This page describes the development of a 3D drawing application for the Fontys Cave, as carried out by a group of students of Eindhoven University of Technology, for the course Interactive Virtual Environments.
Hephaistos is a 3D drawing application, developed for a previous course by two group members. The application is based on so called control points, that define certain shapes. On a high level, all user interaction is based on (re)locating and manipulating control points.
In order to port the application to a format suitable for the cave, all user interaction has to be adjusted. However, the underlying data structure and virtual world can remain unaltered.
Using a 3D application in a 2D environment, like a computer screen, might be non-intuitive and confusing. Complex 3D actions are expected to be easier in a real 3D environment, like the cave. However, some tasks might be performed with a 2D environment easier. Typical actions in this category are planar operations.
In this research it is investigated which Hephaistos tasks are 'easier' in the cave and which are easier on a computer screen. As a measure of hardness of a task, the time needed to complete this task is measured. In order to minimize the effect of learning, each experiment is first repeated a number of times, before measurement starts.
The following tasks are to be performed by a user both on a computer screen and in the cave.
Drag and drop
The user is presented with a ball and a soccer goal. The user has to drag the ball into (or through) the goal.
Connecting the dots
For the second test, a collection of dots have to be connected with a single spline. There are eight dots, positioned on the corners of an imaginary 3-dimensional cube.
The user is presented with a collection of spheres, all of different color and size. The user has to position these spheres such that they are ordered on size.
In the last test, a collection of dots is presented to the user. These dots have the same color and are located at random positions in the scene. The user has to position these dots in such a way that they are collinear.
The Drag and drop and Connecting the dots tasks are expected to take less time in the cave. Reason is that these tasks strongly depend on the three dimensional structure of the environment, which cannot be easily mapped onto a plane.
The Sorting spheres and Collinear dots tasks are expected to take less time on the computer screen. This can be explained by the possibility to divide these tasks into orthogonal subtasks.
As can be seen from the results, both tasks that were expected to be easier in a 2D-environment (Collinear dots and Sorting spheres) definitely are. However, both tasks that were expected to be easier in the CAVE differ not too much in 2D and 3D. This might be caused by several reasons.
- Most test persons are more used to working with a 2D-environment than with a 3D-environment.
- The input device (motion tracked Wiimote) suffered from high latency due to a changed dll file. As the original dll file did not work (for some unknown reason) the experiments had to be carried out with this latency.
- The motion tracker is normally used as a head tracker. Therefore, it has a build-in displacement of several centimetres. Although it should be possible to overcome this problem by adding an offset in the VRJuggler configuration file, this did not work properly. As a result, the user has to deal with a slightly displaced pointing beam.
- The test 'Connecting the dots' was repeated several times with the same input configuration. As a result, test persons came up with a trick to quickly solve this test in the 2D-environment. Therefore, this test could be performed much faster than expected in the 2D-environment.