Path: chuka.playstation.co.uk!news
From: "Christoph Lürig" <CLuerig@aol.com>
Newsgroups: scee.yaroze.programming.3d_graphics
Subject: Re: Perspective projection
Date: Wed, 15 Jul 1998 23:39:22 +0200
Organization: PlayStation Net Yaroze (SCEE)
Lines: 31
Message-ID: <35AD218A.4998D28B@aol.com>
References: <35A25BBF.F272A126@scee.sony.co.uk> <35A5DCF1.10D9E988@immd9.informatik.uni-erlangen.de> <35A5E3E1.488F60AF@scee.sony.co.uk> <35A60516.ED4487A1@immd9.informatik.uni-erlangen.de> <35AA2A37.745C1FF7@ndirect.co.uk>
NNTP-Posting-Host: 167-100-136.ipt.aol.com
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-Mailer: Mozilla 4.02 [en] (Win95; I)
To: aherbert@ndirect.co.uk

The perspektive transformation also maps into three dimensional space (what you
especially need for visibility calculation on a per fragment basis (z-buffer)).
The perspective transformation maps a pyramid, where the top has been cut off,
to a cube. If orthoraphic projection is applied to this normalized viewing
frustum, you get effectively a perspective projection.

Alex Herbert wrote:

> Christoph Luerig wrote:
>
> > What brings me up to the next question: Up to now I assumed (but I did
> >
> > not find it anywhere explicitely), that the z-values used for the
> >
> > ordering tables are the ones used after the application of the LS matrix
> >
> > but before the application of the perspective projection, is that correct
> >
> > (it would be relevant for stability aspects)?
> >
> > Christoph
> >
> Forgive me if I'm wrong, but I'm sure that this has to be correct.
> Surely the perspective projection converts 3D coodinates into 2D
> coordinates and therefore there is no z-axis once projection has been
> performed.
>
> Herbs