Re: Shadow bug

New Message Reply Date view Thread view Subject view Author view

Kenneth B Russell (kbrussel++at++media.mit.edu)
Wed, 13 May 1998 20:02:04 GMT


> The self shadow theory seems unlikely, particularly since only parts
> seem to be occulted and you tried different offsets.
>
> Maybe there is some interaction with pfLayers (stencil or polygon
> offset)
> or other state information and the shadow rendering?
>
> One thing to try is to disable antialiasing and see if the problem
> persists, this might point an accusatory finger ad zbuffer compression
> on iR.

We don't do anything fancy in our application yet; no stencil
planes, no antialiasing, and no manual use of the polygon offset
extension.

The problem is concisely shown in this example from the Performer
2.2 MR.

If you compile /usr/share/Performer/shadows.c from
/usr/share/Performer/src/pguide/libpf/C with the following
command line:

cc -o shadows shadows.c -lpf -lpfdu -lpfutil -lGL -lGLU -lX11 -limage

(we're using the 7.1 compilers patched up to #2072, on an Onyx2
iR running Irix 6.4), and then run the executable by

shadows stool.flt

(where stool.flt is in /usr/share/Performer/data), then there
will be two "blips" as the shadowed light does its orbit; the
shadow suddenly turns into a big black box for a moment. Those
blips look exactly like what I'm seeing in my application.

We have a support contract; should I be reporting this problem in
some other venue than here?

Thanks,

-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 - 17:57:24 PDT

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