Allen Bierbaum

VR Juggler: a virtual platform for virtual reality application development

Today, scientists and engineers are exploring advanced applications and uses of immersive systems that can be cost-effectively applied in their fields. However, one of the impediments to the widespread use of these technologies is the extensive technical expertise required of application developers. A software environment that provides abstractions from specific details of hardware configurations and low-level software tools is needed to provide a common base for the prototyping, development, testing and debugging of applications. This paper describes VR Juggler, a virtual platform for the creation and execution of immersive applications, that provides a virtual reality system-independent operating environment. We focus on the approach taken to specify, design and implement VR Juggler and the benefits derived from our approach.

Virtual reality interfaces using Tweek

Developers of virtual environments often face a difficult problem: users must have some way to interact with the virtual world. The application developers must determine how to map available inputs (buttons, gestures, etc.) to actions within the virtual environment (VE). As a result, user interfaces may be limited by the input hardware available with a given virtual reality (VR) system.

VjControl: an advanced configuration management tool for VR Juggler applications

Virtual reality (VR) installations are often unique; every one is a complex blend of hardware devices, displays and computing resources. The configuration of VR software is therefore a difficult and time-consuming process. VR Juggler, our toolkit for VR application development, addresses these problems with a number of innovations. VR Juggler provides a unique system for organizing the configuration information for a system and minimizing the proliferation of configuration files that many systems suffer. It provides a graphical tool, called VjControl, for editing configurations, which can protect users from many common mistakes. VR Juggler also has advanced capabilities for monitoring and altering the configuration of a running immersive application.

Open source virtual reality

This tutorial provides an overview of VR application development using Open Source software tools. We focus on the VR Juggler suite of tools (see http://www.vrjuggler.org) and various Open Source projects that can work with it. We will demonstrate how Open Source software ensures extensibility and eases integration between software tools. We will also show how leveraging existing and emerging Open Source projects can lower development times and costs while simultaneously increasing portability and stability. In particular, we will discuss several kinds of Open Source VR software: • Development frameworks for immersive VR applications (VR Juggler) • Scene graphs for managing application data. (OpenSG or OpenSceneGraph). • Clustering software for powering low-cost VR systems (NetJuggler). • Collaboration software for writing interactive, networked applications (primarily discussing software currently under development by the VR Juggler team). • Various other tools needed to write compelling applications, such as audio libraries, scripting capabilities, and navigation systems, and how they can be used by VR Jugglerbased applications.

Automated testing of virtual reality application interfaces

We describe a technique for supporting testing of the interaction aspect of virtual reality (VR) applications. Testing is a fundamental development practice that forms the basis of many software engineering methodologies. It is used to ensure the correct behavior of applications. Currently, there is no common pattern for automated testing of VR application interaction. We review current software engineering practices used in testing and explore how they may be applied to the specific realm of VR applications. We then discuss the ways in which current practices are insuficient to test VR application interaction and propose a testing architecture for addressing the problems. We present an implementation of the design written on top of the VR Juggler platform. This system allows VR developers to employ standard software engineering techniques that require automated testing methods.

Product review: an overview of cluster solutions for immersive displays

Consumer graphics cards are now so powerful that many virtual environment groups are considering moving to a cluster solution to drive their immersive installations. The use of off-the-shelf components is attractive for many academic and industrial groups because of the installation, direct maintenance costs, and potentially swifter upgrade path. However, cluster solutions generate their own problems in software distribution and maintenance, and they are not yet common. Several cluster solutions have been built and tested in laboratories, and commercial solutions have started to appear. This review will look at three cluster solutions and compare some of the technical requirements and user experience with them: one mature installation, the Fraunhofer Institute for Industrial Engineering IAO’s HyPI-6; one commercial system, the ARS Electronica ARSBox; and open source solutions based on VR Juggler. Large wall installations have been demonstrated working in a number of different rendering configurations. In sort-first style processes, a single image is partitioned into tiles, each image generator renders one tile, and then all of the tiles are composed into a single image. In sort-last style processes, each image generator is responsible for rendering, with depth, a set of graphics primitives, and these images are subsequently composed with depth to form the final image. Toolkits such as Stanford’s Chromium provide good support for these and other hybrid solutions for the situation in which a single virtual camera can be assumed. In Chromium for example, a large subset of existing OpenGL applications can seamlessly be moved from single-image generator to multiple-image generators without any application-level customization. Immersive systems pose new challenges for cluster rendering systems. With a large wall display, a single viewpoint and third-person view control are sufficient for user interaction. In these conditions, conventional desktop applications can be run unaware of the underlying rendering solution. In contrast, immersive systems usually require more than one display surface and projection system, stereo imagery, and head tracking. With the egocentric and high field of view required for immersive graphics, we can no longer utilize a single desktop viewing configuration because we need to make sure that a complete set of graphics primitives is generated so that any required view can be generated and nothing is culled away because of the application’s reliance on view volume or visibility culling from a single center of projection. Thus, toolkits that target immersion usually work by distributing and synchronizing instances of an application or real-time distribution of a database of graphical primitives. Application instancing has the advantage that it seems simple to set up. However, all but the most trivial applications have interactive capabilities that generate issues with synchronization of behavior across all processes. A common alternative is the scene-graph style distribution. In this mode, an application generates edit events on a scene-graph structure. Enforcement of a particular scene-graph paradigm restricts the application programmer’s freedom, but it does provide a simple framework for the distribution side. A further level of synchronization is required at the image level to make sure tearing and other artifacts are