Differences in Public Perceptions of Recovery in Different Urban Forests Based on Birdsong

Abstract

Urban forests are important public recreation sites and play an important function in urban biodiversity conservation. Recent studies have found that urban forests have a beneficial effect on public physiological and psychological health. Birdsong in urban forests has good physiological and psychological recovery benefits and is also the most frequent natural sound. In this paper, a blank group, visual perception, and audiovisual perception were studied by investigating birds in five kinds of urban forests and simulating their birdsong environment. The results showed the following: (1) The visual perception and audiovisual perception of urban forests are restorative to the public’s physiology and psychology. Forests and urban parks with higher natural elements have relatively strong physiological and psychological recovery benefits, and roadside green spaces with higher artificial elements have relatively weak recovery benefits. However, both natural and artificial urban forests can show better recovery. (2) Birdsong perception preference has a mediating effect on the recovery benefit of heart rate and a moderating effect on the recovery benefit of skin conductance, which can affect the recovery benefits of urban forests on public physiology and psychology. (3) Because people’s perceptions of the timbre of birdsong vary, the benefits of birdsong on recovery are not always favorable. The richness and pleasantness of birdsong in the forest environment are significant contributors to the recovery advantages and the likability and comfort of birdsong in the rest of the setting. Based on differences in the composition of birds in different urban forest environments, this study simulated the birdsong environment with birds based on field research and studied the public perception recovery of the urban forest visual environment and the urban forest and birdsong audiovisual environment in order to explore the influence of birdsong on the public health recovery properties of urban forests. This study provides an optimal strategy for the selection and construction of forest healing functions and promotes the development of urban ecosystems and biodiversity.

1. Introduction

Urban forests play an important function in providing the public with an important ecological place for recreation and urban biodiversity conservation, and regular visits to urban green spaces can reduce the risk of heart disease, respiratory disease [1,2], and obesity [3]. Urban forests have been shown to promote public mental health by lowering stress [4], lowering the risk of mental diseases [5], and even enhancing sleep quality [6]. Urban forests can foster social ties between city dwellers and their neighbors, which is beneficial for urban well-being [7], among other things. Walking in urban forests can dramatically lower public epinephrine levels compared to urban streets and buildings. It also has a positive effect on dopamine, serum cholesterol, and salivary cortisol levels [8,9]. Salivary amylase and salivary cortisol levels can be considerably lowered after 20 min of exposure to a setting with a greater percentage of green space [10]. In addition, recreation in the forest lowers blood pressure and restores heart rate variability and heart rate. It can also increase parasympathetic nerve activity and decrease sympathetic nerve activity so that the human body is in a more relaxed state [11]. The public will see a positive physical and mental health impact when recreating in the urban forest environment, mainly because it contains natural environmental elements such as trees and shrubs, which are beneficial to human health [12,13]. Current research has mostly examined the physiological impacts of urban forests on the public using a variety of indicators, including blood pressure [14,15], heart rate (HR) [14,15], heart rate variability (HRV) [15,16], and skin conductance (SC) [15,16]. Research findings indicate that urban forests have a beneficial effect on public physiological health, reducing blood pressure, restoring heart rate, and skin conductivity [14,15,16]. The State of Anxiety Scale (STAI) [14,15], the State of Mind Scale (POMS) [14], and the Perceptual Recovery Scale (PRS) [17] have been previously used to gauge subjects’ level of attention to examine the psychological effects of urban forests on the public. Neuroscience and EEG measurement technology can also be used [18,19]. In urban forests, people have been shown to be able to control their emotions, lower stress levels [20,21], alleviate weariness, and enhance concentration [22]. Some studies have noted that different landscape scenes have different recovery effects on people because of different landscape characteristics [23]. Different types of urban forest environments have different perceived recovery properties [17,24], and the public’s environmental preferences also affect how urban forest recovery is evaluated [25]. In summary, the urban forest environment has a positive effect on the physical and mental health of the public.

The public primarily experiences the urban forest environment through vision, hearing, smell, taste, and touch [26]. However, current studies on the physical and psychological recovery benefits of urban forests for the public primarily focus on vision, including the visual perception of natural elements [27], landscape morphology [28], tree coverage [29], and other visual perceptions. It has been demonstrated that the most crucial sense during the course of recovery is visual perception [30]. However, compared to visual scenes, in some cases, sound has a more significant effect on people’s recovery from fatigue and worries [31]. Payne et al.’s survey of 400 park users in Sheffield, England, revealed that the soundscape of urban parks was crucial to the restorative experience [32]. In the context of nature-based rehabilitation (NBR), Cerw et al. found that nature sounds were part of a ‘pleasant’ and ‘quiet’ experience, which contributed to recovery [33]. According to research, natural sounds will help people regain their focus, while machine and traffic noises have the opposite effect [34]. Agricultural sounds have the biggest impact on SC, and the sound of flowing water has the biggest impact on HR, according to studies on physiological health [35]. It is apparent that visual and auditory perceptual modalities are not independent but rather interact and enhance one another in a complex way [36]. Visual information influences the auditory perception of both natural and artificial sounds, and audio information influences the visual perception of visual elements [37]. In audiovisual environments, paying attention to visual stimuli decreases the perception of sound, and vice versa [38]. Naturally visual landscapes can often mitigate the negative effects of noise [39]. The interaction of visual and auditory features of the environment can significantly influence environmental evaluation, and audiovisual interactions can lead to a better sense of engagement and immersion in the subject [40]. For instance, coherence between sound and image affects preferences, and sound adds information to the visual information [41]. Acoustic landscape and visual landscape matching combinations can enhance aesthetic preferences [42], and coherence between soundscape and visual landscape usually enhances one’s experience of the place [43]. Anthropogenic activities and their associated sounds are negative elements, while natural visual elements (such as bushes, flowers, and bodies of water) and acoustic elements (such as the sound of flowing water, wind-blown plants, and birdsong) are positive elements that restore benefits and aesthetic preferences. Soundscapes and visual landscapes should be viewed as an inseparable whole, and the combination of audiovisual stimuli typically has better health benefits than single visual stimuli [42]. Current research on visual and auditory landscapes still frequently overlooks the impact of audiovisual interactions in favor of concentrating on the perceptual assessment of single-channel inputs. More research on sensory interactions is required because urban woods tend to provide multisensory experiences for the public.

Birds have strong bioacoustic adaptability in urban environments and can adjust their vocal frequency and pressure for purposes such as communicating, warning, territory calls, and mating calls [44]. From the perspective of the public’s response to bird sounds, bird sounds have good physiological and psychological recovery benefits, and bird sounds are also the most popular natural sound [45]. Research has revealed that the type of street space affects the healing potential, and artificial–natural enclosed and natural semi-enclosed streets are the street types with the best healing effect. When the total sound pressure level was 55 dB(A), the sound combination with birdsong accounting for 70% had a significant positive effect on improving the healing effect of rural community streets [46]. In urban forests, birdsong can improve human health [33], and when compared to insect song or a combination of bird and insect sounds, birdsong most clearly lowers people’s respiratory rates [47]. Additionally, it has been demonstrated that birdsong is the most desired soundscape in biophilic cities, as it tends to boost public health and the pleasant sensations of people [48]. Comparing the effects of a single birdsong versus many birdsongs on mental resilience in natural settings, it was found that both were effective [49]. Furthermore, some researchers have noted that the same sound source in different functional areas, different sources of sound, public perceptual preference, and perceptual harmony have different perceptual characteristics and differences. For example, birdsong in a forest landscape area has a more obvious perception preference advantage than in a human landscape area, and the perception of harmony is also higher in a waterfront tourist area [50]. Although it is still unclear exactly why birds have particular health advantages and contribute to subjective human well-being, birds’ chirping sounds are a key factor [12]. Although the specific reasons why birds contribute to subjective well-being in humans and have some health benefits remain to be explored, birds’ chirping sounds are important components. In natural environments, birdsong has been studied in conjunction with other natural and biological sounds, and its benefits for public preference and health have been investigated. Current research primarily focuses on analyzing the differences in public response between birdsong and other biological, artificial, or mechanical sounds and has found that birdsong has a better physiological and psychological recovery effect. However, most research has concentrated on generalized ‘birdsong’. The characteristics of birdsong are influenced by environmental and vegetation factors, as well as its own response mechanisms [39], but the recovery effects of birdsong on different bird species in different environments remain to be investigated.

In summary, urban forests are important for public physiological and psychological recovery, and bird sounds in urban forests are important in recovery landscape construction and in the enhancement of urban residents’ health and well-being [51,52,53]. However, the environmental characteristics of urban forests are different, and the distribution of birds in urban forests is also different, so studies of bird sounds in different urban forest environments should accommodate the differences in bird composition. The differences in the recovery functions of the audiovisual environments of different urban forests also need to be further studied. This paper fully considers the differences in bird species in various urban forest environments, summarizes bird species and birdsong environments in various environments based on extensive bird species surveys, highlights the results of targeted research on public perceptions in urban forests and bird sound audiovisual environments, and examines the differences in the recovery potential of various types of forest environments to provide optimal enhancement strategies for the selection and construction of sites to enhance a forest’s healing function.

2. Materials and Methods

2.1. Study Area

Kunming is the capital of Yunnan Province and one of the central cities in Southwest China. The altitude of most parts of the city is between 1500 and 2800 m, and the average altitude of the urban area is 1891 m. It is a low-latitude plateau mountain with a monsoon climate. The average annual temperature is about 15 °C, the average annual sunshine is about 2200 h, and the average annual precipitation is about 1000 mm. The rainy season is concentrated in May–November, and the dry and wet seasons are distinct. The weather is sunny and sunny. Kunming is rich in birds, and the birds are mainly from the Eastern region. At present, 266 species and 19 subspecies of birds are found. There are 165 species of breeding birds and their subspecies, 102 species of the Eastern region and their subspecies, and 38 species of the Palaearctic realm and their subspecies.

2.2. Bird Survey of Kunming

According to the natural landscape pattern of Kunming, a 5 km × 5 km grid axis was set up by ArcGIS10.7 with reference to Google Earth’s non-offset satellite images (shown in Figure 1). In the axis grid, five types of urban forest environments—forest, wetland, urban park, street green space, and residential green space—were selected according to factors such as green space area, habitat type, vegetation condition, and land use type. Through field research, it was found that most of the forest and wetland spaces in urban or suburban areas of Kunming have been built into forest parks or wetland parks. A bird survey was conducted in 18 park green spaces, 18 street green spaces, and 15 residential green spaces, and a monthly survey was completed from September 2022 to July 2024. Before the survey, the sample plots were explored on the spot, and the number and location of the survey lines were selected according to the area and shape to ensure that they could represent different urban forest types and included various bird habitat types in the area.

The survey was conducted on sunny and windless days. Considering that birds are less active in the middle of the day, the survey was conducted from early morning to 11:00 a.m. and from 3:00 p.m. to dusk, depending on the size of the sample plots, and the duration of the survey ranged from 20 min to 2 h. During the survey, two telescopes (BOSMA donkey friend II8 × 42 and BOSMA wild wolf II10 × 50) were used to travel along a fixed line at a speed of 2 km/h. The bird species, quantity, behavior, place of stay, and vegetation habitat within 25 m on both sides of the transect were recorded. Bird identification and classification were based on the ‘A Field Guide to the Birds of China’ [54] and ‘A Checklist on the Classification and Distribution of the Birds of China (3rd Edition)’ [55]. The birds were recorded for bird community diversity and distributional characteristics using a Relative Breadth Frequency Index (RB) with the following equation:

RB = (d × S)/D

(1)

where d is the number of days of encountering a bird species, S is the total number of encounters with the bird species, and D is the total number of surveys in all habitats where the bird species was encountered.

Through the analysis and comparison of the research results and the biodiversity index, this study finally chose the following sites closest to the average of the biodiversity index: Jindian Forest Park, Laoyu River Wetland Park, Pubu Park, Xijing Temple Street Green Space, and Huaxia Tianjingwan Residential Green Space. The five sample sites represent the five types of urban forest environments: forest, wetland, urban park, street green space, and residential green space, respectively. The first two are natural urban forest environments, and the last three are artificial urban forest environments. Kunming is one of the primary locations for wintering migratory birds in China. It has many migratory birds crossing the border or overwintering in the fall and winter, with Chroicocephalus ridibundus being the most common, and the structure of bird communities tends to be stable during the summer, with only minor variations in bird species and numbers [56]. Summer birds were ultimately selected as the research object in this study in order to minimize the impact of the season and migratory birds on this experiment. The results of the summer bird field surveys conducted in the five environments mentioned above are shown in

Table 1. Summer bird list of urban forest types.

Environment	Species	Order	Family	Bird Ecotypes	RB
Forest	Passer domesticus	Passeriformes	Fringillidae	Song bird	27.33
	Parus major	Passeriformes	Paridae	Song bird	18.00
	Culicicapa ceylonensis	Passeriformes	Stenostiridae	Song bird	13.00
	Aegithalos concinnus	Passeriformes	Aegithalidae	Song bird	13.00
	Turdus dissimilis	Passeriformes	Turdidae	Song bird	11.33
	Pycnonotus xanthorrhous	Passeriformes	Pycnonotidae	Song bird	8.67
	Phylloscopus yunnanensis	Passeriformes	Phylloscopidae	Song bird	7.33
	Passer cinnamomeus	Passeriformes	Fringillidae	Song bird	5.33
	Motacilla alba	Passeriformes	Motacillidae	Song bird	2.67
	Cisticola juncidis	Passeriformes	Fantail warbleridae	Song bird	2.00
	Rhipidura albicollis	Passeriformes	Fan-tailed flycatcheridae	Song bird	1.00
	Pteruthius aeralatus	Passeriformes	Timaliidae	Song bird	1.00
	Anthus hodgsoni	Passeriformes	Motacillidae	Song bird	1.00
	Pterorhinus sannio	Passeriformes	Leiothrichidae	Song bird	0.67
	Luscinia svecica	Passeriformes	Muscicapidae	Song bird	0.67
	Copsychus saularis	Passeriformes	Muscicapidae	Song bird	0.67
	Aethopyga gouldiae	Passeriformes	Nectariniidae	Song bird	0.33
	Parus monticolus	Passeriformes	Paridae	Song bird	0.33
	Leiothrix argentauris	Passeriformes	Leiothrichidae	Song bird	0.33
Wetland	Pycnonotus xanthorrhous	Passeriformes	Pycnonotidae	Song bird	43.00
	Egretta garzetta	Ciconiiformes	Ardeidae	Wading bird	29.00
	Copsychus saularis	Passeriformes	Muscicapidae	Song bird	9.00
	Prinia inornata	Passeriformes	Fan-tailed flycatcheridae	Song bird	7.00
	Gallinula_chloropus	Gruiformes	Rallidae	Wading bird	6.67
	Upupa epops	Bucerotiformes	Upupidae	Climber bird	6.00
	Ardea cinerea	Pelecaniformes	Ardeidae	Wading bird	5.33
	Zosterops simplex	Passeriformes	Zosteropidae	Song bird	5.33
	Amaurornis phoenicurus	Gruiformes	Rallidae	Wading bird	2.00
	Turdus dissimilis	Passeriformes	Turdidae	Song bird	2.00
	Turdus mandarinus	Passeriformes	Turdidae	Song bird	2.00
	Passer domesticus	Passeriformes	Fringillidae	Song bird	1.67
	Hirundo rustica	Passeriformes	Hirundinidae	Song bird	1.67
	Lanius schach	Passeriformes	Laniidae	Song bird	1.67
	Acrocephalus orientalis	Passeriformes	Acrocephalidae	Song bird	1.33
	Motacilla alba	Passeriformes	Motacillidae	Song bird	1.00
	Aegithalos concinnus	Passeriformes	Aegithalidae	Song bird	0.67
	Acrocephalus stentoreus	Passeriformes	Acrocephalidae	Song bird	0.67
	Pterorhinus sannio	Passeriformes	Leiothrichidae	Song bird	0.33
	Pycnonotus sinensis	Passeriformes	Pycnonotidae	Song bird	0.33
	Ardeola bacchus	Passeriformes	Ardeidae	Wading bird	0.33
	Cuculus canorus	Passeriformes	Cuculidae	Climber bird	0.33
	Phylloscopus inornatus	Passeriformes	Phylloscopidae	Song bird	0.33
	Cyanopica cyanus	Passeriformes	Corvidae	Song bird	0.33
Urban park	Pycnonotus xanthorrhous	Passeriformes	Pycnonotidae	Song bird	77.00
	Pycnonotus aurigaster	Passeriformes	Pycnonotidae	Song bird	53.00
	Egretta garzetta	Ciconiiformes	Ardeidae	Wading bird	23.00
	Copsychus saularis	Passeriformes	Muscicapidae	Song bird	15.00
	Turdus mandarinus	Passeriformes	Turdidae	Song bird	15.00
	Passer domesticus	Passeriformes	Fringillidae	Song bird	8.67
	Spilopelia chinensis	Columbiformes	Columbidae	Terrestrial bird	5.33
	Motacilla alba	Passeriformes	Motacillidae	Song bird	3.33
	Lanius schach	Passeriformes	Laniidae	Song bird	2.67
	Pterorhinus sannio	Passeriformes	Leiothrichidae	Song bird	2.00
	Zoothera dixoni	Passeriformes	Turdidae	Song bird	0.67
	Eophona migratoria	Passeriformes	Aegithalidae	Song bird	0.67
	Sturnia malabarica	Passeriformes	Sturnidae	Song bird	0.67
	Gracupica nigricollis	Passeriformes	Sturnidae	Song bird	0.33
	Alcedo atthis	Coraciiformes	Alcedinidae	Song bird	0.33
Street green space	Zosterops simplex	Passeriformes	Zosteropidae	Song bird	34.00
	Pycnonotus aurigaster	Passeriformes	Pycnonotidae	Song bird	11.00
	Pterorhinus sannio	Passeriformes	Leiothrichidae	Song bird	7.00
	Pycnonotus xanthorrhous	Passeriformes	Pycnonotidae	Song bird	6.00
	Copsychus saularis	Passeriformes	Turdidae	Song bird	5.00
	Motacilla alba	Passeriformes	Motacillidae	Song bird	2.67
	Aegithalos concinnus	Passeriformes	Aegithalidae	Song bird	1.00
	Egretta garzetta	Ciconiiformes	Ardeidae	Song bird	0.67
	Upupa epops	Bucerotiformes	Upupidae	Climber bird	0.67
	Passer domesticus	Passeriformes	Fringillidae	Song bird	0.33
	Spodiopsar sericeus	Passeriformes	Sturnidae	Song bird	0.33
	Spilopelia chinensis	Columbiformes	Columbidae	Terrestrial bird	0.33
Residential green space	Pycnonotus xanthorrhous	Passeriformes	Pycnonotidae	Song bird	13.00
	Zosterops simplex	Passeriformes	Zosteropidae	Song bird	8.67
	Passer domesticus	Passeriformes	Fringillidae	Song bird	5.33
	Egretta garzetta	Ciconiiformes	Ardeidae	Wading bird	5.00
	Turdus mandarinus	Passeriformes	Turdidae	Song bird	5.00
	Pycnonotus aurigaster	Passeriformes	Pycnonotidae	Song bird	2.00
	Motacilla alba	Passeriformes	Motacillidae	Song bird	1.33
	Aegithalos concinnus	Passeriformes	Aegithalidae	Song bird	1.33
	Pterorhinus sannio	Passeriformes	Leiothrichidae	Song bird	1.00
	Ardeola bacchus	Pelecaniformes	Ardeidae	Wading bird	0.67
	Hirundo rustica	Passeriformes	Hirundinidae	Song bird	0.67
	Pica serica	Passeriformes	Corvidae	Song bird	0.67
	Copsychus saularis	Passeriformes	Turdidae	Song bird	0.33
	Acrocephalus stentoreus	Passeriformes	Acrocephalidae	Song bird	0.33
	Spilopelia chinensis	Columbiformes	Columbidae	Terrestrial bird	0.33

Notes. RB is the Relative Breadth Frequency Index for the species.

.

2.3. Experimental Materials

Through the above research on birds, this study combined field research and birdsong perception experience. During the research process, environmental images were taken synchronously as visual stimulation materials for different urban forest environments. After investigation, the birdsong of different birds was obtained by Xeno-canto, and the background noise was small. Yunnan, or Southwest China, was selected as the collection area for screening. The greater the dominance of birdsong, the greater the probability that the public may perceive the birdsong. Therefore, this study selected RB ≥ 5 as the main birdsong environment for this type of urban forest environment and further analyzed the composition of the birdsong soundscape in different urban forest environments. Then, the above birds were summarized, and birdsong in the environment was simulated according to the call habits of the species, combined with the call behavior and actual situation recorded in the survey. Most of the time, the birdsong environment in this study was the common call of several birds. Noise reduction was performed using the ‘Effects’ tools ‘Noise Reduction’ and ‘Hum Removal’ in Adobe Audition 2020, and the birdsong was de-noised and edited into an experimental duration to simulate the birdsong environment in this environment. Finally, the above birdsong and environment images were synthesized into a video file to be used as an audiovisual perceptual stimulus material. The birdsong preference perception material obtained was processed into audios with a duration of 5 s.

2.4. Experimental Subjects

Considering the particularity of physiological and psychological experiments and their experimental design in this study, the sample size of each group was generally selected to be between 30 and 40 people [35,57,58]. Relevant studies generally agree that it is more feasible and representative to select college students for questionnaire surveys, visual behavior, and physiological and psychological change monitoring [59]. In this study, the subjects were randomly selected undergraduate and graduate students of a school. The symptom checklist 90 (SCL90) is widely used as a comprehensive rating of mental health and behavioral problems. Before the experiment, the subjects were tested with the SCL-90 questionnaire. According to the checklist, if the single score is less than 2 points, the number of positive items is less than 43 points, and the total score is less than 160 points, the psychological status is considered good, and the follow-up experiment can be carried out. The final number of valid experiments was 330, including 137 males and 222 females, aged between 18 and 28 years old. The undergraduates were mainly students majoring in landscape architecture, horticulture, urban and rural planning, and architecture, while the graduate students were majoring in landscape architecture. The experiment was divided into 11 groups: blank group (B), forest group (F), wetland group (W), urban park group (U), street green space group (S), residential green space group (R), and audiovisual interaction group (F/W/U/S/R-S) with bird song in the environment; each group consisted of 30 people. The birdsong timbre perception experiment was carried out separately, and the final effective number was 30 people.

2.5. Indicator Selection

2.5.1. Selection of Physiological Indicators

Two physiological indicators—heart rate (HR) and skin conductance (SC)—were recorded in real time through the Eyeso ECGM3 system (Braincraft Technology Co., Ltd. Beijing, China). Before the experiment, the physiological baseline value of each participant in the resting state was tested for 1 min. The study used the relative change rate of physiological indicators, namely the relative change rate of heart rate (ΔHR) and the relative change rate of skin conductivity (ΔSC), to reflect the differences in physiological indicators of the subjects with different environmental changes [60]. When the public is affected by stimulation, excitement, tension, or other emotions, the skin conductance will decrease and the heart rate will rise. When the ΔHR value is positive, the heart rate level after the experiment does not return to the baseline level; the smaller the value, the stronger the recovery ability of the heart rate level. When the ΔSC value is negative, the skin conductivity after the experiment does not return to the baseline level; the larger the value, the stronger the recovery ability of the skin conductivity level rate. The concrete equations of ΔHR and ΔSC are as follows:

ΔHR = (post-experimental HR − baseline HR)/baseline HR × 100%

(2)

ΔSC = (post-experimental SC − baseline SC)/baseline SC × 100%

(3)

2.5.2. Selection of Psychological Indicators

The psychological indicators were streamlined from the State–Trait Anxiety Inventory and the Profile of Mood States, which quantify the underlying psychological state through five evaluative indicators: calmness, stress, tension, agitation, and restlessness [61]. The Perceived Restorative Scale (PRS) was used to analyze the attention healing, and it was revised [62,63]. According to the needs of this study, four first-level indicators, namely, being away (Be A), fascination (Fas), coherence (Coh), and compatibility (Com), and 20 second-level indicators were selected for the questionnaire survey. The above two scales were evaluated by the Likert 5 scale. The respondents scored the relevant descriptions according to their subjective judgments and real thoughts. The lowest score was −2, and the highest score was 2. The specific questionnaire is shown in

Table 2. Psychological indicator selection and questionnaire questions.

Questionnaire Type	Indicators		Question
	1st-Level	2nd-Level
Basic Psychological State Scale	Calmness	Calm condition	Do you feel calm?
	Tension	Pressure condition	Do you feel pressure?
		Tension condition	Do you feel tense?
	Irritability	Anxiety condition	Do you feel anxious?
		Restless condition	Do you feel restless?
PRS	Being away (Be A)	A1	Being here gives me an escape from reality.
		A2	This place gives me a break from the daily routine.
		A3	This place takes my mind off things.
		A4	This place helps me relax.
		A5	This place helps me cut down on unnecessary distractions.
	Fascination (Fas)	B1	The environment here is very attractive.
		B2	There is a lot of interesting things to hold my attention here.
		B3	I am going to want to know about this place.
		B4	There are so many things to explore and discover here.
		B5	I would like to spend more time taking in my surroundings.
		B6	The environment here is fascinating.
	Coherence (Coh)	C1	Everything here is in its place and very harmonious.
		C2	There is plenty of space to explore in many directions.
		C3	There is a lot to concentrate here.
		C4	The arrangements here are well organized.
	Compatibility (Comp)	D1	This is very suitable for me.
		D2	I can do what I like to do here.
		D3	I feel like I belong here.
		D4	I can find a way to enjoy myself here.
		D5	I think I am integrated with the environment here.

:

Among them, the quantitative scores of the four first-level indicators of escape, charm, consistency, and compatibility of the attention recovery perception recovery scale were taken as the final quantitative scores by taking the average of their respective second-level indicators. The mental state takes into account some non-instant psychological effects, reduces the evaluation state into three types, and further summarizes them as calmness (C), tension (T), and irritability (I) [61]. The degree of calmness (C) is the score of calm condition, the degree of tension (T) is the sum of the scores of pressure condition and tension condition minus the score of calm condition, and the degree of irritability (I) is the sum of the scores of anxiety condition and restless condition minus the score of calm condition. In this study, relative changes were also used in the study of public psychological state, including the relative change of calm degree (ΔC), the relative change of tension degree (ΔT), and the relative change of irritability degree (ΔI). The specific change equations are as follows:

ΔC = post-experimental C − baseline C

(4)

ΔT = post-experimental T − baseline T

(5)

ΔI = post-experimental I − baseline I

(6)

2.5.3. Indicators of Public Perceptual Preferences for Birdsong Timbre

This study investigated the public preference for the timbre of birdsong using four indicators. The specific contents of the questionnaire are shown in

Table 3. Birdsong timbre perception preference indicator selection and questionnaire questions.

Indicators	Question
Likability (Lik)	Do you like this birdsong?
Pleasantness (Ple)	Do you think this birdsong sounds pleasant?
Richness (Ric)	Do you think this birdsong sounds rich?
Comfort (Comf)	Do you think this birdsong sounds comfortable?

:

Before data analysis, Cronbach’s alpha coefficient test was performed on all questionnaire data. The basic psychological state scale was 0.702, the attention recovery perception recovery scale was 0.844 (Be A, Fas, Coh, and Comp were 0.685, 0.704, 0.704, and 0.856, respectively), and the public birdsong preference scale was 0.828, all of which were greater than 0.6, meaning they could be further analyzed. In the first stage, 1 min of resting with eyes closed was used as the baseline value, and a psychological state questionnaire was made. In the second stage, the thriller and disaster film fragments were used as simulated pressure stimulation. In order to reduce the fluctuation of physiological data caused by the movement of the subjects, the psychological state was carried out by inquiry. In the third stage, the blank group was conducted by resting with eyes closed. The remaining 10 groups were stimulated with urban forest visual or audiovisual interactive materials. A total of 8 pictures were set up in each group, and each picture was played for 15 s.

2.6. Experimental Design

The experiment in this study was conducted separately in a relatively closed indoor space, and the experimental time of each participant was about 11 min. The audiovisual interaction group wore Honor Earbuds X2 wireless headphones (HONOR, Shenzhen, China) in uniform to listen to birdsong. The entire experiment was carried out in a relatively closed laboratory, and laboratory staff were required to remain silent during the experimental sessions to prevent the influence of other sounds on the study. Before the experiment, the subjects were tested by SCL-90. The result score was less than 60, and the psychological state of the subjects was determined to be normal or better. The physiological data acquisition equipment was worn by the subjects, and the sitting position of the experimental personnel was adjusted to ensure that the head was flat and about 75 cm away from the screen. The experimental process was then introduced to the subjects. The experimental process was divided into three stages: experimental preparation and reference value measurement, simulation pressure, and urban forest simulation experience. In the first stage, the eyes were closed and rested for 1 min as the baseline value, and the psychological state questionnaire was carried out. In the second stage, thriller and disaster film clips were used as simulated stress-inducing stimuli, and efforts were made to reduce fluctuations in physiological data caused by subject movement; the psychological state was questioned. In the third stage, the blank group was performed in a closed-eye resting manner. The video was closer to outdoor perception in landscape perception [64], so the perception group used urban forest visual or audiovisual interactive materials for stimulation. Each group set up a total of 8 pictures, and each picture played for 15 s, pictures of each environment are shown in the Appendix A. The audiovisual interaction group materials were composited in advance into the same MP4 file with video and audio, which were synchronized and played during the experiment. After the experiment, the subjects were guided to fill out the Basic Psychological State Scale and PRS and were assisted to remove the equipment. The evaluation of public perception of birdsong was carried out separately. The specific procedure involved the subjects listening to a certain birdsong for 5 s and then completing the questionnaire for that birdsong within 10 s after the end of the experiment. The next birdsong was then played, and the process was repeated until all 19 birdsongs were played to complete the evaluation. The experimental flow is shown in Figure 2.

2.7. Data Processing and Analysis

Data statistics were calculated using Excel 2023, and data analysis was performed using SPSS 24 (Amonk IBM, New York, NY, USA). In this study, reliability analysis was used to analyze all the scales; Pearson correlation was used to analyze the correlations between psychological indicators and birdsong preference indicators; validity analysis was used to validate the psychological scales, the perceived recovery scales, and the birdsong preference scales; principal component analysis (PCA) was used to streamline the psychological scales and the birdsong preference scales; one-way ANOVA was used to analyze the results of the experiments in the five urban forest environments; mediation effect analysis and moderating effect analysis were used to analyze the mediating or moderating effects of birdsong preference; and partial least squares (PLS) to explore the main factors influencing the physiological and psychological indicators of birdsong preference. The thresholds of statistical significance mentioned above were established at p < 0.05.

3. Results

3.1. Effects of Different Urban Forest Environments on Public Physiological Health Indicators

In the visual perception group, the distribution of ΔHR data was basically normal (the histogram showed approximate normal distribution), but the Levene test showed that the variance was not uniform (p = 0.000, <0.05), so Welch’s ANOVA was used. Shapiro–Wilk test satisfied the normal distribution of ΔSC data (W = 0.998, p = 0.996), and the Levene test showed homogeneity of variance (p = 0.822, >0.05), so one-way ANOVA was used. As shown in

Table 4. Differences in the relative change rates of public physiological health indicators in different urban forest environments.

		Environment Types (Mean ± SD)						F	p
		Blank	Forest	Wetland	Urban Park	Street	Residential
Visual	ΔHR	0.073 ± 0.147	−0.013 ± 0.037	0.010 ± 0.067	−0.008 ± 0.053	0.012 ± 0.025	−0.004 ± 0.023	3.794	0.004 **
	ΔSC	−0.058 ± 0.154	0.068 ± 0.132	0.054 ± 0.128	0.079 ± 0.129	0.019 ± 0.364	0.030 ± 0.142	3.863	0.002 **
Audiovisual	ΔHR	0.073 ± 0.147	−0.017 ± 0.033	0.018 ± 0.057	−0.012 ± 0.021	0.007 ± 0.050	−0.009 ± 0.037	4.137	0.002 **
	ΔSC	−0.058 ± 0.154	0.079 ± 0.142	0.049 ± 0.102	0.100 ± 0.115	0.023 ± 0.127	0.042 ± 0.108	5.675	0.000 ***

Notes. **: p < 0.01; ***: p < 0.001.

, for the HR indicator, the ability of different environments was ranked as forest > urban park > residential green space > wetland > street green space, with the forest environment being the strongest and street green space being the weakest. For the SC indicator, the ability of different environments was ranked as urban park > forest > wetland > residential green space > street-side green space, the urban park environment being the strongest and street green space being the weakest. Both indicators of different urban forest environments showed recovery ability compared with the control group.

In the audiovisual interaction group, the distribution of ΔHR data was basically normal (the histogram showed approximate normal distribution), but the Levene test showed that the variance was not uniform (p = 0.000, <0.05), so Welch’s ANOVA was used. Shapiro–Wilk test satisfied the normal distribution of ΔSC data (W = 0.995, p = 0.789), and the Levene test showed homogeneity of variance (p = 0.272, >0.05), so one-way ANOVA was used. ΔHR and ΔSC were also found to be highly significant in different urban forest environments with birdsong. As shown in Table 4, for the HR indicator, the ability to recover from different environments was ranked as forest > urban park > residential green space > wetland > street-side green space, with the forest environment being the strongest and the street-side green space the weakest. For the SC indicator, the ability to recover from different environments was ranked as urban park > forest > wetland > residential green space > street-side green space, with the urban park environment being the strongest and the street-side green space the weakest. Both indicators of different urban forest environments showed recovery ability compared with the control group.

In this study, we conducted Welch’s ANOVA on 11 sets of ΔHR experimental results and one-way ANOVA on ΔSC experimental results, and the results are shown in Figure 3a,b. They show that there was a certain physiological restoration benefit of the forest landscape compared to both visual and audiovisual perception in the blank group. There were some differences between the resilience of visual and auditory forest environments with birdsong and the resilience of visual perception, but the physiological recovery benefit of the wetland audiovisual environment was not as good as that of the simple visual environment.

3.2. Effects of Different Urban Forest Environments on Public Psychological Health Indicators

3.2.1. Effects of Different Urban Forest Environments on Basic Psychology

The three indicators ΔC, ΔT, and ΔI were analyzed by Pearson correlation, which showed a significant (p = 0.000) correlation between all three indicators. Then, the PCA was used to simplify the above three indicators, which were expressed by the psychological change indicator (ΔP). After transformation, the Kaiser–Meyer–Olkin index was 0.735 (>0.6). The Bartlett sphericity test was p = 0.000 (p < 0.05), and the cumulative variance interpretation rate was 78.728%. Therefore, ΔP was as follows:

ΔP = −0.577 × ΔC + 0.585 × ΔT + 0.570 × ΔI

(7)

Through the above calculation method, it was found that ΔP was inversely proportional to ΔT and ΔI was proportional to ΔC. The histogram test revealed that the ΔP for each perception type and overall the ΔP data for the 11 groups were accepted as normally distributed, but Levene’s test indicated that all were heterogeneous (p = 0.034, <0.05; p = 0.042, <0.05; p = 0.024, <0.05); all used Welch’s ANOVA.

In the visual perception group, the study found that the ΔP of different urban forest environments was significantly different in different urban forest environments. The resilience of the indicators of the basic psychological state was forest > urban park > wetland > residential green space > street green space, and different urban forest environments had significant resilience. The results are shown in

Table 5. Differences in the relative change rates of public psychology indicators in different urban forest environments.

		Environment Type Group (Mean ± SD)						F	p
		Blank	Forest	Wetland	Urban Park	Street	Residential
Visual	ΔP	0.925 ± 2.151	−3.640 ± 2.629	−2.679 ± 3.432	−3.177 ± 2.981	−2.061 ± 2.039	−2.254 ± 2.892	14.023	0.000 ***
Audiovisual	ΔP	0.925 ± 2.151	−3.698 ± 2.952	−2.527 ± 3.957	−3.350 ± 2.522	−2.235 ± 3.009	−2.563 ± 2.997	14.591	0.000 ***

Notes. ***: p < 0.001.

.

In the audiovisual perception group, it was found that the ΔP of different urban forest environments was significantly different in different urban forest environments. The resilience of the indicators of the basic psychological state was forest > urban park > wetland > residential green space > street green space, and different urban forest environments had significant resilience. The results are shown in Table 5.

The results of the 11 groups of experiments were analyzed by ANOVA, and the results are shown in Figure 3c. It was found that compared with the blank group, the urban forest had a certain psychological recovery benefit regardless of whether there was birdsong interaction, but the psychological recovery of the wetland landscape with birdsong decreased.

3.2.2. Effects of Different Urban Forest Environments on Attention Recovery

By histogram verification, Be A, Fas, Coh, and Comp were accepted as having a normal distribution. The Levene test showed that all Fas values in the audiovisual interaction group met the homogeneity of variance except for the uneven variance of Fas. Therefore, Welch’s ANOVA was used for FAS in the audiovisual interaction group, and one-way ANOVA was used for the rest. From

Table 6. Differences in environmental effects on attention recovery in different urban forests.

		Environment Types (Mean ± SD)					F	p
		Forest	Wetland	Urban Park	Street	Residential
visual	Be A	0.593 ± 0.593	0.487 ± 0.716	0.533 ± 0.648	0.407 ± 0.505	0.447 ± 0.706	0.393	0.814
	Fas	0.594 ± 0.592	0.517 ± 0.677	0.606 ± 0.633	0.478 ± 0.448	0.383 ± 0.591	0.708	0.587
	Coh	0.417 ± 0.800	0.217 ± 0.793	0.333 ± 0.772	0.125 ± 0.586	−0.158 ± 0.727	2.716	0.032 *
	Comp	0.467 ± 0.846	0.387 ± 1.066	0.413 ± 0.950	0.300 ± 0.684	0.327 ± 0.835	0.171	0.953
audio- visual	Be A	0.627 ± 0.514	0.473 ± 0.605	0.567 ± 0.523	0.433 ± 0.512	0.467 ± 0.754	0.559	0.693
	Fas	0.600 ± 0.481	0.461 ± 0.696	0.611 ± 0.427	0.489 ± 0.571	0.439 ± 0.395	0.911	0.462
	Coh	0.433 ± 0.698	0.200 ± 0.735	0.383 ± 0.771	0.200 ± 0.877	0.192 ± 0.622	0.740	0.566
	Comp	0.493 ± 0.555	0.360 ± 0.868	0.453 ± 0.615	0.347 ± 0.808	0.340 ± 0.732	0.279	0.891

Notes. *: p < 0.05.

, it was found that except for the obvious difference in the ductility of the visual perception group, the remaining indicators were not significantly different between the groups for visual perception and audiovisual interaction perception. The results showed that when the visual environment of urban forest changed after adding birdsong, the ability of attention recovery changed. Specifically, the consistency experience of residential green space improved after adding birdsong.

3.3. The Effect of Public Perception of Birdsong on the Recovery Benefits of Urban Forests

3.3.1. Evaluation of Birdsong Timbre Based on Public Perception

In this study, Pearson correlation analysis was conducted on the four birdsong timbre perception preference indicators, and the results were p = 0.000 (<0.01). In view of exploring the overall public perceptual outcomes of birdsong and its role in audiovisual perception, this study extracted the bird preference indicators (BPs) by referring to existing studies [65,66] and combining them with the indicators using principal component analysis, which resulted in a Kaiser–Meyer–Olkin index of 0.808 (>0.6), a Bartlett’s sphericity test of p = 0.000 (<0.01), and a cumulative variance explained of 70.934% (>60%).

BP = 0.323 × Lik+ 0.310 × Ple + 0.292 × Ric+ 0.306 × Comf

(8)

Through the above calculation method, the scores of 19 kinds of birdsong in 5 kinds of urban forest environments were calculated, and Welch’s ANOVA analysis was carried out. The scores of birdsong were significantly different (F = 55.736, p = 0.000); the results are shown in Figure 4. The three highest scores were Pycnonotus xanthorrhous (1.467), Passer cinnamomeus (1.367), and Parus major (1.213), and the three lowest scores were Ardea cinerea (−1.940), Egretta garzetta (−1.458), and Prinia inornata (−0.384).

Combined with the composition of birdsong in each urban forest environment, the overall score of birdsong in different urban forest environments was further described, and ANOVA analysis was performed. The results were street green space (0.92 ± 0.77) > forest (0.88 ± 0.80) > urban park (0.26 ± 1.13) > residential green space (0.23 ± 1.17) > wetland (−0.05 ± 1.36), F = 31.101, p = 0.000.

3.3.2. Relationship Between Birdsong Preferences and Urban Forest Resilience

In this study, the mediating effects of ΔHR, ΔSC, and ΔP were analyzed, and it was found that BP had a significant mediating effect on the ΔHR indicator, as shown in

Table 7. Analysis of the mediating effect of BP on the ΔHR indicator.

	c Total Effect	a	b	a × b Mediated Effect Value (95% Boot CI)	c’ Direct Effect
Visual ΔHR → BP → audiovisual ΔHR	−0.028	−1.447 *	−0.026 **	0.037 (0.003–0.093)	−0.065

Notes. *: p < 0.05; **: p < 0.01. In this study, if a and b are significant, c is not significant, and 95% of boot CI does not contain 0, then birdsong preference is a complete mediating factor.

. Then, the adjustment effects of ΔHR, ΔSC, and ΔP were analyzed, and it was found that BP had a significant adjustment effect on the ΔSC indicator, as shown in

Table 8. Analysis of the moderating effect of BP on the ΔSC indicator.

	Model 1			Model 2			Model 3
	Coefficient (SE)	t	p	Coefficient (SE)	t	p	Coefficient (SE)	t	p
const	0.051 (0.010)	4.905	0.000 ***	0.036 (0.013)	2.757	0.007 **	0.045 (0.013)	3.356	0.001 **
ΔSC	0.145 (0.074)	1.976	0.050	0.149 (0.073)	2.049	0.042 *	−0.001 (0.095)	−0.015	0.988
BP				0.050 (0.025)	2.014	0.046 *	0.024 (0.027)	0.902	0.369
ΔSC × BP							0.475 (0.199)	2.387	0.018 **
F	F = 3.905, p = 0.050			F = 4.022, p = 0.020 *			F = 4.667, p = 0.004 **
ΔF	ΔF = 3.905, p = 0.050			ΔF = 4.058, p = 0.019 *			ΔF = 9.85, p = 0.002 **

Notes. *: p < 0.05; **: p < 0.01; ***: p < 0.001. ΔSC × BP shows a significant effect, so birdsong preference has a moderating effect.

.

3.3.3. Effects of Birdsong Preferences on Urban Forest Recovery

In view of the obvious difference in visual restoration benefits in different urban forest environments, this study explored the effect of birdsong on the recovery benefits of different urban forests. The difference between the recovery benefits of the audiovisual interaction group with birdsong and the visual group without birdsong in each environment was used as the effect of birdsong on the recovery benefit of urban forests, which was expressed by ΔHR D-value, ΔSC D-value, ΔI D-value, ΔC D-value, ΔT D-value, and ΔP D-value, respectively.

The six indications were subjected to a histogram test, which showed that they were all essentially regularly distributed. The study employed partial least squares (PLS) analysis to create PLS models between various environmental birdsong preference indicators and physiological and psychological indicators of change because the sample sizes of each environmental group were small and the collinearity analysis showed that the ΔI, ΔC, ΔT, and ΔP D-values had collinearity issues. The PLS model’s R² and Q² values were 0.880 and 0.778 according to the forest environment, 0.908 and 0.861 according to the wetland environment, 0.858 and 0.806 according to the urban park environment, 0.843 and 0.747 according to the street side green space environment, and 0.879 and 0.808 according to the residential green space environment, demonstrating the great reliability of all PLS models. The correlation between each model’s measured and predicted values is displayed in Figure 5 and Figure 6.

Given that VIP > 1 is regarded as a significant influencing factor [67],

Table 9. Standardized regression coefficients (Beta) and variable projected importance (VIP) of each birdsong preference indicator with the amount of change in perceived physiological and psychological benefits.

Indicators			Lik	Ple	Ric	Comf
Forest	Beta	ΔHR D-value	−0.004	−0.019	−0.025	−0.024
		ΔSC D-value	0.011	0.044	0.052	0.059
		ΔC D-value	0.033	0.700	0.534	0.625
		ΔT D-value	−0.100	−0.871	−0.958	−1.515
		ΔI D-value	−0.033	−0.867	−1.298	−0.943
		ΔP D-value	−0.097	−1.238	−1.393	−1.560
	VIP		0.974	1.042	1.144	0.810
Wetland	Beta	ΔHR D-value	−0.035	−0.033	−0.019	−0.035
		ΔSC D-value	0.068	0.051	0.042	0.079
		ΔC D-value	0.325	0.516	0.408	0.709
		ΔT D-value	−1.242	−0.976	−1.004	−0.817
		ΔI D-value	−0.962	−0.704	−1.305	−1.268
		ΔP D-value	−1.427	−1.175	−1.115	−1.159
	VIP		1.060	0.976	0.863	1.086
Urban park	Beta	ΔHR D-value	−0.03	−0.018	−0.014	−0.031
		ΔSC D-value	0.069	0.043	0.049	0.057
		ΔC D-value	0.461	0.289	0.279	0.975
		ΔT D-value	−1.379	−1.189	−1.097	−1.512
		ΔI D-value	−1.745	−0.628	−0.751	−1.630
		ΔP D-value	−2.049	−1.305	−0.814	−1.826
	VIP		1.181	0.985	0.658	1.097
Street green space	Beta	ΔHR D-value	−0.031	−0.023	−0.023	−0.030
		ΔSC D-value	0.119	0.086	0.083	0.088
		ΔC D-value	0.742	0.745	0.375	0.495
		ΔT D-value	−1.525	−0.834	−1.119	−0.968
		ΔI D-value	−0.971	−1.047	−1.267	−1.296
		ΔP D-value	−1.781	−1.270	−1.378	−1.644
	VIP		1.107	0.772	0.863	1.197
Residential green space	Beta	ΔHR D-value	−0.024	−0.02	−0.017	−0.019
		ΔSC D-value	0.097	0.075	0.069	0.102
		ΔC D-value	0.602	0.419	0.749	0.678
		ΔT D-value	−1.820	−0.970	−1.491	−1.742
		ΔI D-value	−2.036	−1.901	−0.738	−1.390
		ΔP D-value	−2.424	−1.953	−1.678	−2.260
	VIP		1.099	0.764	0.804	1.250

Notes. Bold indicates that VIP values greater than 1 are significant influencing factor.

shows that psychological and physiological benefits were significantly impacted by Ple and Ric in the forest environment and by Lik and Comf in the rest of the environment.

4. Discussion

4.1. The Influence of Urban Forest Visual Perception on Public Physiology and Psychology

From physiological recovery, this study used HR and SC as physiological indicators and found that all five different urban forest environments had a positive effect on public physiological health in terms of visual perception, which is in line with the results of existing studies [16,68]. Different urban forest environments and visual perceptions all had good physiological and psychological recovery benefits, but there were some differences between environments. Resilience to HR was forest > urban park > residential green space > wetland > street-side green space; resilience to SC was urban park > forest > wetland > residential green space > street-side green space; and resilience to basal psychological state was forest > urban park > wetland > residential green space > street-side green space. Forests and urban parks had a relatively strong ability to restore HR, while wetlands and street green spaces had a relatively weak ability. Urban parks and forests had a relatively strong ability to restore SC, while residential green spaces and street green spaces had a relatively weak ability. This is in line with the findings of previous research that indicates that, in comparison to non-green environments like streets and buildings, forests and parks significantly improve the human body’s HR, HRV, and other autonomic nervous system markers [15,68,69]. Wetlands, residential neighborhoods, and street green spaces offer relatively limited recuperation advantages for the public psyche, whereas urban parks and forests have relatively strong advantages. According to certain research, green vegetation significantly lowers experimental participants’ systolic blood pressure (SBP), heart rate (HR), heart rate variability (HRV), and sympathetic nerve activity (LF/HF) [68,70]. The comparatively enormous size of the water bodies in the wetland areas employed in this study may have contributed to the results of the experiment, as the participants may have felt psychologically that being so close to the water bodies could make them feel insecure [71].

Through the PRS, we found that there were no significant variations in the environmental outcomes across urban forests, except for the Coh of visual perception, which was not differentiated by the inclusion of birdsong. Given that urban forests that have recovered their environmental conditions typically have similar traits [72], this suggests that birdsong may help to regain public attention to some degree. Urban forests with natural elements, such as forests and urban parks, have better physiological recovery benefits compared to urban forests with higher urban building or artificial construction elements. However, urban forests with higher residential green space, street green space, and other artificial elements still have significant physiological and mental health benefits for the public. According to this study, urban parks and forests with the greatest recovery advantages belonged to artificial and natural urban forest types, respectively. This suggests that while natural-type urban forests do not have better recovery benefits, artificially created urban forest environments that are close to natural-type environments have good recovery benefits. Wetlands and other natural urban forest environments must be judiciously arranged for the public’s activity routes to improve the level of plant communities. The first step in creating street green spaces, residential green spaces, and other artificial urban forests is to think about how to improve plant diversity and multiple plant community levels in the limited space available.

4.2. The Influence of Urban Forest Audiovisual Perception on Public Physiology and Psychology

Based on the visual perception, this study added the birdsong environment and found that the visual and auditory elements in the urban forest had a positive effect on the public’s physical and mental health. It was found that the visual and auditory interaction perception based on birdsong in different urban forests would have a positive effect on the public’s physical and mental health, but there were some differences in each environment. Resilience to HR was forest > urban park > residential green space > wetland > street green space; resilience to SC was urban park > forest > wetland > residential green space > street green space; and resilience to basal psychological state was forest > urban park > residential green space > wetland > street green space. This is consistent with other studies that have shown that exposure to natural visual and auditory stimuli can greatly contribute to stress recovery and physical and mental health [53,73]. While street green spaces had comparatively lesser recovery benefits, urban parks had greater recovery benefits for SC and forests had higher recovery for HR, with both still demonstrating strong recovery advantages in terms of physiological recovery for the public. The benefits of audiovisual perception for psychological recovery were comparatively greater for urban parks and forests and comparatively less for wetlands, residential green spaces, and street green spaces. Under audiovisual perception settings, the physiological and psychological recovery advantages of urban parks and forests were more favorable. It was found that the results of audiovisual perception and visual perception were relatively similar, which may be because the public perceives the natural environment mainly through vision [26]. It is worth noting that there were some differences in the Coh of the Attention Perception Recovery Scale for visual perception in urban forests, but there were no significant differences in the four indicators of the Attention Perception Recovery Scale for audiovisual perception after adding birdsong, which confirms the idea that synergism between visual and auditory elements of natural environments is crucial to public health benefits [74,75]. It has also been confirmed that audiovisual interactive perception can enhance the coherence experience of the environment [43]. This study demonstrates that it is impossible to overlook the soundscape when investigating the impact of urban forests on public health. However, it is also possible that urban parks and forests are better places for birds to live because they have a more diverse plant life and a richer community structure, which draws a greater variety of bird species and improves the listening experience by showcasing a greater variety of birdsongs. On the other hand, following the inclusion of birdsong in this study, the Coh indicator in the Perceptual Recovery Scale of Audiovisual Perception did not differ significantly from visual perception. This implies that birdsong can enhance the public’s perception of urban forests and aid in attention recovery, which is in line with previous research [39]. Therefore, when building urban forests, consideration should be given to improving the restoration benefits of urban forests, increasing public perception of birdsong, and building more bird-friendly habitats.

4.3. Birdsong Perception and Its Effect on Urban Forest Recovery

Three modes of perception were used in this study: psychological, physiological, and public timbre perception. From public timber perception, Ardea cinerea, Egretta garzetta, Passer cinnamomeus, and Prinia inornata were the three least popular birdsongs, while Pycnonotus xanthorrhous, Passer cinnamomeus (1.367), and Parus major were the three most popular. It was discovered that urban forest environments with audiovisual perception of birdsong had superior restoration benefits in terms of both physiological and psychological indicators. However, the outcomes differed slightly from visual environments without birdsong; this is consistent with existing research [76]. But the outcomes were not totally positive. All things considered, the majority of urban forest birdsong was quite nice, but most of the birdsongs that the public did not like were found in wetlands, which led to a low public preference for wetland bird sounds. The recovery benefits of incorporating birdsong in wetland environments were diminished based on a comparison of visual and audiovisual perceptions. Additionally, the differences in bird distribution across different environments indicate that the different birdsongs emitted by different bird species appear to play a significant role in the disparities in public perceptions [12]. This implies that when combined with the visual surroundings, different birdsong preferences will offer varying recovery effects [66].

Regarding physiological and psychological effects, birdsong preference mediated public HR recovery from physiological and psychological effects. When the birdsong environment was added, birdsong preference also acted as a mediating variable that could affect the audiovisual perception of HR recovery benefits, with higher preference translating into better recovery benefits. Birdsong preference moderated the effect of public SC recovery from physiological and psychological effects. When the birdsong environment was added, birdsong preference also acted as a moderating effect variable that could affect the audiovisual perception of SC recovery benefits, with higher preference translating into better recovery benefits. In the research of birdsong timbre perception, the pleasantness and richness of public perception of birdsong were found to be important factors in the restorative benefits of forest audiovisual environments, whereas likability and comfort were found to be important factors in the restorative benefits of urban forest audiovisual environments in other four environments, indicating that the main influence of birdsong on public health benefits varies across environments. It is also possible that forest environments are naturally more aesthetically pleasing, while audiovisual experiences are more likely to require rich and pleasant birdsong environments. The remaining environments should concentrate on birdsong environments that are more comfortable and enjoyable for the public. Different public perceptions of birdsong in terms of the acoustic attribute of timbre can influence the environmental recovery benefits and even negatively affect the recovery effects [77]. Urban forest listening environments with better restoration often require birdsong environments with high levels of pleasantness and enjoyment. Plant community optimization and plant diversity enrichment, on the other hand, can be utilized to establish a habitat for bird species that have a high preference for birdsong in areas with low birdsong preference. This will increase the overall environmental birdsong preference. In addition, urban forest environments with a high preference for existing birdsong can be selected to build urban forest recreation spaces. It follows that increasing the diversity of plants, birds, and habitats can improve the recovery benefits of urban forests.

4.4. Limitations and Future Research Perspectives

This study focused on the perceived physiological and psychological recovery benefits of visual and audiovisual perceptions of five urban forest environments and explored the public perception results of birdsong timbre, which can make a difference in the environmental recovery benefits and even negatively affect the recovery effects. However, there are still some limitations to this study. First, although the number of subjects in this experiment was large, the subjects were all students, and students majoring in landscape architecture and horticulture were predominant. Whether the results of the experiment can represent the public needs to be further discussed, and future experiments need to have strict control over the subjects’ professional backgrounds and occupations [78]. Second, the bird data in this experiment were derived from outdoor surveys and can represent the distribution of birds in different urban forest environments in Kunming, but the birdsong data came from Xeno-canto, which was simulated in audio form. Moreover, the experiment was conducted in a laboratory, not the field, which may have made a difference in the experimental results. Birdsong is mainly emitted by song birds. Egretta garzetta and Ardea cinerea with low birdsong preference scores belong to the real environment of wading birds, and they have relatively few song behaviors, which may lead to some differences between the experimental results and the actual experience. Third, this study only addressed the acoustic attribute of timbre, and the effects of spectral characteristics, tone, loudness, perception times, and even different times of the day on soundscape perception [79] need further exploration. People’s perception of the environment has complex and multi-sensory characteristics. For example, in different seasons, the perception recovery benefits of the same kind of birdsong will be different [76]. Finally, birdsong was found to be a key factor influencing recovery in this study. The birds chosen for this study were from the summer season, when the composition of the bird community is more stable than in other seasons. Therefore, it is important to consider changes in the environment of birdsong brought about by the migration of migratory birds in autumn and winter and the richer diversity of birds during the breeding season in spring [56]. On the other hand, one of the future trends of climate change is global warming and increased precipitation, which will have a significant impact on the distribution pattern of species as well as the structure and function of most ecosystems [80]. As the air temperature rises, birds migrate to high-altitude or polar regions. As temperatures rise, birds move to high altitudes or polar regions, and climate change will also lead to migratory species becoming resident birds or reducing their journey distances [81]. Future changes in birdsong environments will be brought about by changes in bird species. Therefore, subsequent research should strengthen the investigation of bird species and simulate the real outdoor environment as much as possible, considering the temporal nature of the environment and further incorporating the habits of birds to simulate actual birdsong environments. Although there are differences in the auditory perception of different birdsongs, they can be classified into five categories: constant frequency, frequency modulated, broadband pulses, broadband with varying frequency components, and segments with strong harmonics [82]. More in-depth research can be carried out in the future by combining the classification of acoustic features. Moreover, visual perception is still an important way of public perception within urban forest environments; research on audiovisual interactions that place birdsong in visual contexts, such as colors, environmental features, and types of environmental elements, needs to be further discussed.

5. Conclusions

This study explored the physiological and psychological recovery benefits and differences in visual perception in five urban forest environments. The aim of this study was to provide an optimization and enhancement method for the selection and construction of urban forest recovery functions. In particular, the principles of audiovisual perception are essentially the same, and public benefits from urban parks and forests are greater in terms of physiological and psychological recovery than those from street green spaces. The advantages of natural urban forests for recuperation are, however, not superior. The aim of this study was to provide an optimization and improvement strategy for the selection and building of urban forest recovery functions. The rules of visual and audiovisual perception are essentially the same, and the public health benefits from urban parks and forests are greater from physiological and psychological recovery than those from street green spaces. However, artificially created urban forests can also have high recovery benefits. For example, artificially created urban parks have considerable recovery benefits, as do forests. The soundscape of birdsong is one of the aspects of urban forests that the public cannot overlook. Birdsong improves public perceptions of attention recovery levels. However, in contrast to visual perception, birdsong does not necessarily contribute to the recovery benefit advantages of audiovisual perception. It is interesting to note that, in contrast to the ‘birdsong’ public response studies, due to varying public preferences for different birdsongs, the recovery of audiovisual perception will be less than that of visual perception. Recovery benefits will be impacted by the public’s impression of birdsong preferences. The benefits of recovery increase with the preference for birdsong perception. Birdsong preference is a thorough assessment that is based on four different features of birdsong timbre. In particular, the primary determinants of birdsong perception that impact the recovery of audiovisual perception in various urban forest environments vary, impacting the forest environment’s richness and pleasantness and other environments’ comfort and likability. Research on the physical and mental health benefits of urban forest audiovisual perception to the public is enhanced by this study, which also highlights the value of field bird surveys in examining the recovery benefits of birdsong in urban forests for the public. Additionally, the study thoroughly examined the relationship between the public’s perception of the timbre of birdsong and the recovery benefits of birdsong, which offers a fresh perspective on the role of the natural environment and the public, as well as a new idea for improving the recovery properties of urban forests.