Generating an ideal feature list and recommending software for video game procedural urban modelling
Kafami Khorasani, Mohammad (2017)
Kafami Khorasani, Mohammad
Turun ammattikorkeakoulu
2017
All rights reserved
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-2017102616329
https://urn.fi/URN:NBN:fi:amk-2017102616329
Tiivistelmä
The recent developments in computer systems’ processing technology allow graphic artists to create increasingly complex 2D and 3D video games assets. Consequently, the process of generating 3D graphics for video game is becoming more complicated. The purpose of this study is to recommend a list of effective features that accelerate and ease the workflow of 3D urban modelling, and help environment artists to create high-quality and optimized outcomes.
To obtain the desired list of features, the fundamentals of video games environment design were reviewed and the feasibility and advantages of relevant methods were studied. Then, the process of custom urban assets modelling consisting of designing and creating several types of environmental elements was described. At this stage, the applications and software, such as 3D studio Max 2015 and GhostTown 1.0 plugin were examined. Afterwards, two broadly used procedural 3D generating applications, namely Esri CityEngine and Houdini were technically inspected, and their capabilities and respective feature sets were reviewed. As a result, a final feature list was compiled.
The generated feature list provides information on essential capabilities of procedural 3D generating applications in handling a series of consecutive procedures such as: geospatial data entry, road and street networks, parceling, building constructions, texture mapping and finally optimization of the outcomes. Based on the features list, a table was also generated to benchmark the investigated applications.
Despite the user-friendly interface and strong node based façade generator, GhostTown does not deliver game-oriented results and requires major time and effort for optimization of geometry and texture map. Therefore, this plugin is not recommended. Considering the advantages and limitations of CityEngine and Houdini and their complementary quality, a hybrid method is proposed in this thesis. Since CityEngine includes a strong toolset for managing the geospatial data entry, terrain and road network generation, parceling, rule-based building modelling, it can favorably be employed at the beginning phases. While Houdini can collaborate partially in parceling and building generating, it can also perfectly fulfill the rest of the procedure, consisting of models’ topological perfection, optimized texture mapping and creation of the level of details.
To obtain the desired list of features, the fundamentals of video games environment design were reviewed and the feasibility and advantages of relevant methods were studied. Then, the process of custom urban assets modelling consisting of designing and creating several types of environmental elements was described. At this stage, the applications and software, such as 3D studio Max 2015 and GhostTown 1.0 plugin were examined. Afterwards, two broadly used procedural 3D generating applications, namely Esri CityEngine and Houdini were technically inspected, and their capabilities and respective feature sets were reviewed. As a result, a final feature list was compiled.
The generated feature list provides information on essential capabilities of procedural 3D generating applications in handling a series of consecutive procedures such as: geospatial data entry, road and street networks, parceling, building constructions, texture mapping and finally optimization of the outcomes. Based on the features list, a table was also generated to benchmark the investigated applications.
Despite the user-friendly interface and strong node based façade generator, GhostTown does not deliver game-oriented results and requires major time and effort for optimization of geometry and texture map. Therefore, this plugin is not recommended. Considering the advantages and limitations of CityEngine and Houdini and their complementary quality, a hybrid method is proposed in this thesis. Since CityEngine includes a strong toolset for managing the geospatial data entry, terrain and road network generation, parceling, rule-based building modelling, it can favorably be employed at the beginning phases. While Houdini can collaborate partially in parceling and building generating, it can also perfectly fulfill the rest of the procedure, consisting of models’ topological perfection, optimized texture mapping and creation of the level of details.