3D Interior environment optimization for VR
Oksanen, Mika (2022)
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2022122031159
https://urn.fi/URN:NBN:fi:amk-2022122031159
Tiivistelmä
In this thesis, the primary question was to determine if Level of Details (LODs) were optimizing the performance of the VR platform inside Unity. The secondary question was to find other optimization methods and study how they optimize performance for VR. The thesis was commissioned by the research group Futuristic Interactive Technologies of Turku University of Applied Sciences for their multi-user application named TUAS VR Social Platform.
To achieve the objective of this thesis, an altered interior of a real-world building was created. Two copies of the same scene were created where one was not optimized, and the other was fully optimized. The test was carried out by first measuring the unoptimized scene and then comparing it to other optimization methods. The unoptimized scene was then compared to the optimized scene.
The conclusion was that LODs do not significantly affect performance at this scale. The most performance gains occurred with occlusion culling and bake lights. Implementing all optimization techniques in the scene at once resulted in significant performance gains. Nevertheless, the test was ultimately successful as it showed that optimization can improve performance with VR.
To achieve the objective of this thesis, an altered interior of a real-world building was created. Two copies of the same scene were created where one was not optimized, and the other was fully optimized. The test was carried out by first measuring the unoptimized scene and then comparing it to other optimization methods. The unoptimized scene was then compared to the optimized scene.
The conclusion was that LODs do not significantly affect performance at this scale. The most performance gains occurred with occlusion culling and bake lights. Implementing all optimization techniques in the scene at once resulted in significant performance gains. Nevertheless, the test was ultimately successful as it showed that optimization can improve performance with VR.