Undiscovered worlds, real-time procedural generation of virtual three-dimensional spaces

Greuter, S 2008, Undiscovered worlds, real-time procedural generation of virtual three-dimensional spaces, Doctor of Philosophy (PhD), Creative Media, RMIT University.

Document type: Thesis
Collection: Theses

Attached Files
Name Description MIMEType Size
Greuter.7z Greuter.7z Click to show the corresponding preview/stream application/octet-stream 5.38MB
Greuter.pdf Thesis application/pdf 5.38MB
Greuter_readme.txt Source code readme file Click to show the corresponding preview/stream 114Bytes
Title Undiscovered worlds, real-time procedural generation of virtual three-dimensional spaces
Author(s) Greuter, S
Year 2008
Abstract This study attempted to create a game world of enormous proportions that consisted of a visual variety of procedurally generated and distributed objects. A framework was designed to accommodate several procedural generation approaches for the various components of the virtual space.

Game worlds have become increasingly larger and more detailed over the last decade. The approach to manually create all the objects in a game world has become more and more problematic as the increase of workload demands a higher commitment of time and resources and the current solution to “hire more artists” does not scale well with the higher demands of future game worlds.

The approach to reuse a limited collection of objects to fill a game world with content is also problematic, as this repetition can be recognised by the players. As a result, the frequent repetition of geometries and textures becomes more obvious and runs the risk of boring the player. Moreover, such frequent repetition may have a negative effect on the illusion of presence that the player experiences, when playing a game. The experience of the illusion of presence is believed to be an essential factor regarding the usefulness and profitability of computer games.

This research project investigated how three-dimensional virtual spaces such as game worlds can be created more effectively and was guided by two research questions. The first research question asked how three-dimensional objects can be generated in a visual variety using computer software. To this end, this project investigated several procedural generation approaches to generate objects in a visual variety, addressing the problem of repetition in game worlds.

The second research question extended the first research question and asked how three-dimensional virtual spaces that consist of such generated objects in a visual variety can be generated in real-time using computer software. To this end the investigation focussed on approaches to distribute generated objects within a virtual space as well as issues of real-time performance and finished with the development of a framework that combined the findings.

This inquiry used an iterative design approach, which involved the creation of prototypes, which were subsequently, analysed and redesigned to progress knowledge. The prototypes demonstrate and document the results at various stages of this project and are an integral part of this research.

Overall, the study has contributed some weight to the argument that game worlds can be created using a procedural generation approach. Unlike current manual approaches to construct large game worlds which require an increasing number of game artists or the repetitive reuse of objects, this procedural generation approach is potentially faster than the manual creation approach and does not rely on object repetition to fill a virtual space with content. All objects within such a virtual space exist in a visual variety. As such it is believed that especially for large virtual spaces, this approach of procedural generation contributes to a greater illusion of presence experienced by the player.
Degree Doctor of Philosophy (PhD)
Institution RMIT University
School, Department or Centre Creative Media
Keyword(s) Presence
Visual Variety
Procedural Generation
Version Filter Type
Access Statistics: 517 Abstract Views, 462 File Downloads  -  Detailed Statistics
Created: Wed, 30 Mar 2011, 11:26:24 EST by Guy Aron
© 2014 RMIT Research Repository • Powered by Fez SoftwareContact us