Keywords

1 Introduction

It is known that domes, as one type of the IVEs, have prominent advantages of providing highly immersive experience thanks to the seemingly frameless images made possible with wide FOVs [1]. Therefore, domes are favored by the education and entertainment industries. However, fulldome projection systems have limited accessibility due to the technical complexity and high cost to build a dome theater [2]. As a result, to watch or to produce a fulldome movie is not common. Recently, the development and popularization of HMDs have changed the ways of inquiry and production in many media industries. According to the literature review, some researchers have already suggested to use HMD-based software to preview a fulldome movie, but nobody has yet actualized this concept or tested its validity [3,4,5].

Different researchers hold different opinion towards fulldome experience, but most of them agree that it is the immersion experience that makes fulldome unique and attractive [6,7,8]. In order to consider an HMD-based virtual dome as a possible alternative to a physical one, the proof that a viewer can obtain a similar experience, especially a similar level of immersion, is necessary. Therefore, a virtual dome system was developed using HMDs and user immersion it provided was assessed quantitatively in this research.

2 Related Works

2.1 Overview of Dome Display and HMD

Dome Display.

A dome display refers to immersive projections on the inside of a dome. The hemisphere, horizontal or tilted, is filled with real-time or pre-rendered CG animations, live-capture footages or any other kinds of visual content, possibly with surrounding sound. Integrating technologies from fields like domed architecture, planetariums, multi-projector film environments, flight simulation, and Virtual Reality (VR), the dome display environments have many applications in education and entertainment across a wide range of disciplines.

A dome display typically uses single or multiple projector to display a seamless wrap-around image on the inside of a dome structure, with the intention of completely filling its viewer’s FOV. Following the definition of IVEs as environments that perceptually surround users [9], dome displays qualify as an innovative medium through which to present content for a multitude of potential applications [10].

In 2012, Schnall et al. provided a theoretical framework to examine the properties of fulldome environments. The representation of space has featured prominently in IVE research in the past, because the visual elements of the display environment are typically the most prominent difference in regards to the elements of immersion, and this is also a critical element to explore within dome displays [10].

Head-Mounted Display (HMD).

A head-mounted display, abbreviated as HMD, is a display device, worn on the head or as a part of a helmet, that has a small optic display in front of one (monocular HMD) or each eye (binocular HMD). HMDs have great potential in many fields, including gaming, aviation, engineering, and medicine. A head-mounted display is the primary component of VR headsets that is effectively used in simulation systems for virtual experience to improve users’ concentration on images [11].

Aside from much lower prices, the new generation of HMDs also offered better-quality User eXperience (UX). One of the most significant factors affecting the UX of HMDs is FOV. Wearing an HMD, the natural human FOV of 180° is limited both horizontally and vertically and this influences the perceived realism of the VR experience. The new generation of HMDs have FOVs above 100° [12], which greatly enhance the UX they offer.

2.2 Measuring Immersion in Virtual Reality

According to the literature review, immersion has been mentioned and defined many times, but in different fields the definitions of immersion are not exactly the same. In the field of virtual environments, immersion is defined as “a psychological state characterized by perceiving oneself to be enveloped by, included in, and interacting with an environment that provides a continuous stream of stimuli and experiences” [13]. And another more vivid description defined immersion as the “illusion” that “the virtual environment technology replaces the user’s sensory stimuli by the virtual sensory stimuli” [14].

There are numerous studies on measuring immersion in games or Augmented Reality (AR), presence in VR and UX in VR. However, studies focusing on immersion in VR are much less.

According to the literature review, immersion was often conceptualized as a three-level construct composed of: engagement, engrossment and total immersion [15], which was the theoretical model for most of the subsequent questionnaires developed to measure immersion in the field of games [16, 17] and AR applications [18]. In 2017, Georgiou et al. employed totally seven scales of immersion based on Cheng et al.’s assumption of multi-dimensionality within each one of the three immersion levels. The seven hypothetical scales consisted of interest, time investment, usability, emotional attachment, focus of attention, presence and flow [18].

In 2016, Tcha-Tokey et al. reviewed the scales of UX questionnaires and developed the questionnaires with 82 items which can be used in most of the fields of VR [14]. Then they compared the UX effects in a CAVE and with an HMD, and drew the conclusion that CAVEs induced a greater user experience than HMDs with significant difference in presence, engagement, flow, skill, judgement and experience consequence. The results also shew that there was no significant difference in immersion, usability, emotion and technology adoption between CAVEs and HMDs [19]. In 2018, they furtherly developed a model of UX in IVEs [20].

3 Experiments: Measuring Immersion in a Physical and a Virtual Dome

3.1 Environment and Test Material

Physical Dome.

The physical dome used in this research was inclined, 5.5 m in diameter, 45° in tilted angle. It had a 2048 × 2048 resolution achieved by blending 6 XGA projections together (see Fig. 1). Third-party hardware was used for geometric correction and edge blending to achieve seamless rendering on the dome screen. In this research, audiences were seated on a sofa facing the center of the dome screen. The environment was equipped with 5.1 channel speakers.

Fig. 1.
figure 1

Inclined dome display used in the experiment. Left: front view with projection; Right: side view.

Full size image

Virtual Dome.

The virtual dome system was developed using the Unity3D game engine and the Oculus Rift CV1, one of the most popular models on the market. The device comprises a lightweight (0.38 kg) headset with separate displays for both eyes, each with a 1080 × 1200 resolution, yielding a 110° horizontal field of view (FOV) and a framerate of 90 Hz. As shown in Fig. 2, the virtual dome software supported both 360-degree panoramic movies and 180-degree fulldome movies. Figure 3 shows the most significant parameters with which a user may set the diameter, tilt angle of the virtual dome, and the relative position of the virtual viewer by X, Y, Z offsets. With these settings, it is possible to mimic a specific physical dome environment more closely. Finally, a movie file in the dome master format can be loaded and played back (see Fig. 4).

Fig. 2.
figure 2

Choice of movie type. Left: 360-degree panorama movie; Right: 180-degree fulldome movie.

Full size image
Fig. 3.
figure 3

Parameters. Left: parameters for 360-degree dome; Right: parameters for 180-degree dome.

Full size image
Fig. 4.
figure 4

Left: file loading interface; Right: a frame displayed in the virtual dome.

Full size image

As for this research, this virtual dome system was driven by one single GeForce GTX 1070 card running on a workstation. The 180-degree dome was selected and the parameters were modified to match the real dome mentioned above, that is to say, the diameter was set to 5.5 m, tilted angle 45°. The virtual viewer was placed 1 m away from the center of the sphere (Offset Y) and 1.2 m above the ground (Offset Z), which matched the physical world closely.

Contents.

The author prepared five fulldome short movies produced in the College of Arts and Media. Two of them were abstract movies that used the dome only as a curved screen; two were feature movies created for the dome display; the last one was a 3D animation of high presence that aimed to provide a virtual experience of space navigation (see Table 1). All the animations were rendered in 3ds Max with the V-Ray fish-eye lens camera.

Table 1. Types of fulldome short movies.
Full size table

3.2 Immersive Tendency Questionnaire (ITQ)

ITQ was used to measure the differences in the tendencies of individuals to experience presence [13]. The questionnaire began with a participant identification survey asking about personal information (name, gender, age, occupation and major). The following 18 items were derived from Witmer and Singer’s immersive tendency questionnaire [13]. All the questions were in Chinese according to Yu Tian’s Chinese translation [21]. The ITQ used a seven-point scale format that was based on the semantic differential principle [22].

3.3 Tailoring the VRI Questionnaire to the Experiment

This research focused on the comparison of users’ immersion between a physical dome and a virtual one. The first step to build a VRI questionnaire was to identify the scales of immersion. According to the literature review, a three-level construct [15] and seven scales [18] of immersion were used in the VRI questionnaire. Almost all the scales in UX questionnaires [14] can be correspondingly merged into the seven immersion scales, except “simulator sickness” for the evaluation of negative symptoms experienced by users in IVEs. These negative symptoms have been associated prominently with HMD displays, and no systematic study on large scale domes has examined the incidence of these symptoms [10]. However, considering the fact that simulator sickness directly affects a user’s basic experience, it was investigated in this study.

After determining the scales of immersion, items were developed to compile the VRI questionnaire. In this research, all the participants are Chinese, so all the items were developed in Chinese to avoid ambiguity in language translation and comprehension. Items were rearranged to better measure user immersion in VR (see Table 2).

Table 2. Items selected from the existing questionnaires for VRI questionnaire.
Full size table

  • From the Augmented Reality Immersion (ARI) Questionnaire [18], 16 items representing seven immersion scales were selected and adjusted to fit our context (e.g. “The AR application we employed captured my attention” became: “The physical/virtual dome we employed captured my attention”).

  • From the Immersion Questionnaire developed by Jennett et al. [16], 10 items were selected and modified to fit our context. For most of the questions, the words “game” was changed to “physical/virtual dome”. The last question: “How immersed did you feel?” aimed to know a participant’ s subjective judgement of immersion.

  • From the Presence Questionnaire (PQ), item 20 was picked as a reference. It’s known that one advantage of physical dome displays was that the users in it could look at the projected image in an arbitrary direction freely [6]. Meanwhile, in an HMD display, users could also look at the image from multiple angles by rotating their heads. Therefore, question 5 was changed to “I’m interested in watching movies in an arbitrary direction in physical/virtual dome”.

  • From the Achievement Emotions Questionnaire (AEQ) [23], 2 items representing the scale of emotional attachment were picked. One item was positive and activating: enjoyment, the other one was negative and deactivating: boredom.

  • From the Simulator Sickness Questionnaire (SSQ) [24], 9 SSQ symptoms were selected. All these 9 symptoms were combined as a matrix scale in question 29: When you were watching films in the physical/virtual dome, how did you feel about the following symptoms? A seven-point scale (where 1 represented “completely no feeling” and 7 represented “completely feeling”) was employed to evaluate simulator sickness.

Finally, the VRI questionnaire contained 2 items to collect personal information (“What’s your name?” and “Have you ever experienced a physical dome/an HMD?”).

The 30 items to measure immersion were numbered. A seven-point scale (where 1 represented “completely disagree” and 7 represented “completely agree”) was employed for each item. The majority of questions were positive, where a higher score reflected a higher level of perceived immersion (component); only 3 gave negative marks (Q13, Q18 and Q30).

The VRI questionnaire was used to measure user immersion both in the physical dome and the virtual one, so there were two different versions labeled with either “VRIQPD” (VRI questionnaire for the physical dome) or “VRIQVD” (VRI questionnaire for the virtual dome).

3.4 Experiment Design

Firstly, ITQ was used to measure the tendency of an individual to experience presence. Then he/she watched a film (or 2 films) in the physical dome and then with the oculus rift CV1, or in the contrary sequence, according to the random group assignment. After each experience, a VRI questionnaire was employed to measure his/her immersion obtained from it. Finally, data was collected and analyzed to understand the differences.

Participants.

44 participants (25 females and 19 males) took part in the experiment aging from 19 to 30. The main age group was 19 to 22 (70.45%). 41 participants were university students and the other three were teachers. 16 participants majored in animation; 13 participants worked or studied in other field of arts or communication (e.g., MA, MFA, advertising, journalism, radio and TV director); 15 participants worked or studied in the fields of science and technology (e.g., material science and engineering, vehicle engineering, computer science, energy engineering). They were randomly assigned to four groups to watch different test movies, as shown in Table 3.

Table 3. Arrangement and procedure of experiment.
Full size table

Arrangement.

The experiment took place in the Non-planar Screen Lab in the College of Arts and Media, Tongji University. Movies for group A and B, an abstract animation and a space navigation movie, represented two typical and most common types of dome films. They were analyzed to verify whether a virtual dome system can possibly perform as well as a physical one in common scenarios. Group A and C were compared to see whether different content exerted an influence on user immersion. Finally, group D were designed to see whether two environments gave user the same experience of immersion when showing feature movies.

Procedure.

In the experiment, each participant went solo through the following steps:

Firstly, an ITQ was used to measure the tendency of an individual to experience presence. It has been verified that a higher ITQ score reflect a greater tendency to become involved or immersed [13] so that an individual with high ITQ scores tend to report more immersion on a VRI questionnaire. In order to compare the experimental results between groups, it is necessary to avoid big difference in ITQ mean value of each group. As a result, participants with abnormally low ITQ scores shall be excluded from this study.

Secondly, participants watched the same movie(s) once in each environment. In each group, half of the participants experienced the physical dome first, and the other half, virtual dome first in order to avoid the sequential effect.

Thirdly, an VRI questionnaire was required to be filled in by a participant immediately after the experience was over. (There were 2 versions of VRIQs, namely VRIQPD and VRIQVD. The choice was based on the type of the dome the participant just experienced).

Finally, some of the participants were interviewed to know more about which environment they preferred and why. The answers to these questions were collected to extract insights for future studies.

3.5 Collected Data and Analysis

Each of the 44 participants experienced two immersive environments and filled in two VRI questionnaires respectively. Therefore, 44 ITQs and 88 VRI questionnaires were collected. They were all validated and the data analysis was done using SPSS v21.

4 Results

4.1 Overview

The resultant alpha of the 44 ITQs was 0.751, showing a good reliability (alpha > 0.7). After the test for equality of variances (p = 0.283 > 0.05), 44 valid ITQs were examined with one-way ANOVA and multiple comparison to see whether immersion tendencies varied too much among groups. The result indicated no significant difference between each group (p = 0.56 > 0.05) (see Table 4), making it reasonable to compare immersion with data from the VRI questionnaires.

Table 4. The effects of different groups on immersion by method of one-way ANOVA.
Full size table

The VRI questionnaires had a resultant alpha of 0.947 (N = 88), which also suggested a high reliability. Paired-sample test was used to compare the virtual dome with the physical one in each scale of the VRI questionnaires. In order to further understand the influences of other factors, ANOVA was used to analyze variables including visual content, exposure order, prior experience and so forth. Finally, a small random sample from the participants was interviewed to obtain more insights into the actual cause of the differences.

4.2 User Immersion Between the Physical and Virtual Dome in Each Group

Data for the paired-sample tests were analyzed in 5 scales and the results are presented below:

In group A (see Fig. 5), there was no significant difference found in the means of all the scales. However, the data variance from the virtual dome was obviously greater, especially in the scales of engagement and immersion score, which meant that participants’ experience in virtual dome brought different levels of immersion. Two of the participants thought that the immersion experience in the virtual dome was significantly worse. By analyzing the interview results, the reasons were found out to be the low resolution of the HMD and the uncomfortableness of wearing it.

Fig. 5.
figure 5

Results of paired-sample tests of group A (N = 10).

Full size image

In group B (Fig. 6), it’s obviously that all the virtual dome gave better results of immersion. According the paired-sample tests, the user engrossment was higher in the virtual dome (M = 5.07, SD = 0.99) than the physical one (M = 4.19, SD = 1.05), t (10) = −2.612, p = 0.028; the immersion score user rated was also higher in the virtual dome (M = 5.3, SD = 0.67) than the physical one (M = 4.3, SD = 1.06), t (10) = −3, p = 0.015. It’s not difficult to draw a conclusion that user immersion was better in a virtual dome than a physical one when watching space navigation movies.

Fig. 6.
figure 6

Results of paired-sample tests of group B (N = 10).

Full size image

In group C (Fig. 7), there was no significant difference found in the means of all the scales. The virtual dome gave better results in engrossment and total immersion. It’s worth noticing that the lowest scores of the 5 scales of the virtual dome were all given by the same participant. Later interview revealed that the cause was that the HMD was uncomfortable for this participant because of his glasses.

Fig. 7.
figure 7

Results of paired-sample tests of group C (N = 12).

Full size image

In group D (Fig. 8), there was also no significant difference found in the means of all the scales. However, the scores of the virtual dome varied much more especially in engrossment, total immersion and immersion score. This result suggested very different levels of immersion perceived by different participants in the virtual dome. Similar to the last group, the lowest scores were all given by the same participant because the HMD was heavy and uncomfortable.

Fig. 8.
figure 8

Results of paired-sample tests of group D (N = 10).

Full size image

4.3 Other Possible Influential Factors

In addition to the difference between a physical dome and a virtual one, other factors might also influence user immersion. These possible influential factors were also analyzed in this study. In this part, user immersion, the dependent variable was the sum of 5 scales’ mean values, and the independent variables include:

  • Content: different media content used in 4 groups

  • Order: the sequence of user experience (physical dome first or the contrary)

  • Prior Experience: whether a participant had ever experienced a dome, physical or virtual, before

  • Personal Data: gender, age, occupation, major and immersive tendency.

One-way ANOVA was used to test these factors, but no significance was found. For example, Tables 5 and 6 shew the results using order and prior experience as independent variables, but the results are similar and all the p values are much greater than 0.05, suggesting no significant difference.

Table 5. One-way ANOVA test using order as the independent variable
Full size table
Table 6. One-way ANOVA test using prior experience as the independent variable
Full size table

The only exception was the test using content as the independent value as shown in Table 7. Though p values were also greater than 0.5, but the p value of the test for the physical dome was 0.058, pretty close to 0.5. It is possible to conclude with some confidence that in a physical dome, one may experience lower immersion when watching a space navigation animation.

Table 7. One-way ANOVA test using content as the independent variable
Full size table

4.4 Summary of the Interviews

Short interviews following the questionnaires aimed to know more about what the participants subjectively thought about their experiences in different environments. The survey used a sample of 22 people randomly chosen from all the participants. There were only two questions in the survey: Q1. Which do you prefer, the physical dome or the virtual one? Q2. Why?

Table 8 shows the summary of the interviews. The factors directly influenced user immersion are marked with asterisks.

Table 8. The summary of interviews.
Full size table

5 Discussion

This research analyzed and compared two IVEs (Dome and HMD) in terms of user immersion with an adjusted VRI questionnaire covering three components of immersion. By inspecting user immersion, which is the most significant feature of a dome experience, in physical and virtual domes, this research tried to prove that a virtual dome system based on HMDs could provide immersion to a remarkable extent.

The results of paired-sample analysis from each group shew that subjective user immersion was almost the same in the physical dome and the virtual one. Group A (tested with an abstract movie), group C (tested with two different abstract movies) and group D (tested with two feature movies), almost reported the same levels of user immersion. As for group B (tested with a space navigation movie), greater user immersion was reported by virtual dome experiences with significant difference found in engrossment and immersion. Overall, the user immersion brought by a virtual dome is as good as, if not better than a physical one.

The results of the one-way ANOVA across groups shew that the group tested with a space navigation movie reported lower user immersion in the physical dome experience. And other different types of content caused no significant difference in user immersion. Moreover, no significant difference was found to be caused by the other factors either. The participants with no prior dome and VR experience reported a bit higher level of immersion when they experienced the virtual dome first, but the difference was minor.

The results from the random sampling interview shew that half of the subjects preferred the experience in the physical dome, and other half prefer the virtual one. Most of the reasons why participants disliked the virtual dome was caused by the intrinsic problems of HMDs, such as a narrow FOV, a lower resolution and etc. Most of the positive comments about immersion were given to the virtual dome and many interviewees thought that it brought a more immersive experience because it isolated the outside world much better.

6 Conclusion

In this research, experiments were carried out to prove the hypothesis that a virtual dome system built with HMDs could bring a similar subjective user immersion when compared to a physical dome. According to the outcomes from the experiments, the hypothesis was successfully proved. Although the defects intrinsic to a virtual dome system, such as the problems of limited resolution, narrow FOV, uncomfortableness to wear and the ack of shared experience should not be overlooked, a virtual dome system could be used as a low-cost and more accessible alternative to a physical dome to offer an immersive experience of fulldome movie.