Index: [thread] [date] [subject] [author]
  From: joshy <joshy@mindspring.com>
  To  : ggi-develop@eskimo.com
  Date: Thu, 04 Feb 1999 11:00:34 -0500

Re: libggi3d non-clarity

> > > How is radiosity typically done?
> >
> > Beats me ;). I don't think the two steps are intertwined using radiosity, but
> > using, for instance, ray-tracing they may be, because in ray-tracing lighting
> > (as well as other things like hidden-surface line removal) basically comes for
> > free during the rasterization.
> 
> Radiosity doesn't (I just looked it up, but only quickly, so this may be
> wrong), you have to do a low resolution pass over the scene,
> and pick out blocks where it's colour should be altered. Then you ray
> trace, putting in the colour distortion as you go. If you do it properly,
> it's far too slow to be useful.


radiosity is where you divide every surface in your world up into
polygonal patches and then you figure out how much light each patch
reflects on to each other patch.  this is really really slow because if
you have N patches then you've
got N^2 compares to do.  the advantage is that these calculations are
view independant. if you change the view point but don't move any
objects, then you don't have to recalculate anything.   this technique
is usually combined with raytracing or other rendering methods to create
a scene. it's too slow for any kind of realtime rendering, so it's only
used for film work and architectual walk throughs,  or other things
where it's okay to pre-calculate lighting (like quake).

- joshy

-- 

                                         because no one ever suspects
                                             the butterfly

Index: [thread] [date] [subject] [author]