Evaluation of Healthy Acoustic Environments in Industrial Buildings from the Workers’ Perspective: A Mixed-Methods Approach

Abstract

Noise in industrial buildings affects workers’ productivity and can seriously impair their physical and mental health, yet existing studies often overlook workers’ subjective perceptions and rely on a single method. Therefore, this study recruited 263 workers from four industrial buildings in Beijing and adopted a mixed-methods approach. First, 30 semi-structured interviews were analyzed using grounded theory’s three-level coding procedure to construct a conceptual framework of a healthy acoustic environment and its influencing factors. Next, a 30-item subjective questionnaire was developed, and structural equation modeling was conducted on 256 valid responses. Finally, Spearman correlation analysis and multidimensional scaling were used to examine relationships between subjective evaluations and eight physical and psychoacoustic indicators. The results identified nine major dimensions, including Sound Source Localization, Physiological Effects at Work, and Regulatory Control, as well as 15 relational pathways. Compared with existing frameworks, Communication Barrier emerged as a more prominent dimension in industrial building contexts. Structural equation modeling confirmed that 12 pathways were statistically significant. Correlation analysis further showed that only a few objective–subjective associations were significant, indicating that objective acoustic indicators alone cannot explain workers’ multidimensional perceptions. In conclusion, this study developed an evaluation model for healthy acoustic environments in industrial buildings, highlighting the need to emphasize controllability, communication support, and integrated subjective–objective evaluation in acoustic design to better enhance workers’ well-being.

1. Introduction

Noise, as a pervasive environmental stressor, significantly affects human health and overall quality of life. In the fields of architecture and urban studies, the impacts of noise pollution vary across building typologies, primarily manifesting as potential disturbances to occupants’ health, comfort, and work efficiency [1,2,3,4]. For example, Jeon et al. [3] conducted in-home surveys and laboratory-based listening experiments to assess residents’ annoyance responses to various sounds, and found that impact noise from floor slabs had the greatest effect on residential satisfaction. Baloye and Palamuleni [5] measured environmental noise levels across several Nigerian cities using sound level meters and reported that 79% of sampling locations exceeded regulatory limits, with traffic and commercial areas accounting for the highest proportions. Such noise pollution poses substantial threats to residents’ health, sleep quality, and work efficiency.

In addition, an increasing number of studies have focused on the acoustic environment of industrial buildings [6,7,8], where noise generated during production processes adversely affects workers’ physical health. Compared with other built environments, industrial workshops are typically characterized by higher noise intensity and longer exposure durations [9], rendering workers a particularly vulnerable population to noise exposure. Extensive evidence indicates that long-term noise exposure results in cumulative adverse effects [10]. It not only induces pronounced negative emotions, such as irritability and distress [11], but may also lead to a range of physiological disorders, including hearing loss [12], cardiovascular diseases [13], sleep disturbances [14], and depression [15], thereby disrupting normal work performance and overall health [16]. As concerns regarding noise pollution in industrial buildings continue to grow [17,18,19,20,21], related policies have been progressively introduced to promote hearing conservation and prevent noise-induced hearing loss and other adverse health outcomes [22]. However, workers’ subjective evaluations and perceptions of noise within a healthy work environment may not align with the criteria specified in these standards [23]. Even when noise-reduction measures are implemented, they do not necessarily translate into measurable improvements in workers’ health status [24]. Previous studies suggest that research on acoustic environments should move beyond the traditional paradigm of purely physical measurements and adopt a multidimensional and integrative analytical framework [25]. To address this limitation, qualitative approaches have been increasingly adopted in architectural research [26,27]. Moreover, mixed-methods designs and investigations into the relationships between objective acoustic indicators and subjective perceptions have gradually been introduced across various building contexts [28,29].

Grounded theory, as a qualitative research method, is fundamentally focused on inductively generating and constructing theory from empirical data, rather than relying on pre-established theoretical frameworks [30]. In the field of architecture, it offers new perspectives and approaches to address complex design, construction, and management challenges. By generating theory from empirical data, grounded theory bridges grand theories and micro-level narratives. This approach helps to highlight the local characteristics of research phenomena and fosters the development of discipline-independent knowledge systems [31]. In previous studies, grounded theory has been applied to various building types, including residential buildings [29], commercial office buildings [32], and health and wellness architecture [33]. In industrial architecture, there are also studies that apply grounded theory. For example, Monteiro et al. [34] conducted interviews with workers from multiple footwear manufacturing companies in Portugal. After data coding, they developed a three-stage dynamic capability reconstruction model, which offers a roadmap for small and medium-sized enterprises following the loss of cost advantages. Han et al. [35] used grounded theory to conduct in-depth interviews with relevant personnel in the Ningbo region. They identified three conflict scenarios and seven specific types of conflict related to industrial building renovation, developing a theoretical model that provides foundational support for subsequent renovations in industrial architecture.

Conclusions derived from qualitative approaches such as grounded theory can be further examined and validated through quantitative methods. Such a mixed-methods approach strengthens the integration between qualitative insights and quantitative validation. In the quantitative phase, statistical data processing and the construction of a structural equation model can be employed for empirical verification. Structural equation modeling, as a powerful multivariate statistical technique, has been increasingly applied in architectural research. It enables the simultaneous visualization and verification of multiple complex relationships and has demonstrated substantial applications in building management, design processes, user satisfaction, and environmental perception [36,37,38]. Alam et al. [39] employed structural equation modelling to examine the impact of library service quality on user satisfaction in Bangladesh. By decomposing service quality into five dimensions and estimating path coefficients, they found that resources, staff competence, service attitude, and tangible facilities significantly influenced user satisfaction. This study expanded the investigation of structural relationships between service quality and user satisfaction in library buildings. In the context of industrial buildings, the flexibility of structural equation modelling enables the exploration and evaluation of complex interrelationships involving sustainability, energy efficiency, construction management, and safety- and health-related factors [40,41]. For instance, Heravi et al. [42] developed structural equation models for petrochemical industrial plants across three life-cycle stages, reducing the dataset to 19 key indicators. Their results demonstrated that different dominant dimensions should guide decision-making at different life-cycle phases, thereby providing a more comprehensive design framework for such facilities throughout their entire life cycle.

Data used in structural equation modeling are typically derived from subjective questionnaires. Although statistically supported, such data ultimately reflect the individual judgements of different workers. In contrast, parameters such as indoor temperature, humidity, sound level, and illuminance are objectively measurable physical variables. These physical factors are intrinsically linked to workers’ subjective perceptions. In architectural research, the interaction between subjective and objective factors constitutes a critical line of inquiry, addressing the complex relationships between the physical properties of built environments and human perceptual experiences. Although subjective and objective dimensions are inseparable in real-life contexts, they are often treated as opposing approaches in evaluation frameworks [43]. Architects frequently rely on professional analytical methods in architectural design, structural optimization, and green building strategies; however, such approaches may overlook users’ subjective experiences [44]. For example, Li et al. [45] conducted experiments in open-plan offices and examined the effects of informational interventions on occupants’ perceptions of the thermal environment. The results revealed discrepancies between subjective sensations and measured physical parameters, indicating that psychological factors influence individuals’ perception of physical conditions and contribute to divergences between subjective and objective assessments. Similarly, in a study of concert halls, Giménez et al. [28] compared responses from 1284 audience members across 28 questionnaire items with 18 acoustic indicators. Substantial discrepancies were identified between subjective evaluations and objective acoustic measures. Multidimensional scaling analysis further confirmed the divergence, suggesting that the perceived pleasantness of sound is only weakly correlated with its physical acoustic properties.

In summary, noise has a significant impact on industrial buildings and the health-related working environment of employees; however, existing studies have paid insufficient attention to workers’ subjective perceptions in industrial building contexts. Existing qualitative studies on subjective perceptions of acoustic environments have mainly focused on urban building types such as residential buildings, commercial office buildings, healthcare facilities, and schools [32,33,46]. In contrast, research on industrial buildings has predominantly focused on development directions [34] and renovation processes [35], with relatively limited discussion of the acoustic environment itself and workers’ subjective experiences. In quantitative research, structural equation modeling has been applied in fields such as architectural design [47] and building management [39]; however, studies addressing the physical environment of industrial buildings—particularly the mechanisms of acoustic environment influence—remain insufficient. Regarding the relationship between subjective and objective factors, existing studies have explored the associations between subjective evaluations and objective indicators in buildings such as offices [45] and concert halls [28], yet similar analyses for industrial buildings remain scarce. Furthermore, although Chen et al. [29] developed an evaluation model for the acoustic environment of residential buildings based on grounded theory and structural equation modeling, their study did not rely on sufficiently comprehensive face-to-face interviews, nor did it integrate building physical measurements into a combined analysis. Therefore, systematic research on workers’ subjective perceptions in industrial buildings, their influencing pathways, and their relationships with objective acoustic indicators remains to be further developed. Based on the current state of research, although existing studies provide an important foundation for investigating architectural acoustic environments, they are still insufficient to comprehensively reveal workers’ actual experiences of the acoustic environment in industrial buildings and the mechanisms underlying these experiences.

Therefore, in response to the aforementioned issues, this study is grounded in field experiments and employs on-site measurements, semi-structured interviews, scale-based assessments, and statistical data analysis to explore the influencing factors, conceptual framework, and evaluation model of healthy environment assessment in industrial buildings. Furthermore, it analyses the correlations between objectively measured physical parameters and workers’ subjective perceptions within industrial building environments. Specifically, the primary research questions of this study are as follows:

Q1: Based on grounded theory, what are the influencing factors of a healthy acoustic environment in industrial buildings?

Q2: Using structural equation modeling, what are the quantitative relationships among the influencing factors of a healthy acoustic environment in industrial buildings?

Q3: What are the correlations between physical noise indicators, psychological indicators, and workers’ subjective perceptions?

2. Materials and Methods

This study consists of two main components: qualitative and quantitative research. The experimental procedure and data analysis process are described as follows:

1. To explore the constituent elements and theoretical framework of the acoustic environment in industrial workshops, semi-structured interviews were conducted with workers in the workshop. The interview data were coded and interpreted using grounded theory, ultimately resulting in the development of a grounded theoretical model.

2. Based on the key elements identified from the semi-structured interviews, a questionnaire was designed. The collected survey data were analyzed using structural equation modeling to verify the strength of the relationships among variables, leading to the establishment of an evaluation model for a healthy acoustic environment in industrial buildings.

3. During a complete experimental workflow performed by the workers, four physical acoustic indicators and four psychoacoustic indicators were recorded in the workshop. These indicators were matched correspondingly with the questionnaire items, enabling analysis of the strength of correlations between objective measurements and subjective survey responses. The overall research procedure is illustrated in Figure 1.

Figure 1. Research Flowchart.

2.1. Grounded Theory Analysis (Qualitative Research)

2.1.1. Participants

In grounded theory research, purposive sampling [48] was adopted to ensure the collection of rich and in-depth information from interview participants. The qualitative component of this study was conducted through face-to-face interviews with workers employed in four industrial buildings in Beijing. The interview content was transcribed and documented in written form. A total of 30 participants (18 males and 12 females) were interviewed. Their ages ranged from 23 to 65 years (mean = 45.37 years, standard deviation = 11.26 years), covering various job types, working durations, and years of employment within the workshop. Detailed demographic information of the participants is presented in Table 1. The interviews were conducted between September and October 2025, with each session lasting between 20 and 45 min.

Table 1. Demographic Information of Semi-Structured Interview Participants (n = 30).

Item	Category	Percentage (%)
Gender	Male	60%
	Female	40%
Age	18–30	20
	30–40	20
	40–50	26.7
	50–60	26.7
	>60	6.7
Years of Service	1–5	33.3
	5–10	16.7
	10–15	6.7
	15–20	10
	20–25	6.7
	25–30	6.7
	>30	20
Daily Working Hours	1–3	16.7
	3–6	13.3
	6–9	53.3
	9–12	16.7
Whether Wearing Protective Equipment	Yes	50
	No	50

2.1.2. Interview Guide Design

Based on existing interview guidelines, established scales [29,46,49,50], and relevant standards, several items were adapted through translation and contextual modification. An interview guide was then developed covering personal demographic information and six dimensions of the acoustic environment. The detailed structure of the interview guide is presented in Table 2.

Table 2. Interview Outline.

No.	Category	Questions
1	Personal Information	Age, Gender, Years of Service, Job Type, Daily Working Hours, Whether Wearing Protective Equipment
2	Sound Source Identification and Perception	What sounds do you often hear in the workshop?
		Have you noticed that certain sounds become more prominent at specific times?
3	Space and Soundscape Relationship	Do you feel that certain areas in the workshop are noisier or quieter?
		Does the spatial layout affect the perception of sound?
4	Emotions and Reactions	Do certain sounds make you feel irritable, fatigued, or distracted?
		Do you use certain methods to block out these sounds?
		Do sounds affect your cooperation or communication with others?
5	Control Ability	Do you feel that you can control your working environment?
6	Ideal Sound environment	What is your ideal sound environment for the workshop?
7	Health and Behavioral Impacts	Do you feel that the sounds in the workshop affect your physical condition? What are the specific manifestations? What was your most recent experience of decreased work performance due to noise? How long did it last?

2.1.3. Interview Procedure

Semi-structured interviews were employed as the primary data collection method for the grounded theory analysis. The interview content was structured in accordance with Table 2. The interviews began with the collection of participants’ personal information, followed by questions addressing six aspects: sound source identification, spatial relationships, emotional responses, perceived control, ideal soundscape assumptions, and the influence on health-related behaviors. Participants’ responses were transcribed and documented in written form. All participants provided oral consent to participate in the experiment.

2.1.4. Interview Data Analysis

During the data collection and organization stage, when no new information could be obtained from additional interviewees, theoretical saturation was considered to have been reached, and the current phase of the study could be concluded. In this study, theoretical saturation was achieved after the 22nd interview. Subsequently, eight additional participants were interviewed, and no new terms or insights emerged, confirming that the interview process could be formally concluded [51].

The interview data were analyzed using NVivo 15 (Developer: Lumivero; Location: Denver, CO, USA) software and processed through the three-stage coding procedure of grounded theory. First, open coding was conducted to conceptualize, categorize, and label the interview content. Second, axial coding was performed to identify relationships among the previously generated codes and to establish connections between primary and secondary categories. Finally, selective coding was undertaken to integrate the core categories relevant to this study, thereby constructing a comprehensive conceptual framework.

The above coding process was independently conducted by two researchers specializing in architecture to minimize the influence of subjective judgment from a single researcher on the coding results [52]. For items with discrepancies in the initial coding, the researchers carried out item-by-item discussions and repeated comparisons until a consensus was reached. Categories that still allowed for interpretative ambiguity were further reviewed in consultation with an expert panel. Finally, triangulation was employed to further enhance the accuracy and reliability of the coding results [53].

2.2. Structural Equation Modeling Analysis (Quantitative Research)

2.2.1. Participants

In this phase, workers from four industrial buildings located in Beijing were recruited as participants. The experiment was conducted from October 2025 to March 2026, during which a questionnaire survey entitled “Evaluation of the Workshop Acoustic Environment” was administered to the participants. Throughout the experiment, the workers remained inside the industrial building and were continuously exposed to an enclosed environment. A total of 263 individuals participated in the survey. Among the returned questionnaires, 256 were valid and 7 were excluded due to incomplete responses or missing physical indicator records. Consequently, 256 valid questionnaires were included in the subsequent data analysis. The sample size met the minimum requirements for structural equation modeling [54]. Detailed demographic information of the participants is presented in Table 3.

Table 3. Demographic Information of Subjective Questionnaire (n = 256).

Item	Category	Percentage (%)
Gender	Male	52.34
	Female	47.66
Age	18–30	13.28
	30–40	21.48
	40–50	30.86
	50–60	26.95
	>60	7.42
Years of Service	1–5	29.69
	5–10	26.56
	10–15	9.77
	15–20	12.5
	20–25	7.81
	25–30	5.47
	>30	8.2
Daily Working Hours	1–3	23.05
	3–6	12.5
	6–9	37.89
	9–12	26.56
Wearing Protective Equipment	Yes	49.22
	No	50.78

2.2.2. Questionnaire Design

The subjective questionnaire comprised ten sections. The first section collected demographic information, whereas the remaining nine sections were structured according to the conceptual framework derived from grounded theory analysis. These nine elements included Sound Source Localization, Physiological Effects at Work, Communication Barrier, Regulatory Control, Negative Emotions, Passively Accept, Active Regulation, Fatigue Accumulation after Work, and Healthy Acoustic Environments. Each section contained three to four items, yielding a total of 30 questions. The number of items per construct met the minimum requirements for structural equation modeling analysis [55]. All items were rated on a five-point Likert scale, ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). The questionnaire items were primarily developed with reference to ISO soundscape indicators [56] and previous studies [50,57,58]. The original items were translated and adapted to ensure their suitability for the experimental participants and setting of the present study. The detailed outline of the questionnaire items is presented in Table 4.

Table 4. Items of Each Factor in the Subjective Questionnaire.

Factor	Code	Item Description
Sound Source Localization	Sou1	Sound identification
	Sou2	Sound judgment
	Sou3	Spatial distinction
	Sou4	Layout distinction
Physiological Effects at Work	Phy1	Work errors
	Phy2	Continuous noise
	Phy3	Physical discomfort
	Phy4	Decline in mental state
Communication Barrier	Com1	Repetition
	Com2	Communication difficulty
	Com3	Spatial transfer
	Com4	Mishearing
Regulatory Control	Con1	Control ability
	Con2	Right to choose
	Con3	Willingness to adjust
Negative Emotions	Neg1	Irritability
	Neg2	Palpitation
	Neg3	Momentary panic
Passively Accept	Pas1	Continuity
	Pas2	Adaptation
	Pas3	Uncontrollability
Active Regulation	Act1	Active shielding
	Act2	Background sound masking
	Act3	Listening to music for relaxation during breaks
Fatigue Accumulation after Work	Fat1	Difficulty in recovery
	Fat2	Decreased attention
	Fat3	Noise residue
Healthy Acoustic Environments	Hea1	Work impact
	Hea2	Efficiency reduction
	Hea3	Satisfaction

2.2.3. Experimental Setup

In each experimental session, participants were measured under actual workshop acoustic conditions, with each measurement lasting no less than 5 min to meet the minimum time requirement [56]. In this study, the average duration of a single experimental session was 15–20 min. Since workers are already exposed to continuous noise in their daily work, excessively long experimental sessions may increase fatigue and affect the accuracy of subjective evaluations. Therefore, under the premise of meeting standard requirements, the duration of each session was controlled within 30 min.

2.2.4. Data Analysis

The data were analyzed using Amos 31 (Developer: IBM SPSS; Location: Armonk, NY, USA). Reliability analysis was first conducted, followed by exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Subsequently, a structural equation modeling framework was constructed. Given that the data were non-normally distributed, the analysis was performed using maximum likelihood estimation with robust standard errors.

2.3. Subjective–Objective Correlation Analysis (Quantitative Research)

2.3.1. Participants

In this phase, workers from four industrial buildings located in Beijing were recruited as participants. Physical measurements were carried out concurrently with the administration of the “Evaluation of the Workshop Acoustic Environment” questionnaire.

2.3.2. Experimental Design

During a single work shift, each participant was equipped with a HEAD SQobold recorder(Developer: HEAD acoustics GmbH; Location: Herzogenrath, Germany) to measure four psychoacoustic parameters: loudness, sharpness, roughness, and fluctuation strength. Simultaneously, an SC260 sound level meter(Developer: CESVA; Location: Barcelona, Spain) was used to record four physical acoustic indicators: LAeq, LAmax, LAmin, and low-frequency noise (LCeq-LAeq). In this study, the SC260 sound level meter was operated under A-weighting and Fast time-weighting conditions. Each sound pressure level reading represents a time-weighted value with a 125 ms time constant, while LAmax and LAmin denote the maximum and minimum values, respectively, of the time-weighted sound level series during the measurement period. LAeq represents the equivalent continuous sound level over the measurement period, while LCeq-LAeq is used to characterize low-frequency noise predominantly generated by machinery operation, structural vibration, or aerodynamic processes in industrial building environments, expressed as the difference between C-weighted and A-weighted sound pressure levels. All measurement devices were activated before the experiment and stopped after its completion, ensuring that all acoustic parameters were synchronously collected within the same time window. The duration of each experimental session was determined based on the work shift and the specific workflow of the workers; however, all measurements exceeded 5 min, satisfying the minimum duration requirement [56]. In this study, the measurement duration for these physical indicators ranged from 15 to 20 min for each trial.

The experiments in this phase were conducted in four selected industrial buildings. By comprehensively considering the workshop spatial layout and personnel distribution characteristics, a total of 90 measurement points were established. Each measurement point corresponds to a group of workers within its surrounding operational area, and the collected acoustic data reflect the actual noise exposure of individuals in that area. A total of 256 valid datasets and questionnaires were ultimately obtained, and a one-to-one correspondence between spatial measurement points and individual samples was established, enabling an effective linkage between subjective evaluations and objective acoustic parameters. The specific locations of the measurement points are shown in Figure 2.

Figure 2. Distribution of Measurement Points.

2.3.3. Data Analysis

Raw audio files were recorded using the HEAD SQobold device and subsequently analyzed with ArtemiS SUITE 17.0 (Developer: HEAD acoustics GmbH; Location: Herzogenrath, Germany). The mean values of four psychoacoustic parameters—Loudness, Sharpness, Roughness, and Fluctuation Strength—were calculated. An SC260 sound level meter was used to measure four physical acoustic parameters: LAeq, LAmax, LAmin, and LCeq-LAeq. These eight parameters have been demonstrated in previous studies to effectively characterize the relationship between acoustic features and the degree of subjective disturbance [59]. The measurement results for the physical indicators are presented in Figure 3.

Figure 3. Statistical Analysis of Acoustic Indicators.

The maximum, minimum, and average sound pressure levels directly reflect the intensity of noise and its temporal variations. ISO standards consider occupational noise exposure levels and peak sound pressure as fundamental risk parameters [60]. Workers’ subjective responses to noise depend on various acoustic characteristics, such as exposure duration and sound pressure level [61]. Beat et al. [62] found that variations in sound pressure level can influence individuals’ subjective perceptions. Prashanth et al. [63] reported that low-frequency noise can affect workers’ psychological perceptions. Shehap et al. [64] demonstrated that even when complying with conventional standards, low-frequency noise can still cause significant annoyance, fatigue, headaches, ear pressure, and reduced concentration among employees, thereby severely impairing work performance and health. Moreover, traditional A-weighted assessments are insufficient to capture the hazards of low-frequency noise. Therefore, in addition to sound pressure level, the measurement of LCeq-LAeq is necessary to comprehensively characterize the physical properties of noise in industrial buildings. A systematic review by Engel et al. [65] found that five psychoacoustic indicators—loudness, sharpness, roughness, fluctuation strength, and tonality—were the most frequently used in 46 soundscape studies over the past decade. Therefore, both physical acoustic indicators and psychoacoustic indicators are essential for acoustic evaluation [66]. The eight indicators selected in this study can comprehensively reflect the intensity, frequency spectrum, and sound quality characteristics of noise in industrial buildings. Considering that tonality is more discriminative for noise with prominent tonal components, it was not treated as a primary analysis indicator in this study. Overall, the eight physical indicators selected in this study can effectively represent the fundamental characteristics of the acoustic environment in industrial buildings.

Statistical analyses were conducted using SPSS 31 (Developer: IBM; Location: Armonk, NY, USA). Correlation analyses were performed between the subjective questionnaire items presented in Table 4 and the eight objective acoustic parameters. As some variables did not follow a normal distribution, Spearman’s rank correlation was applied. Furthermore, a heatmap and a multidimensional scaling distance model were constructed to further assess the strength and structure of the associations between subjective and objective measures.

3. Results

3.1. Composition of Healthy Acoustic Environments Elements and Conceptual Framework—In Response to Q1

Through grounded theory coding, this study identified nine core elements influencing a healthy acoustic environment: Sound Source Localization, Physiological Effects at Work, Communication Barrier, Regulatory Control, Negative Emotions, Passively Accept, Active Regulation, Fatigue Accumulation after Work, and Healthy Acoustic Environments. An illustrative example of the coding and construction process is presented in Table 5.

Table 5. Example of Partial Coding Process.

Interview Record	Label	Data Analysis	Data Categorization	Category
P04: The noise is very loud, sometimes it makes me feel like I’m not on the same channel when talking to others, sometimes I say the wrong things. If I’m close, sometimes I have to repeat myself and speak loudly to make my colleagues hear clearly. After talking for a long time, I feel my throat getting tired.	a1: Noise a2: Loud a3: Inaudible a4: Repeated communication a5: Throat fatigue a6: Difficult to control equipment a7: No choice a8: Equipment uncontrollable a9: Affects thinking a10: Location a11: Long-term a12: Headache a13: After work a14: Ear pain	aa1: Noise causes me to be unable to hear clearly when communicating with others, and affects my body aa2: Loud noise requires me to repeat communication, causing throat discomfort aa3: Uncontrollable equipment leaves me with no choice aa4: Sound causes me to be distracted and affects my thinking aa5: Long-term noise exposure makes me feel headache aa6: After work, noise still makes my ears hurt	Aa1: Noise affects my communication with others and impacts my body Aa2: I am very passive in controlling sound Aa3: Long-term noise affects my physical health Aa4: Noise has continuous effects on my body	A1: Sound Source Localization A2: Physiological Effects at Work A3: Communication Barrier A4: Regulatory Control A5: Negative Emotions A6: Passively Accept A7: Active Regulation A8: Fatigue Accumulation after Work A9: Healthy Acoustic Environments
P06: Controlling the equipment is difficult because the equipment is very large and cannot be placed in an enclosed environment. Additionally, the possibility of noise reduction is not high, there are not many choices, it can only be as it is now.
P17: Sometimes I feel distracted because I sometimes need to make phone calls to coordinate business, and sometimes I have to think about business matters myself. When thinking, being in such a noisy environment for a long time affects my own thinking, including this kind of sound. When making phone calls, I sometimes can’t hear clearly, so I have to go out to a quiet environment to make the call.
P19: The sound of machine operation, sometimes long-term operation makes me feel a bit headache.
P26: When I just left the workshop after work, there is still a buzzing sound in my ears, sometimes accompanied by ear pain.

3.1.1. Element Composition

Based on their classification, the nine core elements can be divided into three major categories: Influencing Factors, Behavioral Measures, and Results. The category of Influencing Factors includes four element types: Sound Source Localization, Physiological Effects at Work, Communication Barrier, and Regulatory Control. The category of Behavioral Measures comprises Negative Emotions, Passively Accept, Active Regulation, and Fatigue Accumulation after Work. The category of Results consists of a single element: Healthy Acoustic Environments.

Specifically, Influencing Factors refer to a set of variables that act on individuals within the current acoustic environment and regulate or drive their emotional, physiological, and behavioral responses. These factors constitute the antecedent conditions that give rise to behavioral measures. Behavioral Measures denote the emotional, cognitive, and behavioral responses exhibited by individuals as they cope with or adapt to acoustic environmental stressors. Results refer to the final manifestations of individuals’ psychological and physiological changes under the prevailing acoustic environment.

The specific definitions of the nine core elements are provided below.

1. Sound Source Localization refers to workers’ perception of sound-related characteristics, such as sound intensity and sensitivity, which subsequently elicit emotional responses to the acoustic environment.

2. Physiological Effects at Work denote the persistent physiological impacts experienced during working hours as a result of noise exposure.

3. Communication Barrier describes situations in which environmental sounds reduce communication efficiency or cause difficulties in interpersonal communication among workers.

4. Regulatory Control refers to workers’ ability or subjective willingness to regulate or control environmental sounds through personal capacity or intentional actions.

5. Negative Emotions encompass adverse emotional responses such as irritability, anxiety, and fright that arise as primary reactions to the acoustic environment.

6. Passively Accept characterizes situations in which workers are unable to alter the sound environment and are therefore compelled to remain exposed to the existing noise conditions.

7. Active Regulation refers to workers’ attempts to modify the current noise environment by adopting shielding measures or using protective devices, thereby reducing the detrimental effects of noise on their hearing.

8. Fatigue Accumulation after Work denotes the residual physiological discomfort caused by noise exposure after working hours, leading to impairments in both physical capacity and mental state.

9. Healthy Acoustic Environments represents an overall evaluation of workers’ job efficiency, task accuracy, and satisfaction with the working environment.

In the interviews, Sound Source Localization, Physiological Effects at Work, and Regulatory Control were primarily associated with Negative Emotions. Representative interview excerpts are presented below.

P01: Prolonged exposure to this kind of noisy machine vibration makes me feel quite irritable.

P16: Sometimes the air pump malfunctions and releases air suddenly, which startles me. It can be quite frightening and makes me feel very uncomfortable.

In the interviews, Sound Source Localization, Physiological Effects at Work, Communication Barrier, and Regulatory Control were associated with Passively Accept. Selected interview excerpts are as follows.

P28: Communication is not very convenient, but there is nothing we can do. Once the equipment is turned on, it cannot be stopped. The environment is quite noisy, and it is not something we can change just because we want to.

P06: Sometimes I cannot hear clearly because the noise is too loud. In addition, I am relatively old, so younger coworkers have to speak close to my ear for me to hear them. There is really no other option.

In the interviews, Sound Source Localization and Regulatory Control were associated with Active Regulation. A representative excerpt is provided below.

P12: When the noise becomes relatively loud, I can open the window, and it feels like the sound around me becomes somewhat quieter.

In the interviews, Physiological Effects at Work and Regulatory Control were associated with Fatigue Accumulation after Work. Representative excerpts include.

P26: After leaving the workshop, there is still a buzzing sound in my ears, and sometimes I even feel ear pain. This mainly occurs shortly after finishing work.

P14: At night, my mind is filled with the sounds of the workshop, and as a result, I do not sleep very well.

3.1.2. Construction of the Conceptual Framework

Through an analysis of the relationships among the identified elements, a conceptual framework model was developed, as illustrated in Figure 4. According to the model, Sound Source Localization, Physiological Effects at Work, Communication Barrier, and Regulatory Control constitute the fundamental factors influencing the Healthy Acoustic Environments. In contrast, Negative Emotions, Passively Accept, Active Regulation, and Fatigue Accumulation after Work represent the behavioral measures through which this result is shaped. The specific influence pathways are described as follows.

Figure 4. Conceptual Framework Model of Healthy Acoustic Environments.

1. Sound Source Localization—Negative Emotions—Healthy Acoustic Environments: Loud or sudden noises cause discomfort, leading workers to experience irritability and other negative emotions, which ultimately affect their work performance.

2. Sound Source Localization—Passively Accept—Healthy Acoustic Environments: Workers continue to operate in the existing noisy environment without any shielding measures and are forced to remain exposed, which intensifies auditory fatigue and irritability, thereby reducing environmental comfort.

3. Sound Source Localization—Active Regulation—Healthy Acoustic Environments: While working in a noisy environment, workers adopt certain measures to regulate noise, thereby reducing the harmful effects of noise exposure.

4. Physiological Effects at Work—Negative Emotions—Healthy Acoustic Environments: Symptoms such as tinnitus and a sensation of ear pressure during working hours lead to decreased concentration and emotional discomfort, ultimately impairing work efficiency.

5. Physiological Effects at Work—Passively Accept—Healthy Acoustic Environments: Continuous exposure to noise during work without available mitigation or adjustment options forces workers to endure the conditions, resulting in a decline in the overall quality of the work environment.

6. Physiological Effects at Work—Fatigue Accumulation after Work—Healthy Acoustic Environments: Prolonged exposure to noise during working hours continues to affect workers’ physical condition even during non-working time, leading to a decline in overall health status.

7. Communication Barrier—Negative Emotions—Healthy Acoustic Environments: Noise-induced communication difficulties require workers to speak loudly or repeatedly, which leads to irritability or anxiety and may cause work errors or reduced efficiency.

8. Communication Barrier—Passively Accept—Healthy Acoustic Environments: Although noise causes communication difficulties, the lack of effective solutions forces workers to speak loudly, which, over prolonged periods, adversely affects physical health.

9. Communication Barrier—Fatigue Accumulation after Work—Healthy Acoustic Environments: Frequent shouting and repeated verbal communication result in fatigue, accelerate physical exhaustion, and hinder full recovery after work, thereby reducing sustained work capacity.

10. Regulatory Control—Passively Accept—Healthy Acoustic Environments: When workers are unable to control their work environment, they are compelled to endure the conditions, leading to physical fatigue.

11. Regulatory Control—Active Regulation—Healthy Acoustic Environments: Workers attempt to shield or modify noise by using measures such as wearing headphones, opening windows, or adjusting equipment, thereby reducing noise levels and alleviating physical strain.

3.2. Evaluation of the Structural Equation Model—In Response to Q2

The qualitative relationships identified through the grounded theory analysis were further examined and validated using quantitative methods. Specifically, a structural equation model was employed to investigate the relationships among the key elements.

3.2.1. Data Validation

The questionnaire was divided into nine dimensions and subjected to the Kaiser–Meyer–Olkin (KMO) measure and Bartlett’s test of sphericity. The KMO value was 0.853 (KMO > 0.8), and Bartlett’s test of sphericity was significant (p = 0.000, p < 0.005), indicating that the data were suitable for factor analysis.

In structural equation model analysis, EFA and CFA are required [67]. EFA is used to identify the underlying structural characteristics of the measurement data, whereas CFA is employed to test the hypothesized relationships between observed variables and latent variables.

The EFA results indicated that the communalities of all factors exceeded 0.4, suggesting strong associations between the measurement items and their corresponding factors. The cumulative variance explained after rotation reached 71.044%, indicating satisfactory explanatory power. CFA was subsequently conducted. The Cronbach’s α coefficients for all nine dimensions were above 0.7, indicating good reliability. All standardized factor loadings exceeded 0.6, demonstrating adequate measurement relationships between the observed and latent variables. The composite reliability (CR) values were greater than 0.7 and the average variance extracted (AVE) values exceeded 0.5, confirming good convergent validity. The maximum shared variance (MSV) and average shared variance (ASV) were both lower than the corresponding AVE values, indicating satisfactory discriminant validity. Detailed results are presented in Table 6 and Table 7.

Table 6. Exploratory Factor Analysis.

Category and Code	Rotated Factor Loading	Communality	Rotated Variance Explained (%)
Sound Source Localization			9.530
Sou1	0.807	0.711
Sou2	0.741	0.681
Sou3	0.774	0.710
Sou4	0.848	0.786
Physiological Effects at Work			9.449
Phy1	0.750	0.687
Phy2	0.807	0.713
Phy3	0.770	0.704
Phy4	0.765	0.696
Communication Barrier			9.370
Com1	0.788	0.744
Com2	0.704	0.664
Com3	0.769	0.690
Com4	0.798	0.740
Regulatory Control			7.452
Con1	0.799	0.716
Con2	0.754	0.705
Con3	0.757	0.701
Negative Emotions			7.365
Neg1	0.781	0.729
Neg2	0.827	0.771
Neg3	0.683	0.691
Passively Accept			7.226
Pas1	0.757	0.691
Pas2	0.718	0.700
Pas3	0.714	0.667
Active Regulation			6.908
Act1	0.683	0.720
Act2	0.795	0.760
Act3	0.730	0.663
Fatigue Accumulation after Work			6.879
Fat1	0.725	0.711
Fat2	0.805	0.778
Fat3	0.686	0.644
Healthy Acoustic Environments			6.865
Hea1	0.785	0.736
Hea2	0.689	0.673
Hea3	0.757	0.731

Table 7. Confirmatory Factor Analysis.

Category and Code	Cronbach α	Standardized Loading	CR	AVE	MSV	ASV
Sound Source Localization	0.847		0.849	0.584	0.182	0.111
Sou1		0.742
Sou2		0.747
Sou3		0.746
Sou4		0.818
Physiological Effects at Work	0.839		0.839	0.565	0.221	0.143
Phy1		0.728
Phy2		0.741
Phy3		0.765
Phy4		0.773
Communication Barrier	0.845		0.847	0.581	0.287	0.192
Com1		0.791
Com2		0.709
Com3		0.720
Com4		0.823
Regulatory Control	0.755		0.757	0.510	0.221	0.142
Con1		0.755
Con2		0.654
Con3		0.730
Negative Emotions	0.789		0.792	0.560	0.263	0.185
Neg1		0.753
Neg2		0.784
Neg3		0.706
Passively Accept	0.763		0.763	0.518	0.414	0.228
Pas1		0.694
Pas2		0.719
Pas3		0.747
Active Regulation	0.778		0.783	0.549	0.414	0.239
Act1		0.839
Act2		0.738
Act3		0.631
Fatigue Accumulation after Work	0.762		0.763	0.518	0.336	0.227
Fat1		0.694
Fat2		0.727
Fat3		0.738
Healthy Acoustic Environments	0.787		0.788	0.553	0.340	0.221
Hea1		0.754
Hea2		0.741
Hea3		0.736

3.2.2. Structural Equation Model Testing

After confirming that the above data met the required criteria, the structural equation model was tested. All model fit indices satisfied the recommended evaluation criteria, indicating a good level of model fit. Following linear regression analysis, the variance inflation factor (VIF) values for all variables were below 5, indicating no significant multicollinearity. The R² value for Healthy Acoustic Environments was 0.421. Table 8 presents the model fit indices and their recommended thresholds, all of which fall within the acceptable ranges.

Table 8. Model Fit Indices.

Index Name	χ2/df	RMSEA	SRMR	PGFI	PNFI
Recommended Criteria	<3	<0.10	<0.1	>0.5	>0.5
Model Value	2.042	0.064	0.082	0.687	0.696

Based on the data analysis, the structural equation model was established, as illustrated in Figure 5, and the regression coefficients of each path in the model are reported in Table 9. According to the results of the structural equation model and the corresponding tables, 12 out of the 15 hypothesized paths were statistically significant (p < 0.05), indicating that most of the qualitative relationships derived from grounded theory were supported by quantitative evidence. Specifically, Sound Source Localization showed a significant positive relationship with Active Regulation (β = 0.374, p < 0.01), whereas its relationships with Negative Emotions and Passively Accept were not statistically significant (p > 0.05). This may be attributed to the fact that workers are more sensitive to controllable sound sources, while tending to ignore those whose intensity cannot be regulated. Physiological Effects at Work exhibited significant positive relationships with Negative Emotions (β = 0.317, p < 0.001) and Fatigue Accumulation after Work (β = 0.244, p = 0.003), while its effect on Passively Accept was not significant (p > 0.05). This may be attributed to the accumulation of negative emotions during the work process, which may persist beyond the end of the work shift. Communication Barrier showed significant positive relationships with Negative Emotions (β = 0.227, p = 0.008), Passively Accept (β = 0.278, p = 0.003), and Fatigue Accumulation after Work (β = 0.401, p < 0.001). This may be attributed to the fact that impaired communication tends to induce negative emotions, which are difficult to resolve and may persist beyond working hours. Regulatory Control was significantly and positively associated with Passively Accept (β = 0.301, p = 0.002) and Active Regulation (β = 0.356, p < 0.001). Negative Emotions, Passively Accept, Active Regulation, and Fatigue Accumulation after Work all showed significant positive effects on Healthy Acoustic Environments (p < 0.05).

Figure 5. Structural Equation Model.

Table 9. Quantitative Relationships between Elements in Structural Equation Modeling Paths.

Path	Unstandardized Regression Coefficient	S.E.	p	β
Sound Source Localization → Negative Emotions	0.076	0.082	0.355	0.076
Sound Source Localization → Passively Accept	0.101	0.075	0.177	0.111
Sound Source Localization → Active Regulation	0.454	0.089	0.000	0.374
Physiological Effects at Work → Negative Emotions	0.323	0.086	0.000	0.317
Physiological Effects at Work → Passively Accept	0.075	0.083	0.367	0.082
Physiological Effects at Work → Fatigue Accumulation after Work	0.245	0.083	0.003	0.240
Communication Barrier → Negative Emotions	0.196	0.074	0.008	0.227
Communication Barrier → Passively Accept	0.216	0.072	0.003	0.278
Communication Barrier → Fatigue Accumulation after Work	0.345	0.074	0.000	0.401
Regulatory Control → Passively Accept	0.275	0.091	0.002	0.301
Regulatory Control → Active Regulation	0.436	0.094	0.000	0.356
Negative Emotions → Healthy Acoustic Environments	0.220	0.079	0.005	0.211
Passively Accept → Healthy Acoustic Environments	0.184	0.092	0.045	0.159
Active Regulation → Healthy Acoustic Environments	0.268	0.066	0.000	0.311
Fatigue Accumulation after Work → Healthy Acoustic Environments	0.303	0.085	0.000	0.290

The effects of different elements on the Healthy Acoustic Environments were evaluated, and the corresponding effect sizes are shown in Figure 6. Among these factors, Sound Source Localization exhibited the smallest effect size (0.116), whereas Communication Barrier showed the largest effect size (0.208). Physiological Effects at Work and Regulatory Control yielded effect sizes of 0.137 and 0.159, respectively. These results suggest that although the influencing factors demonstrated statistically significant relationships to varying extents, their effect sizes were not entirely equivalent. Specifically, Communication Barrier showed a relatively stronger overall effect than the other factors, whereas Sound Source Localization, Physiological Effects at Work, and Regulatory Control exhibited comparatively weaker overall effects.

Figure 6. Comprehensive Effects of Various Elements on Healthy Acoustic Environments.

3.3. Analysis of Subjective–Objective Correlations—In Response to Q3

3.3.1. Spearman Correlation Heatmap Analysis

Based on Spearman correlation analysis, this study found that among the 72 relationships formed between nine subjective dimensions and eight objective acoustic indicators, 10 pairs were statistically significant, accounting for 13.9%. This result indicates that although there is a certain degree of association between objective acoustic indicators and workers’ subjective perceptions, the overall correlations remain relatively weak and are insufficient to independently explain workers’ multidimensional subjective experiences of the acoustic environment in industrial buildings. As shown in Figure 7 and Table 10, specifically:

(1) Roughness shows no significant correlation with any of the nine subjective indicators, indicating that it is not associated with workers’ subjective perceptions.

(2) SOU is significantly correlated with LAeq, LAmin, LCeq-LAeq, Loudness, and Sharpness, suggesting that the physical characteristics of sound directly influence workers’ recognition and evaluation of it.

(3) Fluctuation Strength shows significant correlations with NEG, ACT, and HEA, indicating that more pronounced Fluctuation Strength is associated with stronger negative emotions toward noise, a greater tendency for workers to actively seek ways to regulate it, and a more noticeable perceived impact on health.

(4) COM is significantly negatively correlated with LCeq-LAeq, indicating that more severe low-frequency noise pollution leads to greater difficulty in communication among workers.

(5) PAS shows a significant positive correlation with LAmax, suggesting that as noise becomes louder, workers are less able to control or change it and are more likely to passively accept it.

Table 10. Spearman Correlation Analysis.

	LAeq	LAmax	LAmin	LCeq-LAeq	Loudness	Sharpness	Roughness	Fluctuation Strength
SOU	0.166 **	−0.029	0.235 **	−0.153 *	0.190 **	0.133 *	−0.055	0.069
PHY	0.042	−0.103	−0.081	−0.091	0.05	−0.039	−0.006	−0.034
COM	0.066	−0.005	0.05	−0.127 *	0.058	0.062	−0.027	0.051
CON	−0.093	−0.011	−0.038	0.042	−0.09	−0.108	−0.017	0.005
NEG	0.021	0.101	−0.023	0.029	−0.007	−0.038	0.019	0.145 *
PAS	0.028	0.130 *	0.106	0.051	0.008	−0.046	0.061	0.112
ACT	−0.017	0.071	−0.011	0.028	−0.053	0.002	−0.046	0.130 *
FAT	0.038	−0.038	0.081	0.018	0.041	0.078	−0.071	0.006
HEA	−0.029	0.005	0.037	0.12	−0.047	−0.025	−0.006	0.170 **

Note. * p < 0.05 ** p < 0.01.

Table 10. Spearman Correlation Analysis.

	LAeq	LAmax	LAmin	LCeq-LAeq	Loudness	Sharpness	Roughness	Fluctuation Strength
SOU	0.166 **	−0.029	0.235 **	−0.153 *	0.190 **	0.133 *	−0.055	0.069
PHY	0.042	−0.103	−0.081	−0.091	0.05	−0.039	−0.006	−0.034
COM	0.066	−0.005	0.05	−0.127 *	0.058	0.062	−0.027	0.051
CON	−0.093	−0.011	−0.038	0.042	−0.09	−0.108	−0.017	0.005
NEG	0.021	0.101	−0.023	0.029	−0.007	−0.038	0.019	0.145 *
PAS	0.028	0.130 *	0.106	0.051	0.008	−0.046	0.061	0.112
ACT	−0.017	0.071	−0.011	0.028	−0.053	0.002	−0.046	0.130 *
FAT	0.038	−0.038	0.081	0.018	0.041	0.078	−0.071	0.006
HEA	−0.029	0.005	0.037	0.12	−0.047	−0.025	−0.006	0.170 **

Note. * p < 0.05 ** p < 0.01.

Figure 7. Spearman Correlation Heatmap.

Figure 7. Spearman Correlation Heatmap.

3.3.2. Multidimensional Scaling Analysis

Multidimensional scaling analysis was conducted on the nine subjective dimensions and eight physical indicators using SPSS 31. The resulting Stress I value was 0.021, which is below the threshold of 0.05, indicating a good fit between the model and the observed data. Accordingly, a multidimensional scaling model was established, as shown in Figure 8, from which the following observations can be derived:

(1) The physical indicators of sound were primarily distributed in the first quadrant, whereas the psychological indicators of sound were mainly located in the third quadrant.

(2) LAmin and LAeq were positioned relatively close to the subjective indicators, suggesting that the minimum and equivalent sound pressure levels have relatively stronger associations with subjective perceptions.

(3) Fluctuation Strength exhibited a pronounced dimensional separation from other subjective indicators, indicating a substantial discrepancy between its perceived characteristics and subjective perceptions.

(4) No significant clustering was observed between objective physical indicators and any subjective perception dimensions, indicating generally weak associations between them.

(5) The objective physical indicators were distributed across extreme values on both dimensions, showing substantial dispersion and indicating large discrepancies between objective indicators and subjective perceptions.

Overall, although two physical indicators were located relatively close to certain psychological indicators, no clear clustering pattern was observed. The remaining objective physical indicators were more widely dispersed, suggesting that objective physical indicators do not form stable correlations with subjective indicators across the two dimensions.

Figure 8. Multidimensional Scaling Model.

Figure 8. Multidimensional Scaling Model.

4. Discussion

4.1. Influencing Factors of the Acoustic Environment and the Conceptual Framework

Regarding Q1:

Through the three-stage coding process of grounded theory, this study developed a conceptual framework for a healthy acoustic environment in industrial buildings. The framework consists of nine core elements, which can be organized into three hierarchical levels: Influencing Factors, Behavioral Measures, and Result. Within this framework, four Influencing Factors activate four distinct Behavioral Measures, which collectively shape workers’ overall evaluation of the Healthy Acoustic Environments.

Regarding the influencing factors, four dimensions were identified: Sound Source Localization, Physiological Effects at Work, Communication Barrier, and Regulatory Control. This classification differs from the characteristics proposed by Chen et al. [29] for residential buildings. Further comparison with the characteristics of healthy indoor acoustic environments and the indoor emotional response model [68] revealed that the Regulatory Control dimension is conceptually similar to the sense of control in that model, whereas the other dimensions are not represented. Previous studies have proposed an overall conceptual framework for healthy acoustic environments applicable across multiple contexts [69]. However, this framework does not include dimensions such as Communication Barrier or Physiological Effects at Work. This further suggests that the assessment of healthy acoustic environments in industrial buildings should not simply rely on acoustic environment frameworks derived from residential or general indoor settings, but should instead be interpreted in the context of specific workplace conditions, including continuous mechanical noise, task-related collaboration, and operational constraints.

Although most dimensions showed inconsistencies across studies, the dimension of “control” has been consistently identified in multiple studies [29,68]. Controllability has been defined as a key mechanism explaining the negative health effects of noise exposure [70,71]. In the present study, Regulatory Control was associated with Active Regulation and Passively Accept, and subsequent analyses revealed significant paths, indicating that the Regulatory Control dimension is present in industrial buildings. In contrast to Chen et al. [29], who suggested that “controllability is a distinctive attribute of indoor acoustic environments in residential buildings,” control in industrial settings is primarily reflected in workers’ ability to make limited adjustments to noise within existing production constraints, rather than in the pursuit of quietness or comfort alone.

4.2. Effects of Sound Source Localization and Communication Barrier

Regarding Q2:

The qualitative findings indicated that Sound Source Localization was associated with Negative Emotions, Active Regulation, and Passive Acceptance. However, the quantitative analysis showed that its relationships with Negative Emotions and Passive Acceptance were not statistically significant. By contrast, more statistically significant pathways were observed for the dimension corresponding to Communication Barrier, which also exhibited the highest overall effect size.

From a general cognitive perspective, perception of sound sources would be expected to be a core factor influencing evaluations of the acoustic environment. However, in this study, two of the three paths associated with Sound Source Localization were not significant, and the corresponding effect sizes were the smallest. This discrepancy may be attributed to several factors. First, prolonged noise exposure may reduce workers’ ability to perceive sound sources. Wang et al. [72] reported that long-term noise exposure impairs the ability to discriminate sound frequencies, thereby affecting the recognition of sound source characteristics. In the present study, approximately 70% of participants were exposed to noisy environments for more than six hours, suggesting that extended noise exposure may be one of the contributing factors. Second, the perceived controllability of sound sources may also influence sound source perception. In the structural equation model, the path between Sound Source Localization and Passively Accept was not significant, whereas the path between Sound Source Localization and Active Regulation was significant. This indicates that workers exhibit weaker perception of uncontrollable sounds but stronger perception of controllable sounds, which in turn further influences evaluations of the acoustic environment. Schreckenberg et al. [73] found that evaluations of control ability may outweigh the influence of noise level itself. In this study, both prolonged noise exposure and perceived control jointly affect sound source perception, and their combined effects may explain the non-significant role of Sound Source Localization observed in the quantitative analysis.

The Communication Barrier dimension exhibited the largest number of significant paths and the highest effect sizes, indicating that communication Barrier plays a crucial role in the evaluation of acoustic environments in industrial buildings. In industrial workshops, noise exposure reduces the efficiency of workers’ communication, which in turn leads to negative emotional responses. As workers often lack the ability to modify such conditions, they are forced to passively accept the situation. Moreover, these effects may persist beyond working hours, exerting a prolonged impact on workers and ultimately resulting in lower evaluations of the acoustic environment. Yang et al. [74] found that higher occupational noise levels were associated with increased perceived noisiness, greater hearing impairment, and increased communication difficulty. In a survey of 133 workers, 34.6% reported that noise severely interfered with on-site communication. The findings of the present study are consistent with these results. Therefore, in noise-affected industrial working environments, communication barrier may represent the most critical factor influencing evaluations of the acoustic environment.

4.3. Effects of the Relationship Between Subjective Perception and Objective Indicators

Regarding Q3:

This study found that there are some significant associations between workers’ subjective perceptions of the acoustic environment and objective acoustic indicators in industrial buildings; however, the overall strength of these correlations remains limited. This result indicates that objective acoustic indicators can, to some extent, reflect workers’ subjective responses to certain characteristics of the acoustic environment, but they are still insufficient on their own to fully explain workers’ overall evaluation of a healthy acoustic environment in industrial buildings.

This finding is broadly consistent with previous studies showing that the relationship between subjective perception and objective measures is not a simple one-to-one correspondence. Giménez et al. [28], in their study of concert halls, pointed out that a mismatch may exist between objective acoustic parameters and subjective perceptual evaluations, indicating that although physical indicators can describe acoustic characteristics, they do not necessarily correspond directly to users’ actual experiences. Studies by Masullo et al. [61] and Chen et al. [75] also showed that different acoustic characteristics, particularly fluctuation, spectral features, and low-frequency noise, may exert different effects on individuals’ physiological responses and subjective annoyance.

The results of this study indicate that SOU is significantly correlated with most physical indicators, suggesting that these objective measures can reflect workers’ overall perception of the acoustic environment. Fluctuation Strength shows significant correlations with certain dimensions, indicating that sound fluctuation characteristics are more likely to trigger workers’ emotional responses and influence their overall evaluations. However, most objective indicators do not exhibit stable and significant relationships with most subjective dimensions, suggesting that workers’ subjective perceptions are influenced not only by sound intensity itself, but also by long-term exposure, the surrounding environment, and individual psychological adaptation.

In addition, noise in industrial buildings mainly originates from the continuous operation of equipment, and workers who are exposed to such environments over long periods tend to perceive it as an integral part of the workplace context. Axelsson et al. [76] and Qin et al. [77] both pointed out that when sounds are consistent with the environmental context, individuals are more likely to accept them rather than perceive them as disturbances. Schreckenberg et al. [73] further demonstrated that perceived control may have a greater influence on individuals’ annoyance evaluations than the noise level itself. Moreover, during a single experimental session, factors such as the occurrence of sudden noises or interpersonal communication may also influence workers’ evaluations of the acoustic environment.

Therefore, this study does not negate the role of traditional physical indicators in acoustic environment research; rather, it further demonstrates that the relatively weak correspondence between subjective and objective measures has important implications for the evaluation of acoustic environments in industrial buildings. If assessments rely solely on objective acoustic parameters such as LAeq and Loudness to determine whether the acoustic environment meets requirements, they may fail to fully capture workers’ actual conditions in terms of communication, emotional responses, and health status. In other words, objective indicators can reflect certain physical attributes of the acoustic environment, but they are insufficient on their own to support a comprehensive evaluation of a healthy acoustic environment in industrial buildings. Furthermore, this weak correspondence between subjective and objective measures may also be related to differences in temporal scale and levels of measurement between the two approaches: objective indicators primarily reflect acoustic characteristics within a single measurement period, whereas subjective evaluations represent workers’ overall assessment of their entire work shift and include more complex perceptual components such as accumulated fatigue; therefore, a simple one-to-one correspondence between the two is difficult to establish. Therefore, in evaluating healthy acoustic environments in industrial buildings, subjective assessments should be integrated with traditional objective measurements to establish a more explanatory comprehensive evaluation framework and to provide more targeted support for subsequent acoustic optimization and design decisions. Future research could also incorporate physiological indicators and related methods to further enhance the correspondence between subjective and objective data.

4.4. Implications for Industrial Building Design

The nine key evaluation factors identified in this study constitute a multidimensional framework for interpreting workers’ perceptions of healthy acoustic environments in industrial buildings. This framework enables designers and managers to move beyond conventional assessments based solely on sound pressure levels by incorporating acoustic characteristics, communication factors, and workers’ psychological responses to the noise environment. In practice, these dimensions provide a valuable basis for the optimization of acoustic environments in industrial buildings.

The findings suggest that, in the optimization of acoustic environments in industrial buildings, greater attention should be given to workers’ perceptions of controllable and identifiable sound sources. Compared with fully enclosed and fixed control approaches, more flexible acoustic regulation strategies may better enhance workers’ sense of environmental control. For instance, more adaptable acoustic interventions could be explored, such as providing highly sound-insulated rest areas, thereby further improving workers’ capacity to manage noise.

Furthermore, Communication Barrier represents a critical dimension that should not be neglected in the acoustic design of industrial buildings. In high-noise workshop environments, limited communication can not only reduce collaborative efficiency but also increase workers’ psychological burden. Therefore, spatial organization should prioritize placing communication-intensive tasks in relatively low-noise areas or providing designated quiet spaces to facilitate more efficient and clearer workplace communication.

Finally, the evaluation of acoustic environments in industrial buildings should not rely solely on single physical acoustic indicators. Given the potential discrepancies between objective measurements and workers’ subjective perceptions, integrating subjective assessment methods—such as questionnaires and interviews—with traditional physical indicators enables a more comprehensive understanding of the actual impacts of the acoustic environment on workers and provides a stronger basis for developing more targeted optimization strategies.

5. Limitations

However, this study still has certain limitations. First, the selected industrial buildings are all coil material workshops, and the research sample is mainly concentrated in the Beijing area. The relatively uniform building type and regional distribution limit the generalizability of the study’s conclusions. Additionally, the environments in which the samples are located are generally constrained by occupational health regulations, with sound pressure levels mainly concentrated in the 70–80 dBA range, lacking data support for higher noise levels (>80 dBA). This somewhat limits the extrapolation of the study’s conclusions to high-intensity industrial sound environments. Second, this study mainly focuses on subjective perceptions on-site, and there is still insufficient exploration of the fatigue accumulation and sustained negative effects caused by long-term exposure to noisy environments. Third, although this study introduces objective physical acoustic indicators, it has not yet integrated physiological indicators for a comprehensive analysis. Therefore, research on the physiological aspects of workers remains inadequate. Fourth, the mechanisms of sound environment perception may vary at different noise levels. This study primarily focuses on “sub-threshold” environments, and there is a lack of systematic exploration regarding changes in perceptual mechanisms under higher noise conditions.

Based on the limitations mentioned above, future research can be conducted in the following areas. First, expand the sample range and scenario types by including more types of industrial buildings and samples from different regions, and covering a broader range of sound pressure levels, in order to further test the applicability and robustness of the conclusions of this study. Second, combine long-term follow-up studies to systematically assess the cumulative effects of noise exposure and its sustained impact on workers’ health-related evaluations of the working environment. Third, introduce physiological indicators in future studies and combine subjective perceptions with objective acoustic measures to construct a more comprehensive industrial building acoustic environment evaluation model from multiple dimensions.

6. Conclusions

This study adopted a mixed-methods approach to investigate workers’ evaluations of healthy acoustic environments in industrial buildings. Based on the three research questions proposed in the Introduction, the following conclusions were drawn:

1. In response to Q1, this study identified nine key evaluation factors for healthy acoustic environments in industrial buildings and, based on these, developed a multidimensional perceptual evaluation framework. This framework does not replace traditional physical noise evaluation; rather, by incorporating workers’ subjective experiences, it complements existing approaches and provides a more comprehensive basis for acoustic environment assessment, architectural design optimization, and workplace environment improvement in industrial buildings. In addition, this framework can provide theoretical support for the future development of simplified evaluation tools for practical applications.

2. In response to Q2, both Communication Barrier and Regulatory Control exert significant effects on workers’ evaluation of the Healthy Acoustic Environments. Among these, Communication Barrier demonstrates the largest effect size, significantly reducing acoustic environment evaluations through Fatigue Accumulation after Work and Negative Emotions. In contrast, Regulatory Control alters evaluations of the acoustic environment through Passively Accept and Active Regulation. Although the structural paths of Sound Source Localization and Physiological Effects at Work are statistically significant, their effect sizes are relatively small, and several pathways within Sound Source Localization are not significant.

3. In response to Q3, Spearman correlation analysis revealed that only 13.9% of the indicators showed statistically significant correlations. This result indicates that although there is a certain degree of association between objective acoustic indicators and workers’ subjective perceptions, the overall correlation remains weak and is insufficient to independently explain workers’ multidimensional subjective experiences of the acoustic environment in industrial buildings. Therefore, in the evaluation and design of acoustic environments in industrial buildings, workers’ subjective perceptions should be further integrated with physical indicators to establish a comprehensive evaluation approach that combines both objective and subjective perspectives.

Overall, this study identified key evaluation factors for healthy acoustic environments in industrial buildings and incorporated workers’ subjective experiences into the evaluation framework based on real workplace contexts. The findings highlight the importance of communication support, environmental control, emotional responses, and the interaction between subjective and objective factors in evaluating healthy acoustic environments. Accordingly, future evaluation and optimization should not rely solely on single physical indicators but should adopt an integrated approach combining objective and subjective perspectives in evaluation, design, and management practices, with particular attention to communication support, environmental control, and workers’ experiences within specific work contexts.