I am creating a city in Three.js. There are two ways of doing this:
a) Via Procedural programming
b) By creating the city in a 3D editor (Three.js editor/ 3DS Max etc.) and importing it as a "scene" in Three.js.
It is easier to design & create the city in an editor and import it as a "scene" in Three.js. Via procedural programming, things get a bit difficult when the city architecture has objects beyond randomly placed buildings. Objects such as parks, buildings with non-cubic architecture, water bodies, terrains of varying heights, and placements of these objects such that its overall architecture resembles that of a real city are easier to create in a 3D editor.
My question is: when using the editor route, other than the disk i/o overhead (due to importing & reading of the objects files), is there any other overhead that will affect the performance (e.g. rendering?) when compared to the procedural route? Especially, when the city is very large and consists of few hundred objects.
Any pointers/references will be helpful. TIA.
