How did quake3 calculated dynamic vertex light on MD3 actor?

on many video card, hardware vertex light is slowly.are there any approximately methods can do this? so that the dynamic objects
can seems better in my lightmaped scenes.
any one sholud help me? thanks a lots.

No, I believe quake3 uses normal per vertex lighting on its models.
Here’s a link to the online version of The Graphics Programming Black Book:- http://www.ddj.com/documents/s=865/ddj0165f/

This gives details on how quake 1&2 did its rendering. Quite out of date, but still a good read.

Q3 uses a ‘light grid’…

But I don’t know any details since I’ve only read about it a few years ago.

Q3 uses in essence a 3D lightmap grid which is very similiar to a 3D lightmap texture but isn’t used as a texture. Instead each point of the grid is simply a directional light source (as determined at map compile time by sampling the lights in the map at each point of the grid). When a model needs to be lit, the engine references those points of the grid that are closest to the vertices of the model and sums them in some manner (probably using dot products of the vertex normal and light vector as a fractional multiplier) that I have yet to fully dissect. The resultant is used as the vertex color. This is done for each vertex of the model.

I should mention, that is how the static lights contribute to light the model. As far as dynamic lights, I assume that light is just added in as needed.

[This message has been edited by DFrey (edited 02-24-2002).]