Jump to content
  • Advertisement

Archived

This topic is now archived and is closed to further replies.

Mille

Modern game render algorithms

This topic is 5318 days old which is more than the 365 day threshold we allow for new replies. Please post a new topic.

If you intended to correct an error in the post then please contact us.

Recommended Posts

I''ve always been wondering what kind of algorithms modern PS2/XBox games uses to render the part of the world/scene which is visible at even given time when playing. For instance I know that Quake/Doom uses BSP trees, but what about all the new games – I mean hot games like Helo, splinter Cell, True Crimes etc. ? Anyone knows if they uses principles as Octree, Protal or a combination ? There must be some pro game programmer here in this forum that can tell – or many just tell me that I’m not allowed to know -, since it’s more or less intern company stuff Thanks..

Share this post


Link to post
Share on other sites
Advertisement
You need to Google. It's complex and there is not one solution for every game :

Space partition:
- BSP, portals (based on convex sets).
- regular grid, quadtree, octree
- various trees based on bounding volumes hierachies.
(Oriented Boxes, AABoxes, Spheres)

Visibility:
- backfacing
- frustrum culling
- occlusion culling (PVS, Portal, HOM, occluders, vectorial map, 1bit depth masks (with perfect sort))

Level of detail (view dependent) :
- static (several meshes per object)
- progressive subdivisions (triangle trees, ex ROAM, displacement)
- implicit surfaces

The general intention behind all these techniques is to render what's sufficient for human vision and to maintain a constant high FPS. This also applies to texture space (fill rate). The ideal would be one texel/pixel.

And else, yes I am pro level, but everything is freely available on the web, I work for myself now. Else don't overestimate the techniques used in most games. Games are developped rather quickly. In the pro world, the challenge is not to implement everything optimally. Efficience is mostly, finish the game and sell it, that is do the minimum required to implement a game design. If a regular grid is enough for one game then no need for an octree. If the world is small, cut into many small levels (possibly an indoor adventure game) then portals or just VF culling will be enough, no complex occlusion or LOD required because the overdraw rate is bounded and sitances are small.



[edited by - Charles B on February 27, 2004 8:16:31 AM]

Share this post


Link to post
Share on other sites

  • Advertisement
×

Important Information

By using GameDev.net, you agree to our community Guidelines, Terms of Use, and Privacy Policy.

We are the game development community.

Whether you are an indie, hobbyist, AAA developer, or just trying to learn, GameDev.net is the place for you to learn, share, and connect with the games industry. Learn more About Us or sign up!

Sign me up!