Perception of a Spatial Implausibility Caused by Seamless Covert

Teleportation

Mathieu Lutfallah

, Dylan Cernadela Pires, Valentina Gorobets

and Andreas Kunz

Innovation Center Virtual Reality, ETH Zurich, Switzerland

Keywords:

Locomotion, Walking, Redirection, Human Perception.

Abstract:

This paper investigates human perception of spatial implausibility through seamless teleportation in circular

and hexagonal closed-loop corridors. Spatial implausibility refers to virtual spaces that deviate from New-

tonian Euclidean geometry rules, while seamless teleportation involves changing the user’s position in the

virtual world without their awareness. In this work, the impression of implausibility is generated by subtle

teleportation of the user within these corridors, thereby allowing them to unconsciously skip certain sections of

them. Different levels of this “implausibility” are presented by varying the percentage of skipping within these

corridors, speciﬁcally 0%, 15%, and 30% of the corridor’s overall length. These implausibilities are assessed

through a within-subject study on naive participants to determine their perception of spatial implausibility. Our

ﬁndings indicate no signiﬁcant difference in the detection rates between the two corridor shapes. Interestingly,

most participants interpreted the manipulation as a change in the environment’s shape or size while only few

could perceive the teleportation and the skip. This paves the way for future research to leverage this technique

for subtle spatial manipulation.

1 INTRODUCTION

A crucial and integral aspect of any virtual reality

(VR) experience is the locomotion technique used

to navigate virtual environments (VEs). Immersive

and engaging VEs come to life through the seam-

less movement and exploration through effective lo-

comotion techniques. By providing users with intu-

itive means to traverse and interact with VEs, loco-

motion systems enhance the sense of presence and

empower individuals to fully immerse themselves in

these VEs. From walking-in place (Wendt et al.,

2010) and on treadmills (Iwata, 1999) to point-and-

teleport methods, joystick navigation, and real walk-

ing approaches, a wide range of locomotion tech-

niques have emerged, catering to diverse user pref-

erences and optimizing the overall VR encounter.

Among these techniques, teleportation remains one of

the most commonly used. It can be achieved in var-

ious ways, allowing the user to either translate or to

both translate and rotate. Additionally, teleportation

can be employed with or without the user’s knowl-

https://orcid.org/0000-0001-7863-8889

https://orcid.org/0000-0002-8615-5972

https://orcid.org/0000-0002-6495-4327

edge. For example, the user might be transported by

a bird to a new location or moved to another section

of the corridor that appears identical to their current

position, without realizing it. In such cases, while the

immediate surroundings remain the same, everything

beyond the user’s current ﬁeld of vision is different.

This method would constitute seamless teleportation.

Real walking, as a locomotion technique, stands

out for its potential to enhance presence and immer-

sion in the VE (Usoh et al., 1999; Cherni et al., 2020)

and provides a more simple, straightforward and nat-

ural user experience compared to other locomotion

techniques. It allows users to physically move within

the virtual space, providing a high level of control

precision and minimizing the occurrence of motion

sickness. However, real walking has one signiﬁcant

drawback, since it requires a considerable amount of

physical space, making it impractical for small track-

ing areas commonly found in home-based VR setups.

Here, redirection techniques come into place that al-

low using smaller physical environments for navigat-

ing large VEs.

Redirection Techniques: take advantage of human

perceptual limitations and enable users to navigate

large VEs while keeping them within the bounds of

their tracking space. These techniques can be ap-

380

Lutfallah, M., Pires, D., Gorobets, V. and Kunz, A.

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation.

DOI: 10.5220/0012573500003660

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 19th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2024) - Volume 1: GRAPP, HUCAPP

and IVAPP, pages 380-389

ISBN: 978-989-758-679-8; ISSN: 2184-4321

plied either to the user or to the VE. For example,

(Langbehn et al., 2020) seamlessly integrated redi-

rection into a game, where the world was turned

around the players while they were concentrating on

a task, i.e. making whipped cream. Various redirec-

tion techniques exist, and efforts have been made to

classify them. In the taxonomy proposed by (Suma

et al., 2012a), redirection techniques are separated

based on their geometric ﬂexibility versus the like-

lihood that they will be noticed by the users. Redi-

rection techniques are divided into repositioning and

reorientation. Repositioning techniques modify the

mapping from points in the real world to the virtual

world to condense a larger VE into a compact physi-

cal environment. Reorientation techniques rotate the

user’s heading away from the boundaries of the phys-

ical space. Concerning the noticeability, the taxon-

omy distinguishes between subtle and overt methods.

Finally, redirection techniques can be implemented

in a discrete or continuous manner. Discrete tech-

niques are applied instantaneously, while continuous

techniques are applied over time. Another taxon-

omy was presented by (Vasylevska and Kaufmann,

2017a) where they distinguish 4 categories of algo-

rithms “Basic Reorientation”, “Rendering Manipula-

tion”, “Sense Manipulation”, and “Scene Manipula-

tion”.

Redirection of a user can be achieved by tech-

niques that modify the VE. Such techniques rely on

impossible spaces that defy conventional logic but

also challenge Newton’s laws of motion and the fun-

damental principles of geometry. An example of

such spaces are overlapping architecture (OA) is a

technique that optimizes tracking space utilization

by overlapping spaces. One example are overlap-

ping neighboring rooms where the goal is to create

the illusion that these rooms occupy separate areas

while, in reality, they share the same physical space.

The overlap occurs when the user traverses a corri-

dor and the wall separating the two rooms shifts. The

maximum allowable overlap before users perceive it

was explored by (Suma et al., 2012b). Additionally,

“Change Blindness” (Suma et al., 2011) refers to the

technique of altering the position of objects within a

space when the user is not noticing it. This allows for

the manipulation of the space and enables the reuse of

the physical area for various segments of the VE.

Prior studies investigating the detection of redirec-

tion thresholds (Suma et al., 2012b; Vasylevska and

Kaufmann, 2015; Steinicke et al., 2010) have primar-

ily involved non-naive participants who were already

familiar with the concept of the redirection applied.

Moreover, existing research papers that analyze the

detection of impossible spaces by the users have pre-

dominantly used a speciﬁc implementation involving

two overlapping rooms. However, there is a lack of re-

search exploring alternative methods relying on seam-

less teleportation to generate such spatial implausibil-

ities. Lastly, the studies that have incorporated por-

tals and teleportation (Koltai et al., 2020; Lochner and

Gain, 2021) to create spacial implausibility have not

examined how users perceive these spaces, particu-

larly in terms of perception or detection.

In this work, we focus on the perception and cog-

nitive map in a new type of spatial implausibilities

where the users skips certain parts of the VE. As

shown in Figure 1, omitting speciﬁc segments of a

corridor enables spatial manipulation, thereby reori-

enting the VE. This technique allows moving por-

tions of the VE previously outside the physical space

boundary into it. This is in contrast to previous imple-

mentations of portals which allowed reusing the same

physical space for different segments of the VE. This

is highlighted in Figure 1 by showing that the small

room shown by the small rectangle could have been

colliding with an obstacle but after the reorientation

the room could be explored. The contributions of this

paper are as follows:

• Presentation of a new type of subtle reorientation

technique relying on corridors and teleportation,

• Investigation of the perception of the users of such

a technique based on various levels of implausibil-

ity and different layouts.

Figure 1: Illustration of spatial manipulation using portals,

by teleporting the user in the corridor between the two rect-

angles (left image) the environment changes orientation as

seen in yellow (right image).

2 RELATED WORK

2.1 Perception for Scene Manipulation

Techniques

A well-known concept of generating spatial implausi-

bilities is named “impossible spaces”, which was in-

troduced by (Suma et al., 2012b). Here the threshold

metric was the overlap between neighboring rooms.

The study concluded that reasonably small virtual

rooms could overlap by up to 56% before users de-

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation

381

tected it. Moreover, based on a qualitative experiment

it was observed that participants who are naive to the

manipulation of impossible spaces tend to experience

a more compelling illusion. Other researchers took

that same spatial compression approach and did fur-

ther research on the inﬂuence of path complexity in

overlapping space detection (Vasylevska and Kauf-

mann, 2015). The research gap they addressed was

the lack of knowledge on the effect of different virtual

architectural layouts on the perception of such over-

lapping spaces. There ﬁndings suggest that path com-

plexity is a variable that has to be taken into consider-

ation when trying to optimize spatial compression. In

a later study by (Vasylevska and Kaufmann, 2017b),

they explored the inﬂuence of layout properties on

the perception of impossible spatial arrangements.

Their study aimed to analyze various parameters as-

sociated with the path connecting two overlapping

rooms. These parameters encompassed the “number

of turns, relative door positions, sequences of clock-

wise and counterclockwise turns, as well as symme-

try and asymmetry”. Among other results, they found

that the absence of landmarks made it harder for par-

ticipants to orient themselves in space. This in turn

is more beneﬁcial for spatial compression. A more

recent study (Eppl

ee and Langbehn, 2022) examined

the effects of using a minimap during exploration of

a VE featuring impossible spaces. (Ciumedean et al.,

2021), on the other hand, tested the effects of different

layouts in impossible spaces: “open rooms”, where

rooms are not separated by walls but only by small

objects to prevent direct crossing from one room to

another without using the corridor; “open corridors”,

characterized by partial walls that allow the user to

view the outside space; and “total inclusion”, the stan-

dard condition. The focus was on how these layouts

impact the detection threshold.

2.2 Teleportation as a Locomotion

Technique

Various studies (Freitag et al., 2014; Langbehn et al.,

2018) compared the use of teleportation as a loco-

motion technique in navigating virtual environments.

Another work by (Langbehn et al., 2018) compared

teleportation with redirected walking techniques, fo-

cusing on cognitive map building, cybersickness, and

user preferences. In their studies, teleportation is ini-

tiated through deliberate user action, as opposed to

being seamlessly integrated. Other research (Bruder

et al., 2009) has explored combining teleportation

with techniques like portals and redirected walking

enhancing the exploration in large virtual environ-

ments. Teleportation can be implemented in two

distinct ways: partial concordance, where it is used

solely for translational movements, and full discor-

dance, where it is employed for both rotation and

translation. The effects of these methods have been

investigated on spatial updating and piloting (Kelly

et al., 2020). Furthermore, a detailed literature re-

view is provided in (Prithul et al., 2021), comparing

teleportation techniques with alternative locomotion

methods. This review also encompasses a discussion

of various approaches to implementing teleportation.

2.3 Spatial Implausibility Using

Seamless Teleportation

Another interesting way of implementing spatial im-

plausibility is by changing the user’s position through

instantaneous and seamless teleportation. The work

by (Koltai et al., 2020) presented a method for creat-

ing scaleable overlapping architecture in VR by “pro-

cedurally generating tile-based mazes that seamlessly

teleport the user using portals”. To implement this

unnoticeable teleportation, they used portals on spe-

ciﬁc locations in the virtual maze. These portals con-

sisted of a rendered plane on which the corresponding

part of the next maze element was displayed. When

the users crossed the portal threshold they were tele-

ported to the maze element that was displayed on the

portal itself. Like this, the users could not notice the

teleportation and an illusion of continuity was cre-

ated. Similarly, (Overdijk, 2023) utilized the same

method, combining it with the game’s narrative to

guide the user, enabling continuous walking in the

virtual world. In the same fashion, portals were intro-

duced by (Lochner and Gain, 2021). They analyzed

different locomotion techniques in impossible spaces,

and used portals and teleportation. In their work, they

used pairs of portals to travel between different sub-

spaces of the VE. The portal pairs all had the same lo-

cal position and orientation so that the user’s rotation

would not need to be altered when teleporting. Only

the users’ position had to be translated to the corre-

sponding subspace. Through this, they were able to

compress several virtual rooms into the space of one

for their experiments.

The aforementioned studies primarily concen-

trated on users’ perception of implausible spaces

while navigating spaces featuring impossible geome-

tries and change blindness. In contrast, other re-

search has focused on teleportation as the primary lo-

comotion technique. However, only a few of these

have seamlessly integrated teleportation. These in-

stances primarily utilized teleportation to overlap vir-

tual spaces rather than to manipulate them. Seamless

teleportation could enable the alteration of the VE’s

HUCAPP 2024 - 8th International Conference on Human Computer Interaction Theory and Applications

382

(a) Circular Corridor. (b) Hexagonal Corridor.

Figure 2: Illustration of the two closed-loop corridor lay-

outs.

orientation in relation to the physical space. This gap

in research has motivated the current work, where

we explore combination of teleportation and corri-

dors and the noticeability of implausible spaces. This

research will explore how different corridor shapes,

speciﬁcally circular and angular with corners, affect

the user’s cognitive map. Additionally, it will ex-

amine the user’s perception of seamless teleportation,

particularly how skipping certain portions of the cor-

ridors inﬂuences this perception.

3 IMPLEMENTATION

Since we want to investigate how users would detect

the implausibility of the space relying on teleporta-

tion, we decided to use two closed-loop corridors in

which participants could walk. One corridor had a

circular shape, and the other a hexagonal one. The

corridors each consisted of a white inner and outer

wall with grey ground and ceiling. The decision to

intentionally omit textures in the scenes was made in

order to avoid visible landmarks, minimize distrac-

tions, and thereby reduce the number of variables that

needed to be taken into account. The size of the corri-

dors was chosen to ﬁll the available physical tracking

space (7m × 9m) with a height of 2.5m. This leads

the corridors to have a width of 1.1m. We also added

a start sign together with a red arrow, positioned on

the inner wall, as a reference point for the users and

to indicate the direction the participant would need to

walk. The sign is also the cue for the user that he is

back to the initial position. Figure 2 shows the corri-

dors with the ceiling removed.

3.1 Seamless Teleportation Design

To create an implausible space within the corridors,

we implemented portals in conjunction with telepor-

tation. The idea was to seamlessly teleport the users

walking in the corridors from one point in the corri-

dor to another point in the corridor. This has the effect

that users skip a section of the corridor and therefore

reach the starting position faster, which would not be

possible in a real-world scenario. The objective of this

approach is to make users believe that they are walk-

ing along the entire corridor while, in reality, they are

skipping a certain section of it. If the users do not no-

tice this illusion, then the implausibility of the space

is successfully unnoticed and the orientation of the

physical space is changed. We included three levels

of teleportation for each corridor shape which would

later represent different experiment conditions. The

ﬁrst condition is the Zero condition where the users

are not teleported at all while walking along the cor-

ridor. We represent these conditions as C0 and H0,

where C represents the circular corridor, H represents

the hexagonal corridor and 0 represents that no sec-

tions have been skipped in the corridor. The follow-

ing conditions use the same terminology. In condi-

tions C1 and H1 a sixth of the corridor is skipped

and in conditions C2 and H2 a third of the corridor

is skipped.

To ensure seamless teleportation, we employed

instantaneous repositioning of the user. We imple-

mented a collision plane located at the end of the ﬁrst

section of the corridor. When the user walks through

this plane, it triggers the repositioning of the player.

The corridor the user is teleported to was already ro-

tated to a speciﬁc degree around its z-axis. The ad-

vantage of using a pre-rotated corridor as the destina-

tion instead of teleporting the player within the same

corridor lies in the fact that we only need to trans-

late the player’s position, without requiring a change

in rotation. Implementing a rotation change of the

player would be considerably more challenging. Con-

sequently, this means that in conditions C1 and H1 the

corridor where the user is teleporting to is rotated by

◦

and for conditions C2 and H2 it is rotated by 120

◦

illustrated in Figure 3.

Portals are generally 2D planes on top of which

the VE is rendered. They are usually placed on walls

or corridor endings where the architectural overlap

of the VE would happen. Consequently, by walk-

ing through the portals, together with teleportation, a

seamless transition from one VE to another is created.

Users perceive this as exploring a new VE, yet they

are reusing a space of the physical tracking space.

We used stencil buffers to create portals which ren-

der parts of a VE that we want to see while hiding the

parts that we do not want to see. The stencil buffer

is a dedicated buffer within the shader that can de-

cide if a pixel of an object is drawn or not. To de-

cide which pixels are to be drawn and which are not,

we can assign certain reference values to objects us-

ing the stencil buffer. By doing a so-called “stencil

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation

383

(a) No skip. (b) Skip of 16.7%.

Figure 3: Implausible space implementation for the hexag-

onal corridors. The green bars indicate the position of the

portals. The patterned sections specify the possible spatial

compression achieved by each condition. The same princi-

ples apply to the circular corridors.

test” the stencil buffer then compares the values of

two objects. If the values coincide the pixels of the

object can be chosen to be drawn, if not the pixels are

discarded. This technique helps us create seamless

transitions between different VEs. To implement the

stencil buffer portal we ﬁrst need to create two sepa-

rate VEs within the same scene in Unity. One is the

VE where our user will be starting and the other is the

VE where he will be teleported to.

3.2 Hardware and Software

The VE was developed using the Unity engine, and

the Meta Quest 2 was used as the headmounted dis-

play (HMD) for presenting the VE during the user

study. The HMD utilizes an LCD panel with a res-

olution of 1832 × 1920 per eye at a frame rate of up

to 120Hz, featuring a horizontal ﬁeld of view (FOV)

of around 104

◦

and a vertical FOV of 98.2

◦

. To pre-

vent participants from peeking through the gap be-

tween their nose and the HMD we installed a little

paper ﬂap. The reason was that we wanted to prevent

the users to be able to orient themselves on the track-

ing space. The VE was streamed from a PC through

SteamVR, the PC had a GeForce RTX3090 graphics

card paired with a 10

generation Intel Core i9 CPU.

To develop our VE and to conduct our user study,

we utilized the Unity game engine (editor version

2019.3.19f1) together with the SteamVR plug-in,

which enables us to make VR applications which can

be directly run on the PC and also streamed to other

HMD, like in our case. Since Unity is limited in its

ability to create custom shapes and 3D objects we uti-

lized the Unity package called ProBuilder. This pack-

age is a 3D modelling and design tool which allows

users to create simple geometries

4 USER STUDY PROCEDURE

Prior to developing our user study, it was necessary

for us to determine the hypotheses we aimed to vali-

date. This allowed us to design our user study in ac-

cordance with these hypotheses. We decided on the

following hypotheses:

• Hypothesis 1: Implausible space detection in-

creases if more sections of a closed-loop corridor

are skipped,

• Hypothesis 2: Implausible space detection is

higher for closed-loop corridors with sharply

curved paths compared to smooth curved paths.

Our ﬁrst hypothesis is straightforward since it is nat-

ural to expect users to notice the implausible space if

the changes are more drastic. In our second hypothe-

sis, we assume that sharp corners can be used as land-

marks (Vasylevska and Kaufmann, 2017b) to help the

users to orient themselves in space and thus help them

understand the spatial impossibility induced by the

teleportation.

Participants for both studies were recruited from

the university community. The sole criterion for par-

ticipation was the ability to walk, ensuring a broad

and accessible participant pool. Prior to commenc-

ing the experiments, we refrained from disclosing the

objective of the user studies in order to prevent par-

ticipants from actively seeking hints about the redi-

rection technique. Each participant was required to

provide their informed consent by signing a form.

Subsequently, the user began by completing question-

naires on demographics and the Simulator Sickness

Questionnaire (SSQ). They were then introduced to

the hardware in the SteamVR room, where they could

walk around and familiarize themselves with the sys-

tem. The user study procedure is shown in Figure 4.

The study then commenced, and the VE featuring

the corridors was launched. The participants started at

the start position in the corridor and were then told to

walk along it until they reach the start position again

shown by the start sign. Following each corridor ex-

perience, participants were presented with a simple

transition scene consisting of a wooden ﬂoor and sky.

They were instructed to follow a ﬂoating hand on

the screen, controlled by the experimenter, until in-

structed to stop. This procedure was conducted while

https://unity.com/features/probuilder, accessed on

02.07.2023

HUCAPP 2024 - 8th International Conference on Human Computer Interaction Theory and Applications

384

the participant wore the head-mounted display. The

dual purpose of this approach was to disrupt the par-

ticipant’s spatial orientation within the tracking en-

vironment and prevent them from seeing their ﬁnal

position in the physical space. Additionally, it facili-

tated repositioning the participant to the starting point

for each corridor iteration. This procedure was con-

sistently repeated for each corridor. Each participant

had in total two series of walking. In each series, the

participant had to walk through three different corri-

dor conditions of the same corridor shape. Half of

the participants started with the circular corridors and

half of them started with the hexagonal corridors. Af-

ter each series, the participant had to execute a draw-

ing task where he had to draw the virtual environment

where he walked. In the walking series, the sequence

of skip levels within the corridors was predetermined

in a counterbalanced order to ensure that each con-

dition occurred uniquely. Consequently, some users

might ﬁrst experience a 33% skip, while others could

begin with no skip or a 16% skip. The purpose behind

this was to prevent learning effects and to be able to

analyze different corridor conditions.

After the study, the true purpose of the user study

was revealed, and we proceeded to explain the par-

ticipants how the corridors operated and provided an

explanation of the potential of spatial implausibil-

ity techniques. Subsequently, users were requested

to provide informal feedback regarding their impres-

sions of the manipulation.

4.1 Drawing Task

In the drawing task, participants had to draw the ﬂoor

plans of the three corridors they have just walked

along. Additionally, they had to indicate on each cor-

ridor drawing the “START” position and to try draw-

ing the corridors to scale. Drawing to scale meant

that if they perceived certain corridors to be smaller or

larger than others, they had to indicate it by drawing

larger or smaller corridors or alternatively describe it

with words. It is important to state that participants

were only told about the drawing task after walking

the corridor to prevent participants from focusing un-

naturally much on the corridor layout. The imple-

mentation of a counterbalanced user study, where par-

ticipants were equally exposed to commencing with

either the hexagonal or the circular condition, con-

tributed to neutralizing potential biases of the results.

Participants were instructed to draw the corridor ﬂoor

plans into boxes by the order they walked them. The

boxes contained a grid to help the participants with

their drawings and to make it easier to draw them to

scale. To extract meaningful data from the drawings

of the participants, we evaluated the drawings by two

measures and compared them to the actual condition

they were supposed to represent.

In order to assess whether participants detected

the implausibility of the spaces, we searched for corri-

dors that were either not closed or had additional start

or stop points drawn. Drawing an non-closed corridor

indicated that the users were aware that they have not

walked a full round. The same holds true for draw-

ing two separate start signs since changing the start-

ing sign to a different section in the corridor would

have the same effect as teleporting the player. Exam-

ples of non-closed-loop corridor drawings and dou-

ble start positions can be seen in Figure 5. The other

measure we evaluated was if the participant drew the

corridors within a series in different sizes or shapes

(SoS), compared to each other. Our reason was that

if they did so, they have not detected the implausible

space but they still detected a change going from cor-

ridor to corridor. Drawings, where only one change

in SoS occurred, were also counted as a detection. If

no change in SoS was detected, we counted it as no

detection.

5 RESULTS

From the 25 participants enrolled in our user study, 16

(64%) were male, while 9 (36%) were female. The

age distribution spanned from 16 to 30 years, with an

average age of 21.76 ± 2.69 years. The participants

consisted of 20 students and 5 non-students. More

than half of the participants (52%) reported having

between zero and ﬁve hours of previous VR experi-

ence. Six participants had no experience, 3 partici-

pants had between 20 and 100 hours of experience,

2 participants had between 5 and 20 hours of experi-

ence, and 1 participant had over 100 hours of experi-

ence. The mean SSQ total difference prior and after

the user study was 1.65 ±9.58. Analysis of the SSQ

responses, administered both before and after the user

study, indicated no signiﬁcant variance among partic-

ipants. Consequently, no participant required exclu-

sion from the study.

5.1 Drawing Task Results

From all participants, four did neither detect the im-

plausibility of the space space nor a change in the SoS

of the corridors. Two participants noticed both, the

implausible space and a change in SoS. The rest of

the participants detected at least either of those.

The overall results from the user study regarding

the detection of the implausibility of the space are de-

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation

385

Figure 4: Procedure of the user study testing detection thresholds for a novel type of impossible spaces. Green boxes indicate

questionnaires for participants, and red boxes denote drawing tasks.

(a) Non-closed corridor. (b) Multiple start positions.

Figure 5: Drawings (a) and (b) count as detection of the

teleportation, while (c) counts as no detection.

picted in Figure 6. This chart shows the detection

rates for each condition (C0, C1, C2, H0, H1, and

H2). The blue columns depict which percentage of

the participants detected the space implausibility, and

the red columns represent the percentage that did not

notice this. The main assumption is that if the user

doesn’t perceive the implausible space, they believe

that they reached the same physical space location af-

ter each navigation task in the corridor.

Hypothesis 1 investigated whether the implausi-

ble space detection increases if larger sections of the

corridors are skipped. In the comparison of detection

rates, a pattern was noted: in conditions C1 and H1,

participants detected the implausible space 3 and 1

times, respectively. In contrast, for conditions C2 and

H2, these numbers increased to 7 and 4 detections,

respectively. However, the application of McNemar’s

test did not yield statistically signiﬁcant evidence (p

Figure 6: Column chart representing the detection of the

space manipulation for each condition. “Yes” represents the

corridor drawings where an implausible space was detected

and “No” represents the drawings it was not detected. For

each corridor condition, n=25 drawings were analyzed.

= 0.25 and 0.125, respectively) to justify the rejection

of the null hypothesis which indicates no difference

in detection of implausibility in terms of the degree

of skip. Thus the hypothesis that increasing the level

of implausibility would make the user more prone to

notice the seamless teleportation effect cannot be ac-

cepted.

Hypothesis 2 investigated whether corridor lay-

outs inﬂuence a user’s ability to detect seamless tele-

portation and alters their cognitive map of the space.

To examine this, we employed McNemar’s test for

repeated measures to assess differences in detection

between the conditions C0,H0; H1,C1; and C2,H2.

However, the analysis faced a limitation due to the

low number of participants who successfully detected

teleportation or a distortion in the spatial layout. Con-

sequently, the p-values obtained (0.375, 0.25, 0.625)

for all three comparisons were signiﬁcantly high.

This outcome negates the necessity for applying a

Bonferroni correction, as the p-values inherently in-

dicate a lack of statistical signiﬁcance in the differ-

ence between the corridor layouts. This ﬁnding does

not align with the previously presented research, af-

ﬁrming the potential inﬂuence of corner landmarks in

HUCAPP 2024 - 8th International Conference on Human Computer Interaction Theory and Applications

386

aiding user’s navigation and layout identiﬁcation.

Participants faced challenges in discerning the

condition without a skip. This is evidenced by 7 out

of the 25 participants identifying an implausible space

in the C0 condition and four in the H0 condition, in

contrast to 3 and 1 participant in the C1 and H1 con-

ditions, respectively. This initial observation under-

scores the success of the seamless teleportation effect.

This ﬁnding was not expected before the realisation of

the study. Out of the 11 implausible space detections

in the zero conditions, 9 of them (82%) started with

either a one-section skip (C1 or H1) or a two-section

skip (C2 or H2). This leads us to believe that those

participants might have taken the ﬁrst corridor as a

reference and compared the other two corridors to it.

5.2 Change in Size and Shape Detection

Since 21 (84%) participants out of 25 did not notice

the spatial manipulation by itself but instead noticed

a change in SoS, we also examined this for all pat-

terns. For both, the hexagon- and the circular-shaped

corridors, 17 participants detected a change in the size

or shape of the environment while 8 did not. This was

expected since they could compare the trials with each

other. However, this supports the previous ﬁnding that

corners did not help in better perceiving the telepor-

tation. However, it should be noted here that such a

detection in the hexagonal shape means the user per-

ceives less turns or corners thus changing the layout

he draw from a hexagonal to a square or pentagon de-

spite the fact that all the angles at the corner are still

equal to 120

◦

. While for the circular corridors, the

change in size refers to perceiving the same layout

a full circle which is smaller thus the cognitive map

would be the same.

To check for the sequence effect on the detection,

we combined the skips into two categories like C0 to

C1 as “1 Section Different” and larger changes like

from C0 to C2 as “2 Sections Different”. There was

a total of 61 “1 Section Different” corridor changes

and 39 “2 Sections Different” corridor changes. This

difference can be explained by the higher chance of

having “1 Section Different” corridor changes since

there are simply more possible combinations. There

is only a small difference in the detection ratios be-

tween “1 Section Different” (49% detected) and “2

Sections Different” (54% detected).

The primary implication of this study is that the

proposed scene manipulation does not cause sickness

or discomfort in users, unlike other similar methods

such as redirected walking. Since users had difﬁ-

culties identifying when the manipulation occurred,

this highlights the effectiveness of the implementa-

tion, suggesting that such scene manipulations are un-

likely to cause discomfort. Another key ﬁnding is the

potential to preserve the overall cognitive map layout

of the environment, despite distortions caused by tele-

portation. This was particularly evident as most users

still perceived a circle, even though parts of it were

skipped. In contrast, in the hexagon experiments,

the change was perceived as fewer turns, altering the

hexagon into a pentagon or square, implying a change

in shape. Therefore, these insights enable us to deter-

mine where this method should be applied, based on

the need to maintain an accurate cognitive map.

6 LIMITATIONS

Our user study includes two separate rounds of walk-

ing, where after each round the participant had to

perform the drawing task. Because of this sequen-

tial approach, the participants might have experienced

a learning effect since in the second walking round

most anticipated that there will be another drawing

task. Due to the limited number of participants, this

could not be avoided. From of the 75 corridor draw-

ings in each round, we see that in the ﬁrst round

of the walking series 12 drawings were attributed to

a correct spatial manipulation detection (16%). In

the second round, it was 21 drawings (28%). This

is an increase of 12% in the second round. By do-

ing the McNemar’s exact test, we get a one-tailed p-

value of p=0.057 which is very close to being a sig-

niﬁcant difference and might indicate a correlation.

This supports that future studies using the drawing

method should focus on having one condition per user

to avoid the learning effect. Another limitations of the

drawing task is that users stated in the informal feed-

back that participants had to remember three corridor

paths at once. During our study, some participants re-

ported not being sure in which sequence the corridor

conditions appeared. By decreasing the number of

corridors a participant has to walk this problem could

probably be mitigated.

7 CONCLUSION

In this work, we created a new architectural layout

which implemented spatial manipulation using por-

tals and teleportation. Our architectural layout con-

sisted of closed-loop circular and hexagonal corri-

dors, in which users skipped up to 33% of the path

through teleportation. To collect the necessary data,

we developed a user study where the main task was

for the participants to walk through corridors with dif-

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation

387

ferent levels of space manipulation and then draw the

ﬂoor plan of these virtual corridors. Overall, the de-

tection of the correct spatial manipulation was rela-

tively low, between 4% and 28% depending on the

corridor condition. Participants could also not signif-

icantly tell apart the three levels of space manipula-

tion. This means that our implementation of spatial

manipulation shows a valid approach to manipulating

the space without the users knowing what exact ma-

nipulation was done. We also did not ﬁnd any ev-

idence proving that there is a signiﬁcant difference

in detection for hexagonal corridors compared to cir-

cular ones. Because we noticed that the participants

perceived the corridors to be of different SoS, we ad-

ditionally analyzed this aspect. We categorized any

detection of change in SoS separately from the detec-

tion of the implausible space. A total of 21 (84%)

out of 25 participants did notice an SoS change going

from corridor to corridor. This is an indicator that the

spatial manipulation was felt to a certain degree, but

the participants did not correctly specify which kind

of manipulation.

7.1 Future Work

Future studies will include a larger number of par-

ticipants and assign each a single drawing task to

mitigate learning effects. Furthermore, investigat-

ing a variety of polygon shapes could provide in-

sights into how different angles inﬂuence the percep-

tion of impossible spaces. Practical applications of

these impossible spaces also require further explo-

ration. Lastly, efforts can be made to quantify detec-

tion in more intricate corridor designs, such as those

with left and right turns, prior to teleporting the user.

REFERENCES

Bruder, G., Steinicke, F., and Hinrichs, K. H. (2009). Arch-

explore: A natural user interface for immersive archi-

tectural walkthroughs. In 2009 IEEE Symposium on

3D User Interfaces. IEEE.

Cherni, H., M

etayer, N., and Souliman, N. (2020). Litera-

ture review of locomotion techniques in virtual reality.

International Journal of Virtual Reality, 20(1):1–20.

Ciumedean, C., Patras, C., Cibulskis, M., V

aradi, N., and

Christian Nilsson, N. (2021). Impossible open spaces:

Exploring the effects of occlusion on the noticeabil-

ity of self-overlapping virtual environments. In 2021

IEEE Conference on Virtual Reality and 3D User In-

terfaces Abstracts and Workshops (VRW), pages 389–

390.

Eppl

ee, R. and Langbehn, E. (2022). Minimaps for im-

possible spaces: Improving spatial cognition in self-

overlapping virtual rooms. In 2022 IEEE Conference

on Virtual Reality and 3D User Interfaces Abstracts

and Workshops (VRW), pages 622–623.

Freitag, S., Rausch, D., and Kuhlen, T. (2014). Reorien-

tation in virtual environments using interactive por-

tals. In 2014 IEEE Symposium on 3D User Interfaces

(3DUI). IEEE.

Iwata, H. (1999). The torus treadmill: realizing locomotion

in ves. IEEE Computer Graphics and Applications,

19(6):30–35.

Kelly, J. W., Ostrander, A. G., Lim, A. F., Cherep, L. A.,

and Gilbert, S. B. (2020). Teleporting through virtual

environments: Effects of path scale and environment

scale on spatial updating. IEEE Transactions on Visu-

alization and Computer Graphics, 26(5):1841–1850.

Koltai, B. G., Husted, J. E., Vangsted, R., Mikkelsen,

T. N., and Kraus, M. (2020). Procedurally gener-

ated self overlapping mazes in virtual reality. In

Lecture Notes of the Institute for Computer Sci-

ences, Social-Informatics and Telecommunications

Engineering, LNICST, volume 328 LNICST, pages

229–243, Cham, Switzerland. Springer.

Langbehn, E., Lubos, P., and Steinicke, F. (2018). Eval-

uation of locomotion techniques for room-scale VR:

Joystick, teleportation, and redirected walking. In

ACM International Conference Proceeding Series.

Association for Computing Machinery.

Langbehn, E., Paulmann, H., Briddigkeit, D., Barnes, M.,

Husung, M., Kirsch, K., Neves Coelho, D., Mayer, T.,

and Steinicke, F. (2020). Frozen factory: A playful

virtual experience for multiple co-located redirected

walking users. In SIGGRAPH Asia 2020 XR, SA ’20,

New York, NY, USA. Association for Computing Ma-

chinery.

Lochner, D. C. and Gain, J. E. (2021). VR Natural Walking

in Impossible Spaces. In Proceedings - MIG 2021:

14th ACM SIGGRAPH Conference on Motion, Inter-

action, and Games, pages 1–9, New York, NY, USA.

Association for Computing Machinery, Inc.

Overdijk, M. (2023). Circulation and narrative in a virtual

environment. In International Conference on Interac-

tive Digital Storytelling, pages 504–516. Springer.

Prithul, A., Adhanom, I. B., and Folmer, E. (2021). Tele-

portation in virtual reality; a mini-review. Frontiers in

Virtual Reality, 2.

Steinicke, F., Bruder, G., Jerald, J., Frenz, H., and Lappe,

M. (2010). Estimation of detection thresholds for redi-

rected walking techniques. IEEE Transactions on Vi-

sualization and Computer Graphics, 16(1):17–27.

Suma, E. A., Bruder, G., Steinicke, F., Krum, D. M., and

Bolas, M. (2012a). A taxonomy for deploying redirec-

tion techniques in immersive virtual environments. In

2012 IEEE Virtual Reality Workshops (VRW), pages

43–46, New York, NY, USA. IEEE.

Suma, E. A., Lipps, Z., Finkelstein, S., Krum, D. M., and

Bolas, M. (2012b). Impossible spaces: Maximiz-

ing natural walking in virtual environments with self-

overlapping architecture. IEEE Transactions on Visu-

alization and Computer Graphics, 18(4):555–564.

Suma, E. A., Seth Clark, S., Krum, D., Finkelstein, S., Bo-

las, M., and Warte, Z. (2011). Leveraging change

HUCAPP 2024 - 8th International Conference on Human Computer Interaction Theory and Applications

388

blindness for redirection in virtual environments. In

2011 IEEE Virtual Reality Conference. IEEE.

Usoh, M., Arthur, K., Whitton, M. C., Bastos, R., Steed,

A., Slater, M., and Brooks, F. P. (1999). Walking ¿

Walking-in-Place ¿ Flying, in Virtual Environments.

In Proceedings of the 26th Annual Conference on

Computer Graphics and Interactive Techniques, pages

359–364, New York, NY, USA. ACM.

Vasylevska, K. and Kaufmann, H. (2015). Inﬂuence of Path

Complexity on Spatial Overlap Perception in Virtual

Environments. In International Conference on Artiﬁ-

cial Reality and Telexistence and Eurographics Sym-

posium on Virtual Environments, ICAT-EGVE 2015,

pages 159–166, Eindhoven, The Netherlands. Euro-

graphics Association.

Vasylevska, K. and Kaufmann, H. (2017a). Compressing

VR: Fitting Large Virtual Environments within Lim-

ited Physical Space. IEEE Computer Graphics and

Applications, 37(5):85–91.

Vasylevska, K. and Kaufmann, H. (2017b). Towards ef-

ﬁcient spatial compression in self-overlapping virtual

environments. In 2017 IEEE Symposium on 3D User

Interfaces, 3DUI 2017 - Proceedings, pages 12–21,

New York, NY, USA. IEEE.

Wendt, J. D., Whitton, M. C., and Brooks, F. P. (2010). Gud

wip: Gait-understanding-driven walking-in-place. In

2010 IEEE Virtual Reality Conference (VR). IEEE.

Perception of a Spatial Implausibility Caused by Seamless Covert Teleportation

389