A comparison of shadow algorithms | Joen Sindholt
| | Abstract | Because shadows are very important in computer generated images, a lot of different algorithms for shadow generation have been proposed. However two algorithms namely Shadow Maps[WILL78] and Shadow Volumes[CROW77] are the mostly used for real-time purposes, and many of the current shadow algorithms are built upon these. While many game-engines choose a single algorithm for shadow generation we look into the possibility of using multiple algorithms, choosing the best based on the current scene at hand. The idea is, that the choice of algorithm might be done automatically at run- time or at least be specified as a parameter at design time.
To do the above, we implement and analyze four di erent shadow algorithms which generates hard shadows, and look into the strengths and weaknesses of the four. We will also look into how to choose the appropriate algorithm depending on scene properties. An application is created allowing to choose among the different algorithms during runtime, enabling easy overview of the differences. It is found, that none of the four algorithms are perfect with regard to all aspects of shadow generation, as they focus on di erent problems. It is also found, that objectively measuring how good a specific algorithm is, seems to be an almost impossible task, and it is concluded that choosing algorithm at design time seems a better approach for choosing the right algorithm. The four algorithm examined in this thesis are the Shadow Map algorithm introduced by L.Williams [WILL78], the Shadow Volumes algorithm introduced by Frank Crow [CROW77], Perspective Shadow Maps by Marc Stamminger and George Drettakis[SD02] and a hybrid algorithm using ideas of both Shadow Maps and Shadow Volumes by Eric Chan and Fr edo Durand [CD04]. | | Type | Master's thesis [Academic thesis] | | Year | 2005 | | Publisher | Informatics and Mathematical Modelling, Technical University of Denmark, DTU | | Address | Richard Petersens Plads, Building 321, DK-2800 Kgs. Lyngby | | Series | IMM-Thesis-2005-37 | | Note | | | Electronic version(s) | [pdf] | | BibTeX data | [bibtex] | | IMM Group(s) | Image Analysis & Computer Graphics |
|