedish preschool [39]. Excessive reverberation in classrooms is, however, clearly unwanted. For example, Astolfi et al. found that long reverberation times (>0.8 s) in first-grade classrooms in Italy were linked to children’s reduced perception of having fun and being happy with themselves [40]. In a review of international classroom standards, room acoustic recommendations were summarised for children at different ages, as well as children with special needs [41]. “Good” conditions for normal-hearing 6–7-year-olds were defined as an unoccupied ambient noise level of L_Aeq < 28 dB, a reverberation time of <0.4 s, and an STI of >0.75.

Since adults and children may perceive room acoustics differently, existing study designs might not use the optimal parameters and measurements to accurately capture children’s needs. As mentioned above, there are physical and physiological aspects indicating that supporting the sound environment for children may require special attention. As a first step towards a child perspective on room acoustics in preschools, we chose to evaluate:

• The influence of measurement height (reflecting a child sitting or standing);
• High frequency content in the room acoustic parameters (better reflecting the resonance frequency of a young child’s HRTF);
• The integration time when calculating speech clarity (since children may have a shorter “integration time of speech” than adults [42]);

The socioeconomic structures of an area may affect how preschools are designed, constructed, and assessed regarding acoustic quality aspects. Furthermore, buildings constructed in different time periods will have been subject to different building standards, leading to a possible variation in acoustic performance. The socioeconomic status (SES) of a neighbourhood is related to early language development, school achievement, and general health [43,44,45]. Studies of SES and children’s acoustic environment are sparse in the literature, but Astolfi et al. found a correlation between socioeconomic parameters (percentage of native-speaking children and district real estate value) and Italian classroom acoustics (reverberation time and noise level in silence) [40]. In the Swedish context, it has been found that preschools vary in educational quality and that there are shortcomings in equality [46]. The extent to which there is a relation between educational quality and the acoustic conditions in Swedish preschools is not known. Additionally, it is also not known whether there is a link between the SES of the area and the acoustic conditions of its preschools.

The aims of the present study were:

• To measure the acoustic conditions in Swedish preschool rooms;
• To examine whether and how the results vary with building year and socioeconomic status of the neighbourhood;
• To measure and analyse common room acoustic parameters in a way that may be more suitable for assessing children’s exposure;

In the study, we found that small preschool rooms have a short reverberation time and a high speech clarity, but also a large sound strength. The trend towards larger preschool rooms increases the reverberation time and decreases sound strength. Preschool rooms in newer buildings have lower unoccupied sound pressure levels. Children may perceive a lower sound strength and speech clarity, but longer early decay time compared with adults in preschool rooms. We found no correlations between socioeconomic status of the neighbourhood and the room acoustics of our investigated preschool rooms.

2. Materials and Methods

2.1. Study Sample

The present study is based on measurements in 57 rooms from 19 distinct public preschools in the Gothenburg area (Sweden). Due to changes in national building regulations and recommendations over time, we adopted a stratified approach for the random selection of preschools. This approach involved categorising them into three distinct periods based on their year of construction: 1980–1994 (#24), 1995–2006 (#11), and 2007–2018 (#22).

To reach an even distribution within each stratum with respect to socioeconomic factors, we used an existing preschool-specific index acquired from the central preschool administration. The index is based on a model from Statistics Sweden (SCB), where school performance after elementary school is linked with several explanatory variables related to the socioeconomic background [47]. When these factors are known for the children at a specific preschool, an averaged index is calculated, centred around 100. Preschools with an index over 100 have a larger share of children with a risk of not qualifying for high school (gymnasium), while preschools with an index less than 100 have a smaller share of children with a risk of not qualifying for high school. The purpose of the index is to prioritise economic support for school units with the highest need, and to increase equity. Most of the rooms in our study belonged to preschools with a high socioeconomic status corresponding to a low index, see Figure 1a. However, compared with the complete set of public preschools in the municipality, our dataset was more weighted towards low socioeconomic status, Figure 1b.

To limit the impact of pedagogic and organisational factors, private preschools were excluded from the study. Temporary buildings for preschools were also excluded, as they are less strictly regulated by the building norms and, hence, not representative. We also avoided preschools where major renovations had been made recently.

Children at preschool are divided into units that are commonly, but not always, based on age. The present study included 31 units catering for older preschool children (4–6 years old). In some of the large preschools, we conducted the measurements in several units. Up to three rooms per unit were chosen, to cover as much of the time spent indoors as possible, focusing on the main playrooms and the meal room. In cases where these rooms were the same, only one room per unit was measured.

2.2. Data Collection

The building year of each preschool and its blueprint was provided by the Gothenburg city facility administration. The measurement procedure included gathering information that might have influenced each room’s acoustic character. Dimensions and shape of the room were noted, a sketch of the furnishings was made, and pictures were taken. The floor and wall construction materials were evaluated and classified into heavy or light weight.

A subjective evaluation of the degree of furnishing was made, and categorised as sparse, normal, or dense. All rooms had acoustic treatment in the ceiling, and the type of material was noted. The porous type was common in newer buildings, whereas various other types such as perforated gypsum boards could be found in older rooms. All acoustic ceilings other than the porous type will be referred to as “other types” in the following. The construction height of the ceiling could only be examined for a limited number of rooms, and has been excluded from the analysis.

Some preschools also had sound absorption on the walls, and the approximate total area of this was noted. The concept of wall absorption is interpreted here as a fabric or porous material mounted on or in the vicinity of a wall which was subjectively judged by the first author as acoustically absorbing. Sound absorption on the floor was noted in terms of the number and types of rugs.

A complete list of the rooms, their identification numbers, and their basic features can be found in

Table A1. Preschool rooms and information collected in situ. The ID number is constructed by combining the building year (1: 1980–1994, 2: 1995–2006, 3: 2007–2018), preschool number, unit number, and room number.

Year and ID	Volume m3	Floor Area m2	Height m	Furnishing	Floor Construction	Wall Construction	Ceiling abs.
1980–1994
1.1.1.1	81	35	2.3	Normal	Heavy	Light	Other
1.1.1.2	55	22	2.5	Normal	Heavy	Light	Other
1.1.2.1	70	28	2.5	Sparse	Heavy	Light	Other
1.1.2.2	63	25	2.5	Dense	Heavy	Light	Other
1.2.1.1	95	38	2.5	Normal	Light	Light	Other
1.2.1.2	45	18	2.5	Normal	Light	Light	Other
1.3.1.1	118	41	2.5	Normal	Light	Light	Other
1.3.1.2	80	32	2.5	Dense	Light	Light	Other
1.3.2.1	110	44	2.5	Sparse	Light	Light	Porous
1.3.2.2	53	21	2.5	Sparse	Light	Light	Other
1.4.1.1	58	21	2.5–3	Normal	Light	Light	Other
1.4.1.2	74	27	2.5–3	Sparse	Light	Light	Other
1.5.1.1	80	24	2.5	Normal	Light	Light	Other
1.5.1.2	58	23	2.5	Normal	Light	Light	Other
1.5.2.1	73	29	2.5	Normal	Light	Light	Other
1.5.2.2	53	21	2.5	Normal	Light	Light	Other
1.6.1.1	73	28	2.4–3.8	Normal	Light	Light	Other
1.6.2.1	127	54	2.4	Sparse	Light	Light	Other
1.7.1.1	81	33	2.5	Sparse	Light	Heavy/Light	Porous
1.7.1.2	91	37	2.5	Sparse	Light	Heavy/Light	Porous
1.8.1.1	79	34	2.4	Normal	Light	Heavy/Light	Other
1.8.1.2	48	20	2.4	Normal	Light	Light	Other
1995–2006
2.1.1.1	58	24	2.4	Normal	Light	Heavy/Light	Porous
2.1.1.2	93	40	2.2–2.4	Normal	Light	Heavy	Porous
2.2.1.1	122	45	2.7	Dense	Light	Light	Porous
2.2.1.2	30	11	2.7	Sparse	Light	Light	Porous
2.2.2.1	120	46	2.6	Dense	Heavy	Light	Porous
2.2.2.2	20	8	2.6	Sparse	Heavy	Light	Porous
2.3.1.1	77	32	2.4	Dense	Light	Light	Other
2.3.1.2	65	27	2.4	Dense	Light	Light	Other
2.4.1.1	81	31	2.2–3	Dense	Light	Light	Porous
2.4.1.2	32	13	2.2–2.7	Sparse	Light	Light	Porous
2.4.2.1	1056	192	4.5–6.5	Dense	Light	Light	Porous
2007–2018
3.1.1.1	123	45	2.7	Normal	Heavy	Light	Porous
3.1.1.2	65	19	2.7–5	Normal	Heavy	Light	Porous
3.2.1.1	73	24	3.1	Normal	Heavy	Light	Porous
3.2.1.2	150	49	3.1	Normal	Heavy	Light	Porous
3.3.1.1	124	40	2.5–3.7	Normal	Heavy	Light	Porous
3.3.2.1	124	40	2.5–3.7	Sparse	Heavy	Light	Porous
3.3.3.1	811	129	3–8.2	Normal	Heavy	Light	Porous
3.4.1.1	525	84	2.7–6.3	Normal	Heavy	Light	Porous
3.4.2.1	110	41	2.7	Sparse	Heavy	Light	Porous
3.4.2.2	95	35	2.7	Normal	Heavy	Light	Porous
3.4.3.1	95	35	2.7	Normal	Heavy	Light	Porous
3.5.1.1	67	25	2.7	Sparse	Heavy	Light	Porous
3.5.1.2	113	42	2.7	Normal	Heavy	Light	Porous
3.5.1.3	107	40	2.7	Sparse	Heavy	Light	Porous
3.6.1.1	60	23	2.6	Dense	Heavy	Light	Porous
3.6.1.2	116	43	2.7	Dense	Heavy	Light	Porous
3.6.1.3	44	17	2.6	Sparse	Heavy	Light	Porous
3.6.2.1	24	9	2.7	Dense	Heavy	Light	Porous
3.7.1.1	32	12	2.7	Normal	Heavy	Light	Porous
3.7.1.2	168	43	3–5.6	Normal	Heavy	Light	Porous
3.7.1.3	51	19	2.7	Sparse	Heavy	Light	Porous
3.7.2.1	167	43	3–5.6	Normal	Heavy	Light	Porous
3.7.2.2	167	42	3–5.6	Normal	Heavy	Light	Porous
3.7.2.3	54	20	2.7	Sparse	Heavy	Light	Porous

the Appendix A.

2.3. Acoustic Measurements

Room acoustic parameters and unoccupied sound pressure levels were measured in each room using a laptop and an external 8-channel sound card (HEAD acoustics SQuadriga II). An omnidirectional sound source with a built-in generation of pink noise (50–20,000 Hz) was used to excite the sound field in the room.

Unoccupied sound pressure levels were measured in accordance with ISO 16032:2004 [48]. In most rooms, the contribution from the ventilation could be estimated by conducting separate measurements with the ventilation unit turned off and on.

Room acoustic measurements were conducted following the precision method in ISO 3382-2:2008 [49]. Three microphones were used simultaneously at different positions in the room with predetermined height intervals (1 ± 0.2 m, 1.4 ± 0.2 m, 1.6 ± 0.2 m). The standard’s recommendation to use natural source positions was implemented by placing the loudspeaker where we estimated that a child’s head would be during sitting and standing activities in the room. In addition, a corner position was always used. Measurements were obtained with at least twelve independent source-receiver combinations, except in small rooms where this was not possible. Impulse responses were calculated with the ArtemiS SUITE 12 software package using the loudspeaker’s electronic signal as the reference.

An extended, non-standard measurement protocol was used in room 2.3.1.1 for detailed examination of how room acoustic parameters varied with microphone height. This room was considered to be a typical preschool room in all aspects apart from the subjective furnishing degree, which was perceived as “dense”, and the acoustic ceiling type, which was categorised as “other”. Photographs of the room are provided in Figure 2. Room acoustic measurements were here performed for nine microphone heights: 0.1 m, 0.3 m, 0.5 m, 0.8 m, 1.0 m, 1.2 m, 1.5 m, 1.8 m, 2.0 m. The lowest microphone positions deviate from the ISO 3382-2 standard due to the vicinity of reflecting surfaces. Each height had two source positions, and each source position had six microphone positions, resulting in a total of 12 measurements per microphone height.

2.4. Acoustic Analysis

Impulse responses and unoccupied sound pressure level spectra were exported to MATLAB. Octave band room acoustic parameters for 125–8000 Hz were calculated according to ISO 3382-1 from the impulse responses using version 8 of the ITA-toolbox [20,50]. The 8000 Hz octave band is outside of the frequency range in the ISO 3382-2:2008 standard but it may still be of relevance, especially for children, and is therefore included in our evaluations. The analysis covered T₂₀, EDT, and C₅₀. An addition to the code was made to evaluate the clarity index with a shorter time (C₃₅) as suggested by Whitlock and Dodd [42].

Sound strength (G) in each measurement position was calculated from the sound power of the loudspeaker (L_W) measured according to the ISO 3741 standard, and the resulting sound pressure level (L_p) at a measurement point [20,51]. The sound power level was measured in an accredited laboratory.

$$G=L_{\mathrm{p}}-L_{\mathrm{W}}+31$$

2.5. Statistical Treatment

Room spatial arithmetic averaging was applied to all parameters. The ISO 3382-1:2009 standard recommends this procedure for reverberation time, and its use has also been suggested for sound strength and speech clarity [25]. Characteristics of the different years of preschool rooms were compared using a t-test with a 5% significance. Correlations between physical room properties and the socioeconomic index were tested with Spearman’s rho.

3. Results

3.1. The Acoustic State of Preschool Rooms, Averaged Results

Table 1. Average acoustic results for 57 preschool rooms. Reverberation time (T₂₀), early decay time (EDT), sound strength (G), and speech clarity (C₅₀) were evaluated in different frequency ranges. Unoccupied background sound pressure levels were evaluated as L_Aeq and L_Ceq.

Parameter	Octave Band Frequency Range	Arithmetic Mean	Standard Deviation
T20 (s)	125 Hz	0.42 s	0.10 s
	250–4000 Hz	0.35 s	0.07 s
	500–1000 Hz	0.35 s	0.07 s
EDT (s)	125 Hz	0.38 s	0.09 s
	250–4000 Hz	0.31 s	0.06 s
	500–1000 Hz	0.32 s	0.07 s
G (dB)	125 Hz	20.2 dB	2.9 dB
	250–4000 Hz	18.4 dB	2.7 dB
	500–1000 Hz	18.7 dB	2.7 dB
C50 (dB)	125 Hz	8.0 dB	2.2 dB
	250–4000 Hz	10.2 dB	2.1 dB
	500–1000 Hz	10.0 dB	2.3 dB
Unoccupied A-weighted equivalent sound pressure level LAeq (dB)	63–8000 Hz	29 dB	5 dB
Unoccupied C-weighted equivalent sound pressure level LCeq (dB)	31.5–8000 Hz	50 dB	6 dB

gives an overview of the averaged single value results and standard deviation of the acoustic measurements in the preschool rooms. Figure 3 shows the arithmetic average in the set of room-averaged parameters as a function of frequency, also indicating the standard deviation in the set. As expected, the reverberation time and sound strength decreased with frequency, while speech clarity increased.

Figure 4 shows the average unoccupied equivalent sound pressure level (L_eq) measured in 1/3 octave bands, as well as the standard deviation. The contribution from the ventilation was estimated by comparing measurements taken when the ventilation system was turned off with those taken when it was turned on. The figure also indicates the low frequency recommendations for educational premises by the Public Health Agency of Sweden in the 1/3 octave bands between 31.5 and 200 Hz [30]. The results show that the equivalent unoccupied sound pressure level, on average, was dominated by the contribution from the ventilation in the investigated rooms.

3.2. Room Acoustic Parameters of Importance from a Child Perspective

In this section, we examine the possibility of evaluating and adapting current room acoustic parameters to better consider children’s perceptions and responses. Assessed variables relate to differences between children’s and adult’s receiver positions and hearing function [52], and include frequency content, height of the microphone, and integration time.

3.2.1. Frequency Content

Room acoustic design and regulations usually use single number quantities to simplify frequency dependent room acoustic parameters.

Table 2. Common current single number room acoustic parameters and a novel single number parameter with an emphasis on very high frequencies.

Parameter		Arithmetic Average in Measurement Set
T20	125 Hz 1	0.41
	250–4000 Hz 1	0.36
	500–1000 Hz 2	0.36
	4000–8000 Hz 4	0.32
EDT	500–1000 Hz 3	0.32
	4000–8000 Hz 4	0.27
G	500–1000 Hz 3	18.5 dB
	4000–8000 Hz 4	16.3 dB
C50	500–1000 Hz 3	9.8 dB
	4000–8000 Hz 4	11.7 dB

¹ SS 25268:2007 + T1:2017. ² For example, the Italian room acoustic standard UNI 11367 (2010). ³ ISO 3382-1:2009 [20]. ⁴ Novel parameter with an emphasis on very high frequencies.

shows the most common definitions of the single number quantities and corresponding calculated values in our measurement set of preschool rooms. The table also contains a high-frequency single number definition: the arithmetic average of the parameter in octave bands between 4000 and 8000 Hz. As expected, this high-frequency single number definition yielded slightly lower values for reverberation time and sound strength, and higher values for speech clarity.

3.2.2. Microphone Height Dependency

When comparing room acoustic parameters acquired with the microphone placed at 1.6 ± 0.2 m (adult position) with the results from the microphone placed at 1 ± 0.2 m (child position), T₂₀ did not change. However, EDT tended to be slightly longer between 125 and 500 Hz in the low microphone position (Figure A7 and Figure A8 in the Appendix A). Sound strength and speech clarity, on the other hand, seemed to be more affected by height (Figure 5 and Figure 6). Sound strength is sensitive to propagation path, and the low microphone height resulted in lower G in the mid and high frequencies. The effect of microphone height on speech clarity was small, but there was a tendency of a “hump” around 500 Hz in the result for the high microphone. This increase was a general trend and not due to a specific case.

The results of the extended measurement protocol in room 2.3.1.1, where nine different microphone heights were used, can be seen in Figure A9, Figure A10, Figure A11 and Figure A12 in the Appendix A. The room-averaged acoustic parameters are shown here as a function of microphone height for the 125 Hz octave band and the arithmetic averages in octave band intervals 250–4000 Hz, 500–1000 Hz, and 4000–8000 Hz. No clear trends can be seen in the variations with microphone height for the mid and high frequency results. This is as expected, since the first node at a quarter of the wavelength is located 0.34 m or less above the floor (for perpendicularly incident sound). Therefore, the 6 dB boost close to the room boundaries will typically not be noticeable here, except for the lowest bands at the microphone positions closest to the floor. For the octave band of 125 Hz, however, the first node is located 0.68 m above the floor. Therefore, a substantial increase in sound pressure level can be expected at the lowest microphone position (0.1 m) and a decreasing level upwards. Consequently, G is expected to vary in the same way. More variation can also be expected in EDT, since that parameter is heavily influenced by early prominent individual reflections which will interact in various ways depending on factors such as incidence angle and frequency. The measurement results did show such dependency of height, as well as a greater variation, indicated by the wide confidence interval. This was especially noticeable for EDT and sound strength, but speech clarity also showed a great variation. At a height of 0.5–0.8 m, the EDT was slightly longer, and sound strength and clarity were slightly lower than at positions above 0.8 m.

3.2.3. A Shorter Integration Time of Speech Clarity

Following Whitlock and Dodd’s suggestion that speech clarity should be evaluated with a shorter integration time for children [42], we also analysed this parameter using an integration time of 35 ms. In this time, sound travels around 12 m compared with 17 m for a time of 50 ms. With a shorter integration time, C₃₅ will be smaller than C₅₀. This was also what we found when comparing the averages for the complete measurement set (Figure 7). In other aspects, the two measures varied in a similar way, indicating that C₃₅ does not add any extra relevant information that can be expected to be of value from a child perspective. For regulatory purposes, a specific limit on C₅₀ that applies to children would have the same effect as introducing C₃₅ in order to set a child-specific limit.

3.3. Influence of Building Year

This section describes our analysis of differences in room acoustics between the project’s three building year strata (1980–1994, 1995–2006, 2007–2018). We first examine the physical room properties and then assess how the room acoustic parameters themselves varied with building year.

3.3.1. Physical Room Properties

Newer buildings tended to have larger rooms, with the average volume increasing from 76 m³ to 144 m³ between 1980 and 1994, and 2007 and 2018, but the difference was not statistically significant (p = 0.08). The three very large rooms in the dataset (>500 m³) were all in the 1995–2006 and 2007–2018 strata. Figure A2 in the Appendix A shows boxplots of how volume and floor area varied between our three strata.

Building constructions and acoustic treatments varies over time; for example, lightweight floors were found to be much more common in older buildings (Figure A3 in the Appendix A). Acoustic ceilings were found in all measured rooms, but the type varied, with newer buildings having the suspended porous type while older buildings had sundry other types (Figure A4 in the Appendix A). Wall absorbers were more common in large rooms and in newer buildings. There was no clear tendency of differences in subjective furnishing degree between building years.

3.3.2. Room Acoustic Parameters

T₂₀ was similar between the different strata except in the 125 Hz band, where newer buildings had a slightly longer reverberation time (Figure A5 in the Appendix A). Figure 8 compares T₂₀ for buildings from 1980 to 1994 and from 2007 to 2018, giving a clearer picture of the difference in the 125 Hz band.

Room-averaged sound strength differed significantly between the strata for all but the low frequency bands. Newer buildings have a lower G than their older counterparts (Figure 9 below and Figure A6 in the Appendix A).

C₅₀ does not vary significantly between the strata. However, it has a slight increase around 500 Hz for newer buildings; see Figure 10.

Unoccupied A- and C-weighted equivalent sound pressure levels L_Aeq and L_Ceq are shown in Figure 11 together with the estimated contribution from the ventilation. The level is significantly lower in newer buildings and most rooms in buildings from 1995 fulfil the national requirements of 30 dBA and many fulfil the 50 dBC limit. Most rooms in the oldest buildings exceed the national guidelines, both for the A- and the C-weighted level.

3.4. Room Acoustics and Socioeconomics

Our study explored the connection between room acoustics and the socioeconomic index. After an analysis of our dataset, we found no statistically significant correlations for any of the measured parameters, as illustrated in

Table 3. Linear correlation coefficients (Spearman’s rho) for the socioeconomic index and the room acoustic parameters, calculated using the corr function in MATLAB. No significant correlation was found for any of the parameters.

Linear Correlation between Socioeconomic Index (SE Index) and Room Acoustic Parameters
	Spearman’s rho	p-Value
T20 (125 Hz)	0.11	n.s. (0.41)
T20 (250–4000 Hz)	0.07	n.s. (0.63)
T20 (500–1000 Hz)	−0.001	n.s. (0.99)
G (500–1000 Hz)	−0.07	n.s. (0.60)
C50 (500–1000 Hz)	0.02	n.s. (0.90)
Unoccupied A-weighted level (dBA)	0.10	n.s. (0.46)
Unoccupied C-weighted level (dBC)	0.02	n.s. (0.87)

.

4. Discussion

4.1. Averaged Results

The reverberation time in our complete dataset was short, and fulfilled the demands of the Swedish standard [18] even in the 125 Hz octave band. It was also short in an international (classroom) perspective [41]. The average sound strength in our dataset was in the same range as previously reported for classrooms [53]. However, the average speech clarity was several dB higher than previously reported for classrooms [25,28,53]. This difference may be related to the bigger room volumes and longer reverberation times of the classrooms in previous studies compared to the rooms in our dataset.

The average unoccupied sound pressure level fulfilled the Swedish requirements, but the standard deviation was high, with 40% of the rooms exceeding the L_Ceq limit of 50 dB and 31% exceeding the L_Aeq limit of 30 dB. An unoccupied level above 30 dBA corresponds to a “Bad” rating in primary school classrooms for children with hearing or language impairments according to Mealings [41]. As preschool children are even younger and many have a non-native language background, masking effects from background noise may impair language acquisition and the consequences of unoccupied noise levels above the recommendations should be further investigated.

The low frequency target values for educational premises in Sweden [30] were exceeded by more than 1 dB in 18% of the dataset. Our measurement results also revealed that the unoccupied sound pressure level, on average, was dominated by the contribution from the ventilation. External sources such as road traffic noise hence seemed to be of less importance in these rooms, compared with an international perspective [40,54].

4.2. Supportive Room Acoustics for Children

When we examined differences in room acoustic parameters related to microphone height, in the analyses of the complete dataset, there was a tendency toward a higher sound strength (200–1000 Hz) and a higher speech clarity (around 500 Hz). Additionally, we observed a slightly shorter EDT (125–500 Hz) in the adult/high microphone position compared with the child/low microphone position. The consequences of these findings are still an open question.

Conversely, the findings from the special measurement set that was applied in one preschool room showed greater differences in the 125 Hz band. Below 1 m, corresponding approximately to the height of an early age preschool child or a child who is sitting down, children may find themselves in positions where there is a large spread and variation in the room acoustic parameters. This means that they can experience either higher or lower values than adults, whose heads are mostly in a zone where the variations are smaller.

Considering previous research indicating preschoolers’ sensitivity to high-frequency sounds and their HRTF and ear canal characteristics [11,52], we also analysed the single value parameters with more emphasis on high frequencies. This revealed a shorter reverberation time, a lower sound strength, and a higher speech clarity compared to the standard frequency intervals.

The results of the above tests imply that children may be exposed to a lower sound strength in preschool rooms than adults, and that due to their height, adults may have better speech intelligibility conditions in preschool rooms than children. How this affects the children is an open and intriguing question. In particular, the results suggest that the most common room acoustic parameters may not give an accurate picture of the acoustic qualities of a room as perceived by preschool children.

A shorter integration time to calculate speech clarity (C₃₅) did not add any relevant extra information in the present study of the complete dataset. It may, however, add perspectives in future studies where room acoustic parameters are compared with perceptual and subjective data from the children in the preschools.

4.3. Building Year

The trends in acoustic characteristics of the preschool rooms in our dataset can be summarised as follows: rooms in newer buildings had lower unoccupied sound pressure levels and lower mid- and high-frequency G. The latter can be expected for a diffuse sound field, as the average room size has increased over time, almost doubling between the 1980–1994 and 2007–2018 strata in our dataset. The use of a porous type of acoustic ceiling and wall absorbents further decreases G and could be part of the reason why we did not see an increase in reverberation metrics between our strata as would be expected for larger rooms.

It is, however, not certain that the tendency to build preschools with larger floor space and volume is advantageous for a child-supportive sound environment. For example, Mealings et al. found that large open plan kindergarten classrooms did not provide supportive learning environments due to the occurrence of high intrusive noise levels from parallel activities in the rooms [55,56].

4.4. Room Acoustics and Socioeconomic Index

We did not see a similar connection as was found by Astolfi et al. between socioeconomic parameters (native-speaker percentage and district real estate value) and classroom acoustics (reverberation time and sound pressure level during silent activities) [40]. Gothenburg is considered a highly segregated city [57]. Our findings suggest that the central facility administration, responsible for the maintenance and construction of preschool facilities, at least, does not contribute to the inequalities in the city.

4.5. Limitations

The lack of correlation between room acoustics and socioeconomic status found in our dataset may not apply to preschools in other municipalities or parts of Sweden as the facility management may differ. There is also a risk that the study sample may have been skewed towards preschools with good room acoustics, as participation required some engagement from the principal and the staff. Preschools with a strained work environment, possibly with very bad room acoustics, may have been reluctant to participate. On the other hand, preschools with bad acoustic conditions may have been attracted to the project, in hopes of drawing attention to the problem.

5. Conclusions

The measurements of the selected room acoustic parameters (T₂₀, EDT, G, C₅₀) in these 57 unoccupied preschool rooms showed no major deviations from previously reported measurements in similar settings. Compared to current Swedish regulations for preschools, the reverberation time in the measured rooms was lower than the requirement. The equivalent sound pressure levels for unoccupied rooms were mostly within the guidelines, but the variation between rooms was large and the levels in some rooms were too high, especially for the low frequencies. The dominant noise source for unoccupied rooms was related to the ventilation system.

The attempts made in the present article to evaluate how room acoustics may be supportive for preschool children’s perception and response provided some conclusions and some questions that remain to be further investigated. In short, children may be exposed to a lower sound strength than adults, but may also have less optimal speech intelligibility conditions than adults. We did not find any variation in T₂₀ that could be related to child-relevant height exposures. However, the question remains of how the variations in strength and speech intelligibility might affect children’s wellbeing, hearing, or language acquisition. This will be investigated in ongoing and future studies where perceptual and subjective aspects are considered. It is important to strive for an optimal room acoustics, especially to facilitate language acquisition for the relatively large group of children with a foreign language background, and children with a language disorder and/or hearing impairments.

The construction year of the buildings had a significant impact on some of the room acoustic parameters; rooms in newer buildings had lower unoccupied sound pressure levels and lower G for mid and high frequencies. Newer buildings also had a longer reverberation time at low frequencies and a slight increase in C₅₀ around 500 Hz. The main explanation for these differences is likely that rooms in newer buildings in our sample tended to be larger (more volume) and more often had porous ceilings.

For the analysis regarding socioeconomic status, the calculated index showed no relation to any of the room acoustic parameters. This indicates that the 57 preschool rooms were of equal room acoustic quality regardless of their socioeconomic situation.