Comparative analysis of spatial experience in physical and virtual environments: The case of anadolu hisarı

Abstract

Spatial experience constitutes the aggregate of sensory, cognitive, and physical interactions an individual establishes with their environment. The human mind perceives, interprets, and mentally reconstructs space not solely through visual inputs but through the simultaneous operation of multisensory channels, including touch, hearing, smell, and bodily movement. Physical spatial experience is a tangible process wherein the user is directly embodied within the environment, simultaneously perceiving surface textures, climatic conditions, olfactory cues, and acoustic properties. Virtual Reality (VR), conversely, is a simulation technology that isolates the user from the physical environment, immersing them in a digitally generated setting. While its three-dimensional visualization and immersive capabilities allow for the experience of diverse spaces, this experience relies predominantly on the mimicry of visual and auditory stimuli. Despite technological advancements, the extent to which virtual environments can replicate physical spatial experience remains a subject of debate. In this context, the representational power of virtual spaces and how they shape the 'sense of presence' constitute a significant area of research. Furthermore, VR holds the potential to transcend physical limitations, offering accessible and alternative experiential possibilities. A review of the existing literature reveals that VR research is heavily concentrated on technical performance, healthcare rehabilitation applications, and architectural education. The majority of these studies are conducted in controlled laboratory settings; research addressing the sense of presence, sense of place, and exploratory behaviors in spaces experienced for social purposes remains limited. Moreover, studies predominantly focus on enclosed spaces, leaving semi-open spaces, which combine interior and exterior characteristics, insufficiently examined. To address this gap, the Anadolu Hisarı Museum, a semi-open cultural heritage site used for socialization, was selected as the study area. Accordingly, a mixed-method group experiment was conducted to examine the perceptual and sensory differences in spatial experience between physical and virtual environments. Adopting a multi-dimensional attitudinal framework, the research design deconstructs spatial experience into its cognitive, emotional, and behavioral components to enable a holistic and structural comparison. Participants were divided into a physical experience group (n = 9) and a virtual experience group (n = 8); all subjects were selected from first-year architecture students. While the physical group experienced Anadolu Hisarı on-site, the virtual group experienced a 1:1 scale model of the space, created using Blender and Unity, via a Meta Quest 3 headset. The process was documented using video recordings for the physical group and screen recordings for the virtual group. Post-experience data collection included semi-structured interviews, cognitive map drawings, and the simulator sickness questionnaire (SSQ) and igroup presence questionnaire (IPQ) for the virtual group. Interview transcripts were subjected to sentiment analysis using AI-assisted Natural Language Processing (NLP) techniques, and the resulting data were evaluated using both qualitative and quantitative methods. Results indicated no statistically significant difference in general sentiment levels between the two groups; however, the components constituting the experience diverged significantly. The physical experience facilitated a strong connection with the site's historical and architectural identity through tactile, olfactory, and environmental interactions. In contrast, the virtual experience was defined primarily by visual representation, immersion, technological novelty, and a gamified sensation. Notably, while undefined circulation paths in the physical environment evoked anxiety, the same ambiguity in the virtual environment was perceived as a sense of freedom and exploration. Despite this, the virtual experience provided a cognitive structure for grasping spatial organization comparable to that of the physical experience. In both groups, the view emerged as the strongest common experiential component. Behavioral and duration analyses revealed that participants in both environments spent the most time in the terrace area. In the physical experience, the average time spent was 125.11 seconds on the terrace and 95.33 seconds on the ground floor; in the virtual experience, these durations were 74.25 seconds and 44.75 seconds, respectively. While frequent pauses to examine surfaces and historical details were observed in the physical environment, a faster and more linear circulation pattern was detected in the virtual environment. Participants in the virtual space tended to move towards the initial visual focal point perceived upon entry, indicating that the experience was shaped by a specific directional orientation. An examination of cognitive maps revealed that while the physical experience resulted in richer and texture-laden mental representations, the virtual experience led to more schematic and geometry-based recalls. Despite this disparity in detail, both groups successfully reconstructed the general spatial organization and circulation paths with comparable accuracy. This finding is corroborated by the IPQ results, which indicate that virtual participants developed a strong sense of presence by accepting the space as a valid navigable volume, yet rated its realism significantly lower due to the lack of sensory fidelity. Finally, SSQ results indicated that discomfort among virtual participants was primarily related to nausea and discomfort, with experienced VR users reporting significantly lower levels of physical distress. In conclusion, this study demonstrates that the virtual experience cannot fully present the physical space in terms of fostering a complete sense of presence, particularly regarding emotional and sensory engagement. On the other hand, the results reveal that there is no significant difference between the two mediums in terms of cognitive and spatial learning. It must be acknowledged that the limited sample size and specific participant profile may affect the scope of these findings. It is anticipated that future research, designed to include larger and more diverse participant groups, will enhance the universal validity of the findings and contribute more comprehensive, generalizable data to the literature.