Please use this identifier to cite or link to this item:
|Title:||Managing and rendering large environments in games||Authors:||Hazrat Pradipta Ranjali.||Keywords:||DRNTU::Engineering::Computer science and engineering::Computing methodologies::Computer graphics||Issue Date:||2010||Source:||Hazrat, P. R. (2010). Managing and rendering large environments in games. Master’s thesis, Nanyang Technological University, Singapore.||Abstract:||Terrain rendering is one of the most challenging interactive visualizations in a game application. It contains a large number of triangles which may slow down the rendering process if there is no special algorithm implemented to reduce the triangles number. There are a lot of dynamic terrain algorithms to speed up the rendering process. Most of them have only one concern: to reduce the number of triangles to be sent to GPU for processing. However, game applications nowadays are not only concerned with frame rate, they also aim to achieve good visual quality and simulate the real world. This report presents a comparison, analysis, and characterisation of various type of terrain rendering algorithms which support terrain collision detection. In GeoMipMap algorithm, data streaming is improved by pre-computing necessary data. Visual artefact is fixed by considering triangle angles during triangulation. In GeoClipMap algorithm, GPU terrain collision is implemented to handle large number of queries. The performance of these improved methods is then tested using heightmap data from one commercial game engine. The number of data streaming invocation calls and the time required decrease significantly. Visual artefact does not appear in the rendering. And the performance of GPU terrain collision detection is almost unaffected by the number of queries.||URI:||http://hdl.handle.net/10356/36281||Fulltext Permission:||open||Fulltext Availability:||With Fulltext|
|Appears in Collections:||SCSE Theses|
Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.