Effect of birdsongs and traffic noise on pedestrian walking speed during different seasons

Walking pace and environmental sounds

Current investigations provided some evidence that noise exposure might influence the walking pace. A recent study by Maculewicz, Erkut & Serafin (2016) clearly showed that the sound characteristics of specific environments affect walking pace in those environments. The participants were asked to listen to the sounds of a seashore, a busy street, a restaurant, and busy offices and simultaneously walk on an aerobic stepper. The results revealed that the seashore and restaurant sounds elicited a slower pace than the sounds of streets and offices. Interestingly, the results documented not only the effect of traffic noise on the increase in walking pace but also indicated that listening to nature sounds may result in a decrease in walking pace. In our previous study (Franěk, 2013) participants were asked to walk along a route located in an urban area with different environmental features. The participants tended to walk statistically significantly faster in areas without greenery and with more traffic and higher perceived noise than in sections with greenery and less traffic and perceived noise. Our previous study (Franěk et al., 2018) found that listening to a soundtrack with traffic noise while walking on an urban route increased the average walking speed. Conversely, listening to a soundtrack with natural sounds (birdsongs) decreased the walking speed in contrast to the control condition, but this difference was not significant. Moreover, under both conditions, the walking speed was lower in areas with a greater amount of perceived naturalness than in areas with less greenery.

Auditory and visual interactions

There is a large body of studies on auditory and visual interactions. One line of research, based on the subjective evaluation of a combination of photographs of diverse environments and sounds, showed that certain sounds can influence the evaluation of physical environments. In a pioneering study, Anderson et al. (1983) showed that natural and animal sounds had enhancing effects on evaluations of wooded natural and residential sites, while other sounds had detracting effects on evaluation of the same sites. More recently, Gan et al. (2014) found that anthropogenic sounds (vehicle alarms, motorcycle rumbling, the roar of engineering machinery, adult voices) had a negative impact on landscape preference, as opposed to biological and geophysical sounds. The authors also showed that acoustic preference played a much more important role in landscape evaluation than visual preference. Furthermore, a set of studies has examined the effect of anthropogenic sounds on the ratings of landscapes in natural parks. Mace, Bell & Loomis (1999) examined the effect of helicopter tour noise on evaluations of scenic overviews in a national park. The negative experiential effect was more strongly associated with soundscapes that included helicopter noise than with purely natural soundscapes. Benfield et al. (2010) reported that anthropogenic soundscapes (helicopter noise, airplane overflights, motorized ground vehicles, and human voices) were each responsible for detriments to the visual assessment of the landscapes shown. Weinzimmer et al. (2014) replicated previous studies by showing that motorized recreation noise had detrimental effects on the ratings of both aesthetic and affective dimensions.

Conversely, another line of research showed that the evaluation of sound environments can be affected by co-occurring visual settings. Viollon, Lavandier & Drake (2002) found that the more urban the visual setting, the more negative the sound ratings of the environment. Lee, Hong & Jeon (2014) explored how participants assessed rural soundscapes featuring high-speed train noise. The noise from the high-speed train was rated as less annoying if the sound was presented with a picture containing a higher percentage of natural features. The same effect documented also surveys of residents exposed to noise in their neighborhood. Li, Chau & Tang (2010) studied the effects of annoyance modifiers on residents of high-rise buildings overlooking urban parks and wetlands. Their results indicated that the perception of greenery considerably reduced noise annoyance and concentration problems Having a window facing a yard and a view of water or green space is associated with substantially reduced noise annoyance and concentration problems (Bodin et al., 2015). Consistently, Van Renterghem & Botteldooren (2016) reported that the extent to which vegetation is visible through the living room window was a strong predictor of self-reported noise annoyance among residents with high exposure to traffic noise. Leung et al. (2017) found that views of the sea, urban river, or greenery could reduce the probability of invoking a high noise annoyance response from residents living in high-rise buildings, while views of a noise barrier could increase the probability. Views of greenery had a stronger noise moderation capability than views of the sea or urban river. Cassina et al. (2017) investigated tranquility ratings in urban areas judged on the basis of visual and auditory elements. The authors showed that the positive sound sources increased statistically significantly the perceived tranquility, while negative sound sources decreased it. However, visual elements had only negative effects on the tranquility score.

Van Renterghem (2019) summarized potential explanations for these effects. First, it was shown that nature sounds are preferred over anthropogenic ones, namely over vehicle and construction sounds (Hong & Jeon, 2013; Krzywicka & Byrka, 2017; Medvedev, Shepherd & Hautus, 2015; Ratcliffe, Gatersleben & Sowden, 2016; Yang & Kang, 2005). Several studies documented that after inducing psychological stress, physiological recovery of sympathetic activation is faster during exposure to pleasant nature sounds than to sound perceived as less pleasant (Alvarsson, Wiens & Nilsson, 2010; Annerstedt et al., 2013; Medvedev, Shepherd & Hautus, 2015). Natural sounds are considered the most complex and informational sound types. They can provide a large amount of information pertaining to species, season, and temporality (Pijanowski et al., 2011), and signify an actual living or vital environment (Ratcliffe, Gatersleben & Sowden, 2013). Importantly, natural soundscapes can provide restorative benefits independent of those produced by visual stimuli (Benfield et al., 2014). Bird sounds are ranked at the top of the desired natural sounds in an urban environment; they outrank the other sounds in terms of their pleasantness (for review, Van Renterghem, 2019). Bird sounds are most commonly associated with perceived stress recovery and attention restoration (Ratcliffe, Gatersleben & Sowden, 2013), and they are associated with green spaces, spring, and summer (Ratcliffe, Gatersleben & Sowden, 2016). Second, some studies have documented that the invisibility of a source of noise among vegetation decreases perceived noisiness (e.g., Bangjun, Lili & Guoqing, 2003; Sun et al., 2018; Watts, Chinn & Godfrey, 1999). Finally, vegetation is often a source of natural sounds. Moreover, natural sounds increase one’s feeling of the presence of nearby nature, even when it is not directly visible.

Restorative environment

Environmental psychology presents a large body of evidence that direct and indirect exposure to nature positively affects humans. There is considerable evidence that living in green areas or even having a view of nature results in decreased stress. It was observed that the residents of neighborhoods with a high amount of greenery have relatively low chronic stress (e.g., Hartig et al., 2011; Nilsson & Berglund, 2006; Ward Thompson et al., 2012). Stress recovery was observed to be more rapid in people who viewed natural scenes than in people who viewed urban scenes (e.g., Ulrich et al., 1991). Furthermore, a large body of studies has shown that viewing a surrogate image of nature (photographs, videos, slides, window views, and virtual nature scenes) resulted in decreased stress and increased positive emotions (e.g., Brown, Barton & Gladwell, 2013; De Kort et al., 2006; Felnhofer et al., 2015; Hartig et al., 1999; Jiang, Chang & Sullivan, 2014; Ulrich, Simons & Miles, 2003; Valtchanov, Barton & Ellard, 2010).

As previously mentioned (e.g., Anderson et al., 1983; Gan et al., 2014), natural and animal sounds had enhancing effects on the evaluations of the restorative potential of wooded natural and residential sites. It is known that people use various commercially produced recordings of natural sounds for their relaxation. Recently, virtual reality relaxation applications have also been available (e.g., Lindner et al., 2019). There is research evidence of the effects of listening to natural sounds on relaxation. Such studies documented enhanced stress recovery during exposure to pleasant nature sounds (e.g., Alvarsson, Wiens & Nilsson, 2010; Annerstedt et al., 2013; Medvedev, Shepherd & Hautus, 2015). Although diverse natural sounds may be used (e.g., ocean, birds, rain, nighttime in the jungle, water, and waterfall sounds), bird sounds have specific effects. Several studies have documented that this type of natural sound is most commonly associated with green spaces, spring, and summer (Ratcliffe, Gatersleben & Sowden, 2016) and with attention restoration and perceived stress recovery (Ratcliffe, Gatersleben & Sowden, 2013). Urban settings with birdsong are more highly evaluated than urban settings alone, and birdsong contributes to the positive ratings associated with urban green space (Hedblom et al., 2014).

Congruence of the visual and auditory features of the environment

The interaction of audition and vision in the perception of an environment may also be shaped by the perceived congruence of the visual and auditory features of the environment. “Congruence” between the two types of stimuli means that their combination makes sense because people expect specific sounds to occur in each environment that is congruent with the physical features of the environment (Bruce & Davies, 2014). In the study by Carles, Bernáldez & Lucio (1992) visual stimuli (village, stream, park with children, steppe, empty park, and residential neighborhood) and corresponding sound stimuli were presented in varying combinations. The authors found that the congruence between sound and image influenced the preferences. Brambilla & Maffei (2006) showed that in parks, the more that sounds were congruent with expectations, the less they evoked annoyance. Hedblom et al. (2017) reported that people gave higher evaluations to natural sounds in areas that they considered highly natural, such as urban woodlands, compared with less natural areas, such as parks, allotments. and lawns. A study by Zhao, Xu & Ye (2018) demonstrated that appropriate combinations of congruent acoustic and visual stimuli increased the perceived restorative potential of an environment. For instance, a landscape containing natural water and high plant coverage matches the visual association with a bird singing, and combining wind sounds with a landscape with high vegetation coverage increases the restorative quality of that environment. Furthermore, some studies suggest that evaluations of natural environments are more sensitive to visual-auditory incongruence than evaluations of urban environments. Ge & Hokao (2005) found that the sounds of transportation were more disliked in a natural landscape than in ordinary urban street environments where they were more congruent. Consistently, Jahncke, Naula & Eriksson (2015) examined the combined effect of various acoustic stimuli (nature sounds, quiet broadband noise, office noise) and visual settings (office and urban nature environments) on perceived restoration. They found that a picture of nature was more sensitive to the influence of auditory stimuli than a picture of an office.

The current study

The objective of the present study was to analyze the combined effect of diverse sounds and visual environmental features on walking speed in certain outdoor environments. The present study aimed to enhance these findings by conducting two experiments in different environments and seasons.

In a previous study (Franěk et al., 2018), participants were asked to walk along an urban route while listening to soundtracks with intense crowded city traffic noise (noise from motorized vehicles, engine sounds, automobile horns, and human voices) or birdsongs from headphones, or they walked without listening to any acoustic stimuli in the control condition. The participants listening to traffic noise walked significantly faster on the route (mean walking speed was 1.65 m/s) than those listening to forest birdsong sounds (mean walking speed was 1.53 m/s). The participants listening to the birdsong walked slightly slower than those under the control condition (mean walking speed was 1.58 m/s), but this difference was not statistically significant. Simultaneously, the walking speed in all conditions was influenced by the environmental features of the route. In areas with more natural features, the participants walked more slowly than in areas with less greenery. Consistent with previous findings, the participants liked the environment more in the absence of noise or in the presence of birdsongs. Importantly, the previous experiment not only documented the negative effect of noise but also showed that relaxation sounds may decrease walking speed. This study also raised questions regarding the congruence between sounds and the visual environment. Do nature sounds (forest birdsong) have an effect on the deceleration of walking speed only in congruent environments? Specifically, the question we sought to answer was whether listening to birdsongs that are commonly associated with spring and summer had similar effects in May and November. In the previous study (Franěk et al., 2018), the walking route was located in green areas with a relatively small amount of traffic. The first part of the route was a street with driving cars located along a small park or a meadow with threes; the second part was a dense oak alley that led out into the street with traffic. What is the effect of spring birdsongs when walking along a busy road or when walking in the fall?

The current study continues the previous investigation (Franěk et al., 2018) by using an identical methodology but in different locations and in two diverse seasons. While the previous experiment was conducted in a relatively calm area with a greater amount of greenery, in the current study, the participants were asked to walk along a route that contained areas near both a busy road and a calm alley to test, whether the observed associations between sounds and walking behavior would also occur in a different environment. Specifically, our objective was to determine whether the effect of nature sounds (birdsongs) also occurs in an environment with a high level of traffic, which may not be perceptually congruent with nature sounds. Furthermore, to study the effects of the environmental features and perceived congruence between the sounds and the environment more precisely, the experiments were conducted on an identical route during two vegetation periods - in November and in May. As in the previous study, we used three experimental conditions: (1) listening to a soundtrack with birdsongs, (2) listening to a soundtrack with crowded city noise, and (3) listening to no soundtrack (control condition). However, walking without listening to a soundtrack cannot represent a true control condition in the current study because, in contrast with the outdoor environment used in the previous study, high levels of noise occur in some areas. On the other hand, it is almost impossible to have a perfect control condition in a real outdoor environment. Thus, the control condition in the present experiment represents the real outdoor noise. In summary, the aim of the current study was to investigate the effects of listening to diverse environmental sounds on walking speed while walking on a route in a real outdoor urban environment in two different seasons. Consistent with previous findings, it was presumed that listening to the soundtrack with intense crowded city noise would increase walking speed, while listening to the soundtrack with nature sounds would decrease walking speed. We also supposed that listening to the soundtrack with intense crowded city noise may increase walking speed in comparison to that when listening to real outdoor noise. Furthermore, the effect of congruence between the environment and sound was examined. It was hypothesized that the effect of bird sounds would be more salient in the spring than in the fall because of congruence between the sounds and the environment. Finally, the effects of both soundtracks on the evaluation of the walking route and the walking experience were examined. It was hypothesized that the walk would be evaluated as more pleasant while listening to the soundtrack with birdsongs than when listening to the soundtrack with traffic noise and under the real outdoor noise condition.

Stimulus material

There were three conditions. The participants listened to a soundtrack with relaxation nature sounds or a track with traffic noise; they did not listen to any soundtrack under the real outdoor noise condition. The soundtrack from the video “Forest Birdsong - Relaxing Nature Sounds - Birds Chirping”, available on YouTube (https://www.youtube.com/watch?v=Qm846KdZN_c), was selected as the nature sound. The track consists of the sound of birds singing. The track lasted approximately 2 h. The soundtrack from the video “Hectic Kolkata (Calcutta)—India”, available on YouTube (https://www.youtube.com/watch?v=IFc2KhKLiho), was selected as the crowded city noise. The track contains noise from motorized vehicles, engine sounds, automobile horns, and human voices. The track was modified using the software Audacity because of its short length to have a length of 38 min and 43 s, corresponding to the length of the participants’ walk. The sound levels of the tracks were adjusted to a comfortable level and did not differ among participants. The mean sound pressure levels were as follows: forest birdsong—56 dB(A) (a relatively high magnitude of the mean sound pressure level reflects higher differences between thesound levels of a silent background and the birdsongs), and crowded city noise—53 dB(A). The participants listened to the tracks, which were played on a cell phone.

The sounds were listened to through lightweight headphones Genius HS-M200C. For safety reasons, headphones did not entirely mask sounds from the outside. Participants assigned to the real outdoor noise condition did not wear headphones while walking.

Walking route

The experiment was conducted in Hradec Králové in the Czech Republic. This city is located in the northeastern part of the Czech Republic and has approximately 100,000 inhabitants. The walking route was a circuit with a length of approximately 1.8 km (see Fig. 1). The route was located in areas with minimal movement of other pedestrians; thus, walking speed was not affected by the flow of other pedestrians. Eight sections were chosen for the analysis of walking speed. The sections were chosen to provide a direct route and avoid crossroads or other obstacles. The lengths of the sections ranged from 45 to 100 m. The first part of the route was an oak alley along a river (sections 1–3). There was no street with traffic close to this area. Traffic noise in these sections was approximately Lday = 55–60 dB(A) (Hlukové mapy, 2012). Section 4 was located still in the oak alley, but it was closer to a road with dense traffic. Traffic noise was approximately Lday = 60–65 dB(A) (Hlukové mapy, 2012). Sections 5 and 6 went along a busy road, where cars, buses, and trucks were driving. On one side of the route, there was a busy road, while on the other side, there were low trees and open areas with a lawn. Traffic noise in these sections was approximately Lday = 75–80 dB(A) (Hlukové mapy, 2012). The final part of the route (sections 7 and 8) went along a different street with less traffic. On one side of the route was a street, while on the other side, there were low trees and open areas with a lawn. The traffic noise in section 7 was approximately Lday = 65–70 dB(A), and in section 8, it was approximately Lday = 55–60 dB(A) (Hlukové mapy, 2012).

Measurement of walking speed

The participants walked with a small video camera (i.e., a Sony Bloggie MHS-PM5K) on a belt around their waist (size 19 × 108 × 55 mm, weight 110 g). The environment, the participant’s feet, and the participant’s arms were captured through a fish eye lens. Beginning and end of each section of the route was indicated by a line drawn on the sidewalk. An evaluator marked two frames of the video recording to create the beginning and end of the annotation for each particular track section in the software Elan (see https://tla.mpi.nl/tools/tla-tools/elan/). Every annotation represented the entire section of the track, so that the extent of time subjects spent there could be determined. This enabled to calculate the average speed reached by the participants in all sections.

Evaluation of walk experience

The participants rated their experience during their walk and their enjoyment of the environment using the following four items: (1) I was fine during the walk, (2) It was a pleasant time, (3) I liked the route I went through, and (4) The sounds I listened to from my headphones bothered me. They were required to rate agreement or disagreement with these items using a 7-point Likert-type scale with anchors 1 = absolutely disagree and 7 = absolutely agree.

Data analysis

The mean walking speeds were calculated for specific sections of the route. A two-way mixed analysis of variance (ANOVA) was conducted to access the effects of the condition (forest birdsong, crowded city noise, real outdoor noise), and the section of the route (1–8) on walking speed. Greenhouse-Geisser correction was applied based on the test’s Epsilon (ε) where assumption of sphericity was violated as assessed by Mauchly’s test of sphericity. Differences between reported evaluations of the walk experience under particular conditions were compared by using a one-way ANOVA or t-test for independent samples. Some participants did not answer to questions related to the evaluations of the walk experience. They were excluded from the statistical analysis. Statistical analyses were conducted using IBM SPSS Statistics 24.

Ethical statement

Ethical approval for the experiments was obtained from the Committee for Research Ethics at the University of Hradec Králové (No. 2/2018). All participants provided written informed consent. The participants signed a consent declaration in which they declared that they voluntarily participating in the experiment and that they were informed about the experimental procedure. They could decide to stop being a part of the research study at any time without explanation. There were no known risks for participants in this study. The collected data were anonymised and used for the research purposes only.

Methods

Results

Analysis of walking speed

The mean time of walking for total route was 12.90 min (SD = 1. 28). The results revealed an overall faster walking speed under the crowded city noise condition (mean = 1.62 m/s, SD = 0.16) and the real outdoor noise condition (mean = 1.60 m/s, SD = 0.13) compared with the forest birdsong sounds condition (mean = 1.56 m/s, SD = 0.17). The mean walking speeds for specific sections of the route are shown in Table 2 and Fig. 2.

Table 2:

The mean walking speeds (m/s) in specific sections of the route in Experiment 1.

Section	Forest birdsong		Crowded city noise		Real outdoor noise
	Mean	SD	Mean	SD	Mean	SD
1	1.53	0.18	1.61	0.20	1.59	0.14
2	1.54	0.20	1.60	0.16	1.58	0.15
3	1.55	0.19	1.59	0.19	1.59	0.13
4	1.56	0.20	1.61	0.18	1.60	0.14
5	1.58	0.17	1.63	0.16	1.61	0.15
6	1.59	0.18	1.65	0.22	1.61	0.13
7	1.58	0.17	1.62	0.17	1.60	0.12
8	1.59	0.20	1.64	0.17	1.61	0.13

DOI: 10.7717/peerj.7711/table-2

The mean walking speeds were calculated for specific sections of the route. A two-way mixed ANOVA was conducted to analyze the effects of acoustic conditions (forest birdsong, crowded city noise, or real outdoor noise) and the route’s environmental properties (the section of the route) on walking speeds. The acoustic condition was selected as the between-subjects factor, and the section of the route was selected as the within-subjects factor. The walking speed was selected as the dependent variable. The dependent variable was normally distributed for nearly all combination of the levels of the between-subjects and within-subjects factors (acoustic condition and section walked), as assessed by Shapiro–Wilk’s test (p > .05), except for section 6 under the crowded city noise condition (p = .007).

A two-way mixed ANOVA indicated a statistically nonsignificant between-subjects main effect of the condition (F_2,84 = 0.886, p = .416, η2 = .021). Furthermore, the ANOVA indicated a statistically significant within-subjects main effect of the section of the route (F_{3.209,269.592} = 5.874, p < .001, η2 = .065, ε = .458). A post hoc analysis with a Bonferroni adjustment showed that the participants’ walking speeds differed statistically significantly in the following pairs of sections: 1–6, 1–8, 2–5, 2–6, 2–8, 3–5, 3–8, 4–8, and 7–8. There was no statistically significant interaction between the acoustic condition and the section of the route (F_{6.419,269.592} = 0.604, p = .738, η2 = .014, ε = .458).

Discussion

The results showed some differences in average walking speeds under specific conditions, but these differences were small and statistically nonsignificant. The participants who listened to birdsong sounds walked slightly slower than the participants under the crowded city noise and real outdoor noise conditions, but in contrast to the previous experiment (Franěk et al., 2018), these differences were not significant. We assumed that the small effect of these acoustic stimuli was caused by incongruence between spring or early summer bird sounds and the appearance of the environment in late fall. The small and nonsignificant differences between the crowded city noise condition and the real outdoor noise condition likely reflect the fact that the outdoor environment was generally noisy; thus, listening to the crowded city noise soundtrack and hearing the ambient noise under the control condition did not result in a substantial difference.

The statistical analysis revealed statistically significant differences in walking speeds only among particular sections of the route that reflected both their environmental character and noise level. In general, there was a slower walking speed in sections 1–4, which were located in the oak alley, with a low level of noise. In contrast, participants walked faster in sections 5 and 6, which had a high level of noise and traffic. Finally, the participants statistically significantly increased their walking speed in the last section of the route. This increase may simply reflect the fact that they had already seen the end of the route and decided to quickly finish their task.

Although many differences in walking speed among the consequent sections are visible in the graphical representations of the data (Fig. 2), the post hoc test did not indicate statistical significance (except for the differences between sections 7 and 8). This suggests that these differences in walking speed may reflect tiny combinations of the effects of visual features of the environment, acoustic conditions, immediate changes in atmospheric conditions and individual variables. The difference between sections 7 and 8 could also be explained by the participants’ intention to quickly finish their task.

The results also revealed that the type of acoustic stimulus to which participants listened during the walk had only a small effect on their evaluation of the walk experience because significant differences were found only for the statement “It was a pleasant time” between the forest birdsong and the crowded city noise conditions. These findings also reflect the effects of incongruence between the sounds and the environment.

Methods

Results

Discussion

In contrast to Experiment 1, the statistical analysis revealed a significant effect of the acoustic condition on walking speed. The participants who listened to birdsong sounds walked statistically significantly more slowly than the participants under the crowded city noise condition. Although we found that the mean walking speed was slower under the condition of birdsong sounds than under the real outdoor noise conditions, this difference was not statistically significant. The data plotted in Fig. 2 show clear differences between Experiment 1 and Experiment 2; specifically, there were substantially greater differences in walking speed under the birdsong sound condition and the other two conditions. Because this experiment was conducted in the spring, it is assumed that the birdsong sounds were perceived as more congruent with the spring environment and therefore had a greater relaxing effect than the same sounds presented in the fall, which resulted in a considerable decrease in walking speed on the route.

The analysis also showed the significant effects of the sections of the route. As expected, the walking speed was slower at the beginning of the route in the section located along an oak alley with a low level of traffic noise. The acceleration of the walking speed under all conditions in section 2 may be explained in terms of a possible negative reaction to a specific place in this section—a damaged fence painted with graffiti on the left side of the route. Although this location had the same appearance in both experiments, it may have been less visually congruent with the surrounding environment in beautiful spring vegetation. Consistently, in a study by Franěk, Van Noorden & Režný (2014), in which the participants were asked to walk along a different route, they sped up close to a house with a damaged facade. As expected, in sections 5 and 6, which had the highest level of traffic noise, the walking speed increased, while in sections 7 and 8, which had lower levels of noise, the walking speed decreased. In contrast to fall, the participants did not speed up along the final section of the route. This difference might reflect the effect of the season; in springtime, because of the warmer temperatures and lively vegetation, the walk was more pleasant than that in fall; thus, the participants might not have intended to quickly finish their task.