RE: [Algorithms] Bullets on walls

[Stephen J B. <sj...@li...>]

On Sun, 23 Jul 2000, Brian Marshall wrote:

> 	The N64 supports this op. It was used for the bullet holes in Perfect Dark
> (and Goldeneye.). It still doesn't give a glitch free result. It does 'clip'
> to the edges on polygons when the decal goes over the edge of the polygon
> its sat on, but its really just scissoring with Z to do that.
> 	One thing to remember is that the error in the z slope calculations on the
> small decal polygon is higher than that on a larger (in screen space)
> polygon. So you are always going to have some accuracy worries. Using one
> decal polygon the same size as the surface it's resting on may work very
> well. With individual decals someone will always get them all on one wall if
> you let them - which is a big hit with individual decals, small with one big
> decal.
> 	Unless you can use precisely the same polygon for generating Z values you
> will need to add Z-bias to minimize glitching.
 
Strictly speaking (in OpenGL at least), even making the polygons the
same size won't guarantee no Z-fighting.  The OpenGL spec allows implementations
to generate different Z values even when vertices are shared when certain GL
states are not shared.

The reason for that is that you might (for example) have to do a software
fallback for one set of states and not for the other - since software and
hardware rendering are unlikely to generate the exact same Z values, you
can't guarantee no Z-fighting.

However, you'll probably get away with it on most machines.

Steve Baker                      (817)619-2657 (Vox/Vox-Mail)
L3Com/Link Simulation & Training (817)619-2466 (Fax)
Work: sj...@li...           http://www.link.com
Home: sjb...@ai...       http://web2.airmail.net/sjbaker1