C02 - Carrasco y Chen

Automation in Construction 126 (2021) 103677
Contents lists available at ScienceDirect
Automation in Construction
journal homepage: www.elsevier.com/locate/autcon
Application of mixed reality for improving architectural design

comprehension effectiveness
Moisés David Osorto Carrasco *, Po-Han Chen
Department of Civil Engineering, Construction Engineering and Management Division, National Taiwan University, Taipei, Taiwan
A R T I C L E I N F O A B S T R A C T
Keywords: Extended Reality (XR) technologies such as AR, VR and MR have influenced many industries, including archi
Mixed reality (MR) tecture. Even though they are all capable of creating immersive digital worlds, the only one capable of merging
Architectural design review the real world with a holographic 3D modelby letting the user interact intuitively and naturally with the project
Holograms
is MR. In this paper, 42 participants were divided into two groups and analyzed an original architectural
Sketch up viewer
Microsoft HoloLens
renovation design. They assessed the effectiveness of design review using Mixed Reality (MR) versus traditional
Design effectiveness 2D methods. The results show that MR based design review can effectively communicate 85% of the information
to the client versus 70% provided by 2D media. At the same time, it has the potential to enhance the client’s
comprehension of the aesthetic characteristics of materials, giving the possibility to replace physical samples
during the finishing stage of construction.
1. Introduction several physical models to analyze multiple design options for a single
project, architects who wish to save time in tedious explanations to non-
In the last decade, the most common way of representing architec architect clients about certain project details, contractors who are
tural design ideas has been using traditional 2D orthographic projection looking to eliminate the creation of physical mockups when solving
drawings (i.e. elevations, sections and floorplans), along with realistic design issues in the construction site, or firms wanting to reduce the
images of digital 3D models called renderings. These renderings were amount of paper waste they produce while designing a project. MR’s
drawn by hand in the past but, due to technological advancements, they ability to allow the user to see the real world while displaying a holo
are now made digitally in a computer using different software such as graphic 3D model merged with the physical environment [9], gives a
Sketchup, Revit, AutoCAD, or ArchiCAD. These tools have given the more intuitive and natural interaction with architectural design. It can
construction process many benefits like: communication improvement, provide just the critical, spatially referenced information, that augments
fluid development of design ideas and problem identification in early an individual’s knowledge of the environment [4] better than printed
stages of the project [1]. Even though these representations have precise rendered images, computer screens and VR headsets produced for the
measurements and defined materials, they can only be seen through same purpose [5].
two-dimensional means such as printing or computer screens, which The rapid advancement of technology and the potential applications
make the process less intuitive. of MR both in the office and the construction site, make it a technology
However, in recent years architectural rendering has received the that might revolutionize the way architects and other construction
influence from Extended Reality (XR) technologies. This is a term professionals work in the next couple of decades. That is why the re
referring to wearable devices and computer-generated graphics that searchers decided to explore MR’s possibilities. They designed an
allow the creation of real-and-virtual environments with which the user experiment based on an original architectural design, in order to
can interact [2]. XR includes many types of technologies such as determine if the use of Mixed Reality in the architectural design review
Augmented Reality (AR), Virtual reality (VR) and Mixed Reality (MR). process is more effective than traditional representation methods for
With XR immersive models, it could be possible to address many existing communicating the architect’s proposal during the development stage of
problems in the current design review process which traditionally have design. This will hopefully aid in expanding the existing body of
inconvenient solutions. For example, design teams in need to produce knowledge that relates XR and architecture, give more concrete results
* Corresponding author.
E-mail addresses: [email protected] (M.D. Osorto Carrasco), [email protected] (P.-H. Chen).
https://doi.org/10.1016/j.autcon.2021.103677
Received 10 September 2020; Received in revised form 14 March 2021; Accepted 16 March 2021
Available online 31 March 2021
0926-5805/© 2021 Elsevier B.V. All rights reserved.
M.D. Osorto Carrasco and P.-H. Chen Automation in Construction 126 (2021) 103677
on how this relationship can be sustained and possible methods on how 2.2. Review of related research with XR
XR, specifically MR solutions, can be applied pragmatically in the field
or the office. At the same time, this research could also help pro Recent years have witnessed the emergence of a newly created in
fessionals, who are unsure of whether MR is a good fit for their company, terest for XR. Technologies like VR and AR are now in the forefront of
to make a decision that fits their specific needs. change for many industries including construction, seeing the numbers
of XR related research almost double in the 2007–2017 decade [11]. The
2. Background emergence of robust XR applications has allowed AR and other tech
nologies to enter people’s daily lives and become a major trend for the
2.1. Extended reality future due to their portability and ubiquity [12]. These characteristics
help improve usability and maintain functionality in the field, such as
Extended Reality (XR) refers to all real and virtual combined envi providing information contained in drawings for comparison with the
ronments as well as human-machine interactions generated by computer real situation on site. Recent years have also seen VR and MR headsets
technology and wearables [2]. All these technologies can create become more readily available at varying retail prices. This has given
immersive digital worlds to various extents, and each offers specific the opportunity for many researchers to explore and evaluate the value
tools that can allow the user to achieve different goals. Hence to discuss and usability of XR for architecture in different areas. Hence, a brief
about XR, it is necessary to first discuss what is known as the Reality- summary of related research is presented in hope of expanding the
Virtuality Continuum [6] (see Fig. 1). In one end, AR happens when reader’s knowledge on the topic and give a general idea of what has
the real world is enhanced with digital content. That is like pointing a already been done with this technology.
smartphone’s camera to a specific place or object and later getting in In the year 2006 a group of researchers performed an experiment to
formation about it on top of the displayed image. A recent example of evaluate how VR could benefit the design review of courthouses in the
this technology is the “Google Glass” head mounted display (HMD) United States [13]. This research was used as a pilot project by the
which provides users with glanceable, voice activated assistance [7]. On government to determine wether VR design review would succeed as a
the other end, VR happens when the real environment is completely shut substitute to the current court house review process. The original
out, and a user is immersed by a completely digital environment. One method involved renting a space with a similar size to the intended
example of this technology is Facebook’s Oculus Quest which is already courtroom and creating a plywood mock-up model of the design.In that
widely used by the videogame industry [8]. MR stands i in the midpoint year, VR was still in an early stage, so no head mounted display was
of this “flow”. It refers to a continuum where computer generated con used. Instead, to create the immersive enviroment researchers used a
tent is blended in varying proportion with an individual’s view of the system known as CAVE (Computer Assisted Virtual Environment),
real-world scene [4]. Currently the most popular headset in this category which used enviornmental projectors directed to the walls of a cube
is the Microsoft HoloLens. According to Microsoft itself, Mixed Reality is shaped room to create a VR experience. Then, in order to create geo
a blend of the physical and digital worlds where humans, computers and metric data for display, Autodesk Revit BIM software was used. Results
the environment can interact[9]. showed a significant reduction in review time, which went down from 8
There are no devices today that can provide the user with an expe h with the plywood mock-ups, to 3 h with the CAVE VR. At the same
rience across the entire spectrum, thus users must first clarify their main time, since the review team was viewing the model together, they could
goal for using XR and choose a device accordingly. However, XR tech focus on one issue at a time, and discussions also proved to be more
nology could be useful in the learning process of design [10], since efficient.
ambiguities in communication could be eliminated, ultimately In 2018 researchers from the Georgia Institute of Technology carried
improving the design review process. out an experiment to analyze the user interface and human factors
Fig. 1. Differences between AR, MR and VR. Reference: [6].
2
involved in immersive VR platforms for design review [14]. They a construction project, since it is present in almost all of its phases.
developed an experiment were participants had to identify elements in a Changes in a construction project are inevitable [18], and when they
VR model (number of doors, shape of windows, column location, etc.) happen change orders appear. These are formal documents that notify
and then answer a question which evaulated their overall comprehen changes in the design [19], and for many authors [37–39], architectural
sion of the space. For this research the Oculus Rift DK2 headset and design related problems are a main source for the creation of change
Autodesk Revit were used to create the VR display and geometric data orders. They agree that the lack of communication fluidity caused by
respectively; the sample size was of 5 participants, all of them being variations during construction can lead a project to incur in cost over
professionals in the architecture or engineering fields. The results runs, delays in completion times, as well as causing rework and legal
showed immersive VR was able to provide users with excellent cognitive claims and disputes [20]. Hence, good communication during design
performance. They reported that VR immersive simulations can leverage review is very important to improve the effectiveness of handling
spatial perception and improve design comprehension as opposed to changes in a project.
traditional media. Design review depends on communication, and the range of people
On the same year, a group of researchers used the Microsoft Hololens that are consulted for it is always extensive [35]. During the process,
to enhance user experience in House Selling [15]. The team used Unity participants must view many graphics at the same time, and verbal
software to create their geometric data and import it to the HMD. Their critiques are usually supported by simple sketches [21,36].Nevertheless,
experiment had a random sample of 12 people, and consisted of two in 2D drawings, spatial conflicts among different specialties must be
groups divided according to the media to be used (Hololens or 2D). Users addressed by “skilled” users who rely on their own human perception to
would explore a 3D model or 2D plans, and then answer a questionnaire understand drawings, in order to avoid changes and rework [22]. The
about what they saw and provide feedback; their times were also problem is that these representations are abstract and fragmented: Ab
recorded. They concluded that potential clients could be influenced by stract because drawings of building components are represented
models that help them to quickly and clearly understand and familiarise through lines and symbols, and fragmented because relevant informa
themselves with the design. They also reported that participants who tion is distributed in different drawing sheets [17]. That means re
used the Hololens could intuitively understand the designer’s ideas and viewers must go through many drawings scattered around different
save time during the process of information exchange, with lower error pages just to get a clear picture of the design. Sometimes, even if the
rates compared to the users of 2D. They also mention that MR tech person looking at the plans is a skilled worker in the architecture field,
nology has potential for the future house selling market, as consumers questions about the design will always arise during the first inspection.
show higher purchasing inclination when they can quickly understand The researchers of this paper have construction working experience
the information being presented. in Honduras and Taiwan. Working in two places which are intrinsically
In 2019, a researcher from the University of Applied Sciences Upper different sheds some interesting light onto how similar the problems that
Austria developed a research to analyze the potential of VR on engi architects face can be, and give theoretical value to the exploration of
neering design review. [16] He used an HTC Vive HMD and Unity with ways to use available technology to make design review more efficient.
Autodesk 3DS Max for preparing the VR display and geometric data While working on-site for a Honduran consulting company and then in-
respectively. 16 participants with different working backgrounds were office for a Taiwanese architecture firm, many challenges related to
asked to review and find faults in the 3D model of two power units. The design review were identified. Most of them during the construction
goal was to compare the effectiveness of VR supported design review phase of the building and the preliminary stage of design review.
versus conventional approaches with CAD software support (2D). Par First, in the case of the consulting company, the appearance of
ticipants were divided in teams, were one person would use the HMD change orders during construction can cause many parts of a building to
and the rest would look at the model through a TV screen; then the team be redesigned or modified. These changes then have to be explained and
did another review with an alternative 2D method. Later, a question clarified to the contractors and in certain cases directly to site workers.
naire was used to assess how many “flaws” participants could find inside When this happens 2D workshop drawings are usually taken to the site,
the VR model. The results show that VR-supported design review allows and the team attempts to clarify how the new modifications would look
users to see more faults in a 3D engineering model than CAD software- like on the intended space. Other times, for example when building a
based design review approaches, due to its more intuitive way of looking hand railing, physical mockups need to be produced to make sure the
at a 3D object. activity will be carried out according to the specification requirements.
Finally, a group of researchers carried out a study which finalized in In these cases, the ability to see a design placed on site, in 3D and full
2019, where 13 design review meetings were analyzed over the course scale would make those situations smoother in terms of explanation and
of 28 months, focusing on projects inside the campus of a large uni visual aid for all the parties involved.
versity [17]. They recorded every design meeting and used software to In the case of an architecture firm, when developing the preliminary
extract the data. Results show that VR models describe details more design for a client, the team has to come up with multiple design options
concretely, and viewers are able to understand better as they can that later need to be transferred to a computer based 3D model, a
simulate their workflow during occupation and operation, making it physical model, or both, in order to showcase the design to the stake
easier to identify mistakes. They also mention that VR is good for holders. During this process, firms usually produce large amounts of
conveying a correct perception of size, scale, volume and depth of a paper waste for reviewing design options and also spend many hours
space more accurately, and that it benefits design comprehension when cutting cardboard or other materials to create a model that will be used
dealing with irregular, complex and curved shapes. for one meeting and then will probably be thrown away. Under both
The related research presented above might lead to the conclusion scenarios, long, sometimes tedious explanations are needed to ensure
that XR techonology will dramatically improve design comprehension. every participant in the review meetings understands the different parts
However, recent research also asserts that technologies like VR should of the design.2D drawings are not always capable of offering enough
be treated with equal importance to 2D drawings and BIM (Building information for = professionals in different fields who oversee the
Information Modelling, like Autodesk Revit). Adducing that VR should construction, to allow full understanding of a new design or modifica
complement the shortcomings of traditional media (like drawings), and tion. Under this light, XR technology can aid to improve this issue, since
not completely replace them [17]. as previous research shows, its use is more intuitive and contributes to
the effectiveness of this process by allowing users to observe all the re
2.3. The architectural review process and design change lationships of a particular design simultaneously, and engages the users
by reducing the effort needed to understand a proposal [17].
The design review process is a very important part of the lifecycle of
3
2.4. Choosing the XR: VR, AR or MR? increase the accuracy of the measurements and procedures applied in
this research.
In the last decade, XR became more readily available at affordable Kevin Lynch, in his book “The Image of the City”, defines the
prices, giving individual users the chance to start exploring the possi apparent clarity or “legibility” of a space as the ease with which its parts
bilities that the related devices could offer for different fields. By the can be recognized and organized into a coherent pattern [27]. Some
year 2015, smart glasses or head-mounted displays (HMD) had already authors [40,41], have already defined the process a person undergoes
gained traction as next generation mainstream devices [23], and in while trying to understand a physical space. These processes are sum
2016, many VR and MR HMD were released to the market. The HTC Vive marized in Table 1.
(VR), Oculus Rift (VR) and the Microsoft HoloLens (MR) are some of the Previous research [42,43,45] also shows that Immersive Virtual
most prominent examples. All of these HMD’s were capable of rendering Environments (IVE) seen through XR technology, are a satisfactory
immersive 3D environments, each with its own specific way of inter representation of real physical environments. That means that IVE’s are
acting with the real-world. During that time, the use of plastic handheld percieved in the same way as real environments. The difference is that
controllers along with a tethered HMD device was the best way to IVE’s permit the acquisition of essential information about the client’s
interact with a digital environment, which was completely shut out from preferences during the design phase, by allowing the architect to test
the real-world. However, the HoloLens innovated the user-3D model different alternatives of design which will normally not be possible due
interactions by offering the possibility to mix the virtual and the real to constraints in time and resources. Hence, MR and VR technologies
worlds through its MR (mixed reality) technology [3,29]. The device have a great potential for improving the effectiveness of the design re
was untethered, and was capable to realistically integrate 3D informa view process, as they display environments that people can understand
tion into a user’s perception of the real-world, which no other intuitively, just like real ones.
commercially available technology could offer [5].
By the year 2019, many other companies had created untethered, or 3. Materials and methods
standalone head mounted displays, predominantly for VR. Through an
analysis of the manufacturer’s websites, only three might be comparable 3.1. General information
to the HoloLens, due to a function Oculus HMD developers call “Pass
through”. This function is generally advertised as a safety measure for In an attempt to find a method for combining 2D and XR in design
defining a “walking area” where the headset will be used. It allows a VR
HMD to temporarily step outside of VR and see a real-time view of the
Table 1
world around the user through the use of integrated grayscale cameras
Analysis of how people perceive a real space and the related cognitive process.
on the device [24]. According to the information gathered at the
Reference: [40,41].
beginning of the year 2020, the devices that included this function were:
Process Dimension Type of Description
The Oculus Quest, the HTC Vive Focus, and Lenovo Mirage Solo. All of
information
these headsets allow users to see the outside world, however the func obtained
tions they offer to interact with it are very limited. Moreover, even
1 Exploring Cognitive Determine fixed Does the user
though some HMD’s like the Oculus Quest include a hand tracking
The individual Dimension or given layout know what
function [8], VR technologies still highly rely on the use of handheld will explore The individual and spatial activities and
plastic controllers in order to move through the space, even while using physical areas in thinks about, boundaries functions can
the “passthrough” function. order to orient organizes and
Environment
be performed
himself and keeps in the
In contrast, the HoloLens uses hand gestures and eye tracking to Experience and
develop means of information that environment?
control its interface. Its holographic display also allows the user to locomotion and allows him to
definition
remain in the “real world” while still looking at the 3D model and not communication make sense of the
lose his sense of location. The HoloLens also works without cables or within the given environment.
other attached devices, has stereo sound and a Holographic Processing space.
Unit CPU that allows people to instantly manipulate data [15]. It has 2 Categorizing Affective Environmental Does the user
color display, gaze tracking, two simple control gesture inputs, 4 mi The individual Dimension Sensory know if the
develops Involves our information spaces in the
crophones, and spatial sound and voice support [25,44]. In a way, it
categories of feelings and how environment
resembles the use of a smartphone because the user requires no prior information or a they influence feel narrow,
training to interact with the 3D model. It basically becomes a small, taxonomy of the our perception wide, low,
wearable computer capable of generating high resolution, spatially environment, the environment high, dark, or
located 3D content in real time. The absence of prior required training through which he and vice versa. illuminated?
tries to classify
means the designer can load a 3D model beforehand and the clients only Interpretative General and Does the user
characteristics in
need to put on the HMD and walk around instinctively. The Hololens a space in order to
dimension specific know what are
The individual information of the things in
could allow the users to interact with 3D models as if they were “real” understand it.
will rely on the environment the
physical models placed on a table, or walk around a 3D holographic memory points environment
space placed on site. It also becomes really useful in the construction for comparison made of? Their
site, where being aware of one’s surroundings is crucial to avoid po with newly materials,
tential hazards. Therefore the characteristics of the HoloLens make it a experienced textures, and
stimuli. size.
suitable option for architectural design review, which is why it was the
chosen headmounted display for carrying out this research. 3 Systematizing Evaluative Environmental Is the user
The individual Dimension coherency and capable of
systematizes the It incorporates symbolic describing the
2.5. Environmental perception and recognition environment values and meaning environment in
through analysis preferences, and a coherent way,
The perceptual aspect of design is complicated and non-obvious, of environmental the explaining its
contingencies determination of parts and
when the designer gives out a message, the receptor tries to decode its
(events good and bad different
meaning, based on what he knows and is capable of understanding [26] happening in the elements in the elements?
For this reason, concretely defining the process of how people look and environment). environment.
understand their environment, both virtual and real, is essential to
4
review, the authors of this paper decided to design and perform an 3.3. System description
experiment in hopes of shedding some light on the subject [28]. This
experiment was performed in Taipei, Republic of China (Taiwan) during The system used for this research was composed of three main ele
the first half of the year 2020. During this time, the world was facing a ments: A desktop computer with an i7–4790 CPU, 8GB RAM memory,
pandemic known as COVID-19, which caused many countries to halt Nvidia GeForce 750 GPU, and a Microsoft “HoloLens One” Head
work and normal activities for a period of time. Nevertheless, Taiwan mounted Display (HMD). For the Geometric data, Sketchup software
never went into complete lockdown, which allowed this research to be along with the Sketchup Viewer App for the HoloLens were used for
performed with complete normality and be delivered on time. producing the holographic 3D model. The 2D data and initial 3D model
was produced with Autodesk REVIT 2018 along with AutoCAD 2018
3.2. Expert interviews software.
The Expert interviews provided valuable input for setting the bases 3.4. Preparation of geometric data
for the experiment design in this research [28].The interviews were
made in a conversational manner, with a set of predetermined questions For creating the hologram, the researchers used Sketchup and its VR/
about how the experts carried out design review with their clients. The MR function for the Sketchup Viewer App, since it simplified the process
goal was to extract information that would give the researchers a general in just two steps: Have a SketchUp model, then upload into the HMD
idea of what methods have been used for design review in recent years using the app.
according to normal practice, and also to serve as a starting point for The Sketchup viewer app is capable of supporting “full scale size
beginning the questionnaire and experimental treatment design. The adjustment”, which allows the user to adjust the size of the 3D model to
researchers also wanted to identify which aspects of architectural design match the exact size of the space it is placed on top of. With its “Table-
were the most complex for clients and other users while trying to un Top view” function [30] it is possible to anchor the model to a fixed
derstand design for the first time. There were four interviewees from the point inside a room and scale it up to 1:1. If the 3D model and the room
architecture and civil engineering fields. Two civil engineers from are the same size and shape, it is possible to create a holographic life-size
Honduras with background in project management, with 16 and 14 mockup of the intended space, which made the app suitable for the
years of experience respectively. One architect from Nicaragua with 53 experiment in this research.
years of experience and one architect from Taiwan, specialized in BIM
software and 3D models, with 15 years of experience. 3.5. Variable definition
While recording the answers, the researchers noticed there were
some aspects all the interviewees naturally agreed upon: The experiment aimed to measure the percentage of comprehension
(dependent variable) of a client while looking at an architectural design,
1. The office architectural typology is the easiest for clients to under by understanding how the absence or presence of MR (independent
stand, since it has the least number of elements to be explained. The variable) affects the client’s comprehension effectiveness.
most complex typology is residential as the people involved are more For this reason, participants were randomly divided into two groups:
numerous and have different backgrounds.
2. At the same time, comprehension of architectural details (false 1. The control Group (2D group): This group looked at the design using
ceilings, moldings, bases, railings, etc.), and the overall 3D geometry a printed 2D drawing set that included elevations, sections, 3D ren
of elements are the hardest aspects for clients to understand and derings, floor plans and architectural details.
visualize. 2. The Experimental Group (MR group): This group looked at the design
3. They also agreed that while explaining design, methods such as: using the HoloLens and a 3D hologram of the design proposal placed
using physical samples and models, as well as size comparison with on top of the real site.
similar existing physical spaces are really helpful for increasing client
comprehension. 3.6. Sampling
The Taiwanese BIM expert also gave very helpful suggestions for this While working, an architect might encounter many different kinds of
research. He pointed out that MR and 2D might actually work together clients, who have different culture, race, gender, age, and levels of
to explain design, rather than thinking one would replace the other. He knowledge. Clients can come from anywhere, and this last characteristic
suggested both methods should not be analyzed separately, instead they is especially important as it suggests that the sample for this research
should be analyzed together in order to understand things like: with needed to be heterogeneous in terms of cultural and academic back
which method participants are able to see specific elements better? or grounds. Thus, the researchers did the random sampling process by
Which one is better for showcasing general ideas? advertising the experiment in different places around Taipei where
Based on the gathered information, and the review from related people with the desired characteristics could be found; these places
research, the researchers agreed upon the following: were: Chinese language centers, local companies, sports teams, and
universities (bachelor’s, master’s, and PhD levels). The sample included
1. A questionnaire would be the instrument used to measure the per both Taiwanese people and foreigners from different nationalities,
centage of comprehension of the design proposal. regardless of their background and older than 18 years old. People who
2. The office typology would be used for the creation of the experi arrived at the laboratory for experimental treatment participated on
ment’s design proposal. This decision was taken because the exper their own accord, and in full possession of their mental and physical
iment needed to be simple in order to avoid long treatment times and faculties. For a detailed list of participants and their backgrounds please
not lose the participants’ attention or enthusiasm in the process. look at Table 5.
3. The questions in the questionnaire should move from the general to
the specific aspects of design, in order to evaluate how good MR is for 3.7. Data collection
visualizing and understanding architectural elements (false ceilings,
moldings, bases, etc.…) and their properties (thickness, material, 3.7.1. Interview method
texture, etc.). Expert interviews suggested that regular interaction with clients was
done in a conversational manner, and since the experiment needs to
replicate as much as possible the real conditions of an architect’s work
5
environment, it was decided to use a questionnaire in the format of a overcomplicating the situation, and it was also suitable to receive par
Semi-structured interview. According to previous research, this personal ticipants comfortably since it has bus and bicycle stations in its
interview method can overcome poor response or non-response issues, proximity.
where good respondents are those who appear comfortable while
interacting with the interviewer and provide solid answers with good 3.7.3. The architectural design
detail [31]. Semi-structured interviews usually have a questionnaire In order to know what questions to ask, the researchers had to start
with predetermined questions, but their conversational nature allows by creating an architectural design that that would provide the elements
users to explore issues they consider important [32]. At the same time, it to be evaluated in the questionnaire according to the aspects of
gives the questionnaire a certain degree of flexibility to overcome the perception defined on Table 1. This design was meant to challenge the
language barrier that certain participants might have. participant’s ability to achieve clarity or “legibility” of a space.The
evaluated aspects and included characteristics are as follows (see
3.7.2. Experiment site Figs. 3,4 and 5):
To keep the process simple, short, and easy to standardize, the studio
office typology was chosen. This is a kind of small office destined to 1. Exploring:
carry out business and handle processes, with a size that can vary from Aspects to be evaluated:
one single room to an entire floor of a building [33]. Thus the re 1. Determination of spatial boundaries and layout of the envi
searchers decided to use a room provided by the National Taiwan Uni ronment; 2. Identification of where activities can be performed; 3.
versity (see Fig. 2), with an area of 66 m2, located inside the Civil Identification of new and existing elements (demolition plan); 4.
Engineering Research Building (CERB). The size of the room was Identification of visible and non-visible architectural elements (in
adequate to create a reasonably complex design without this case non-visible refers to elements covered by false ceilings like
Fig. 2. EMCA lounge at the 4th floor of the Civil Engineering Research Building at NTU, Taiwan. Existing floor plan (top), view 1 (left) view 2 (right).
6
3.7.4. The questionnaire

The final version of the questionnaire included 24 questions about
the aspects of design, and the last question (number 25) was “open-end”
to assess the participant’s level of environmental coherency; there were
also 5 questions regarding the interaction with the HoloLens. As
mentioned before, the questionnaire was based on the literature review,
user trials, and expert interview recommendations. The number of
questions accurately answered represent the percentage of comprehen
sion the participant had according to the group he was assigned to; each
question had a value of 4% to sum up to 100%. At the same time, there
was a device interaction test to measure comfort while using the Hol
oLens, since the researchers initially considered it could be a significant
variable in the comprehension process [28]. Table 2 summarizes the
tests applied to the participants:
The responses were then recorded using the evaluation form
included in the appendix Table A.1. All the interviews were recorded on
video and the experiment treatment was applied identically to every
Fig. 3. Isometric Diagram of the New Floor Plan. 1.Work Area; 2.Reception; 3.
participant to ensure proper results. For a detailed description of the
Meeting Room; 4.kitchen; 5. Meeting Hall.
Standard Operating Procedure please go to Table 4.
Once the data was collected, the results were tabulated and counted
beams, or elements wrapped by other materials or walls).
using the evaluation form, in order to obtain averages and percentages
Characteristics:
needed to understand the results. Table 3 provides the questions asked
The design has many spaces, concretely 5 (see Fig. 3), with
during the interview, classified according to the category they belong to.
different functions and varying floor levels and ceiling heights. This
way the participant was able to understand the space in a three-
3.7.5. Standard operating procedure
dimensional aspect and do some exploration to put the design
In order to have a clear process while performing the experiment, a
together.
standard operating procedure was developed. This procedure served as a
2. Systematizing:
protocol while applying the experiment to ensure every participant
Aspects to be evaluated:
received the same and equal treatment (see Table 4).
Environmental coherency (ability to describe the environment in a
coherent way and identify its general function).
4. Results
Characteristics:
The design includes complexity of materials (see Fig. 5). This
4.1. Participants
means it has many surfaces with different textures to understand if
the participant can notice differences in texture, color and size. At
Altogether, 42 participants (Male: 26; Female: 16) took part in the
the same time, since the design was placed on top of an existing
experiment, with 20 as part of the control group (2D), and 22 as part of
space, a demolition plan was also created.
the experimental group (MR). The average age was 26.36 years and they
3. Categorizing:
ranged from 20 to 58 years. Regarding academic background, 30 par
Aspects to be evaluated:
ticipants were professionals with either a bachelor’s or a master’s de
1.Identification of specific information about materials used in
gree, and 12 participants were students from different university majors.
furniture and surfaces; 2. Identification of inherent properties of
Then, 16 out of 42 people had previous knowledge reading plans, either
materials (in terms of smoothness, roughness, warmness or cold
from the industry, empirical experience obtained from buying real es
ness); 3. Perception of size and 3D space (in terms of high, low,
tate, or personal house renovation projects. Finally, regarding the par
narrow and wide).
ticipant’s cultural background, the sample belonged to 5 different
Characteristics:
regions in the world: Africa (1), Middle East (1), Asia (15), North
The design included some polyvalent spaces (see Fig. 4). This was
America (3), South America (14) and Europe (8). (see Table 5).
meant to challenge the participants while they were creating a
coherent idea of the space, since some areas had different functions
4.2. Presentation of results
that were understood after looking attentively.
Effectiveness is defined as the capability of a person or object of
Once the architectural design was finished, the researchers used it as
producing a decided, decisive or desired effect [34]. In this sense, the
the base to write and classify the questions in the questionnaire and
results above indicate that MR supported design review is more effective
create a measuring tool. The questionnaire draft was tried out, modified
than traditional 2D methods of representation. When a client comes in
and improved through a trial process which included 4 participants.
contact with a design proposal for the first time, MR makes it possible to
These trials gave out the following conclusions:
receive design related information more intuitively by 15%. This is
based on the cognitive test carried out by the experimenters, which
1. The amount of time needed to explore the design with MR was 5 min.
measured that on average, MR helped participants understand 85% of
2. The plans for the 2D group had to be printed in full color (A2 paper
the total information when using the HMD, while 2D users only obtained
size to show the desired scale) in order to ease reading.
70% of the total information in the questionnaire on the same amount of
3. A specific seat was chosen in the room so that the orientation of the
time (5 min).
plans would coincide with the orientation of the physical space to
While looking at the results in Table 5 the researchers noticed a
make comprehension more straight forward.
direct linear relationship between the use of MR and a higher percentage
4. The draft of the evaluation form was finalized and ready to be
of comprehension. Thus, it was decided to carry out a linear regression
implemented (see appendix Table A.1).
to test this relationship. The initial variables considered in the estima
tion were: Previous experience reading plans, gender, age, years of
professional experience, and comfort level of the HMD (measured in the
7
Fig. 4. (Up) Renderings: 1 Kitchen and Meeting Hall, 2 Working Area, 3 Kitchen, 4 Meeting Room and circulation space; (down) New Floor Plan.
last five questions of the questionnaire). The variables years of profes The estimation also passed the respective residual and stability di
sional experience and comfort level of the HMD were omitted as they agnostics meaning that the model was neither heterogeneous nor mis
turned out to be insignificant while running the first estimations for this specified. The tests had the following results: Heteroskedasticity test
model. The final estimation included the variables mentioned on eq. 1, BPG F-statistic of 0.20 > 0.05; White F-statistic of 0.62 > 0.05; RESET
which used an alpha level for the p-value of 0.05, and obtained a Test F-statistic of 0.51 > 0.05(see Tables A.2 to A.4 in the Appendix
resulting Adjusted R-squared value of 0.57 (See Table 6 and Eq. (1)). section). The estimation results show that previous experience reading
plans is not significant while using MR. This might suggest that
Gcomp = 83.35 + 14.42MR + 0.93PlanExp − 8.36Gen − 0.38Age + u
regardless of the initial comprehension level of the person looking
VariableDefinition : Gcomp(GeneralComprehension); MR(MixedReality);
(bachelors, masters, PhD, architect or non-architect), MR will help the
PlanExp(PreviousPlanexperience); Gen(gender); Age(age) user understand the design better by 14.42% on every case. This value is
1
close to the estimated 15% mentioned earlier. This supports the
8
Fig. 5. Interior space with material specifications.
people used MR and 21 people used 2D, then both media were consid
Table 2
ered to be equally effective for reading the information asked in the
Metrics for the experiment tests. Reference: [14].
question. If the difference between both media was less than 30% (5
Test Metric Expected results points) it was also considered both were equally effective. Finally, if
MR 2D there was a difference greater than 30%, one medium was preferred over
Comprehension Questionnaire (24 85–95% of 55–75% of the other. The results indicate the following:
Level questions) comprehension comprehension
(25 questions)
Overall Score 4 or higher Score 4 or less
1. Both MR and 2D are suitable for: Identifying spaces and general
comprehension (1 layout, identifying were activities can be performed, and identifying
question) heights.
Device Ergonomic Quality < 4 on all 5 Does not apply 2. 2D plans are suitable for: identifying specific measurements of the
Interaction questionnaire aspects space (Length and width), understanding the demolition plan, and
(5 questions) identifying countable elements in the design like number of lamps,
switches or sockets.
3. MR is better for understanding how elements in the space interact
statement of previous authors that the level of intuitiveness during the
with each other, I.e. if the columns are covered by a specific material
interaction with any given display is what determines ease of compre
or if the false ceiling will cover certain beams. It was especially useful
hension, and not so much the person’s background or experience. In
for quickly identifying the specific materials and textures of the
addition to this, the coefficients for Gender [− 8.35] and Age [− 0.38]
design (Ceiling, floor, walls, and kitchen), visually understanding
show that younger people tend to understand design better by 0.38% for
size in terms of width, and understanding innate properties of ma
every one-year difference. Also, women seem to have the tendency of
terials like roughness, smoothness, warmness or coldness.
understanding 8.35% less than men while looking at architectural
design for the first time.
Finally, after re-watching the recordings of the process the experi
The researchers also used the evaluation form to record and count
menters were able to identify certain behavior patterns for the partici
what type of information (floor plan, notes, section, elevation, and iso
pants in each group. First, the MR participants tended to begin the
metric) was used more frequently by the users in the 2D group while
exploration process by noticing the smaller, tiny details like the deco
trying to understand the design. This was done by counting how many
rations on the tables or the content of the computer screens. Then, they
times 2D participants used a specific kind of drawing to give answers
made a quick scan to identify the spaces in the layout, and finally moved
during their experiment session. The results indicate that the isometric
into the broader aspects of the design such as the ceiling and floor ma
diagram is the most useful and easiest to read as it can quickly convey
terials, illumination, and movable/fixed furniture.
information to participants. The second place is for plan drawings (floor
Meanwhile, people who analyzed the design using 2D drawings were
plans and elevations), third place is for sections, fourth place for ren
able to identify some measurements, and the general layout along with
derings, and last come the notes (see Table 7).
its related activities, mainly by comparing the isometric drawing to the
Regarding how to combine MR and 2D for design review, the re
floor plan. Nevertheless, they mostly failed to understand specifics of the
searchers recorded which media (MR or 2D) participants used to answer
design, such as the aesthetics of any particular material or certain
each question, in order to conclude which one was more useful while
inherent material properties such as thickness and texture.
trying to find the different types of information asked in the question
naire (see Table 8). The greatest amount of points each media could get
5. Discussion
on every question was 42, implying that every participant used either
only MR or 2D to answer. If the difference was zero, in other words, 21
We see that MR’s advantage is the ability of allowing the user to
9
Table 3 Table 4
Questionnaire (MREXP-002). Experiment standard operating procedure.
Exploring Experiment standard operating procedure (MREXP-004)
Determination of Layout and spatial boundaries
Group A (2D) Group B (MR)
01 What are the spaces included in this renovation?
02 What is the overall length and width of the room? Group A will be provided only with the Group B will be provided only with the
03 Which walls serve as a boundary for the space to be renovated? renovation’s Ready to Submit (RTA) MR model and HoloLens. The researcher
Environmental experience and definition plans and renderings. The researcher will follow the steps as described below:
04 In what area of the renovation can you eat your lunch? will follow the steps as described
05 In what area of the renovation can you sit and listen to a presentation? below:
06 In what area of the renovation can you work? 1. The researcher will read a brief
07 In what area of the renovation can you take a nap? introduction to the experiment and
08 Is there a space for storing books? 1. The researcher will read a brief the research.
introduction about the experiment 2. Then the researcher shall make sure
Categorizing
and the research. the model is in place and make sure
General information
2. The plans shall be explained by the the participant knows how to adjust
09 What elements of the existing space will be demolished or taken away, and
researcher to make the participant the MR gear.
which ones will remain after the renovation is finished?
familiar with the content. 3. Participants explore the model for 5
10 What structural elements (columns, beams, slab) are going to be left exposed
3. Participants look at the plans for 5 min.
(not covered by any kind of material except for paint) and which ones are going
min. 4. After 5 min, the researcher will go
to be covered?
4. After 5 min, the researcher will go through the questions, marking
Specific information
through the questions, marking which ones the participant is able to
11 How many types of false ceiling are included in the design, what is their
which ones the participant is able to answer. (Annex Table A.1, Form
material?
answer. (Annex Table A.1, Form MREXP-003)
12 Are there any curtain or glass walls in this design? How many?
MREXP-003) 5. Percentage of questions answered
13 How many floor types are included in the design? What is their material?
5. Percentage of questions answered with MR for this round should be
14 What is the material of the countertops (both kitchen and printing area)?
with 2D for this round (R1) should be recorded to this point in Section 4 of
15 How many types of lighting fixtures are included in the design?
recorded to this point in Section 4 R1 form MREXP-003.
16 What is the material of the kitchen backsplash?
of form MREXP-003. 6. Afterwards, the participant shall be
Sensory information 6. Afterwards, the participant shall be given 2 more minutes to go through
17 With the current layout, a person presenting in the meeting room can move given 2 more minutes to go through the information again in order to
around the space while all the chairs are occupied? the information again in order to answer missing questions.
18 Could a person walk to the reception area, through the workspace while answer missing questions. 7. Percentage of questions answered
everyone is seating down working? 7. Percentage of questions answered with MR for this round (R2) should be
19 Are the stairs at the meeting hall wide enough to seat 10 people in each row? with 2D for this round (R2) should be recorded to this point in Section 4 R2
20 Is the space for the mezzanine area enough to allow a person with a height of recorded to this point in Section 4 R2 of form MREXP-003.
1.65 m or more stand up? of form MREXP-003.
21 If you are taller than 1.85 m, will you be able to stand in the kitchen? The sum of the percentages obtained during R1 and R2 shall be added to obtain the
22 Is the floor used in the meeting room smooth or rough? total percentage of comprehension presented in Table 5.
23 Are the materials used for the kitchen countertop and backsplash warm or cold? Since the experiment was advertised as a comparison between paper drawings and
holograms, the following steps (8–11) were taken to ensure the participants could
Systematizing interact to all the media available in the experiment and had the chance to
Environmental coherency experience MR.
24 What will be the overall function of the space in this renovation? (A place to
work, relax or entertain?)
25 Please describe in your own words the renovation that is to be undertaken in 8. If there are questions left 8. If there are questions left
this space. (For example, mention what are the main spaces, materials, unanswered, the participant will use unanswered, the participant will
functions, and feelings the space will have. You can add any other information the MR headset for 5 min to look at look at the 2D RTA plans for 5 min.
you consider relevant.) the 3D model. 9. The questions that are still
Device interaction (based on Likert scale) 9. The questions that are still unanswered will be asked again.
26 The HoloLens was heavy unanswered will be asked again. 10. Percentage of questions answered
27 The HoloLens was difficult to adjust 10. Percentage of questions answered with MR for this round (R3) should
28 Your eyes felt sore while using the HoloLens with 2D for this round (R3) should be recorded to this point in Section 4
29 You had motion sickness or dizziness while interacting with the MR model. be recorded to this point in Section 4 R3 of form MREXP-003.
30 You were lost or disoriented while exploring the MR model R3 of form MREXP-003. 11. Questions 26–30 should be asked in
11. Questions 26–30 should be asked in order to assess user experience with
order to assess user experience with the HoloLens.
explore a given space in a way that accurately mimics reality (see Fig. 6). the HoloLens. End of the session
Since MR headsets, like the HoloLens, do not need prior training, it End of the session
becomes clear that mixed reality can become a helpful aid for the design
review process and make it more efficient. MR allows users to interact
such as texture, thickness and appearance was accurate and immediate.
with a 3D model while still being able to interact with the real world and
Under this light, MR HMD’s could be capable of aiding architects
read specific information from a printed drawing at the same time.
solve several problems that derive from design-based change orders in
Based on the results, it is possible to say that if users are provided
the construction site. First, MR’s 1:1 scale display ability can aid in the
with isometrics and floorplans along with a 3D hologram they can
creation of virtual mock-ups. This function can allow stakeholders to
interact with, their comprehension can be boosted by at least 15% more
visualize things that usually need a lot of detailing like hand railings,
while looking at a design for the first time.
doors, and other small elements, such as bases and crowns, on site
It is also possible to assess that the type of interaction the participant
without having to carry any cables or heavy equipment; hence, the
has with the design is also an important factor for determining the
approval process for design could become more effective.
comprehension level. In other words, the more accurately a display can
Additionally, MR has the potential to replace physical material
realistically simulate the interaction between the user and the new
samples. Its ability to accurately represent an element both in thickness
space, the better it will be understood.
and appearance is very valuable for clients and contractors to under
Correlatively, the correct use of textures can also aid comprehension.
stand the choices and decisions made by the architect. However, in order
By looking directly at materials placed on site, with correct widths and
for this process to be successful, the architect must make sure to obtain
scale, MR’s ability to provide information about material properties
detailed, high-resolution textures to ensure the hologram can be an
10
Table 5
Participant background and results from cognitive tests according to treatment group.
Treatment group: MR (Experimental)
N Academic background Age Gender Prof. Exp. (years) Previous 2D Exp. Nationality Test result %
2D MR
01 Student 20 M 0 No Honduras – 88
02 Bachelor’s Mass Communication 26 M 3 No Spain – 92
03 Student 20 M 0 No Honduras – 84
04 Civil Engineering 31 F 2 Yes Mozambique – 88
05 Student 21 F 0 No Japan – 76
06 Bachelor’s International Studies 28 M 0 No United States – 96
07 Student 22 M 0 No Belize – 84
08 Bachelor’s Business and Finance 29 F 0 No United States – 72
09 Master’s psychology and statistics 26 M 5 No Spain – 92
10 Bachelor’s Chemical Engineering 26 M 3 Yes Guatemala – 88
11 Ph.D Civil Engineering 33 F 3.5 Yes Taiwan – 72
12 Student 21 F 0 No Guatemala – 92
13 Bachelor’s Architecture 24 M 1 Yes Japan – 80
14 Bachelor’s Civil Engineering 23 M 0 Yes Taiwan – 88
15 Master’s Architecture 46 M 20 Yes Taiwan – 80
16 Master’s Construction Management 58 M 30 Yes Taiwan – 72
17 Bachelor’s Physics 25 M 1 No Iran – 88
18 Bachelor’s Electrical Engineering 24 M 0 No Spain – 96
19 Bachelor’s Applied English 32 M 0 No Taiwan – 92
20 Bachelor’s Computer Science 29 F 4 No Costa Rica – 80
21 Bachelor’s Nutrition Science 26 M 3 No Taiwan – 84
22 Master’s Architecture and urban design 26 M 5 Yes Nicaragua – 92
Treatment group: 2D (Control)

N Academic background Age Gender Prof. Exp. (years) Previous 2D Exp. Nationality Test result %
2D MR
23 Master’s Economics and Finance 31 M 7 Yes Belgium 88 –
24 Bachelor’s Teaching and Translation 28 F 5 No Mexico 56 –
25 Student 23 M 0 No St. Vincent and the Grenadines 72 –
26 Bachelor’s Agrobusiness 26 M 3 No Guatemala 80 –
27 Bachelor’s Atomic Physics 24 F 0 No Germany 72 –
28 Student 20 F 0 No Korea 68 –
29 Student 22 M 0 No Mexico 72 –
30 Bachelor’s Mass Communication 23 F 3 No Philippines 60 –
31 Student 21 M 0 No Guatemala 76 –
32 Student 20 M 0 Yes Spain 84 –
33 Student 21 F 0 No Honduras 56 –
34 Master’s Integral Security 30 M 7 Yes Spain 72 –
35 Bachelor’s Architecture 26 F 2.5 Yes Taiwan 68 –
36 Bachelor’s Civil Engineering 24 F 0 Yes Taiwan 60 –
37 Student 23 F 0 No Japan 80 –
38 Bachelor’s Civil Engineering 23 M 0 Yes Taiwan 64 –
39 Bachelor’s Computer Information Systems 25 F 0.5 Yes United States 80 –
40 Master’s Biotechnology 28 M 1.5 No Spain 64 –
41 Student 23 F 0 No Malaysia 60 –
42 Bachelor’s Criminology and Philosophy 30 M 9 No Puerto Rico 72 –
accurate representation of the intended material.

Table 6
Finally, during the experiment implementation, the researchers
Estimation output for the experiment.
noticed three major limitations regarding the use of MR software:
Dependent variable: GComp
Method: Least Squares 1. Pixelated low-quality textures with low resolution, can hinder the
Date:06/30/2020 Time:15:58 process of comprehension. For example, floor materials like tiles,
Sample (Adjusted): 1 42
could be perceived as carpet due to the size of the pixels.
Included Observations: 42 after adjustments
Variable Coefficient Std.Error t-Statistic Prob. 2. The size of the images used for textures can make the 3D model really
MR 14.42535 2.412412 5.979640 0.0000 big in terms of file size, making it difficult for the processor in the
PLANEXP 0.925219 2.484834 0.372346 0.7118 HoloLens to render the image quickly.
GENDER − 8.356152 2.437746 − 3.427819 0.0015 3. Bright illumination could affect the correct visualization of holo
AGE − 0.378674 0.182241 − 2.077870 0.0447
grams, since they may look transparent, and according to the level of
C 83.35069 4.866258 17.12829 0.0000
R-squared 0.612502 Mean dependent var 78.09524 brightness, may become hard to see.
Adjusted R-squared 0.570610 S.D. dependent var 11.28740
S.E. of regression 7.396388 Akaike info criterion 6.951204 6. Conclusions
Sum square resid 2024.142 Schwarz criterion 7.158069
Log likelihood − 140.9753 Hannan-Quinn Criter. 7.027028
F-statistic 14.62109 Durbin Watson stat 3.019079 The work describes the design, and implementation of an experiment
Prob (F-statistic) 0.000000 which used the holographic capabilities of the Microsoft HoloLens to
evaluate its effect on design review effectiveness during the develop
ment stage of design. For this purpose, an original architectural design
11
Table 7
2D plan information according to frequency of use per treatment session.
2D group participant Times information was used per session
number Rendering Isometric Notes Plans Section
23 8 4 0 7 0
24 1 11 0 1 0
25 1 6 1 10 0
26 5 5 0 9 4
27 0 4 1 9 2
28 1 7 4 2 0
29 3 6 1 5 2
30 0 7 3 6 1
31 2 7 2 2 0
32 1 7 3 8 3
33 3 4 1 2 3
34 2 2 1 9 5
35 0 6 1 7 2
36 4 3 1 4 3
37 2 11 1 3 3
38 2 4 1 6 2
39 0 7 3 6 3 Fig. 6. Representation of user using the HoloLens for looking at the design
40 2 6 1 3 3 proposal integrated with the real environment.
41 0 5 1 4 3
42 0 6 1 5 5
Average times Information 1.85 5.9 1.35 5.4 2.2
limitations for this research as the loading times, due to the size of the
was used per session model, were long. This prevented the researchers from adding more
people to the sample due to time constraints. The authors want to
encourage other researchers to try out this experiment with a bigger
Table 8
sample, and using Microsoft’s Hololens Two as it has better processing
Questions according to media used for answering (MR or 2D). power and a wider range of view. At the same time a small interior space
with limited sunlight was used for carrying out this research. Further
Media used to answer Questions
study in larger spaces will be of great value to corroborate the infor
mation in this paper. The authors also believe there could still be more to
Times media Times media
Questi was used to Preferred Questi was used to Preferred
on answer medium on answer medium learn by analyzing how MR interacts in outdoor environments and how
2D MR 2D MR to overcome the transparency factor mentioned above. Future research
01 21 21 Both 15 32 10 2D on this topic might be oriented towards the following directions: Is it
02 41 1 2D 16 07 35 MR possible to use MR to place a 1:1 model of a building on-site? To what
03 23 19 Both 17 16 25 MR extent can this technology be used to make a preliminary site analysis in
04 19 23 Both 18 15 27 MR the design phase of a project? How much could this help to sell a project?
05 12 30 MR 19 17 25 MR This would complement the information discussed in this paper, and
06 20 22 Both 20 19 23 Both would present yet another application of MR for architecture.
07 20 22 Both 21 21 21 Both Altogether, the study suggests that MR-based design review process
08 20 22 Both 22 04 38 MR is more effective and significant than 2D drawing-based methods for
09 28 14 2D 23 0 42 MR communicating the architect’s design proposal effectively. The evalua
10 17 25 MR 24 20 22 Both tion estimated that MR can boost the comprehension level of any given
11 05 37 MR 25 20 22 Both participant by 15% regardless of his or her academic background. It also
12 19 23 MR provided valuable input regarding how to combine MR and 2D in a way
13 09 33 MR in which the strengths of each one compensates the weaknesses of the
14 13 29 MR other. It is possible to say that the MR technology of the HoloLens, which
does not require previous training and allows the user to interact with
the real and virtual worlds at the same time, proved to be a very good
way of displaying an architectural proposal at initial design review
proposal was created, in order to provide a base line that would allow meetings. Finally, it is capable of aiding professionals to avoid long and
the comparison of participants. They were divided into two groups: a tedious explanations while presenting a design proposal, and could also
control group, which looked at the design with 2D drawings, and an replace physical samples and mock-ups, especially when the visual
experimental group, which looked at the design with MR. The quanti properties of materials need to be analyzed.
tative measurement was made with the use of a questionnaire based on
aspects of environmental perception provided by the literature review.
The questionnaire allowed the calculation of a percentage of compre Declaration of Competing Interest
hension according to how many questions any participant answered
according to the assigned group. The results were recorded in an eval The authors declare that they have no known competing financial
uation form, and then were tabulated and counted to produce results. interests or personal relationships that could have appeared to influence
The processing capacity of the “Hololens One” provided many the work reported in this paper.
12
Appendix A. Appendix
Table A.1
Evaluation form for experiment application.
13
Table A.2
RESET test for equation 01.
Ramsey RESET Test
Equation: 01
Specification: GCOMP MR PLANEXP GENDER AGE C
Omitted Variables: Squares of fitted values
Value df Probability
t-statistic 0.667123 36 0.5089

F-statistic 0.445053 (1,36) 0.5089
Likelihood ratio 0.516045 1 0.4725
F-test summary:
Sum of Sq. df Mean Squares
Test SSR 24.71805 1 24.71805
Restricted SSR 2024.142 37 54.70655
Unrestricted SSR 1999.424 36 55.53956
Unrestricted SSR 1999.424 36 55.53956
LR test summary:
Value df
Restricted LogL − 140.9753 37
Unrestricted LogL − 140.7173 36
Unrestricted Test Equation:
Dependent variable GCOMP
Method: Least Squares
Date: 06/30/2020 Time: 16:00
Sample: 1 42
Included observations: 42
Variable Coefficient Std. Error t-Statistic Prob.
MR 46.62040 48.32071 0.964812 0.3411
PLANEXP 3.144145 4.163104 0.755240 0.4550
GENDER − 26.71351 27.62661 − 0.966949 0.3400
AGE − 1.272228 1.351943 − 0.941037 0.3530
C 183.9881 150.9325 1.219009 0.2308
FITTED^2 − 0.014219 0.021314 − 0.667123 0.5089
R-squared 0.617234 Mean dependent var 78.09524
Adjusted R-squared 0.564072 S.D. dependent var. 11.28740
S.E. of regression 7.452487 Akaike info criterion 6.986536
Sum squared resid 1999.424 Schwarz criterion 7.234775
Log likelihood − 140.7173 Hannan-Quinn criter. 7.077526
F-statistic 11.61044 Durbin-Watson stat 3.027053
Prob (F-statistic) 0.000001
Table A.3
Heteroskedasticity test for equation 01 (BPG).
Heteroskedasticity Test: Breusch-Pagan-Godfrey
F-statistic 1.592135 Prob. F (4,37) 0.1969

Obs*R-squared 6.167574 Prob. Chi-Sqaure (4) 0.1870
Scaled explained SS 3.212285 Prob. Chi-Sqaure (4) 0.5229
Test Equation:
Dependent Variable: RESID^2
Date: 02/18/21 Time: 23:01
Sample: 1 42
C 57.07801 36.15071 1.578890 0.1229
MR − 24.97414 17.92145 − 1.393533 0.1718
PLANEXP 25.92974 18.45947 1.404685 0.1685
GENDER 21.72174 18.10966 1.199456 0.2380
AGE − 0.529473 1.353844 − 0.391089 0.6980
Adjusted R-squared 0.054614 S.D. dependent var. 56.51151
Sum squared resid 111,708.1 Schwarz criterion 11.16881
Prob (F-statistic) 0.196858
14
Table A.4
Heteroskedasticity test for equation 01 (White).
Heteroskedasticity Test: White
F-statistic 0.825959 Prob. F (11,30) 0.6160

Obs*R-squared 9.763018 Prob. Chi-Sqaure (11) 0.5518
Scaled explained SS 5.084916 Prob. Chi-Sqaure (11) 0.9270
Test Equation:
Dependent Variable: RESID^2
Date: 02/18/21 Time: 23:01
Sample: 1 42
Collinear test regressors dropped from specification
C − 96.28102 160.9945 − 0.598039 0.5543
MR^2 100.0678 141.0442 0.709478 0.4835
MR*PLANEXP − 42.40450 42.11503 − 1.006873 0.3220
MR*GENDER 1.445328 44.58816 0.032415 0.9744
MR*AGE − 4.298211 5.528824 − 0.777419 0.4430
PLANEXP^2 46.37534 140.2354 0.330696 0.7432
PLANEXP*GENDER − 29.17956 47.81926 − 0.610205 0.5463
PLANEXP*AGE 0.508110 5.619150 0.090425 0.9286
GENDER^2 90.81293 154.6430 0.587242 0.5614
GENDER*AGE − 2.173174 6.594215 − 0.329558 0.7440
AGE^2 − 0.035661 0.171421 − 0.208029 0.8366
AGE 5.942935 9.442626 0.629373 0.5339
Adjusted R-squared − 0.048981 S.D. dependent var. 56.51151
Sum squared resid 100,499.2 Schwarz criterion 11.68602
References Construction Research Congress 2018, New Orleans, United States, 2018,
pp. 419–428, https://doi.org/10.1061/9780784481264.041.
[15] Y.H. Xu, Z. Fan, L. Hou, H. Guan, C. Mao, Utilising AR HoloLens to enhance user
[1] Autodesk, Architectural Rendering [Online]. URL, https://www.autodesk.com/s
experience in house selling: An experiment study, in: Proceedings of the 17th
olutions/architectural-rendering, February 2020 [Accessed 22 February 2021].
International Conference on Computing in Civil and Building Engineering (ICCCBE
[2] Å. Fast-Berglund, G. Liang, D. Li, Testing and validating extended reality (xR)
2018), Tampere, Finland, 2018, pp. 1–7. URL, https://icccbe2018.exordo.com/file
technologies in manufacturing, Proc. Manuf. (2018) 31–38, https://doi.org/
s/papers/84/final_draft/Utilising_AR_HoloLens_to_Enhance_User_Experience_in_H
10.1016/j.promfg.2018.06.054.
ouse_Selling_An_Experiment_Study_1_.pdf.
[3] E. Miller, S. Paniagua, Get Started with HoloLens (1st gen), Microsoft [Online].
[16] J. Wolfartsberger, Analyzing the potential of virtual reality for engineering design
URL, https://docs.microsoft.com/en-us/hololens/hololens1-hardware, 6 April
review, Autom. Constr. 104 (2019) 27–37, https://doi.org/10.1016/j.
2019 [Accessed 22 February 2021].
autcon.2019.03.018.
[4] P.S. Dusnton, X. Wang, Mixed reality-based visualization interfaces, J. Constr. Eng.
[17] Y. Liu, F. Castronovo, J. Messner, R. Leicht, Evaluating the impact of virtual reality
Manag. (2005) 1301–1309, https://doi.org/10.1061/(ASCE)0733-9364(2005)131:
on design review meetings, J. Comput. Civ. Eng. 34 (2020), https://doi.org/
12(1301).
10.1061/(ASCE)CP.1943-5487.0000856.
[5] M. Kalantari, P. Rauschnabel, Exploring the early adopters of augmented reality
[18] O.A. Sunday, Impact of variation orders on public construction projects, in:
smart glasses: the case of Microsoft HoloLens, Augment. Real. Virtual Real. (2017)
Proceedings of the 26th Annual ARCOM Conference, Leeds, UK, 2010,
229–245, https://doi.org/10.1007/978-3-319-64027-3_16.
pp. 101–110. URL, https://www.arcom.ac.uk/-docs/proceedings/ar2010-0101-0
[6] P. Milgram, H. Colquhoun Jr., A Taxonomy of Real and Virtual World Display
110_Sunday.pdf.
Integration, in Mixed Reality: Merging Real and Virtual Worlds, Springer-Verlag
[19] P. Keane, B. Sertyesilisik, A.D. Ross, Variations and change orders on construction
Berlin Heidelberg, New York, 1999, pp. 5–30, https://doi.org/10.1007/978-3-642-
projects, J. Leg. Aff. Disput. Resolut. Eng. Constr. 2 (2010) 89–96, https://doi.org/
87512-0_1.
10.1061/(ASCE)LA.1943-4170.0000016.
[7] Google, Glass, Google [Online]. URL, https://www.google.com/glass/start/, 2020
[20] A.H. Memon, I.A. Rahman, F.M. Abul Hasan, Significant causes and effects of
[Accessed 09 November 2020].
variation orders in Construction projects, Res. J. Appl. Sci. Eng. Technol. 21 (2014)
[8] Facebook Technologies, Oculus, Facebook [Online]. URL, https://www.oculus.com
4494–4502, https://doi.org/10.19026/rjaset.7.826.
/quest-2/, 2020 [Accessed 9 November 2020].
[21] J.N. Davidson, D.A. Campbell, Collaborative Design in Virtual Space - GreenSpace
[9] B. Bray, What is Mixed Reality?, Microsoft [Online]. URL, https://docs.microsoft.
II: A Shared Environment for Architectural Design Review, Design Computation:
com/en-us/windows/mixed-reality/discover/mixed-reality, 26 August 2020
Collaboration, Reasoning, Pedagogy: ACADIA Conference Proceedings, Tucson,
[Accessed 9 November 2020].
1996, pp. 165–179. ISBN 1-880250-05-5.
[10] A.Z. Sampaio, O.P. Martins, The application of virtual reality technology in the
[22] P.S. Dunston, X. Wang, User perspectives on mixed reality tabletop visualization
construction of a bridge: the cantilever and incremental launching methods,
for face-to-face collaborative design review, Autom. Constr. 17 (2008) 399–412,
Autom. Constr. 37 (2014) 58–67, https://doi.org/10.1016/j.autcon.2013.10.015.
https://doi.org/10.1016/j.autcon.2007.07.002.
[11] S.S.-C. Chen, H. Duh, Interface of mixed reality: from the past to the future, CCF
[23] I. Hong, K. Bong, D. Shin, S. Park, K. Lee, Y. Kim, H.-J. Yoo, 18.1 A 2.71nJ/pixel
Trans. Pervas. Comput. Interact. 01 (2019) 69–87, https://doi.org/10.1007/
3D-stacked gaze-activated object-recognition system for low-power mobile HMD
s42486-018-0002-8.
applications, in: 2015 IEEE International Solid-State Circuits Conference - (ISSCC)
[12] S.-C. Kang, H.-L. Chi, X. Wang, Research trends and opportunities of augmented
Digest of Technical Papers, San Francisco, United States, 2015, pp. 1–3, https://
reality applications in architecture, engineering, and construction, Autom. Constr.
doi.org/10.1109/ISSCC.2015.7063058.
33 (2013), https://doi.org/10.1016/j.autcon.2012.12.017, 116-112.
[24] Facebook Technologies, Oculus/Support: What is the Passthrough for Oculus
[13] T. Majumdar, M.A. Fischer, B.R. Schwelger, Conceptual design review with a
Quest? Facebook, 2020 [Online].URL, https://support.oculus.com/609948
virtual reality mock-up, in: CIB W078 23ndJoint International Conference on
326077538/#faq_609948326077538.
Computing and Decision Making in Civil and Building Engineering, in-house
[25] G. Evans, J. Miller, M. Iglesias Pena, A. MacAllister, E.H. Winer, Evaluating the
publishing, Rotterdam, 2006, pp. 2902–2911. URL, http://www.irbnet.de/daten/
Microsoft HoloLens through an augmented reality assembly application, in: SPIE
iconda/CIB21113.pdf.
Defense + Security, Anaheim, United States, 2017, pp. 1–16, https://doi.org/
[14] D. Paez, J. Irizarry, A Usability Study of an Immersive Virtual Reality Platform for
10.1117/12.2262626.
Building Design Review: Considerations on Human factors and User Interface,
15
[26] S.A. Yazdanfar, A.A. Heidari, N. Aghajari, Comparison of architects’ and non- [37] C. Arun, Knowledge-Based Decision Support Tool for Duration and Cost Overrun
architects’ perception of place, Proc Soc. Behab. Sci. 170 (2015) 690–699, https:// Analysis of Highway Construction Projects, Proceedings of the Inaugural
doi.org/10.1016/j.sbspro.2015.01.071. Construction Management and Economics ‘Past, Present and Future’ conference,
[27] K. Lynch, The Image of the City, The M.I.T. Press, Cambridge, Massachusetts, 1960. Reading, UK, 2007, pp. 1713–1722. ISBN:9780415460590.
ISBN: 0 262 62001 4. [38] J.M. Assbeihat, G.J. Sweis, Factors affecting change orders in public construction
[28] M.D. Osorto Carrasco, Application of Mixed Reality for improving Architectural projects, Int. J. Appl. Sci. Technol. 5 (2015) 56–63. ISSN: 2221-1004.
Comprehension Effectiveness (Master´s Thesis), National Taiwan University, Taipei, [39] A.S. Alnuaimi, R.A. Taha, M.A. Mohsin, A.S. Al-Harthi, Cuases, effects, benefits,
2020, https://doi.org/10.6342/NTU202002391. and remedies of change orders on public construction projects in Oman, J. Constr.
[29] Y. Huang, S. Shakya, T. Odeleye, Comparing the functionality between virtual Eng. Manag. 136 (2010) 615–622, https://doi.org/10.1061/(ASCE)CO.1943-
reality and mixed reality for architecture and construction uses, J. Civ. Eng. 7862.0000154.
Architect. 13 (2019) 409–414, https://doi.org/10.17265/1934-7359/ [40] M. Carmona, S. Tiesdell, T. Heath, T. Oc, Public Places, Urban Spaces: The
2019.07.001. Dimensions of Urban Design, Second ed., Architectural Press, Burlington, United
[30] Trimble, SketchUp Viewer: Interacting with a Model in Tabletop View, SKetchUp, States, 2003, pp. 111–132. ISBN-13: 978-1-85617-827-3.
2020 [Online]. URL, https://help.sketchup.com/en/sketchup-viewer/interacting- [41] W.H. Ittelson, Environmental perception and urban experience, Environ. Behav. 10
model-tabletop-view [Accessed 10 April 2020]. (1978) 193–213, https://doi.org/10.1177/0013916578102004.
[31] K.L. Barriball, A. While, Collecting data using a semi-structured interview: a [42] O. Ranjbar Pouya, A. Byagowi, D. Kelly, Z. Moussavi, The effect of physical and
discussion paper, J. Adv. Nurs. 19 (1994) 328–335, https://doi.org/10.1111/ virtual rotations of a 3D object on spatial perception, in: 6th International IEEE/
j.1365-2648.1994.tb01088.x. EMBS Conference on Neural Engineering (NER), San Diego, CA, 2013,
[32] R. Longhurst, Semi-structured interviews and focus groups, in: N. Clifford, pp. 1362–1365, https://doi.org/10.1109/NER.2013.6696195.
S. French, G. Valentine (Eds.), Key Methods in Geography, Sage Publications Ltd, [43] A. Heydarian, J.P. Carneiro, D. Gerber, B. Becerick-Gerber, T. Hayes, W. Wood,
London, 2016, pp. 143–156. ISBN-13: 978-1412935098. Immersive virtual environments versus physical built environments: a
[33] A. Plazola Cisneros, A. Plazola Anguiano, G. Plazola Anguiano, Edificios de oficina, benchmarking study for building design and user-built environment explorations,
in: Enciclopedia de Arquitectura Plazola vol. 8, 1996. Plazola Editores, Mexico Autom. Constr. 54 (2015) 116–126, https://doi.org/10.1016/j.
City. ISBN-13: 978-9687478074. autcon.2015.03.020.
[34] Merriam-Webster, Inc, Dictionary Entry: Effective, Merriam Webster, 2020 [44] Microsoft, HoloLens 2: Get to know the New Features and Technical Specs
[Online]. URL, https://www.merriam-webster.com/dictionary/effective [Accessed [Online]. URL, https://www.microsoft.com/en-us/hololens/hardware, 6 April
13 November 2020]. 2020.
[35] M. Perry, D. Sanderson, Coordinating Joint Design Work: The Role of [45] A. Schmelter, P. Jansen, M. Heil, Empirical evaluation of virtual environment
Communication and Artefacts, Design Studies 19, 1998, pp. 273–288, https://doi. technology as an experimental tool in developmental spatial cognition research, in,
org/10.1016/S0142-694X(98)00008-8. Eur. J. Cogn. Psychol. 21 (2009) 724–739, https://doi.org/10.1080/
[36] J. Verlinden, I. Horváth, T.-J. Nam, Recording augmented reality experiences to 09541440802426465.
capture design reviews, Int. J. Interact. Des. Manuf. (IJIDeM) 3 (2009) 189–200,
https://doi.org/10.1007/s12008-009-0074-8.
16

C02 - Carrasco y Chen

Uploaded by

Copyright:

Available Formats

C02 - Carrasco y Chen

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

C02 - Carrasco y Chen

Uploaded by

Copyright:

Available Formats

Automation in Construction 126 (2021) 103677

Contents lists available at ScienceDirect

Application of mixed reality for improving architectural design

Fig. 1. Differences between AR, MR and VR. Reference: [6].

3.7.4. The questionnaire

Fig. 5. Interior space with material specifications.

Treatment group: 2D (Control)

accurate representation of the intended material.

t-statistic 0.667123 36 0.5089

F-statistic 1.592135 Prob. F (4,37) 0.1969

F-statistic 0.825959 Prob. F (11,30) 0.6160

You might also like