Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cwjcc!gatech!prism!vsserv!geomag!prem
From: prem@geomag.fsu.edu (Prem Subrahmanyam)
Newsgroups: comp.graphics
Subject: Re: How does radiosity work?
Keywords: radiosity
Message-ID: <119@vsserv.scri.fsu.edu>
Date: 25 Aug 89 18:51:17 GMT
References: <1540@ndmath.UUCP> <273@nap1.cds.wpafb.af.mil>
Sender: news@vsserv.scri.fsu.edu
Reply-To: prem@geomag.UUCP (Prem Subrahmanyam)
Organization: Florida State University Computing Center
Lines: 29


    I am not a radiosity expert, but here is my 2 cent's worth from reading
    former SIGGRAPH radiosity papers and similar papers in other graphic
    magazines.  

    Each object is composed of numerous polygonal patches.  The algorithm
    sums up the contributions of light from all the other patches, some
    reflect diffusely,  some are light sources, etc.  Now, the paper I've
    read most thoroughly mentioned some kind of iterating, where it goes
    through several (hundred) times and recalculates the light contributions.
    It can easily be seen why if we happened to sum up all the contributions
    to a particular patch before a patch near a light source had been
    calculated properly, thus forcing us to recalculate once we had a better
    value for a patch near a light source.  Radiosity inherently has a large
    storage need, as the color values of all the polygonal patches in the
    scene need to be stored.  Once the picture has been calculated properly,
    displaying it is a "simple" matter of hidden line removal display of
    a number of colored polygons, thus we have "view independency"...all
    we have to do is redisplay all the polygons from a different viewpoint.

    This is in no wise a technical report on radiosity, but an attempt to
    explain simply the underlying algorithm.  Other factors need to be 
    included, like determining whether or not a particular patch contributes
    to the patch we are considering due to the fact that another patch might
    be in the way (shadowing), etc.  In any case, I've tried my best to
    "explain in about one paragraph" what radiosity does.

    Ray-tracing is more my cup-o'-tea.
    ---Prem Subrahmanyam