Level of detail

From Valve Developer Community
Revision as of 06:52, 14 October 2007 by Beeswax (talk | contribs) (→‎See also)
Jump to navigation Jump to search

Level of Detail. To reduce the number of polygons that need to be rendered, the Source engine will render lower-quality models if the model is very far away. If it moves closer to the player's point of view, it will change to a higher-quality model, with more polygons and more textures. As it continues to come closer, it will gain more detail until it reaches perfect quality.

The advantage of using LOD is obvious: Objects that are far away, where they cannot show much detail anyways, will not use up nearly as many memory and processing resources for the player. In extreme cases, they may not even be rendered at all. It is up to the developer then, to decide how many levels of quality are appropriate for a model, and at what distances they should be "activated" at. For example, in Half-Life 2, the G-man character model had five different models.

Disadvantages:

Carefully consider the view distances of any models making use of levels of detail in a map. If, for example, the player is staring at the model when the level of detail changes it can be distracting.

See also