Path: chuka.playstation.co.uk!scea!peter_alau@playstation.sony.com
From: "Charles Nicholson" <chnicholson@vassar.edu>
Newsgroups: scee.yaroze.programming.3d_graphics
Subject: Re: hhhelp mme . . I'mmm.... shrinking!!!??!!
Date: Tue, 26 Jan 1999 21:22:06 -0500
Organization: SCEA News Server
Lines: 30
Message-ID: <78lsk8$6d43@scea>
References: <78l3c2$gpd7@chuka.playstation.co.uk>
NNTP-Posting-Host: chnicholson.vassar.edu
X-Newsreader: Microsoft Outlook Express 4.72.3155.0
X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0

I could be wrong, but it sounds to me like you're keeping you're using
cumulative rotation (i. e. simply adjusting your rotation matrix instead of
recalculating it every frame from a position VECTOR and a rotation SVECTOR).
When you modify a pre-existing matrix a few thousand times, the fixed-point
rounding errors accumulate until the vertices of your TMD are irreparably
distorted.

Instead, hold your rotation in an SVECTOR, your position in a VECTOR, and
point your object's coordinate system to your own GsCOORDINATE2 structure.
Calculate your local coord system with RotMatrix(&<your rotation
SVECTOR>,&<your temporary in-function MATRIX>), and then use the newly
calculated temp MATRIX to set the coordinates.

That solved my problems, anyway- I couldn't figure it out when that first
happened to me.  I was implementing a fixed-point 3d program with a cube.  I
finally got it spinning and was so proud that I left it running when i went
to grab a bite to eat.  I came back and it had crumpled itself to the point
where it was nowhere near being recognizable as a cube....:)

hope that helped...
-charles

>my 3d models shrink in the x,z plane when rotated a lot!!??
>
>this is probably due to dogy code, but if anyone can help, please do so
>
>rob