Abstract
As projects become more globally dispersed, site visits and analysis become challenging, often leading to the use of secondary information (e.g., photos, plans, and videos). Immersive technology offers embodied, visual, and spatial perspectives, providing unique information about a site that could be beneficial. Our research examines how virtual environments (VEs) can help landscape architects understand a site by exploring immersive technology for a remote site visit in a joint landscape and architecture studio. Students explored an informal settlement (favela) in Brazil first using a VE through three separate technologies:HTCVive,MobileVR, and WebVR, and then in person. Students’ responses helped identify perceptions toward technology and future improvements to the VE. Therewere four key findings. (1) VE establishes familiarity with a site; (2) VE is used for checking details; (3) walking is desired over realism; and (4) control of the VE experience is enjoyable. The findings suggest that VE cannot replace an in-person experience but provides familiarity when used alongside common secondary materials. Future research is needed to discern what VE features generate site familiarity.
INTRODUCTION
Landscape architecture practice increasingly involves globally distant locations, requiring travel to experience a project site, which is not always feasible (because of cost and safety concerns, for example). Site visits elicit a site’s characteristics (e.g., scale, color, light, and areas where people congregate) through direct experience and confirmation of information to guide intervention. Before a site visit, landscape architects often review secondary information (such as photographs, satellite imagery, or GIS data). During a site visit, they record primary information using various media (e.g., photos, sketches, and video) to capture characteristics of a site. When site visits are not possible, designers rely only on secondary information, which can be limited in the amount available, viewpoints offered (George, 2016), and surrounding context. Immersive technology can supplement traditional secondary information through embodied, visual, and spatial means by enabling multiple virtual site visits and different viewpoints.
Immersive technologies, including virtual reality (VR) and augmented and mixed reality systems, engage a user’s perceptual senses (Suh & Prophet, 2018) to convey information. As immersive technologies become more common, easier to use, and more affordable, landscape architects might offset an inability to visit a site by using these technologies, just as they have explored using such technology for design (Freitas & Ruschel, 2013). We refer to immersive technology by its ability to provide a virtual environment (VE).
New forms of VE substantially immerse a user in virtual content through sound, sight, and spatial dimensions. VEs can communicate important landscape characteristics (such as landform, vegetation, human elements like buildings and walkways) (Orland, Budthimedhee, & Uusitalo, 2001) and support landscape perception (Bishop, Ye, & Karadaglis, 2001). For instance, George (2016) found that aerial 360°videos can be used to analyze a site and later found 360°imagery helps spatial encoding of a site (George, 2018). These works suggest that immersive technology enables an understanding of a site by providing embodied, visual, and spatial perspectives. To address how VE can help landscape architects understand a site, we selected three common VEs to investigate students’ perceptions of a site. Our selection focused on not limiting any potential benefit of a VE, such as embodied perspectives from an HTC Vive, and ensuring that all students would have access to a VE (at the time of the study, there was only one HTC Vive available). We piloted the VE in a semester-long joint landscape and architecture studio studying the Santa Marta favela in Rio de Janeiro, Brazil (Figure 1). The Santa Marta favela presented unique visual characteristics, a spatially complex remote location, had a steep topography, and safe access concerns.
Santa Marta favela, Rio de Janeiro, Brazil. Photo courtesy of Debora Verniz. Aerial image 2020 Maxar Technologies, Google Earth.
In this study, we focus on exploring how VEs can help landscape architects understand a site, particularly a remote one, through student perceptions of site feature recognition to help their design interventions. First, we present how the fields of landscape architecture and architecture define characterizing a site and the site visit and analysis process. We discuss VE in landscape architecture and its use for conveying spatial information. Then we present our case study with student responses to using three different VEs as a remote site experience. We conclude with a discussion of the findings and implications for further research.
LITERATURE REVIEW
Contemporary landscape architectural and architectural discourse often stresses the importance of place for designing and consequently states that site visiting is a fundamental task in the initial design stages. As such, there is a significant body of literature addressing how sites are perceived, which site features are relevant to construct a notion of place, and how the process of apprehending such features unfolds. This section provides a brief overview of site analysis and the application of VEs in design disciplines to convey spatial information.
Characterizing the Site for Intervention
Landscape architects and architects share a common goal of creating a sense of place through design intervention. One of the first steps of the design process is to characterize the intervention site, which depends on an initial understanding of the location (Pressman, 2012). Landscape architects gather enough information about a site to understand and characterize it by identifying the analytical and perceptual components obtained by the sensory experience from a site visit. Analytical components include the “visual and external appearance” of a site (Jiven & Larkham, 2003: 69)—the more tangible aspects. Perceptual components are more subjective and rely on experience and observation to garner meaning. Landscape architects process these forms of information through the design process to actively test ideas (Lyle, 1985; Pressman, 2012). This process occurs when a landscape architect searches for scenarios for intervening, finds alternatives for intervening (proposes—subjective), evaluates these alternatives against the project needs (disposes— objective), and then collects, reviews, and reflects on both types of components (data) of a site (Lyle, 1985; Williams, 2014).
Analytical information comes from an increasingly technical perspective, where information comes from different types of in-person site experiences (such as GIS, demographics of the area, imagery, transit analysis, satellite imagery of impervious surfaces, vegetation, and water basins) (Makstutis, 2010; Baudoin, 2016). Since the advent of digital media in design, GIS provides much of this information before even going to a site to enhance the analysis of 2D representations and 3D visualizations (Lange & Schmid, 2000). Such information provides details that help reduce the amount of information needed when first visiting a site. But not all information can be gathered analytically.
Perceptual or subjective components relate to experiencing and observing a site, the visceral response to processing one’s senses. The senses provide five main exteroceptive (external) senses as categories of sight, sound, smell, taste, and touch (Damann, Voets, & Nilius, 2008). Humans also have interoceptive (internal) senses that provide information about the body (Damasio, 2003). Landscape architects focus on a fundamental understanding of how the human body reacts to different objects external to itself (Franck & Lepori, 2007). The most relevant interoceptive sense is proprioception, the sense of position and orientation of one’s body (Slater, Usoh, & Steed, 1995), which aids in perception of spatial legibility (Koseoglu & Onder, 2011). These senses do not hold equal roles in understanding a site, often varying between different designers and sites.
The Site Visit and Analysis
Being able to propose interventions begins with becoming familiar with a new place through the site visit and analysis (Leach, 2002). Depending on the discipline (architecture or landscape architecture), this process is referred to as the site visit or site analysis. For clarity, we use the term site experience to refer to both. The site experience consists of two approaches. First, is a subjective approach, where the designer is a sightseer experiencing the place (Jakle, 1987) gathering perceptual information. Then there is the objective approach, with purposeful deconstruction of the site (Antrop, 2000) to comprehend it (Beauregard, 2005) using analytical information.
The approaches to the site experience serve different purposes. The first is to familiarize the designer with the site, referred to as landing, the “first act of site acknowledgment” (Girot, 1999: 61) at the start of a project where a designer simply walks through a site (Pressman, 2012). After the initial experience, a designer will begin deconstructing the site, the process of grounding, which is “more about reading and understanding a site through repeated visits and studies” (Girot, 1999: 62).
Once conceptualization has taken place, the site experience remains relevant through the process of finding, which “entails the act and process of searching as well as the outcome, the thing discovered” (Girot, 1999: 63). After a design is put together, a designer will revisit it in its context, founding, which is “always a reaction to something that was already there,” (Girot, 1999: 64). Founding is about revisiting the site after a design is complete to double-check decisions or convince clients (Milburn & Brown, 2003). The site experience can play an integral role throughout the design process to inform the design and verify, validate, and communicate decisions (Figure 2).
Aligning the site experience stages from Girot (1999) to the design process.
Familiarity derives from engagement with the site, which requires data gathering and analyzing biophysical information on existing conditions through multiple engagements over time. This physical (in-person) survey is combined with cultural analysis to include many visual, spatial, and cultural details. The survey is often conducted through media documenting (including photographs, videos, painting, and sketching) and is frequently used in combination with interviews with locals. This documentation becomes the tool for unraveling the meaning or dialogue with a site (Pressman, 2012; Baudoin, 2016). For clarity, we refer to the physical survey as an in-person site experience.
VE in the Design Disciplines
VE represents changes in how people understand the world through visual communication or “visiting” or “going to” locations (Portman, Natapov, & Fisher-Gewirtzman, 2015). This technology offers an audiovisual representation of physical environments (examples can be found in Bishop, Ye, & Karadaglis, 2001; George, Sleipness, & Quebbeman, 2017; Lombardo, 2018) which supports the designers’ needs to experience the place to understand it. VE can offer different types of visual representations of a site and for this article, we are focusing on analog models (Mitchell, 1975) that offer a picturesque representation (Clarke, 2005) of a site because the favela is visually complex and hard to interpret in other forms of representation.
The strength of VE lies in its ability to engage users visually, spatially (Hill et al., 2019), and increasingly auditory means (Echevarria Sanchez et al., 2017) and communicate information of varying complexity (Parke, 2005). By engaging users, information is not limited to a single format or period, thus offering the ability to represent spatial and temporal future projections (Lange, 2001). Although early VE had several limitations due to cost, usability, and technical issues, VE has steadily improved (Suh & Prophet, 2018) and has become more widely available.
Regarding current practices, the move from drawing by hand to generating digital 3D models has been explored as a means to connect VE to design, producing comparable results in spatial awareness as if going to a real site (Hill, George, & Johnson, 2019). VE has important attributes (such as illustrative, immersive, interactive, intuitive, and intensive) that make it a successful tool for designers (Orland, Budthimedhee, & Uusitalo, 2001). Challenges in this context relate to the time needed to create the VE (Lombardo, 2018) and the need for specific skills (Ghadirian & Bishop, 2008). To minimize the effects of these challenges, visual fidelity is often sacrificed. Visual fidelity (realism) relates to how well a VE aligns with reality (how authentic the visual experience is) and is highly dependent on the context of the content displayed (Gilbert, 2016). In architecture, realism is valued for conveying a project’s spatial qualities for interpretation of a proposed design (Drettakis et al., 2007).
There are many ways to capture site data for immersive purposes, such as 3D models, hypertext, multimedia, GIS, imagery (from satellite, digital cameras, 360°cameras, and so on), and laser scanning (Orland, Budthimedhee, & Uusitalo, 2001). The strength of VE lies in the ability to integrate and overlay these types of information to provide a context in many formats through immersive technology. Although realism is often lost in 3D environments in favor of interactivity, other content formats (such as 360°images) offer increased realism but lower interactivity. This trade-off comes at the cost of development time, where 360°images are faster to input into a VE than fully rendered 3D models (Lombardo, 2018). With much of the use of VE in design focusing on 3D models, little is known about other more rapidly developed content formats, such as the 360°images.
Panoramic or 360°images in VE (virtual tours) provide a fixed point in a sphere where a person can look around (Lombardo, 2018). They are explored across several disciplines, from landscape architecture to geography, which seek to replace the need for in-person site visits to teach about a place (George, 2018; Klippel et al., 2019). Virtual tours offer a limited set of capabilities compared with 3D environments. Creating 360°imagery and panoramas is cost-effective compared with building 3D models. More advanced tours embed multimedia information (George, 2018) and can connect multiple 360°images to simulate navigation through a space created of photos. It is not known how well these different approaches to VE enable a designer to comprehend an in-person site visit from the virtual experience.
Spatial comprehension and learning through VE have produced mixed results (Bullinger et al., 2010; Balakrishnan et al., 2012). Deriving an understanding of space through VE often overlooks the nature of the technology used and the format of the content. As an example, early studies on spatial comprehension used less immersive technologies, such as desktop-based VE, including avatar-based systems with completely 3D-modeled environments (Milgram & Kishino, 1994). Later studies used highly immersive technology, such as large-scale, room-based VE systems, to determine spatial comprehension in 3D modeled environments (Balakrishnan & Sundar, 2011; Balakrishnan et al., 2012; Castronovo et al., 2013). More recently, the use of VE headsets was examined for factors of spatial awareness (George, 2016), which can be highly immersive.
An additional attribute of VE is the ability to place a user inside the virtual space, where the user’s body becomes a point of reference for spatial comprehension. This attribute, known as embodiment, occurs when users are perceptually engaged through a combination of immersive and interactive means, much like the feeling of being in a virtual space (sense of presence) and controlling a body (Balakrishnan & Kalisperis, 2009; Kilteni, Groten, & Slater, 2012). When a user feels embodied, factors about the experience such as memory and cognition are reported to improve (Balakrishnan et al., 2012). Schutte and Stilinovic (2017) explain that embodiment in VE “has the potential to increase engagement and through engagement increase characteristics such as empathy” (711). In addition, Shin (2018) suggests that “the immersion quality is established explicitly and implicitly by users, and then empathy starts to kick in” (70).
In summary, landscape architects use the site experience as a means to characterize a site through a combination of analytic and perceptual information. When an in-person site experience is not possible, there has been success in using immersive technology, and accompanying VEs, to successfully provide an understanding of a site. The caveat in the literature is the extent to which the VE was used and the complexity of the site. In the cases where VE successfully enabled landscape architects to encode spatial information, the sites were often simple and focused only on identifying specific information. This presents an opportunity to explore more complex sites, such as favelas in Brazil, from a remote location using VE.
METHOD
This research uses a qualitative holistic case study design (Yin, 2018) about a graduate and undergraduate design studio (the Rio Studio) consisting of 12 students at a U.S. university. This studio focused on favelas, which emerge as a result of population growth and uncontrolled urbanization. In the studio, students were asked to intervene in a specific favela, Santa Marta (Rio de Janeiro), using advanced digital technology to reinforce structural stability, strengthen social activity based on expert insights, create and enhance open spaces, and develop the landscape as ecological infrastructure. Ten students agreed to participate in the study, three from landscape architecture and seven from architecture. Eight participated in the follow-up interview.
The geographical distance of Santa Marta from the U.S. university, combined with safety issues (e.g., infrastructure and political tensions between the local government and favela), discouraged instructors from taking students to visit the site early in the semester. To address this problem, VE was used as an intervention to provide a remote experience alongside other secondary media (including images, videos, GIS data, a 3D physical model, faculty experts, and survey responses from the community). Specifically, a 3D physical model of the site helped students gain a basic understanding of the site’s slope and topography, complementing the GIS information.
We piloted three different VEs—WebVR, HTC Vive, and MobileVR—to provide an experience of Santa Marta. Each VE provided visual and audio content, a partial or completely immersive experience, and stereoscopic content projection or 360°head tracking. Officials in Rio provided the favela site plan and topographical information, while the researchers gathered 360°media (93 images and 34 videos) with a Ricoh Theta S camera on two occasions: September 2015 and June 2016 (respectively spring and winter in the Southern Hemisphere). As a tropical country, Brazil does not experience four distinct seasons, although differences in pluviometry are represented in the span of media collected. On both occasions, media collection took one weekday (from 9 am to 6 pm), capturing different activity periods. Access on weekends was discouraged due to safety concerns. Media were captured during both dry and rainy periods.
The Santa Marta favela can only be accessed by pedestrian routes, and it has three points where it connects to neighboring streets (referred to as entrances). The content followed major pedestrian routes from the entrance to the upper region and was mapped to the site plan and presented in each VE (Figure 3). All three VE applications used the same point-and-click functionality, where a user clicked a point on the site plan and the related 360°content opened. Each application provided a slightly different perspective of the same experience (see Oprean et al., 2018 for more description of the applications).
The immersive applications, user views, and respective details.
The VE applications were available to students throughout the semester. As students used the VE, they could report issues and improvements. The studio was offered in the spring 2017 and lasted 16 weeks, with activities divided into six main steps (see Figure 4). We align these steps to Girot’s (1999) four trace concepts to illustrate the use of the VE applications. Students used the three VEs throughout the semester and visited the favela in person twice at the end of the semester.
Timeline with course activities and data collection.
Steps 1 and 2, presentation of the case study (intervention site and design problem) and the VE applications, occurred in the first two weeks. Students practiced using each VE on their own. Step 3 corresponds to the class visit by a Brazilian faculty member who lectured on Rio and favelas and worked with the students for a week. Step 4 represents the monthly virtual meeting with all the Brazilian faculty members involved in the studio to comment on students’ intermediate work. In step 5, students developed their designs and used Lumion to generate 360°images. These images were uploaded to the MobileVR application for the final presentation. In step 6, students traveled to Santa Marta and present their proposals to the residents, visiting the favela twice.
Data collection on student perspectives occurred during the semester through four surveys, observations from the studio and trip to Santa Marta, and follow-up interviews. For a sample of the questions from each of the surveys and the interviews, see Table 1. The first survey, a two-minute open-ended survey, was filled out throughout the semester every time a VE application was used. It gathered the purpose of using each application and included an area for making suggestions for improvements. The second survey was administered twice after each visit to Santa Marta. The survey contained both a 10-point Likert-scale (where 1 was strongly disagree and 10 was strongly agree, with no neutral option) and open-ended reflective questions gathering details on the in-person site experience, interaction with residents, and perceptions of site familiarity. The final survey, at the end of the semester, was completed after returning from Brazil. The last survey contained the 10-point Likert-scale and open-ended questions to capture the general outlook toward using VE for site experiences. Participants also took part in a one-hour semi-structured interview. The interview included open-ended and 10-point Likert-scaled questions focused on gathering student perspectives toward the in-person site experience and perceived differences with the VE applications.
Select Survey and Interview Questions
RESULTS
Data from all surveys, observations, and interview transcripts were analyzed with a qualitative method supported by averages of Likert-scale data. Two independent researchers collated the data from the surveys, observation notes, and interviews for a single qualitative evaluation. The data were coded using a line-by-line method using MAXQDA. The approach examined responses with a mixed deductive and inductive thematic analysis (Fereday & Muir-Cochrane, 2006) coded in two intervals, first for deductive codes and then inductive. The deductive codes were not mutually exclusive, as lines found for the deductive codes could be coded into the inductive codes; the inductive codes were mutually exclusive. Several iterations of code refinement helped derive consensus. We cross-checked our deductive codes (kappa of 0.96) and our inductive codes (kappa of 0.90). The Likert-scaled responses were treated as interval data due to the large 10-point value (Wu & Leung, 2017) for averaging. We calculated the mean for each of the individual Likert-scale ratings to provide descriptive information to help clarify the results from the coded themes.
The data analysis is presented using four themes derived from thirteen codes (Table 2), one graph of the frequency of codes, and four graphs with averaged responses for Likert-scale data. The first theme focused on the trace concepts (refer to Girot 1999) for site experiences. The second theme focused on the need for a more authentic site experience by engaging other senses. The third theme focused on control of the site experience. The last theme grouped perceptions on the resulting familiarity of the site. These themes form the basis of our analysis.
Themes, Codes, and Simplified Code Definitions
Based on the qualitative analysis, the theme for Girot’s trace concepts consisted of four preexisting codes: landing, grounding, finding, and founding. This theme included responses toward the student designer’s traditional site experience and illustrated changes to that process from using the VEs. The second theme that emerged from the data, authenticity of site experience, consisted of four codes: analytical information, perceptual information, format of information, and interaction needs. These codes focused on the use of the VEs for the site experience and the information experienced. The third emergent theme, control over site experience, is defined by three codes: access to applications, amount of information, and technical issues. This theme focused on the student’s perceptions of the amount of control over the site experience and the VE. The last theme, familiarity of site, comprises two codes: remote site experience and success of applications. This last theme focused on how students perceived the remote site experiences helped them understand the site. Figure 5 shows the frequency of each code from all student response sources.
Frequency of codes across all participant responses.
Averaged ratings from the Likert-scale questions were calculated to provide more context for the codes and themes. Given the small sample size, the averages are all considered close and not meant for determining the statistical significance of differences. Four sets of ratings were calculated to illustrate the perceptions of the students on preference for which VE application to use, sources of information to connect with a site, the perceived role of the senses in a site experience, and confidence in understanding the site through VE.
Overall, the students preferred to use the WebVR application over the other options. The HTC Vive was rated as second-most preferred, leaving the MobileVR application as the least preferred (Figure 6). The preference for the least immersive yet more accessible application, WebVR, holds implications for how students developed a sense of familiarity of the site. For technological experiences, the students elaborated that control over which tools they use and where they use them (e.g., how they access them) highly influenced the use of the VE.
Average participant ratings for the preference of using of each VE application on a 10-point scale (1 strongly disagree-10 strongly agree).
From the Likert-scaled responses, sources of information for understanding a site differ depending on the type of project. According to the participants, talking with the people and recognizing natural features were rated the highest for helping them connect with a site during an in-person visit (Figure 7). However, all forms of information rated highly, supporting the idea that the need and use of information about a site depends on the specific project.
Average participant ratings for information sources used to connect to a site on a 10-point scale (1 strongly disagree-10 strongly agree).
The role of the senses collectively in a site experience was rated highly (Figure 8), but the breakdown of each sense indicates varying importance. The students reported being primarily interested in visual information from the site with audio considered second. The interpretation of the other senses, however, focused on how residents would potentially experience the site (e.g., the student was putting themself in the resident’s shoes to interpret the unpleasant experience of smelling trash or sewage every day). Sight and sound were the senses rated higher. Further research would help us better understand the role of each sense in a site visit and how immersive technologies can portray them. For instance, several students mentioned the lunch they had at a family restaurant in the favela as an important experience for them to comprehend the local culture and understand the site. Such an experience involves other senses (smell and taste), and students’ comments stress the importance these have in understanding a site and the limitations of the VEs mainly focused on visual, audio, spatial, and embodied experiences.
Average participant ratings for the importance of the senses during a site experience on a 10-point scale (1 strongly disagree-10 strongly agree).
Last, students were asked about their understanding of the site by using the VE. Before the trip to the favela, the students were confident that the VE was helping them understand the site. Following the trip, the in-person site visit was rated as very different from the VE experience, but the amount of familiarity on the behalf of students with the site was still high in both types of visits. Overall, students perceived they had a high degree of familiarity and understanding of the site after using the technology during the semester. All of this was despite the students indicating a notably high degree of difference between the VE experiences and the in-person site visit (Figure 9).
Average participant ratings for how the VE helped with understanding the site on a 10-point scale (1 strongly disagree-10 strongly agree).
Our results show that the VEs used in this studio can help people understand a site. Students considered themselves intuitively familiar with the favela before the in-person site visit. However, the VEs still have limitations, as pointed out by the students, such as the reduced perceptual information and overcoming expectations of the type of information and how it should be accessed. The role of the VE in this project aligned primarily in the grounding stage, as supported by student responses, where students were seeking analytical information to check details.
DISCUSSION
As projects become more globally dispersed, there is a need to understand how VEs can bridge the gap for landscape architects to understand a remote site. The Rio Studio offered the opportunity to test VE applications for remote site experience. Over the 16-week studio, 10 students provided various forms of feedback about the experience. We found four key findings: (1) VE site experience focused on grounding; (2) authentic interaction surpasses visual realism; (3) controlling the VE site experience matters; (4) VE can establish familiarity with a site. Overall, the findings suggest that VE applications cannot replace an in-person site experience but can lead to a certain degree of confidence in understanding a site. Therefore, VE can supplement in-person site experiences, with some limitations. These qualitative findings are explained in more detail below.
VE Site Experience Focused on Grounding
As a focus on the overall process, we begin by revisiting how the VE site experience occurred in terms of the trace concepts (Girot, 1999). This first finding derives primarily from the coded data, where grounding was the most prominent of the steps among student responses (Figure 5). Interestingly, the students understood the site experience as primarily grounding, the analysis stage (Hill, George, & Johnson, 2019) where inventories are completed to inform the conceptual design stage. As one student noted “I try to visit the site several times to get me familiar with all aspects of it” and “exploring and understanding the density of Santa Marta.” Grounding further defines site information by synthesizing predesign research with direct experience (Milburn & Brown, 2003), enabling designers to test potential ideas (Pressman, 2012).
However, the students reported some use of the VE as they were developing their design ideas, which corresponds to Girot’s finding stage, the second most frequent code from the trace concepts theme. Finding is a more refined step that synthesizes and tests information against a design idea, as one student noted, “Seeing for the potential of roof gardens and how accessible they are and the buildings that look most structurally safe.” Students reported that the VE alone did not provide the information the students felt was needed to build an inventory of the site, with one stating, “I usually combined it with Google Earth/Maps.”
One possible consideration of how VEs were used is to consider the locations students worked and subsequently influencing the preferences for specific VE applications in their ratings. During an in-person site experience, designers make active decisions on what to document by mentally comparing what they know to what they are experiencing. With a VE in a remote experience, such preparation may or may not occur depending on whether other information is readily available. The students indicated they preferred to have the materials close to the VE application so “I would rather sit at my desk and have my 360 in front of me than be separated from all my sketches and all my pin-ups and you know just record my experience there for like five minutes or ten minutes then come back up to my desk.” Although this is one potential reason for how VE was used as the site experience, it is speculative. George, Sleipness, and Quebbeman (2017) similarly noted that a VR intervention changed how students approached the design process, suggesting that this change may require further investigation. This finding, unlike the rest, was based on exploring how the VE aligned with Girot (1999)’s trace concepts to better understand the context in which the rest of the findings align, which primarily falls in the grounding/analysis stage.
Authentic Interaction Surpasses Visual Realism
Remote site experiences should present a high degree of visual realism to convey information about site characteristics accurately, but what is perceived by a user may be grounded more in the overall expectations toward the experience (Gilbert, 2016). The use of realistic 360°images was considered more of an extension of regular images rather than the experience of the site. As a result, there were requests for more information formats and modes of interaction to be integrated into the VE.
Overshadowing the realism of the 360°imagery, from the surveys and interviews there was a desire to overlay and integrate more technical and ambient information. There was an overall sense that although multiple media forms were available (as would be found in a typical studio), there was a desire to have the media more integrated into the VE, with students suggesting things like “maybe including other types of information in the site: population, demographics, sites of interest, other sources” and “general information site, GIS data, better slope analysis data, more site images.” When creating the VE for this studio, we opted to use the 360°imagery for its cost, time efficiency, and realism, which follows the suggestion of Lombardo (2018) that the time needed to optimize 3D models for VE is an obstacle to its implementation.
It was further noted that some technology worked better than others at giving a sense of authenticity to the site. One student noted, “The Vive provided a sense of direct perspective that one could get used to working with.” The HTC Vive was the most embodied experience but rated second to WebVR (as seen in Figure 6). If there is a correlation between immersion and embodiment, and embodiment is suggested to improve memory and cognition (Balakrishnan et al., 2012), would the exclusive use of the most immersive technology change the students’ expectations toward VEs?
One last aspect found among the student responses was a desire to explore a different form of interaction, particularly in the more embodied VE, like the HTC Vive. The passive act of looking around and clicking through the images was not seen as authentic to the experience, with one student suggesting, “[adding interaction like] in Call of Duty, the game, there’s a favela as one of the locations and I think I had more like mental archive from my experience … because I was able to like free roam around the site.” The idea of a free walk as opposed to clicking through images builds on the idea that movement is a fundamental part of a landscape (Relph, 1976).
Controlling the VE Site Experience
Traditional secondhand data can bias a design by omitting the surrounding context (Milburn & Brown, 2003; Pressman, 2012; George, 2016). Using 360°imagery reduced this bias in the content; as one student noted, “you can see everywhere because you know sometimes people have the videos but just from their own viewpoint.” But the students pointed out that omission of data from the whole favela biased where projects were focused, stating, “that made us focus on the parts that we had more data on and leave out parts that were sort of like inaccessible in the first part.” This connects to the findings in Figure 7, where students reported that most information was equally important for a site visit, yet different information types are preferred based on the specific project, which could support the students’ claim of bias in the information. The bias introduced was not necessarily problematic, as the same student pointed out that “this bias gave us a direction, so everyone’s work was sort of focused on the more dense areas informationally and it was helpful in the way it was structured information and organized it in a way that was accessible.” Nonetheless, students wanted more 360°imagery, stating that “you know that the HTC can’t go everywhere in the site,” and “I would focus on parts that were left out,” referring to areas that were not covered with the 360°imagery. This need for more information related to having more control in the experience and over the experience was considered.
Students appreciated having multiple VE options to use while recognizing that each option offered a different way to experience the same information. The preference for using the WebVR suggests that the accessibility of the applications influenced the frequency of their use over what may have been better. One student noted, “I think I just used the website because all the work I was doing was on the computer, [it] was on my computer anyways. So, I just go on the website you know versus switching devices.” Students also pointed out barriers to using other VEs, such as difficulty with setting up, discomfort (headaches), and the need to seek help from the researchers. “I had to borrow somebody and their time [to set the VE].” Overall, there was a desire to maintain and improve all three VE applications as the students indicated the capabilities of each was important, especially if the barriers were addressed.
Students also suggested that having more interaction in the VE would improve their encoding. Statements like “sometimes we had difficulty because we couldn’t zoom in and snap … so maybe I [would] add a zoom-in feature in the map with the points” and “[about the HTC Vive] you wanted to sometimes stop because you were moving” suggest the students want to be able to control the imagery. George (2018) says that the control of the interaction with the landscape is critical to contextualize what is viewed. while the ability to learn an object in the VE seems to increase when the user has control of the view. This suggests that control over the experience is directly linked to developing familiarity with a site and should be further investigated.
VE Establishes Familiarity with a Site
The familiarity with a site comes from multiple engagements over time that helps a designer develop an understanding of a place (Leach, 2002). From a student’s perspective, VE can enable a degree of understanding of a site’s characteristics; students could identify characteristics in the site and recognized this based on their ratings from the survey, but this was considered vastly different from the inperson experience (Figure 9). When asked directly, one student said, “when I went there [in person] I kind of felt the same, like it did not strike me as something very different.” Throughout the in-person visit, the students were observed orienting themselves within the favela, recognizing the main areas, and navigating to where each of their intervention sites was located, demonstrating spatial legibility over reading a site (Koseoglu & Onder, 2011). One student confirmed this: “ I recognized a lot of the places we went and I felt like I knew where I was going.” This aligns with the findings of George (2018) affirming that students can encode “the virtual tour site in a similar manner to physical sites” (417) but may not completely read a site well enough for design from a VE experience alone.
Despite the familiarity, students were surprised by how different some unique characteristics of the site during the visit were then first perceived in the VE applications. Several students noted differences in scale (size of elements) and degree of slope. One student noted that “the different levels, it was way more different or difficult than first perceived,” and another commented that “360s were biased a lot in the sense of scale so they made the space look bigger.” This aligns with George (2016), where students had a challenging time judging slope differences correctly with only 360°imagery, building on the notions from Lange (2001) and Bullinger et al. (2010) that VE using 360°imagery alone may not present adequate information for detailed evaluation. For instance, it seems that the sense of gravity is not one of the traditional five senses but is important for perceiving the slope and how it affects people’s daily routine. Steep terrain is a characteristic of favelas in Rio. Despite the use of the 3D physical model, VE applications, and GIS referring to the slope as the main site feature, the students were still impressed when they visited the site in person, mentioning it often, as the in-person experience enabled them to understand how it affected people living in the favela.
Familiarity through VE applications considers perceptions of understanding and recognition of key characteristics of the in-person site. As one student noted, “overall [the in-person site was] a very familiar yet exploratory experience” at the end of the trip. Similarly, another student noted, “I did not find much of a difference because in VR [VE] I kind of knew some of the places…some of the places were very familiar so when I went there I kind of felt the same like did not strike me as something very different.”However, in support of the significantly different rankings in Figure 9, students also noted as a key difference that “[Through the VE] we didn’t engage in any of our other senses through visual data, so pretty much everything we heard and what we smelled or we touched was feeling new.” This suggests that some of large distinction between the VE and in-person visit came from the lack of perceptual information. Despite the differences indicated in Figure 9 between the in-person visit and the VE visit, the students were able to use the VE with other secondary data to design their projects and gain an intuitive understanding of the site. Overall, this lack of impact of the somewhat substantial difference between the experiences could be attributed to factors from all of the key findings.
VE for Supplementing Site Experience
Overall, the students had a positive impression of the VE applications for understanding the site before the visit and in showcasing their final projects. The VE enabled a perception that site information was being communicated although there were distinct differences in how students recognized the site. Even though the students were familiar with the idea of the site as represented by the VE, this ultimately did not equate with the site as visited in person (see Figure 9). This suggests the importance of having the VE together with (not in place of) an in-person site visit (see Figure 8). One student mentioned that “there are certain aspects of the site that you only understand fully from a design perspective when you go there.” Another stated that the “difference in visual aspect to that of digital is you see the work and hands-on details that the people used to build the favela. This difference is hard to portray in videos or pictures.” The VEs came with considerations for interactions over realism and for control over the technology. Technical and information limitations were prominent influences on how the students engaged with the VEs to learn the site.
In terms of the timeframe with the frequency of survey responses, students were primarily using the VEs to finish their design concepts rather than experience the site, suggesting that they did not consider the VE a site experience. One potential consideration was that the students did not have to leave their research and other information as they visited the site and could do so any number of times. This new way to consider the site experience could change how designers engage with a site. The fact that students developed a certain level of familiarity with the site over time suggests that more integrated information and additional data could improve their understanding of the site, therefore improving users’ familiarity. However, consideration should be made for how much time and effort it would take to add more information and interaction into a VE.
Last, the students considered the senses universally important in understanding a site, relating how senses enabled them to be empathetic to a site’s users. The VE focused on only a few of the senses depending on which application was used, but the less clear perceptual information from the site experience may seem more important to students when an inperson visit is not possible. As the students used the VE, they were able to become more comfortable with the site, indicating they were not as fearful of being in the settlement as one student explained after the trip, “it’s at those moments that I think ‘oh these are people and they have a normal life.’” For example, a visiting faculty member pointed out that in the 360°videos, a man with a gun was present although the researchers did not notice him while filming. From another perspective, students noted the importance of additional perceptual information not included in Figure 8, such as taste, to understand the culture of the residents for whom they were designing. Such occurrences in an actual site could be distracting, but after multiple exposures in a VE, a landscape architect could become more empathetic to the residents and change their emotional reaction to what happens in the site (Schutte & Stilinović, 2017). Emotion in the environment is considered a part of being able to “see” a site (Relph, 1976), making a need to distinguish between the designer and user of the place to understand the appropriate emotion. The distinction on the role of emotion as a means to perceive a site can be both essential and distracting from what is happening on a site, particularly when the site is highly complex.
The site experience is one of learning, where there are aspects of gathering information, but that information is internally compared to preexisting knowledge of the site, the end users, or even precedents of similar projects. Much of the site experience is understood as an information-gathering activity. In the studio case presented, the sample, though small, provided insight on how a remote site can be apprehended by students through the immersive sensorial experience of VEs.
Limitations
The approach used in the case study focused on capturing feedback from the students every time the VEs were used. This enabled data to be collected in off-hours away from the studio, but it also depended on the students entering the data on their own. To supplement the survey information, observations were noted during and outside of the studio when possible but were not captured every time the VE applications were used. Along with the observations only occurring during studio time (inclusive of the trip), the measures for capturing students’ understanding of the site were focused more on confidence and perception of familiarity. More objective measures of site understanding (tests or graded design exercises) would provide deeper insights in a future study to compare with the confidence and perceptions of familiarity.
Another limitation relates to the accessibility of all the applications. A single HTC Vive located outside of the studio space was considered an obstacle for the students. Similarly, the MobileVR application was exclusively for Android phones, making the application only accessible to part of the student group. As a work-around, many of the students shared the MobileVR application among their peers.
The variety of applications enabled every student access to the 360°content in at least two ways. There was minimal disruption from technical issues throughout the semester. Technical issues were addressed as feedback was received from students. Last, developing the VE applications had additional limitations. Attempts at re-creating a 3D version of the favela proved costly and difficult, so we opted for 360°images as an affordable and manageable solution. The 360°imagery was all from eye-level view, as the favela was near an airport and drones could not be used to capture aerial data. This provided a unique first-person perspective of the site. In addition, the 360°imagery was collected from only the main routes in the favela, which were perceived as safer for the researchers to collect media during the three visits. Despite the limitations, this study provides insights into what about the VE applications was successful in generating familiarity of the site.
CONCLUSIONS
Immersive technology offers the ability to bring remote sites to the designer. However, it cannot be considered a replacement for an actual site experience, given the state of the technology we used, the timing of the project, and the budget for the studio. Rather, VE can supplement an actual site experience reasonably well, as found in our study. Specifically, the trade-off for realism versus interaction should be considered based on the complexity of the site.
Because VE can combine multiple forms of information (such as images, 3D models, and GIS) into a single experience, consideration should be given to better understanding how VE as a site experience changes the design process. In addition, the more complex a VE becomes, the more time it takes to create, which reduces time to work on the actual project. With this notion, there is a need for more empirical research in (1) understanding how combined perceptual and analytic components inform learning the site, (2) realism versus interaction, and (3) fundamental changes to how site experiences integrate into the design process. Such research can improve practitioners’ familiarity of a site and enable better analysis. In addition, the study revealed the importance and role senses beyond vision and audio play in someone experiencing and understanding a site.
The outlook for the use of VE for site experiences is promising but not without limitations. In the Rio Studio case, the VEs succeeded in generating a general familiarity with the site and provided an idea of what aspects enabled that familiarity to develop. More investigation should be done to fully understand how these findings would transfer to other remote studio cases, particularly when more information about the location may be available.
AUTHOR CONTRIBUTION
Danielle Oprean provided her expertise in organizing and creating the web-based immersive experience and providing guidance on the other immersive experience development. She contributed to the research design, data collection, and the analysis and interpretation findings.
Debora Verniz was the graduate teaching assistant of the World Studio course and contributed to gathering the imagery used in the immersive experiences. She further assisted with students’ use of the immersive experiences, data analysis, and interpreting the findings.
Jiayan Zhao, a graduate research assistant, assisted with developing the HTC Vive application and provided the images and details of the HTC Vive application. He made improvements to the HTC Vive application throughout the studio based on student feedback.
Jan Oliver, an independent researcher with ChoroPhronesis, developed the mobile application and updated it for the studio’s visit to Rio de Janerio. He further provided details on how the mobile application functioned.
Tim Baird was the co-instructor of the World Studio course who contributed his expertise in landscape analysis and design, specifically about inventory and site analysis. He provided insights that guided the data analysis.
Jose Duarte was the co-instructor of World Studio and the coordinator of the research project Decoding and Recoding Informal Settlements, within whose context the current research evolved. He gathered the imagery used for the VR systems and contributed to their design, as well as to the design of the research and the analysis and interpretation of data.
Alexander Klippel provided his expertise in creating and collecting immersive experiences throughout the design of the Santa Marta experience. His lab provided the equipment for the data collecting. He collaborated and supervised the implementation and provided feedback on the study design and write up.
PEER REVIEW STATEMENT
This article was subjected to a blind review by at least two external peer reviewers.
ACKNOWLEDGMENTS
The work of Debora Verniz was partially funded by grant 207028/2014-1 from the Science without Borders Program administered by the Brazilian National Council for Scientific and Technological Development (CNPq).
The authors would like to acknowledge the work on the immersive experiences created through ChoroPhronesis at The Pennsylvania State University. Funding for this research comes from the Stuckeman School of Architecture and Landscape Architecture and the Stuckeman Center for Design Computing.
REFERENCES
Jump to section
Related Articles
Cited By...
- No citing articles found.