Re: Shadow bug

New Message Reply Date view Thread view Subject view Author view

Kenneth B Russell (kbrussel++at++media.mit.edu)
Tue, 12 May 1998 22:18:40 -0400 (EDT)


> Have you looked at what you are drawing to the shadow map channel?
> Try checking this to see if there is anything undesirable appearing
> there.

I haven't checked this yet, though I'd guess the bogus shadow is
showing up in the shadow map. If I cover part of the window
during rendering, for example, I see a bogus shadow corresponding
to the occluded region; so I would guess that something (the
hardware, or some Performer internals) think there is a piece of
geometry there.

> depth buffer clearing is also important. Are there any transparent
> polygons in the scene which might be causing the occultation in the
> depth buffer image?

No, not in that scene. Again, the reason it's strange is that it
only occurs for some camera positions and viewing angles. From a
little testing, it looks like it's happening when the camera is
either facing (close to) the direction of the shadow frustum, or
is orthogonal to it. Zooming in/out changes the behavior as well.

Incidentally, I saw this problem the first time with the stock
shadows.c example when loading certain files out of
/usr/share/Performer/data; I'll have to check which ones. In that
example the camera wasn't moving but the light source was, and
occasionally a big dark square would appear in place of the
shadow. I can probably provide a screen shot tomorrow.

-Ken

__________________________________________________________________________
Kenneth B. Russell Synthetic Characters Group, MIT Media Lab
kbrussel++at++media.mit.edu http://www.media.mit.edu/~kbrussel
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
            Submissions: info-performer++at++sgi.com
        Admin. requests: info-performer-request++at++sgi.com


New Message Reply Date view Thread view Subject view Author view

This archive was generated by hypermail 2.0b2 on Mon Aug 10 1998 - 17:57:23 PDT

This message has been cleansed for anti-spam protection. Replace '++at++' in any mail addresses with the '@' symbol.