Association between perceived sound type dominance and overall assessment of the acoustic environment using ISO 12913 soundwalks^a) Open Access

In order to transform cities into sustainable and health-promoting places (WBGU, 2016), it is necessary to also consider and develop high quality acoustic environments in cities. So far, the urban acoustic quality has hardly been addressed beyond the sound pressure level, basically in terms of noise. However, noise is not only a risk factor to be controlled, but also an important resource for a health-promoting environment. This approach provides a new perspective that emphasizes health-promoting urban sound quality through a positively perceived acoustic environment. As a result, the acoustic environment, especially in urban environments, can be understood as a designable quality (Maag, 2016), making it an important resource for a health-promoting city.

In recent years, there has already been a shift towards a proactive approach to managing urban acoustic environments in response to societal pressure, moving away from purely noise control-focused policies (Asdrubali, 2014; Brown, 2012; Kang, 2017; Kang and Aletta, 2018). In particular, the concept of “soundscapes” plays an increasingly important role. The concept aims to support health and well-being through the promotion of positive acoustic environments especially in urban environments (Aletta , 2018; Andringa and Lanser, 2013; Bild , 2018; Kang and Schulte-Fortkamp, 2016). In recent years, an increasing number of publications (Kang , 2016; Kang and Aletta, 2018), policy documents, guidelines, and recommendations refer to the soundscape approach [European Parliament and Council (2002), COST TUD Action TD-0804 (Kang , 2013), and European Environment Agency (2021)].

In 2014, the international standardization series 12913 of the soundscape concept was launched, establishing a common understanding of terminology and concepts, as well as technical measurement procedures and data evaluation methods against the background of the already existing concepts and approaches [International Organization for Standardization (ISO) (ISO, 2014, 2018)]. According to this, a soundscape is “an acoustic environment that is perceived, experienced, and/or comprehended by a person or a group of persons in context” [ISO (2014), p. 1].

Overall, the ISO 12913 soundscape concept allows human perception to be used to capture essential situational and contextual factors, providing a holistic view of environmental noises in the urban area, beyond mere noise assessment (Lercher and Schulte-Fortkamp, 2003). One of the key instruments of the soundscape concept is the soundwalk. The ISO 12913 defines it as a “method that implies a walk in an area with a focus on listening to the acoustic environment” [ISO (2018), p. 2]. This instrument is a participatory tool in which the acoustic environment is experienced, systematically captured, and evaluated in situ at defined locations, together with the participants.

Following the soundscape approach, many studies gathered data using surveys and acoustic measurements in situ. Thus, large datasets have been sampled. Some used mainly acoustic monitoring, like studies conducted in Montreal (Steele and Guastavino, 2021), London (Aletta , 2020), Rome and Milan (Pagès , 2020), or Granada, Spain (Manzano , 2021). Subjective ratings have also been the focus of soundscape research ranging from examining mental health in (primarily) Bulgarian students (Dzhambov , 2021) to the perception of changed acoustic environment during the COVID-19 lockdown in Argentina (Maggi , 2021). Other studies combined a survey approach and acoustic measurements following the ISO 12913 and calculated a predictive model based on a large data set collected in London and Venice (Mitchell , 2021).

Meanwhile, a large number of studies have been carried out using the ISO 12913 questionnaire. Most of the locations analysed were urban parks (Evensen , 2016; Jeon and Hong, 2015; Liu , 2014; Song , 2018), city centers (Sudarsono , 2016), university campuses (Mancini , 2021), public places (Aletta , 2019), historical sites (Huang and Kang, 2015; Jalil , 2023; Pérez-Martínez , 2018), tourist sites or a combination of these (Adams, 2008; Bahalı and Tamer-Bayazıt, 2017; Davies , 2013; Jeon , 2013; Kogan , 2017). So far, soundwalks are rarely carried out in residential areas or everyday places where people do not explicitly go to spend their leisure time (Engel , 2018). However, as people spend a considerable amount of time in their neighborhoods, meaning in residential areas, it is important to address acoustic conditions here as well. In addition, most studies are based on only a small number of soundwalks and only on specific urban location. Therefore, studies aiming at repeated measurements to estimate the variance of soundwalk assessments at specific urban locations are lacking. The importance of quantifying the association between sound types and overall assessments has been acknowledged in the soundscape literature. Aletta (2016) highlighted the need for more quantitative research to deepen our understanding of how specific sound types contribute to soundscape perception. There is also a need for studies that allow comparisons to be made between different urban areas. For example, to identify the dominant types of sound and assess their impact on the overall assessment of acoustic quality.

Therefore, the main objective of our study is to analyse the perception of dominant sound sources in two different urban areas and their impact on the overall assessment of the surrounding sound environment using the soundwalk approach. To this end, we conducted repeated soundwalks in two areas in the city of Essen, a city in a highly urbanized metropolitan region of the Ruhr area in Germany with the corresponding questionnaire according to ISO 12913.

First, we describe and compare the two study areas in terms of the perceived sound type dominance and overall assessments of the surrounding sound environment as measured by ISO 12913. Second, we investigate the relationship between the perceived sound type dominance and the overall assessment and whether this relationship differs between the study areas.

Our study design is based on the soundwalk method according to the German version of ISO/TS 12913-2 “Acoustics—Soundscape, Part 2: Data collection and reporting requirements” (ISO, 2018). A soundwalk is a method in which an acoustic environment is experienced and evaluated in situ by the local population. The acoustic environment was measured both longitudinal and cross-sectionally conducting repeated soundwalks between July 2022 and September 2023. An overview of the survey phases can be found in the Appendixes.

The soundwalks are part of the research and practice project Be-MoVe (participation-based transformation of active mobility for health-promoting urban and transport infrastructures, aiming to test co-created alternative forms of mobility and designs of public spaces in neighborhoods (City of Essen, 2024). Within the framework of Be-Move, two distinct intervention neighborhoods were selected. Our original intention was to test the effects of the Be-Move interventions on the acoustic environment. For this reason, we used these study areas and selected the respective routes accordingly. The longitudinal timeframe for the soundwalks was also based on the planned participatory interventions. However, due to logistical reasons, the time frame between planned participatory interventions and the soundwalks could not be synchronized. In this respect, we cannot provide evidence of effects of the interventions on the acoustic environment. However, this is, to the best of our knowledge, one of the most comprehensive studies gathering data using the ISO 12913 soundwalk method. It enables robust and comprehensive evaluations of the questionnaire data over a long period of time at many different urban locations.

As part of the quality management, an expert from the TU Berlin who was not part of the project trained the staff on the soundwalk method during a three-day on-site training course. The staff were researchers and students from the Institute of Urban Public Health. Additionally, a study design was implemented, including standard operation procedures (SOPs) for all study procedures.

The study was approved by the institutional review board of the university clinics of Essen, Germany. All participants were informed of their rights and had to consent in writing to participate.

Essen Holsterhausen is located in the southwest of the city of Essen, close to the city center and is bisected by the A40 highway, which divides the city into north and south. The district has a total area of 2.97 km², of which 67.7% is built-up and 5.0% is recreational and open space. The built-up area is dominated by apartment buildings. With approximately 26 000 inhabitants and a population density of 87.4 persons per hectare, the district is one of the most densely populated in Essen (Stadt Essen, 2021).

Essen inner city has a total area of 0.90 km², of which 49.2% is built-up and 1.6% is recreational and open space. The building development is mixed-use as typical for a city center. The area is characterized by a range from a shopping street to a shopping center to a purely residential areas with a high density of multi-family dwelling. With around 4000 inhabitants and a population density of 45.6 people per hectare, the district is above the average population density of 28.0 people per hectare in Essen (Essen Statistics Office, 2021).

According to ISO 12913 we selected six listening stations per route (ISO, 2018). As explained above, the selection of the listening stations was originally and primarily based on the planned mobility interventions in the Be-MoVe project. This meant that the locations were already predetermined by the researchers before starting the soundwalks and before feedback from participants had been given. Within the locations, we took the following criteria into account. On the one hand, the listening stations of the soundwalks should represent typical everyday places of the study areas, which are familiar or at least known to residents or people working on site. On the other hand, the listening stations should reflect various urban situations in the district (e.g., main street, side street, public square, public facilities such as schools, residential, commercial, or leisure functions).

Using ArcGIS pro software, a buffer with a radius of 50 m was placed around each listening station to characterise the spatial conditions.

The routes were walked in both directions to control sequence effects. A map with the corresponding routes and listening stations is shown in Fig. 1.

Along Route I, station A is located on a side street in a residential area. The following stations B and C are located on major roads with railroad facilities and are characterized by a variety of land uses such as mixed-use areas, developed areas serving housing as well as parking areas and the University Hospital Essen. Listening stations D, E, and F are again located on residential and development streets and are characterized by mixed-use land, green space, and parking areas. The mixed-use areas at station E and F are characterized by first floor retail and residential uses on the upper floors. Building heights at these two stations range from 5 to 6.5 m. An exception to this is the public use area at station F, where there is a church of 51 m in height with a church forecourt opening onto the street. Following the route K, the listening stations G and C are located on major roads with railroads and have mixed-use areas, built-up areas used for housing and parking areas. The height of the buildings varies between sixteen and 20 m, with the exception of a gas station at listening station G. In the buffer of stations H, I, and J, there are both residential and development roads, as well as major roads. In addition to smaller green areas and parking areas, there are common areas for educational institutions, kindergartens, and nursing homes. The building height in these three listening stations is between 16 and 25 m.

The inner city of Essen is characterized by a main pedestrian zone as well as surrounding major four lane roads. Route I runs along mixed-use areas on the edge of the inner city. Stations A to D on route I have public spaces and are adjacent to the pedestrian zone. Station D is directly located in front of a big shopping mall (Limbecker Platz). Stations E and F are located on major roads, but border on green spaces and parking lots. The height of buildings at all of these listening stations ranges from 4.5 to around 10 m. Route K starts at listening station G, which lies in front of a theatre (Grillo-Theatre). Stations H and I are on public squares adjacent to churches (as is station L). Station J also represents a public square with a street and parking spots around it. Station K is on a cross section in the pedestrian zone with shops and small restaurants around it. As with route I, mixed-use development is present in all buffers of the stations along route K, as well as pedestrian areas and parking areas.

Our target group included local residents of the selected neighborhoods and people with links to the study area, e.g., through work or leisure, but ultimately there were no restrictions on participation. We used mixed recruitments methods, distributing flyers (1500 per phase and area) in homes, local shops and in two citizen workshops as part of the overarching Be-MoVe project. The flyers were also forwarded in digital form to local stakeholders, i.e., associations and other civic initiatives, with a request for dissemination. In addition, we published press releases in local newspapers, which were also available online. The soundwalks were also advertised on our institute's own homepage. In these recruitment methods, we consistently directly addressed the local residents to recruit them for the study. However, there was no exclusion in the case of non-local residence.

Except for one soundwalk at 10 am, all soundwalks were performed from 6 pm to around 7 pm. All soundwalks were conducted by a moderator who gave the verbal instructions and guided the people through the soundwalk and an assistant who was responsible for providing questionnaire material and for making the microphone recordings. After all participants arrived at the meeting point, they were briefed about the research project and the background and process of the soundwalk. They were given a sticky board with the personal information questionnaires, including a statement about whether they were local residents or not, the soundscape questionnaires, an explanation about data protection and the corresponding consent form and had the opportunity to hand in the personal information and take a first look at the questionnaire. The instruction for the soundwalk included the request not to make loud disturbing noises during the listening time and the advice was given that they could slowly expand their listening circle from themselves to further away or even close their eyes to better concentrate on listening. The extent to which this was practiced by the participants was up to them. Any questions raised were answered before the group walked from the meeting point to the first listening station. Participants positioned themselves and the assistant gave the signal from when the three-minute listening time and the audio recording would begin. After the three minutes, the assistant announced the end of the listening time. The participants now started to fill in the questionnaire. Afterwards the next station was headed for. Following the final listening station, participants were thanked and provided with an opportunity to share brief feedback on the soundwalk or exchange experiences and impressions. As a little incentive they were invited to a local café for an ice cream.

The questionnaire's initial section, “Identification of Noise Sources,” prompts participants to evaluate the dominance of traffic, natural, human, and other noise sources. The ISO standard emphasizes that the term “noise” should not be interpreted as a value judgement. The second section, “perceived affective quality” includes the evaluation of the attributes pleasant, chaotic, vibrant, uneventful, calm, annoying, eventful and monotonous. The last two sections relate to the overall assessment (question C.3.1.4) and the appropriateness (question C.3.1.5) of the surrounding sound environment. All variables are rated on a 5-point Likert scale. A question from Method B was added asking “what is going through your mind” to capture any impressions possibly not covered by the previous scales. All questionnaire items were derived from the German version of ISO/TS 12913-2 “Acoustics—Soundscape, Part 2: Data collection and reporting requirements” (ISO, 2018).

During the 3-min listening period, the wind speed and temperature were measured with the SKYWATCH BL 400, a mobile weather measuring device.

For the binaural recordings, a 3Dio FS Pro II with a ZOOM H4nPro stereo recording device was used. The stereo-recordings had a length of three minutes according to the listening time, a sampling rate of 44 100 Hz and a bit depth of 24. In a follow-up workshop, selected recordings were played and discussed with the participants. We deviated from the specifications for binaural miking and analysis of ISO 12913-2, Annex D, because our main study aim was to investigate the relationship between perceived sound type dominance and overall assessments. The only reason we recorded during the soundwalks at all was to be able to play back recordings in our planned workshops or to check for any events that the participants might have referred to in the questionnaire.

Ahead of analysis and the process of selecting the statistical procedures that align with our objectives and data, we engaged extensively with our data. The variance was checked for differing from zero and also on multicollinearity. Variance of the errors was analysed for homoscedasticity. Data on errors were tested for independence/autocorrelation and the made assumption of their normal distribution. Linearity of the relationship of the outcome variable and predictors was evaluated. Furthermore, following the choice of using a mixed model a normal distribution of random intercepts around the overall model was evaluated. We considered the requirements for the tests described below to be fulfilled. We conducted descriptive statistics by calculating the mean and respective variance. Given the presence of multiple ratings from each participant and multiple ratings for each of the 22 listening stations, a two-level hierarchical structure for the ratings was designed. The design operates on the assumption that ratings within one participant/listening station are likely to exhibit greater similarity compared to ratings across participants/stations, respectively. We applied a mixed model incorporating both fixed and random effects to examine the relationship between overall ratings and sound sources. A random intercept was calculated for participant ID (for each person) and for each station. Estimators' slopes were treated as fixed effects. The dependent variable (overall assessment), was treated as a continuous variable even though it is defined by a 5-point Likert scale, which is supported by Norman (2010) and Zumbo and Zimmerman (1993). The same logic applies regarding the independent variables of perceived sound type dominance.

Finally, stratification by study area (model II and model III) was performed based on model I and the adjusted model, respectively.

All analyses were performed using R version 4.1.1. (R Core Team, 2021). For mixed model analysis the R-package “lme4” version 1.1.27.1 was used (Bates , 2015).

A total of n = 91 residents (women 66%, 34% men, 0% of diverse gender) participated in our study. Given that n = 29 participants completed several soundwalks, ranging from two (n = 15), three (9), four (2), to six (1) times, we collected a total of n = 143 questionnaires (79 study area A, 64 study area B). Furthermore, of the 91 participants n = 10 participated in both study areas. The distribution of the number of participants per soundwalk is presented in Table I. The ISO 12913-2 recommends less than five people per soundwalk to prevent effects arising from the group size. This recommendation was generally followed. However, for organizational reasons (scheduling preferences, weather-related cancellations), it was not possible to adhere to the recommendation across the board. We did not exclude data from participants in soundwalks with larger group sizes for reasons of case numbers. When asked, the soundwalk instructors were likewise unaware of any major differences in the questions and answers with regard to group size. Overall, we covered a wide age range with n = 15 participants aged between 18 and 29 years, n = 21 aged 30–44, n = 21 aged 45–59, n = 26 aged 60–75, and n = 8 aged over 75 years. One of the 91 participants was excluded from the analysis because they had not provided any ratings for “overall assessment.”

Figure 2 shows the average ratings for the overall assessment and Fig. 3 the perceived dominance of the four sound types at the listening stations in Holsterhausen. With the exception of listening stations D and F, the traffic noise was the dominant perceived sound source. This is consistent with the lowest overall ratings for these stations (all located at major roads). Accordingly, natural sounds were generally perceived as less dominant. Our results are consistent with the general assumption that places with more noticeable natural sounds are rated higher overall (see stations A: side road in a residential area; D and H: both side roads near a school). It can also be seen that the stations where human sounds were perceived as dominant were rated better overall [see stations D, E, and F (both located near a shopping street with tall residential buildings)]. At all stations where natural and/or human sounds were perceived as relatively dominant and the overall assessment was high, traffic noise was rated lower in the descriptive evaluation. Other noise was not perceived as dominant at any of the stations. In general, the overall ratings for these stations were mediocre to poor.

The trend of mostly mediocre to poorly rated stations continues to show in the second study area, the inner city (Figs. 4 and 5). However, in this study area, traffic noise was the dominant perceived sound source at only five of the 10 listening stations. The strong dominance of traffic noise at stations E and F is to be expected, as they are located directly on a major road. The overall rating here is correspondingly poor. At station D, which is located in front of a shopping center, traffic noise is perceived as less dominant, similar to stations H and I, which are located in a pedestrian zone. Overall, natural sounds were perceived as less prevalent in the city center. Noticeable is the dominance of human sounds, e.g., at station C a public square with restaurants. The overall ratings were mediocre in the city center and poor at two stations. Compared to Holsterhausen, inner city stations appear to have been rated slightly better, possibly due to the often less dominant traffic noise in the inner city.

Table II shows the effect sizes (β) and corresponding 95%-confidence intervals of the associations between sound types and the overall assessment of the acoustic environment. Model I represents the crude overall model, in which the sound sources as well as the persons and stations were included as random effects. Models II and III are the individual models for Holsterhausen and the inner city, respectively. Adjustment by gender, age categories, wind speed and temperature did not change the estimators for the sound sources (see  Appendix B).

The analysis of the overall model, which includes data from both study areas (Holsterhausen and inner city), revealed associations between sound types and overall assessment of the acoustic environment. Traffic noise showed a negative association with overall assessment, with an estimated β of −0.44 (95% CI: −0.50 to −0.38), indicating that higher levels of perceived traffic noise were associated with lower overall assessments. On the other hand, nature sounds indicated a positive association (β = 0.13; 95% CI: 0.08 to 0.19), suggesting that the presence of nature sounds was linked to higher overall assessments. Human sounds and other noise sources had negative associations, but with much smaller and virtually non-existent effect sizes compared to traffic noise and nature sounds (Human sounds: β = −0.07, 95% CI: −0.12 to −0.01; Other noise: β = −0.02, 95% CI: −0.07 to 0.04).

In Holsterhausen, similar to the overall model, traffic noise had a negative association with overall assessments, which was even slightly stronger than in the overall model. Nature sounds showed a positive association. As in the overall model, the estimated effects of human sounds and other noise sources are so low that we assume no correlation here.

In the inner city, traffic noise exhibited a negative association with overall assessments, while nature sounds show an even smaller estimate compared to the other models with a confidence interval containing zero. Interestingly, the effect of traffic noise actually appears to be smaller here than in Holsterhausen, which is supported by the non-overlapping confidence intervals of the estimators. As in Holsterhausen, human sounds and other noise show no association with the overall assessment.

Aim of this study was to investigate the relationship between perceived sound type dominance and overall acoustic environment assessments, comparing two urban areas in the city of Essen, Germany. One of the central findings of this study is that higher levels of perceived traffic noise are linked to lower overall evaluations. This aligns with previous studies that identified traffic noise as a major contributor to noise pollution and annoyance and its adverse effects on acoustic quality and well-being (Clark and Paunovic, 2018; Ohrström, 2004; Yang and Kang, 2005). Conversely, the positive association between nature sounds and higher overall assessments suggests that the dominance of natural sounds contributes positively to the perceived acoustic environment, also aligning with findings from current literature (Abbott , 2016; Hedblom , 2017; Krzywicka and Byrka, 2017). Bird songs in particular seem to contribute positively to the perception of the soundscape (Ma , 2021). However, such a specific differentiation is not made in the ISO 12913 questionnaire we used. Therefore, we cannot draw conclusions about specific effects of bird songs in our study, which would likely fall under the “natural sounds” we asked about. Alternatively, the ISO 12913 suggests a listing of all the sound sources that participants notice, which participants fill out themselves [see ISO (2018), Method B, Fig. C8]. In that way, a more specific differentiation of sound sources noticed down to the subordinate level is possible (Guastavino, 2018).

Although these results are not unexpected, our study revealed interesting aspects. For instance, we were surprised to find that there was no evidence of an association between gender, age, wind speed, and temperature on the overall assessment. In current studies personal factors were found to have an effect on the perception of the acoustic environment, e.g., personal viewpoints on meaning of tranquility or personality traits (Aletta , 2018; Filipan , 2017; Lercher , 2016; Lindborg and Friberg, 2016). In our study, however, we did not find any major differences between the different demographic groups in terms of the perception of different sound types, such as traffic. This suggests that the impact of sound types on the perceived acoustic environment is relatively consistent across different demographic groups and environmental conditions.

We observed that the negative effect of traffic noise on the overall assessments varied across the different study areas, suggesting possible variations in the perception of traffic noise in different urban contexts. Bruce and Davies (2014) showed in their study that the expectation of the acoustic environment influences the evaluation in the actual experience of an acoustic environment (Bruce and Davies, 2014). Expected sounds are more likely to be accepted than unexpected sounds, meaning they are tolerated and not rated negatively, especially if they have fundamentally negative connotations. However, from our own experience of talking to our participants, we assume, that people might expect the inner city to have more motorized noise as well as much more overall noisy environment. Perhaps this explains the lower impact of traffic noise on the overall assessment we found in our study. Additionally, participants in Holsterhausen reported that traffic noise is a major concern. We therefore assume that this situation has resulted in a focus on traffic noise during the soundwalks, so that traffic noise has a strong negative impact on the overall assessment.

The question of the appropriateness of the acoustic environment, which is also asked in the ISO 12913 questionnaire, could provide further information. However, we found that participants struggled to classify the listening station according to the appropriateness of what they heard. In other words, whether it was appropriate in the sense of how the acoustic environment here would be optimal or good, or appropriate in the sense of the extent to which it corresponds to the own expectations of the actual listening experience at this location. For this reason, we did not include this part of the questionnaire in our analyses. We suggest adapting this question accordingly and, above all, explaining the context and implication for future studies more explicitly.

One of the strengths of our study is that we used a standardized questionnaire released by International Organization for Standardization (ISO, 2014, 2018), providing the advantage of a better comparability of our results with past and future studies. In addition, we trained our staff extensively to perform standardised soundwalks. The drafting of a study protocol, including SOPs, also ensured the quality of the study. A further strength is that we have been able to achieve a reasonably large sample size with our repeated measures approach. Our study contributes to the research by employing a quantitative approach to examine the association between sound types and overall assessments of the acoustic environment in specific urban settings. This advancement contributes to the growing body of literature advocating more robust quantitative methodologies in soundscape studies. Our research quantified the strength of the association between the perceived dominance of specific sound types and the overall assessment using the ISO 12913 questionnaire. It offers several advantages, including the ability to control for confounding variables, assess the strength of associations, and compare the impact of different sound types within and between study areas.

Alongside strengths, some limitations are to be mentioned. First, we excluded the analyses of the affective qualities provided by the questionnaire, because in our discussion with the participants it became obvious that these items were subject to quite different ways of understanding. As we were unable to adjust for these different interpretations, we considered omitting these constructs and focusing on less discussed ones, which are also quite tangible when it comes to the design of urban environments, for example. Second, we have based our evaluations on the intra-questionnaire correlations. We are aware that additional data from noise measurements are commonly used for such evaluations. However, our aim was to apply the ISO 12913 questionnaire to compare two different study areas, which is why we did not use the noise measurement data in the first run here. Third, we did not integrate important small-scale data of the built environment into our analyses. We intend, however, to address these important aspects in future studies.

In conclusion, our findings show the negative association between perceived traffic noise as well as the positive association between nature sounds and the overall acoustic environment assessments. We suggest that the impact of sound types on the perceived acoustic environment is relatively consistent across different demographic groups and environmental conditions.

Overall, our comprehensive study not only provides a base for further analyses, such as the investigation of temporal and spatial differences and their impacts on the perceived sound environments, but also for future studies using the ISO 12913 soundscape methods. Our study contributes to the understanding of soundscapes and provides a more detailed and quantitative understanding of the relationship between sound types and the overall assessments of urban acoustic environments. Given that our data collection actually took place within a mobility intervention project, we still consider it important to examine such urban interventions in terms of acoustic perception as well.

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.

Informed consent was obtained from all participants.

The data from the soundwalks are available from the corresponding author upon reasonable request.

See Fig. 6.

See Table III.