Research on the Healing Potential of Urban Parks from the Perspective of Audio-Visual Integration: A Case Study of Five Urban Parks in Chengdu

Abstract

In recent decades, rapid urbanization has been linked to negative impacts on people’s mental health. However, the healing potential of urban parks as central features of cityscapes has not been fully explored from the perspective of audio-visual integration. This gap limits designers’ ability to leverage parks’ healing systems to their full potential. To address this gap, this study used the Mindwave monitoring and recovery component scale to evaluate the healing function of urban parks in Chengdu, specifically focusing on audio-visual integration. Based on audio-visual scene combination samples collected through shooting and recording, we found that: (1) The visual and audio qualities of birdsong (the sky above the blue space and the green space) can significantly differ. (2) Birdsong and conversation seem to provide healing effects from seemingly contradictory dimensions of “quiet” and “social”, with gender differences as the primary influencing factor. (3) Visible children laughing at low levels (30% and 50%) has a more negative impact than invisible traffic noise at the same sound pressure level. (4) Audio-visual interaction does not always have a positive effect, with the visibility of the sound source as the primary influencing factor. (5) An increase in audio proportion did not necessarily correspond to a linear increase in the healing effect. Through exploring the influence of audio-visual combination scenes on healing effects in urban parks, this study provides an empirical basis for park design and planning that considers audio-visual healing effects. It offers insights into designing healing systems for parks and urban areas as well.

1. Introduction

In recent decades, intensive urbanization has led to isolation and disconnection between humans and nature, resulting in mental health issues. Global mental health issues have worsened over the past decade, with a dramatic 25% increase in the global prevalence of anxiety and depression following the COVID-19 pandemic [1]. The economic downturn, lockdowns, and quarantines resulting from the post-COVID-19 era have placed people under considerable pressure to survive [2,3,4,5,6,7,8]. Facing mental health issues, mental health systems worldwide are not adequately equipped to address the current burden of mental disorders. Traditional psychiatric medications have established side effects [9,10,11]. However, pro-nature non-invasive preventive, complementary, or alternative therapies are widely accepted, such as Forest Bathing [12], Aromatherapy [13], Horticultural Therapy [14], and Five Senses Therapy [15]. Numerous studies have shown that these methods bring many benefits to humans through contact with or perception of nature, such as reducing stress and depression, helping to recover from fatigue [16], improving emotional state, shortening recovery time after surgery [17], promoting physical activity [18], reducing anxiety, and improving mental health [19].

Wilson et al. proposed the “pro-life hypothesis”, suggesting that the natural environment is vital for physical and mental health and that humans are born with green genes [20]. Although there is evidence that Forest Bathing, Horticultural Therapy, and Vertical Greening effectively improve human physical and mental health, few people today benefit from these environments because forests are less accessible and less frequently used [21]. Home gardens need to work better in densely populated low-income cities. Vertical greening is still in its infancy for cities [22]. Many scholars have begun to notice the role of urban healing systems in spiritual healing and healing systems. Liu et al. proposed taking the “prevalence rate of mental disorders among people aged 18 and above” as the evaluation index of park city in constructing the park city evaluation system [23]. Li et al. have systematically expanded traditional healing environments into healing streets, healing buildings, virtual healing, and other areas as part of the healthy city healing system [24]. The team of Ha also proposed in the connection between cities and health that more than urban blue and green space is needed. The positive impact on mental health requires active planning in the spatial distribution of multiple landscapes [25].

Vision and hearing are the primary sources of information for the human body. From the perspective of audio-visual integration, exploring ways to improve the healing potential of city park systems can enhance the city’s prevention and healing capabilities against public mental disorders and reduce the burden on the mental health system. Furthermore, studies on the healing potential of environments through visual perception alone cannot provide comprehensive guidance for actual landscape design [26]. The soundscape serves as an innovative design tool that holds great potential in concealing negative ambient sounds and enhancing positive auditory experiences within a given environment. In contrast to visual stimuli, which experience momentary interruptions during each blink, resulting in the loss of approximately 450 milliseconds of information per blink and approximately 6 s per minute, the auditory channel remains continuous and uninterrupted. Consequently, many scholars began to investigate the relationship between visual and acoustic landscapes. For example, better therapeutic effects have been observed when natural water bodies and dense vegetation landscapes are complemented by birdsong [27]. This includes Alvarsson et al.’s work, which showed that natural sounds help with recovery from sympathetic activation after psychological stressors [28]. Through research, Renterghem found that the level of noise perception reduction with green vision was easily more than 10 decibels [29]. Natural-related audio-visual stimuli are generally believed to have greater health benefits compared to singular visual stimuli [30]. However, simultaneously, Li et al. discovered that audio-only stimuli may produce more therapeutic EEG responses than audio-visual stimuli [31]. Annerstedt et al. demonstrated that adding natural sounds to a virtual reality environment can even promote stress reduction [32]. Furthermore, in park settings, demographic factors such as age, education, and stress levels have been found to significantly influence the therapeutic effects of bird songs [33]. In a recent study, it was discovered that individual traits have a greater impact on the potential restorative effects of urban sounds compared to innate characteristics [34]. In summary, the acoustic environment is a critical factor that affects overall comfort. Visual and audio elements are always in the same interaction, highlighting the importance of studying audio-visual blending environments, including soundscapes, to improve the impact of the physical environment on people’s bodies and minds [35,36]. However, research on the mechanism of the effects of audio-visual composite scenes on the human body in urban parks needs to be more robust.

To enhance the therapeutic potential of urban parks, this study aims to explore the expected relationship between visual and auditory stimuli and their therapeutic potential by measuring people’s brainwave signals when experiencing different park visual and auditory scenes and evaluating the psychological relaxation effect of restorative components. This study intends to provide guidance for improving and optimizing the therapeutic function in urban park planning and design based on evidence of visual and auditory therapeutic potential.

2. Materials and Methods

2.1. Research Area

Chengdu, as the pioneer of park cities, has achieved initial results in urban park construction over the past four years [23]. Based on careful consideration of construction year, construction scale, distance from the city center and functional themes, five urban parks were selected as research objects (

Table 1. Overview of urban parks.

Park Names	Year of First Opening/Years of Recent Renewal	Park Features	Construction Scale (km2)	Bearing from the City Center/Distance (km)
Guixi Ecological Park (A)	2016/2021	Athletic vitality	About 0.933	Due South/10.5
Water Park (B)	1997/2022	Ecological	0.024	Northeast/2.7
Qinglonghu Wetland Park (C)	2016/2020	Demonstration	About 20.000	Due East/12.7
Tianfu Art Park (D)	2021/2022	The Beauty of Art	About 2.022	Northwest/8
Xinglonghu Wetland Park (E)	2016/2022	Large volume	About 3.573	Due South/29.4

). The five urban parks, Guixi Ecological Park, Water Park, Qinglonghu Wetland Park, Tianfu Art Park, and Xinglonghu Wetland Park, are labeled in order from A to E.

2.2. Selection of Research Samples

2.2.1. Visual Sample

Street photography for research purposes offers the advantages of easy operation and strong experimental control. Previous studies have shown no significant difference between street photos and actual landscape views [37]. For this study, a total of 969 photos were taken to capture changes in the spatial morphology and functional positioning of each park, with 44 candidate photos (Appendix B 

Table A2. The shooting position and space type of urban park photos.

Name	Shooting Position	Space Type	Scene
Guixi Ecological Park		Modeling and ornamental space A-1
		Lawn recreation space A-2
		Park rest space A-3
		Square recreation space A-4
		Waterfront recreation space A-5
		Interactive amusement space A-6
		Interactive amusement space A-7
		Interactive amusement space A-8
		Modeling and ornamental space A-9
Water Park		Underforest recreation space B-1
		Park rest space B-2
		Park rest space B-3
		Underforest recreation space B-4
		Square recreation space B-5
		Square recreation space B-6
		Underforest recreation space B-7
		Underforest recreation space B-8
		Underforest recreation space B-9
		Park rest space B-10
		Square recreation space B-11
Qinglonghu Wetland Park		Interactive amusement space C-1
		Interactive amusement space C-2
		Modeling and ornamental space C-3
		Waterfront recreation space C-4
		Waterfront recreation space C-5
		Underforest recreation space C-6
		Lawn recreation space C-7
		Square recreation space C-8
		Interactive amusement space C-9
		Lawn recreation space C-10
		Lawn recreation space C-11
Tianfu Art Park		Underforest recreation space D-1
		Underforest recreation space D-2
		Waterfront recreation space D-3
		Underforest recreation space D-4
		Waterfront recreation space D-5
		Modeling and ornamental space D-6
		Waterfront recreation space D-7
Xinglonghu Wetland Park		Waterfront recreation space E-1
		Waterfront recreation space E-2
		Underforest recreation space E-3
		Waterfront recreation space E-4
		Waterfront recreation space E-5
		Park rest space E-6

) selected after removing unqualified and duplicate photos. These photos were taken on clear and smog-free days (PM2.5 < 100) in August 2022 between 9:00 a.m. and 11:30 a.m. and 1:00 p.m. and 4:00 p.m., including weekdays, weekends, and holidays.

Building upon existing research [38], this study selected seven categories as the influencing factors for urban parks, including factor richness, space openness, waterside, greening degree, activity (i.e., capacity for walking, resting, and playing), functional type, and green space form (i.e., accessory form and natural form), while considering the integrity and diversity of the landscape. Each factor was rated on a scale of 0 to 1, with 1 indicating satisfaction with the characteristics of the influencing factor (Appendix A 

Table A1. Scores of urban park categorized factors.

Number	Factor Abundance	Playability	Space Openness	Capacity		The Vegetation Area Exceeds the Total Area by 0.5	Waterside	Function	Nature	Paving
				Walk	Rest
A-1	0	0	1	1	0	1	0	0	0	0
A-2	0	0	1	1	1	1	0	0	1	0
A-3	0	0	0	1	1	0	0	1	0	1
A-4	1	0	1	1	1	0	0	1	0	1
A-5	1	0	1	0	0	1	1	0	1	0
A-6	0	1	1	1	0	0	0	1	0	1
A-7	0	1	1	1	0	0	0	1	0	1
A-8	0	1	1	1	0	0	0	1	0	1
A-9	0	0	1	1	0	1	0	0	0	1
B-1	1	0	0	1	1	1	0	1	1	0
B-2	1	0	0	1	1	1	0	1	0	1
B-3	1	0	0	1	1	1	0	1	0	1
B-4	1	0	0	1	1	1	1	0	1	0
B-5	1	0	1	1	1	0	0	1	0	1
B-6	1	0	1	1	1	0	1	1	0	1
B-7	1	0	0	1	1	1	1	0	1	0
B-8	1	0	0	1	1	1	0	0	1	0
B-9	1	0	0	1	1	1	1	0	1	0
B-10	0	0	0	1	1	1	0	1	0	1
B-11	1	0	1	1	1	0	0	1	0	1
C-1	0	1	1	1	0	0	0	1	0	1
C-2	0	1	1	1	0	0	0	1	0	1
C-3	0	0	1	1	0	1	0	0	0	1
C-4	1	0	1	0	0	1	1	1	1	0
C-5	1	0	1	0	0	1	1	0	1	0
C-6	1	0	0	1	1	1	0	0	1	0
C-7	0	0	1	1	1	1	0	0	1	0
C-8	1	0	1	1	1	0	0	1	0	1
C-9	0	1	1	1	1	0	0	1	0	1
C-10	0	0	1	1	1	1	0	0	1	0
C-11	0	0	1	1	1	1	0	0	1	0
D-1	1	0	0	1	1	1	1	0	1	0
D-2	1	0	0	1	1	1	0	0	1	0
D-3	1	0	1	0	0	1	1	1	1	0
D-4	1	0	0	1	1	1	0	0	1	0
D-5	1	0	1	0	0	1	1	1	1	0
D-6	0	0	1	1	0	0	1	0	0	1
D-7	1	0	1	0	0	1	1	1	1	0
E-1	1	0	1	0	0	1	1	1	1	0
E-2	1	0	1	0	0	1	1	0	1	0
E-3	1	0	0	1	1	1	1	1	1	0
E-4	1	0	1	0	0	1	1	1	1	0
E-5	1	0	1	0	0	1	1	0	1	0
E-6	0	0	0	1	1	0	1	1	0	1

). The clustering spectrum diagram of urban parks was obtained through systematic cluster analysis, Euclidean distance, and the farthest neighbor method (Appendix A Figure A1). According to the figure, the 44 photos were categorized into seven categories, namely waterfront recreation space, lawn recreation space, underforest recreation space, square recreation space, park rest space, interactive recreation space, and modeling and ornamental space, marked with T1–T7, respectively. Based on the significant differences in photographic quality and landscape space, each landscape type was independently voted on by five designers with more than three years of work experience, and one representative photo that they believed could represent the space was selected. Seven photos were selected based on the voting results and marked with V1–V7 (Figure 1). Any irrelevant text, symbols, and other factors were removed from the photos using Adobe Photoshop, and the brightness, color saturation, and contrast of each photo were adjusted for consistency to eliminate any perceptual bias caused by the photos themselves.

2.2.2. Audio Sample

Firstly, a field study was conducted to capture the sounds of birdsong, children laughing, traffic noise, flowing water sounds, and conversation. The original mono audio was recorded near the source, without being disturbed by other sounds. The ZOOM H5 digital recorder was placed on a tripod 1.5 m from the ground and more than 3.5 m away from other reflective surfaces. The collection was conducted without rain, snow, lightning, or wind speeds exceeding 5 m/s.

Secondly, nine sampling points were established to obtain physical sound level guidance for urban parks, with a measurement time of 10 min per sampling point [39]. All sampling points were located within 5 m of the research object. The average of the nine sampling points was taken as the environmental sound pressure level (SPL) of an urban park, with a Laeq value of 55 dB (A).

The audio was then imported into Adobe Audition software to produce clear and complete audio with consistent frequencies, adjusting the sound pressure level to 55 dB (A). Five experts were invited to identify the sound sources and select the 20 s audio clips that were most representative of the various audio samples found in city parks. Finally, the sound pressure levels of certain audio samples were adjusted to 70%, 50%, and 30% of their original levels, respectively. We obtained three audio samples with different sound pressure levels and combined them with the most likely sound types based on the visual environment. After combining the samples, they were readjusted to a sound pressure level of 55 dB (A). This process resulted in 14 composite sounds, along with one original sound, as audio samples, each lasting for 20 s.

To calibrate the sound pressure levels of the audio, a Megawar electronic artificial ear with a SONY noise-canceling headset was used. During calibration, Adobe Premiere was used to adjust the volume of all audio signals to approximately 55 dB (A).

2.2.3. Audio-Visual Samples

The seven visual samples and 15 audio samples extracted from the survey were combined one by one, and finally, 21 groups of 20 s audio-visual scene videos were obtained as experimental samples (Figure 2).

2.3. Experimental Design

2.3.1. Subjects

This study used the Social Readjustment Rating Scale [40] to investigate the stress levels of participants and recruited university students who had been under chronic stress for a long time as the research subjects. The inclusion criteria were: (1) aged between 19 and 26 years old; (2) physically and mentally healthy, with no bad habits such as staying up late, alcoholism, or smoking; (3) no major recent life changes; (4) normal vision and hearing; (5) right-handed. A total of 20 participants were recruited, including 10 males and 10 females, with an average age of 22 years old. The participants were required to participate one after another in this experiment.

2.3.2. Experimental Environment

The experiment was conducted in the laboratory of the College of Arts at Sichuan University, China. The experiment was conducted for ten days starting on 15 September 2022. The room was kept quiet during the experiment to prevent noise from interfering with brain waves, and any non-essential electronic equipment that could cause electromagnetic interference was eliminated. The laboratory was maintained at a constant temperature of 24 °C throughout the experiment.

2.3.3. Experimental Process

The experimental conclusion will be drawn by measuring physiological data and filling in a subjective scale. Participants first viewed seven photos separately and then listened to 15 audio clips in sequence, with a 10 s eye closure break between the two pieces of information and a one-minute rest period at the end. They then experienced 21 simulated scenes with random audio-visual combinations, with a one-minute rest period between the two scenes to fill out the subjective scale (Figure 3).

The brainwave experiment lasted approximately 51 min, which included three minutes of experimental content display, five minutes of electrode assembly and testing, three minutes of rest, 38 min of audio-visual stimuli and recording of brainwave data, and two minutes of electrode removal.

2.4. Physiological Indicators

The experiment computed the meditation values of 20 participants to derive the overall perceived sense of healing for each scenario, i.e., the degree of relaxation, using the Mindwave brain wave testing machine (Figure 4). This device consists of a single dry electrode (12 mm × 16 mm) placed on Fp1 and employs a pea-sized (~0.8 mm diameter) electrode clipped to the left earlobe as a reference, according to the international 10–20 system, which inputs data to a TGAM1 (ThinkGear ASIC Module) integrated circuit. These two elements are mounted on a light headset (90 g). The device samples data at 512 Hz. Mindwave generates a meditation value every second through the eSense algorithm, ranging from 0 to 100. The availability of Mindwave devices and eSense parameters has been empirically supported [41,42]. The values are classified as follows: 0–20 (extremely low-value area), 20–40 (low-value area), 40–60 (general value area), 60–80 (high-value area), and 80–100 (extremely high-value area). The higher the value, the more relaxed the subject is.

2.5. Statistical Methods of Subjective Scale

This study employed the scale designed by Fu et al. [36], which consists of three parts: the Recovery Component Scale (RCS), the Visual Aesthetic Quality (VAQ) of Visual Perception, and the Tranquility Rate (TR) of audio perception (Appendix C 

Table A3. Restorative components scale used in this study.

Dimension		Description	Scales
RCS	Being-away(B)	B1 This allows me to temporarily forget the troubles of work and daily life.	0 1 2 3 4
	Extent(E)	E1 There are many beautiful associations here.	0 1 2 3 4
	Fascination(F)	F1 There is a lot here that appeals to me.	0 1 2 3 4
	Compatibility(C)	C1 This gives me the opportunity to do what I love to do.	0 1 2 3 4
TR		The environment here makes me feel very quiet.	0 1 2 3 4
VAQ		The environment here is beautiful and striking.	0 1 2 3 4

). The RCS scale, developed by Laumann et al., consists of four dimensions: being away, extent, fascination, and compatibility [43]. Four items suitable for environmental assessment of urban parks were selected as a short modified questionnaire [44]. “Beautiful and striking” was used to describe visual aesthetic quality (VAQ) [45], and “quiet” and “noisy” were used to describe the serenity rate (TR) [46]. Participants rated their agreement with each question on a scale of 0 (strongly disagree) to 4 (strongly agree), based on their actual perception. The experiment calculated the average score of 20 subjects to obtain the score of each dimension.

2.6. Data Statistics

In this study, a one-way analysis of variance was conducted to examine whether there were significant differences in the meditation level of brainwaves between visual stimuli and audio stimuli. Additionally, a multifactor analysis of variance was performed to investigate whether there were significant differences in the relaxation level of brainwaves among visual stimuli, audio stimuli, the interaction between visual and audio, and gender. When testing the reliability and validity of the questionnaire, the entropy method was used to re-assign weight values to the RCS scale. The Kruskal–Wallis test was used to test whether significant differences existed between different visual and audio samples on the subjective scale.

3. Results

3.1. Effects of Visual Sample and Audio Sample on Meditation

The results of the one-way analysis of variance indicated that the visual stimuli in different urban park spaces showed heterogeneity of variance for the eSense “meditation level” index of subjects’ brainwaves, while the audio stimuli had a significant impact on the index (p < 0.05) (

Table 2. Significance Levels of Homogeneity of Variance Tests and One-Way ANOVA for Visual and Audio Stimuli.

	Homogeneity of Variance				ANOVA
	Levene’s Statistic	df 1	df 2	p	F	p
Visual stimuli	3.197	6	133	0.006 **
Audio stimuli	1.369	11	228	0.189	27.298	0.000 **

** p < 0.01.

). The results of the multifactor analysis of variance revealed significant differences in the meditation level of brainwaves for the visual stimuli, audio stimuli, interaction of visual and audio, and gender (p < 0.05) (

Table 3. Significance Levels of Multifactor ANOVA for Visual Stimuli, Audio Stimuli, Interaction of Visual and Audio, and Gender.

	Visual Effect		Audio Effect		Audio-visual Effect		Gender Effect
	F	p	F	p	F	p	F	p
Visual & Audio-visual (+S0)			42.035	0.000 **
Audio (V1) & Audio-visual (V1)	15.538	0.000 **	32.408	0.000 **	4.431	0.014 *
Audio (V4) & Audio-visual (V4)			18.362	0.000 **	7.058	0.001 *	5.882	0.017 *
Audio (V6) & Audio-visual (V6)	25.478	0.000 *	30.696	0.000 **	4.233	0.000 **

* p < 0.05; ** p < 0.01.

).

Figure 5 demonstrates that after the addition of the S0 (100% birdsong) audio stimuli, the meditation level improved significantly except for the waterfront recreation space, indicating that S0 can effectively increase the comfort of the scene. However, the meditation level of the waterfront recreation space showed a significant decline. The S0 does not seem to be suitable to be used in waterfront recreation space.

Figure 6 shows that among audio stimuli, S0, S11, and S12 audio samples adapted to the waterfront recreation space, among which S12 > S11 > S0. The meditation level of the visual stimuli of the waterfront recreation space was taken as the control group. It was found that the meditation level of the waterfront recreation space was significantly improved after the addition of S11 and S12. In contrast, the meditation level of the waterfront recreation space significantly decreased after the addition of S0. Figure 7 presents the results of the audio samples (S41, S42, S43) adapted to the square recreational space, where gender determines some of the differences. The meditation level of audio stimuli in male subjects was S43 > S42 > S41, while that in female subjects was S41 > S42 > S43. The meditation level of the visual stimuli in the square recreation space was taken as the control group. The meditation level of male subjects decreased after the addition of S41, while it increased significantly after the addition of S43 and S42. On the other hand, the meditation level of female subjects decreased significantly after the addition of S43, increased after the addition of S42, and also increased significantly after the addition of S41.

Figure 8 illustrates the results of ten audio samples of S0, S61, S62, S63, S64, S65, S66, S67, S68, and S69 adapted to the interactive amusement space. The meditation level of the first five audio stimuli (S0, S61, S62, S64, S65) was significantly greater than that of the last five. The top five audio samples contained 50% or more birdsong, but there was no positive correlation between birdsong and meditation level. When birdsong was matched with children laughing, S61 > S62 > S63, the birdsong ratio was positively correlated with meditation, while children laughing was negatively correlated. We also observed an interesting phenomenon where there was no significant difference between S64 and S65, as well as between S68 and S69. This indicates that when the proportion of traffic noise is either 30% or 50%, there is no significant difference in the impact of noise on the healing effect. Additionally, we found that S65 > S68 and S64 > S69, indicating that the meditation level when paired with birdsong is significantly higher than when paired with children laughing, for both 30% and 50% proportions of traffic noise. However, when the traffic noise reaches 70%, there is no significant difference between S66 and S67, suggesting that even with 30% birdsong inclusion, it does not have a positive effect. In this regard, children laughing shows a different pattern, where regardless of the proportion of children laughing (S61, S62, S63, S67, S68, S69), pairing it with birdsong always results in significantly higher meditation levels compared to pairing it with traffic noise.

The meditation level of the visual stimuli of the interactive amusement space was taken as the control group. It was found that audio-visual stimuli showed the same trend as pure audio stimuli. The first five meditation levels were significantly greater than the last five. It was found that the meditation level after adding the first five pieces of audio was significantly higher than that of the visual stimuli. The meditation level after adding the last five audio pieces was significantly lower. It is worth noting that the increase in meditation was not related to the proportion of birdsong in the audio. Although both S63 and S66 contain 30% birdsong, the audio stimuli still have a negative effect on the scene. The combined audio-visual stimuli of S0 and S64 results in a higher meditation level compared to pure visual or pure audio stimuli. The meditation level of the audio stimuli S61 is higher than that of S64; however, when visual elements are introduced, the increase in meditation level for the audio-visual stimuli of S61 is significantly lower than that for S64.

3.2. Effects of Visual Samples and Audio Samples on the Restorative Component Scale

Reliability and validity analyses were used to test the internal consistency of the 20 restorative scale data obtained in the experiment. The results indicated that the overall Cronbach α coefficient was 0.925, with the RCS (Recovery Component Scale) scale at 0.901, the VAQ (Visual Aesthetic Quality) scale at 0.904, and the TR (Tranquility Rate) scale at 0.915, all of which were higher than 0.9. Additionally, the KMO and Bartlett’s tests were passed, indicating that the data quality of the scale was very high. After a standardized entropy analysis of the RCS scale data, the weights of the four dimensions were being away (19.21%), extent (25.5%), fascination (26.58%), and compatibility (28.71%). The weighted average of these dimensions provided the RCS value, while the TR and VAQ scales were averaged as the TR and VAQ values, respectively. The Kruskal–Wallis’ analysis showed that different combinations of visual and audio scenes significantly affected the RCS, TR, and VAQ values of urban park scenes, respectively.

As shown in Figure 9, when the audio conditions remained unchanged (S0-100% birdsong), with the exception of the interactive amusement space, the average RCS score in all other spaces was significantly greater than 2, indicating a good overall healing effect of urban parks, except for the interactive amusement space. The lawn recreation space had the highest RCS value (3.41), followed by the underforest recreation space (3.14), while the interactive amusement space had the lowest RCS value (2.01). The TR value and VAQ value showed a similar trend. Notably, the TR value of the park rest space and square recreation space was higher than that of the waterfront recreation space.

In the classification discussion, the visual conditions were unchanged, and audio-visual scenes were represented using audio sample coding for easy reading. The visual conditions were waterfront recreation space, interactive amusement space, and square recreation space. In the waterfront recreation space, the RCS values of the composite audio S11 and S12 were significantly greater than that of the original audio S0, with no significant difference between S11 and S12. The TR and VAQ values showed similar trends, but S11 had a significantly larger TR value than S12, with the difference being mainly provided by women (Figure 10). In the square recreation space, the data difference of the RCS questionnaire was mainly provided by women, and the data of women and men showed opposite trends. The combined questionnaire was not representative, so gender classification was discussed (Figure 11). In the results of the female RCS value, S0 > S41 > S42 > S43, with the decrease in the RCS value being related to the decrease in the proportion of birdsong. In the male RCS value results, S43 > S42 > S41, with the decrease in the RCS value being related to the decrease in the proportion of conversation.

As depicted in Figure 12, the RCS values of all audio-visual scenes, except for S0 (2.01), in the interactive amusement space were less than the scale mean value of 2. Similarly, TR and VAQ values also showed poor healing perception in this audio-visual space. These results suggest that the overall healing degree of the interactive amusement space is not satisfactory and optimization plans are necessary.

In the ten audio-visual scenes of the interactive amusement space, subjective rating data revealed features the same as physiological data:

(1)

The RCS, TR, and VAQ values of the first five audio-visual combinations were significantly higher than those of the last five audio-visual combinations.

(2)

In audio-visual stimuli, S64 was significantly greater than S61, and S65 was significantly greater than S62. Children laughing at 50% or less had a more negative impact on the scene than traffic noise at 50% or less.

(3)

In audio samples containing 30% and 50% traffic noise, S64 was greater than S69, and S65 was greater than S68. However, the audio samples containing 70% traffic noise (S66, S67) showed no significant difference when paired with birdsong or children laughing.

4. Discussion

4.1. Effects of Audio-Visual Interaction on Physiological Data

According to Figure 4, physiological data indicated that birdsong enhanced meditation in six types of urban park spaces, which was consistent with previous studies [47].

4.1.1. The Impact of Waterfront Recreation Space

When birdsong was added to the waterfront recreation space, meditation levels significantly decreased, despite water bodies being considered restorative landscapes [27]. Birdsong is considered the natural sound that enhances healing perception the most [47], but the decrease in meditation after combination with water seemed to suggest that the visibility of the sound source was a crucial factor in the success of audio-visual interaction. This is similar to the notion that “subjects were surprised by the presence of sound (birdsong) because there was no visible source in the image at all”. Although previous studies have reported green vegetation and sky as visible sources of birdsong [48], our study shows that ample, calm water and the sky above it are not visible sources of birdsong. This finding may be related to the fact that the scene differs from the perception in traditional park images, with more than 90% of the image covered by water and sky. The overall calm tone and lack of greenery may lead to the perception that few forest birds are present near the bay or in the middle of a large lake, resulting in a mismatch of audio-visual interaction. Additionally, the birdsong used in this study was collected from the park, which is quite different from seabird song. Matching the visual with seabird song may generate better meditation. The premise that the sky becomes the visible sound source of birdsong should be related to other factors in the image. The sky above the blue space and the sky above the green space may significantly differ in the perception of birdsong. The increase in meditation with the increase in the proportion of flowing water sounds also indicates the effect of sound source visibility.

4.1.2. The Impact of Square Recreation Spaces

Regarding the square recreation space, the audio-visual analysis showed that males and females responded differently to birdsong and conversation (Figure 7), presenting opposite trends. This gender difference was rarely reported in previous studies. Liu Fangfang reported in her study that gender difference was not statistically significant, but subtle differences existed. For example, women were more sensitive to sound comfort and noise and were more easily disturbed by sound than men [49]. Wang, Q. reported that women preferred quiet landscapes [50]. In Wortzel’s study [51], “Quiet” and “social” factors were a pair of seemingly contradictory influences that gave people a sense of healing from different starting points. Birdsong provides a masking effect that creates a sense of quietness [35], and it is speculated that conversation may provide a sense of sociality. This may explain why the meditation level in a square recreation space is positively correlated with two seemingly contradictory factors. It should be noted that because the population in the experiment only consists of young students experiencing chronic stress, gender differences may only exist within this same population.

4.1.3. The Impact of Interactive Amusement Spaces

Our study on the interactive amusement space differs from Fu’s research [36]. We found that 50% or more of birdsong significantly impacts the healing potential of the space, while 30% of birdsong has no significant impact. The positive effect of 50% or more is unrelated to the proportion of birdsong but is closely related to the matching sound source. This situation has never been reported in previous studies, indicating that reasonable audio combinations can produce superior meditation. However, the meditation level of most audio-visual stimuli is not as high as that of pure audio stimuli, which is consistent with Li’s conclusion [31]. The interactive effect of audio-visual stimuli may not always be positive, and in some cases, pure auditory stimuli may be more effective than audio-visual stimuli. Nevertheless, the data from S0 and S64 demonstrate that audio-visual interaction can produce a level of meditation that is not possible with visual or audio stimuli alone. This finding is consistent with previous studies [27], underscoring the need for audio-visual interaction scene design. Regarding S61, S62, and S63, the meditation level produced by audio-visual stimuli and audio stimuli is opposite, suggesting that children laughing in S61, S62, and S63 may interact with the visible sound source, reducing the meditation level of the scene. Visible children laughing, when present at 50% or lower intensity, may disturb the site’s meditation level more than invisible traffic noise. This finding confirms Bangjun’s conclusion [52] that the visibility of poor audio sources can have a negative impact on the sense of healing. It also highlights the importance of consciously designing audio-visual interaction scenes, such as first considering reducing the probability of unrelated tourists seeing children and using sound-absorbing materials to reduce the decibels, thereby solving the fundamental problem of affecting the therapeutic effect of the site, and then considering using green plant environments or soothing sounds to enhance the healing potential of urban park sites.

4.2. Effects of Audio-Visual Interaction on Subjective Rating Data

The subjective rating data showed significant correlations among different dimensions and significant correlations with physiological data, providing basic support for the physiological data results. In the waterfront recreation space, women demonstrated a greater tendency than men to have an increase in the proportion of birdsong to enhance their subjective healing feelings, which was consistent with Luo’s findings [53]. In the square recreation space, RCS and VAQ values exhibited a gender trend like the physiological data, where females preferred birdsong and males preferred conversation. However, males showed a positive correlation with birdsong in TR values. Although both “quiet” and “social” were significantly associated with healing potential, it is evident that the two are separate dimensions [54]. When healing potential is caused by the “social” dimension, as in the male data in the square recreation space in this study, it is normal for TR values to show a trend opposite to RCS values. In the interactive amusement space, the subjective healing experience of the venue was excellent when the proportion of birdsong was 50% or higher. The combination of (50% and below) children laughing appears to induce more therapeutic interference compared to traffic noise at the same sound pressure level, which may be related to the visibility of the sound source. Additionally, we found that the RCS value of any one source was not positively correlated with the increase of its proportion but was closely correlated with the combination of the other half of the source.

4.3. Factors Influencing the Effectiveness of Audio-Visual Interaction in Park Therapy

Urban parks play a vital role in mitigating urban ailments in cities. However, there is a need to understand the mechanisms of how different park sounds and visual scenes interact with people and their positive or negative effects. This study investigated the impact of audio-visual interaction scenes of “what can happen in the park” on people’s healing experiences. Although birdsong is considered a significant sound source, [47] it may not always be suitable for use as a healing sound. The study found that the proportion of birdsong is not linearly related to healing potential but is instead closely related to the combination of audio and visual. Therefore, to improve the healing potential, it is not recommended to only increase the proportion of birdsong, but to study the visibility of sound sources and audio matching. The interaction between different sound sources and healing potential is not straightforward and requires further investigation.

High levels of children laughing can significantly damage people’s healing experiences, while low levels of children laughing combined with birdsong can enhance them. Traffic noise has a significant negative impact on healing potential [55], but in this study, since there was no visible source of traffic noise provided, and a visible source of children laughing was provided instead, the interference caused by the audio-visual interaction of traffic noise on meditation effectiveness was even smaller than that of children laughing. Controlling the visibility of sound sources is an important factor in controlling the therapeutic effects of audio-visual interactions. In addition, trees can provide sound absorption which can effectively weaken the impact of traffic noise on the therapeutic environment [56].

The data of the conversation in the square recreation space indicate that the human voice is not necessarily a negative factor. Both physiological data and RCS data (Figure 5 and Figure 9) indicate a positive inclination of males towards conversational sounds in terms of therapeutic perception. Social interaction is essential for city dwellers [57], and according to Peschardt’s study [54], “Quiet” and “social” were the most crucial factors associated with people’s perception of restoration in urban parks. Additionally, tranquility characteristics seem to align with natural features [58]. The sound of conversation matches social traits, which provides new evidence for the positive effects of human voices. Adding visual stimuli reinforces this trend, as previous reports have shown [59]. Furthermore, artificial landscapes increased the acoustic comfort of human-related sounds compared to natural landscapes. Based on the current situation, the sound of conversation does not show significant adverse effects and may even have positive effects in some cases. However, further investigation is needed to explore the healing properties of conversation and even the human voice.

In contrast to previous design propositions, the study does not recommend simply adding natural sounds to urban park environments to increase the healing potential. Instead, it emphasizes the need for excellent construction and planning of urban park systems that consider the urban context and the inevitable soundscapes of human voices. The vitality of the healing system of urban parks lies in the delicate intercommunication and disciplinary complementarity of research processes that take into account the unique characteristics of urban parks and their potential for therapeutic benefits in the park city and urban park healing system.

4.4. Limitations and Prospects

While the study presented some valuable insights that answer some of the expectations and perplexities of recent research on complex soundscapes [27,30,33,36,53], there are also several limitations that need to be considered. Firstly, the trial was short-term, and therefore, the long-term healing potential of urban park therapy could not be assessed. Additionally, as in most previous works [33,34,36], the subjects were predominantly young, and it is unclear whether the findings can be generalized to different age groups. Furthermore, the need for control variables meant that the portability of the experimental equipment was not fully utilized, and direct field experiments with the equipment may lead to new ideas [31].

Given the current policy environment and the mental health crisis in China [7], it is crucial to further investigate the role of park therapy and the transformation of urban development modes [24]. This can be achieved by improving the overall sample representativeness and population segmentation and utilizing interdisciplinary research to explore the mechanism of park therapy on the five senses of the human body. While there has been progress in unveiling the mystery of the interaction between the senses, there is still much to be done. In particular, research on the audio-visual and olfactory senses is still ongoing, and the limitations of experimental conditions and the complexity of the experiments pose significant challenges. Therefore, further analysis of the interaction principles of the five senses for healing is crucial to gain a deeper understanding of the role of urban park therapy in promoting human well-being.

5. Conclusions

This study investigates the therapeutic effects of urban parks on students under chronic stress through Mindwave monitoring and the Recovery Component Scale. The following conclusions were drawn: (1) There may be significant differences in the visual sky above the blue space and the visual sky above the green space when interacting with birdsong. (2) Birdsong and conversation from seemingly contradictory dimensions of “Quiet” and “social” seem to provide healing, with gender differences among young students being a key factor, providing new evidence for their healing mechanisms. (3) Visible children laughing at low levels (30% and 50%) has a more negative impact than invisible traffic noise at the same sound pressure level. (4) Depending on the visibility of the sound source, audio-visual interaction may not necessarily have a positive effect. (5) In the experiment, no single sound source was found to have a positive correlation with the RCS value as its proportion increased. Instead, it was closely related to the matching sound source, indicating the complexity of acoustic landscape therapy.

In fact, multi-sensory design, such as audio-visual integration, has great potential in enhancing healing effects, enabling urban parks to play a role in “substitute therapy” and reducing the burden on the mental health care system. Therefore, future park designers should consciously consider strategies for improving the healing potential from the audio-visual perspective by controlling the visibility of sound sources to reduce the negative effects of children laughing and traffic noise, enhancing the positive effects of birdsong and flowing water sounds; by adjusting the matching of sound sources to transform children laughing, which is originally a negative effect, into an overall positive effect; and by enhancing positive sound sources that may appear in visual stimulation to achieve healing, thereby maximizing the audio-visual healing effect of urban park spaces.