Elsevier

Computers in Human Behavior

Volume 27, Issue 6, November 2011, Pages 2186-2191
Computers in Human Behavior

A comparative study of four input devices for desktop virtual walkthroughs

https://doi.org/10.1016/j.chb.2011.06.014Get rights and content

Abstract

This paper presents the results of an experiment measuring the effect of four different input devices on overall task performance for desktop virtual walkthroughs. The input devices tested are: a keyboard, a mouse, a joystick and a gamepad. The results indicate that the participants completed the tasks in significantly less time and distance travelled with the mouse than with the three other input devices. The use of the mouse also significantly reduced the number of collisions, while the use of the gamepad resulted in significantly more collisions.

Highlights

► We compare four travel techniques for desktop virtual walkthroughs. ► Each travel technique uses a different input device and we used a directed search task. ► The four input devices are: mouse, keyboard, joystick, gamepad. ► The mouse interface reduces time, travel distance compared to the other techniques. ► The gamepad interface increases the number of collision with the environment.

Introduction

Virtual walkthrough is a common travel metaphor used for viewpoint control in virtual environments (VEs) (Brooks, 1986, Usoh et al., 1999). Today, a large proportion of users perform walkthroughs on desktop VEs (i.e. displayed using a monitor) as opposed to immersive displays (e.g. HMDs, CAVEs) (Ruddle et al., 1997, Santos et al., 2009). Virtual walkthroughs can be divided into a set of cognitive tasks and a set of physical tasks that a person performs to get from one location to another (Bowman et al., 2005, Sebok et al., 2004). While many studies have investigated the cognitive aspects (orientation and wayfinding) of walkthroughs (e.g. Ruddle, 1999, Ruddle et al., 1998, Ruddle et al., 1997) we are focusing on travelling itself, which is the physical tasks that allow a navigator to move from one location to another by using the input devices (a.k.a. control devices) of the user interface (Bowman et al., 2005, Sebok et al., 2004). The input device design space for virtual walkthroughs is quite large given the possible combinations of commands that can be issued from a number of input devices. However, very little of this design space has been investigated empirically. As with other user interfaces, empirical evaluation of possible interaction techniques is important to improve our understanding and to develop theoretical models and to assist designers in building more usable systems (Fröhlich et al., 2006, Lampton et al., 1994, Lindeman, 2006, Newell and Card, 1985).

In virtual walkthroughs, the number of Degrees Of Freedom (hereafter “DOF”) of movement varies between 2 and 4, out of a potential 6. Three of the six DOFs involve translation: forward/back, left/right, and up/down (which is typically not needed in virtual walkthroughs). The other three DOFs involve rotation: rotation around the axis perpendicular to the travel plane (yaw axis), and rotation around each of the axes forming the travel plane (pitch and roll axes), though the last one is not typically needed in virtual walkthroughs. Each DOF can potentially be controlled by a different input device and a different input command from that device. Moreover, several travel techniques are possible by using one or two-hands.

To date, the research shows that, in the case of 3D interface, there is still not an input device that demonstrates its superiority for accomplishing basic 3D tasks such as navigation, manipulation and selection. This contrasts with the case of the 2D graphical interface where the computer mouse established itself as the de facto standard input device (Zhai, 1998).

In the case of maze travelling tasks, research indicates that bimanual travel control is quite feasible and even outperforms the status quo mouse-mapping interface (Zhai, Kandogan, Smith, & Selker, 1999). Also, two studies did not find significant differences in performance when travelling with a user interface using two or three DOFs, when the third DOF is a strafe movement (i.e. a side translation of the viewpoint) (Lapointe and Savard, 2009, Lapointe and Vinson, 2002). Finally, another study suggests that the use of either velocity or position control techniques for viewpoint orientation does not have a large effect on travel performance (Lapointe & Savard, 2007).

For desktop walkthroughs, commonly found control interfaces in video games use a keyboard and mouse combination for PC games and two joysticks (on a gamepad) for console games. The use of these specific combinations appears to be driven by availability of the input devices and is not backed by published scientific analysis of the performance or usability of these interfaces.

To place the design and use of walkthrough interfaces on a more empirical footing, and to contribute to the exploration of the design space, we conducted an experiment in which we compared the user’s performance and preference when using four different input devices, using either two or three DOF to travel in desktop virtual walkthroughs. We employed four input devices that are commonly used for desktop virtual walkthroughs: a keyboard, a mouse, a joystick and a gamepad.

Hence, this paper describes the results of an experiment that aims to characterize the usability benefits of those control devices for doing desktop virtual walkthroughs.

Section snippets

Method

To evaluate the different travel techniques, we used a maze-like virtual world made of a complex trail offering an open view, so that users could always look around and/or keep an eye on the end point of the maze while they travelled (Fig. 1). This open-view maze offered more incentives to use all the available DOF than a traditional maze environment. A traditional maze has high walls in which only views along the center line of a corridor offer interesting visual cues to spatially orient the

Quantitative results

We first determined the effect of practice on performance. This allowed us to exclude practice effects from our analyses of the travel technique performance effects. Fig. 7 shows that practice-related performance improvements on completion time disappeared after a few trials. Consequently, we restricted our travel technique performance analyses to the last three trials in each block.

Fig. 8 illustrates the mean task completion times for the 4 travel techniques (average of the last three trials).

Discussion

The results of this experiment indicate that, for this specific virtual environment, the mouse travel technique offers better performance than the three other travel techniques tested, namely the keyboard, the joystick and the gamepad travel techniques.

Accordingly, the subjective ratings indicate that the mouse travel technique also provides a better user experience in regard to ease of use, speed, accuracy and overall preference.

It is important to note that the mouse travel technique uses its

Conclusion

In conclusion, the mouse input device seems well suited as a single-handed device for simple walkthrough tasks in a desktop virtual environment.

Moreover, further experimentation is required to confirm the superiority of the mouse in other virtual environments and for other users.

Future work

Given the superiority of the mouse and previous findings showing high performance levels for bimanual techniques, travel could potentially be further improved by combining the mouse with a keyboard as is often the case in first-person (egocentric) 3D computer games. Of course, this hypothesis awaits empirical confirmation.

Another future research direction could look at other input devices and interaction techniques such as those provided by sensor-based input techniques such as the Wii and

Acknowledgements

The authors thank all the volunteers who participated in this study. This experiment has been approved by the research ethics board from the National Research Council of Canada. This research has been funded by the Institute for Information Technology of the National Research Council of Canada.

References (21)

  • S. Zhai et al.

    In search of the ‘magic carpet’: Design & experimentation of a bimanual 3D navigation interface

    Journal of Visual Languages and Computing

    (1999)
  • D.A. Bowman et al.

    3D user interfaces: Theory and practice

    (2005)
  • Brooks, F. P. (1986). Walkthrough – a dynamic graphics system for simulating virtual buildings. In Proceedings of...
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • D.B. Duncan

    Multiple range and multiple F tests

    Biometrics

    (1955)
  • B. Fröhlich et al.

    On 3D input devices

    IEEE Computer Graphics and Applications

    (2006)
  • R.J.K. Jacob et al.

    Integrality and separability of input devices

    ACM Transaction on Computer–Human Interaction

    (1994)
  • D.R. Lampton et al.

    The virtual environment performance assessment battery (VEPAB): Development and evaluation

    Presence: Teleoperators and Virtual Environments

    (1994)
  • Lapointe, J. -F., Savard, P. (2007). Comparison of viewpoint orientation techniques for desktop virtual walkthroughs....
  • Lapointe, J. -F., Savard, P. (2009). A comparative study of three bimanual travel techniques for desktop virtual...
There are more references available in the full text version of this article.

Cited by (40)

  • Exploring staircases as architectural cues in virtual vertical navigation

    2020, International Journal of Human Computer Studies
    Citation Excerpt :

    Since travel and wayfinding are closely related, the method of travel can have an effect on the ability to perform wayfinding tasks, in setting one's viewpoint position and in determining spatial orientation. In desktop VEs, the most common travel metaphor used for viewpoint control and navigating architectural buildings is virtual walkthrough (Lapointe and Savard, 2007; Lapointe et al., 2011). Bowman et al. (1999) categorized virtual travel techniques by the amount of control the users have over their motions as discrete target, continuous and one-time specifications.

  • Navigation in virtual environments using head-mounted displays: Allocentric vs. egocentric behaviors

    2018, Computers in Human Behavior
    Citation Excerpt :

    Altogether, they made a number of findings. First, the mouse was the best one-handed interface for navigating the desktop VE (Lapointe, Savard, & Vinson, 2011). Second, more experience in gaming led to superior performance in virtual navigation (Murias, Kwok, Castillejo, Liu, & Iaria, 2016).

  • The Influence of Method of Control and Visual Aspects on Exploratory Decisions in 3D Video Games Environments

    2022, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
View all citing articles on Scopus
View full text