Which is the best method to process a 3D World...
Author
Hello I am begining with a project similar to Froger 2 3D or Pacman 3D, where the game is viewed from a Fixed Top-down camera view, and I am using OpenGL for it and I would use 3D Studio MAX as my level designer exporting my 3D World to the ASE File Format, but my trouble is Which could be my best choice for store my 3D World?, I really appreciate some addvice about it. The Game Will look somthing like Crash Bandicoot, which is a Fixed view, so I want to avoid any unnecesary polygon drawing.
PS: I Really apreciate so much if any could give some URLS about 3D Engines design and how to store and manage simple 3D Worlds.
Thanks in addvice!!
I assume you have your file format and whatnot, and just want a way to sort out which parts of the level not to draw.
I myself am a devout follower of both the benefits and the difficulties with the "tile-based" and "track-based" sorts. So, first, the "tile-based".
On map load, take all the polygons (or all the data chunks, however you do it), and chunk them into an array, a grid. This grid should be sized so that a single "sector", a grid square, takes up anywhere from two-thirds to the entire screen. Then, when you render, you take only the polygons/data in the current "sector" and all of the sectors adjacent to it. Easy enough?
The "track-based" is much simpler, and would be what they used in Crash Bandicoot. You store all your data in horizontal "strips". These strips are an array that extends "forwards", with each strip of data going side-to-side. Then, easy enough, you render from the strip the camera is in, and all of the strips a reasonable distance out. So, with a 100-meter viewing range, and strips that are 5 meters "long" (And however wide), you would render 20 strips out from the camera.
I hope this helps! If not (or if almost
), Email me at (zlaj@hotmail.com).
I myself am a devout follower of both the benefits and the difficulties with the "tile-based" and "track-based" sorts. So, first, the "tile-based".
On map load, take all the polygons (or all the data chunks, however you do it), and chunk them into an array, a grid. This grid should be sized so that a single "sector", a grid square, takes up anywhere from two-thirds to the entire screen. Then, when you render, you take only the polygons/data in the current "sector" and all of the sectors adjacent to it. Easy enough?
The "track-based" is much simpler, and would be what they used in Crash Bandicoot. You store all your data in horizontal "strips". These strips are an array that extends "forwards", with each strip of data going side-to-side. Then, easy enough, you render from the strip the camera is in, and all of the strips a reasonable distance out. So, with a 100-meter viewing range, and strips that are 5 meters "long" (And however wide), you would render 20 strips out from the camera.
I hope this helps! If not (or if almost
), Email me at (zlaj@hotmail.com).
This topic is closed to new replies.
Advertisement
Popular Topics
Advertisement
