Archived

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

caffeineaddict

Question

Recommended Posts

Hi, I was recently faced with a lot of time on my hands(in a car for 6 hours to go on vacation) anyway, i got to thinking, in todays 3d games, load times are a concern for the developers as well as frame rates, i had an idea of how to reduce load time for some games like 1st person shooters maybe. I would like feedback on ideas as if this would be possible or practical. To my knowledge, now, most fully 3d games load all the 3d models that they need for that level in the initialization stage of the program, all of this loading takes time, and players don''t like to wait, well i had the idea of loading the entire 3d room model, then texturing all of the surfaces to what they should look like, what i''m thinking though is this, instead of loading complex shapes for the textured surfaces, load just boxes maybe, then when the player gets to a certain distance from an object that you want to be dynamic, ie, open doors etc. loads the more complex shapes, and information for the object. then when the player is a certain distance away from the dynamic object it just loads the simple shape in place again, this would cut down on the amount of geometry being used at one time, and would in my mind make better frame rates. Well thats the basic idea, if you have any ideas or questions please post them, i posted this at 5 in the morning and though everything made sense in my head, i''m not sure if it was conveyed in my typing so if you need me to clarify something let me know.

Share this post


Link to post
Share on other sites
Something similar to this is already in use in Quake engine based games: except, for textures. Each texture has 3 predefined "sub-mips", that is, 3 versions of the full sized texture progressively smaller. At large distances, the smallest sub-mip is used, and at close range, the original texture is rendered. The main reasoning behind this is that the player shouldnt be able to see such detail at long distances, so why keep it in memory.

To apply this to architecture, one may have to manually specify a similar set of circumstances for varying distances; whether this would need to be done manually or perhaps some AI based determination system could come up with these pre-definitions (ie. this room can be simplified down to a block, the polygons making up this clynder structure could also be a block, etc...).

Share this post


Link to post
Share on other sites