EnigmA Amiga Run 1997 July/EnigmA AMIGA RUN 20 (1997)(G.R. Edizioni)(IT)[!][issue 1997-07 & 08][EAR-CD IV].iso/lightwave/cgplwlist/cgplw00251-00500.txt/[e-mail]_Sun Dec 4 09:47:38 PST 1994.msg

Text File | 1994-12-25 | 4KB | 77 lines

Article: 359 of comp.graphics.packages.lightwave
Xref: netcom.com comp.graphics.packages.lightwave:359
Path: netcom.com!ix.netcom.com!netcomsv!netcomsv!sunspot.nosc.mil!newshub.nosc.mil!ihnp4.ucsd.edu!library.ucla.edu!europa.eng.gtefsd.com!howland.reston.ans.net!math.ohio-state.edu!news.acns.nwu.edu!uicvm.uic.edu!news.eecs.uic.edu!bert.eecs.uic.edu!bdupras
From: [e-mail] (Brian Dupras)
Newsgroups: comp.graphics.packages.lightwave
Subject: Re: Shadow Mapping
Date: 1 Dec 1994 04:19:58 GMT
Organization: University of Illinois at Chicago
Lines: 60
Distribution: na
Message-ID: <3bjite$[e-mail]>
References: <[e-mail]>     <[e-mail]>
NNTP-Posting-Host: bert.eecs.uic.edu
X-Newsreader: TIN [version 1.2 PL2]

Bryan Ewert ([e-mail]) wrote:
> On 19 Nov 1994 Brian Dupras orated:

>  BD> Shadows can be raytraced or "mapped".  Mapped shadows are extremely fast,
>  BD> but not as accurate.  Sometimes, however, the mapped shadows are more
>  BD> desireable visually that traced.

> OK -- you're someone who likes shadow mapping...

> We have tried to use shadow mapping in our animations.  However, we find
> that the results are not worth it.  The map most often ends up with large
> blocky artifacts.

> For example, create a large flat plane on the X & Y axes.  Place a smaller
> plane, extruded just a little, in front of this plane in Layout with a
> small gap between the two. Pivot the box on the Z axis so it is banked
> about 200 in the camera view.  Render this with shadow mapping.

> What we experienced was a "stair-stepped" shadow on the rear surface.  This
> was even more undesirable when animated.  Increasing the Shadow Map Size
> decreased the size of the "stair-steps", but we ran out of 32MB of RAM
> before it was acceptable.

> Are we missing something?  Or is this a fact of life with shadow mapping?
> And how are you getting effective results from it?

>       -- Bryan Ewert
>           JUUL: Animation, Ltd.
>            Member of the Amiga Users of Calgary
>             Look for AMUC's Centennial Scrolling Demo '94 -- coming soon!
>              Internet: [e-mail]

Disclaimer : I'm speaking from memory.  ;)

Well, the key to getting Mapping to work to your advantage is remembering
that the whole map has a fixed "resolution" that you must work with.
There are things you can do to hide the jaggies like "Shadow Fuzziness".
The best trick, though, was shown to me a few months back and I'll try to
re-explain it as best I can.

There's an option for the shadow map that either uses the cone angle of
the light to "stretch" the map across the view of the light, or that
allows you to define a square (from the light's point of view) where the
shadow is mapped.  Here's my example :  We created an off-shore oil rig,
and wanted to use shadow mapping with our "sun" light.  Since SM only
works with Spotlights, we had no choice but to use a spotlight REALLY far
away to mimic a distant light.  But, when we looked through the light,
the oil rig (the only important thing in the scene) was merely a spec on
the screen.  When the SM rendered, we got a shadow composed of about ten
or so fuzzy squares.  So, we turned off the "use cone angle" feature and
manually positioned the shadow map box (from the light view) over the
small area that the rig was in.  This cleared up the problem.

Again, this is from emory, so I might be a bit confused.  Any LW-Pro
editors out there that can offer more insight?  Or readers for that
matter...

Brian
[e-mail]