CWI Box groep1
This describes the construction of an immersive virtual reality application used for quantifying the immersion experienced by the user in a CAVE. Since the application is based on an existing application, a porting process is detailed first. In additionion various virtual reality features are added on top. A user study on immersion is conducted from which the results are compared to existing conclusions in the literature on immersion. In addition, a test is conducted on the effects of various lengths of artificial delay. The findings are concluded, discussed and finally some recommendations are given. Eventually some future work is briefly mentioned.
This project focuses on the immersive experience of the user residing in CAVE type virtual reality environment setting. The actual cave setting used is located at the Fontys Virtual Reality Lab which is operational since late 2008. From now on we refer to this setting as the CAVE. As a starting point the BOX Application from the Centrum Wiskunde and Informatica in Amsterdam is taken. With "BOX" we refer to this application in original state.
The BOX application is a simple application to demonstrate several aspects of virtual environments. The box consists of colored paper airplanes flying a particular route inside a solid box, such that the planes are evenly distributed within the box (the starting locations and velocities are randomized within a certain range). If desired, a trail is left by the planes to indicate the route that was taken by the planes. The solid box has a checker board pattern on one side to increase visibility of the planes. The box as a whole is scalable and the number of airplanes is controllable. Scaling the box also scales the path of the planes accordingly. During the porting of the BOX application all these existing characteristics are mostly preserved.
Since the CAVE is built using the cross-platform virtual reality framework VR Juggler the first step forward is porting this application such that it uses VR Juggler API as a abstraction from all underlying hardware. The virtual reality features added are: head-tracking, hand-tracking, surround sound, interaction with the planes and a skybox.
These features are switched on and off in particular sequences to quantify the impact of the features on the immersive experience of the user during the user study. To place the results of this study into context, comparisons are made against existing conclusions in literature on immersion.
After concluding the results of both tests, a final discussion is given along with some future extensions.
Immersion and presence are relatively new subjects, however both are seen as very important to the experience in interactive virtual environments. Little research is done on parameters influencing immersion and presence. Since interactive virtual environments make use of all sorts of parameters that can have an influence on immersion and presence it is very interesting to look deeper into the mutual effects.
Head tracking and 3D audio are both assumed to have an influence on immersion and presence and there is some literature that backs these assumptions up. Interaction through hand tracking and realism by use of a background are also assumed to have influence on immersion and presence. Research gives an indication that this is true, however there is no direct research done to back this up. This is why it is important to study whether interaction and realism have influence on immersion and presence. We study the effect of interaction through hand tracking and realism by use of a background on immersion and presence when used in a CAVE.
Delay can be a serious issue in virtual environments. Especially in case the system is distributed as is the case in the Fontys CAVE. A delay is always present due to the amount of time it takes to render complex scenes. Additionally, in a distributed system, data needs to be distributed over a network to keep all clients synchronized. We study the effect of delay on the performance of a user when executing simple tasks in a CAVE, especially the Fontys CAVE.
Delay can be interpreted in several ways, such as a slow reaction on the input of the user or a delay between two consecutive updates of the rendered scenes of the virtual environment. The latter one is tested more thoroughly in this experiment. Several tests are conducted to test the effect on the performance of a user to execute a task, while having delays of different lengths.
The participant is handed over a WII-controller with a tracking device and glasses for the CAVE. After this the participant is led into the CAVE.\\
In conditions without interaction the participant is told to look around in the setting. In conditions with interaction the participant is told to look around in the setting, push the A-button on the WII-controller to grab a helicopter, drag it and finally let it go.\\
After 1 minute and 30 seconds the setting is turned of and the participant is escorted to a table and chair and asked to fill in a game questionaire. When finished the participant is prepared for the next condition, until all conditions have taken place.
In order to test the effect of delay on the performance of the user in the cave, the following test is conducted.
Consider the box application in the Fontys cave. Both the skybox and the sound are turned on. All other settings are set to their default value. The person, conducting the test, is positioned in the center of the cave, wearing the stereographic glasses and holding the handtracker. Before the experiment starts, the person conducting the test, is instructed on how to perform the test.
At the moment the test starts, ten white helicopters and four red helicopters fly through the cave. The goal of the person in the cave is to touch the four red helicopter using the hand-tracker. For each of the artificial delays, it is tested how long it takes to touch all the red helicopters. If a red helicopter is touched, it turns green to indicate it has already been touched.
In this study, we set out to research the influence of both realism in backgrounds and hand-tracking based interaction on immersion in a CAVE setting. The results show (for immersion as well as flow) that the addition of hand-tracking did increase experienced immersion, while the addition of a background did not. There was no interaction effect between the hand-tracking and background additions; the combination of hand-tracking and background was not significantly more immersing than hand-tracking alone.
As mentioned before the delays are accumulated to the base delay in reality. According to the assumption of at least 50 miliseconds of base delay all the delays need be corrected with at least 50 miliseconds. Total delays above 100 miliseconds are considered harmful to the user performance, as stated by one of the found articles. Consequently, it is reasonable to expect a decrease of the user performance when adding an artificial delay of 25 or even 50 miliseconds.
An analysis of the results shows that the average performance decrease is linear in the added artificial delay. According to one of the found articles of Meehan and co, a significant decrease in performance is noticable in case the delay is above 50ms. Although a slight decrease in performance is visible, it is not "significant".
Concluding, either the base delay is a lot higher than the estimated 50ms and thus the user's performance is bad from the start, thus even with 0ms of artificial delay, or the results depend on the test that is performed.