EnigmA Amiga Run 1997 July/EnigmA AMIGA RUN 20 (1997)(G.R. Edizioni)(IT)[!][issue 1997-07 & 08][EAR-CD IV].iso/lightwave/lwmlist/95.lightwave-07/000326_owner-lightwav…bcom.webcom.com_Tue Jul 18 23:01:07 1995.msg

Internet Message Format | 1995-08-06 | 3KB

Received: by webcom.webcom.com
    (1.37.109.15/16.2) id AA247423667; Tue, 18 Jul 1995 23:01:07 -0700
Return-Path: <[e-mail]>
Received: from access1.digex.net by webcom.com with ESMTP
    (1.37.109.15/16.2) id AA247343661; Tue, 18 Jul 1995 23:01:01 -0700
Received: (from erniew@localhost) by access1.digex.net (8.6.12/8.6.12) id CAA23067 ; for ; Wed, 19 Jul 1995 02:05:58 -0400
Date: Wed, 19 Jul 1995 02:05:57 -0400 (EDT)
From: Ernie Wright <[e-mail]>
To: [e-mail]
Subject: Re: Scaling the universe
In-Reply-To: <9507181339.AA18418@itgmsm>
Message-Id: <[e-mail]>
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: [e-mail]
Precedence: bulk

Concerning the limited precision of digital floating point numbers,
[e-mail] wrote:

> The following issue is being discussed regarding the Reality Labs
> Rendermorphics system.  I'm curious as to whether or not any LW user
> may have come accross a similar problem?  I've allready thought of
> an alternative solution (which only visually solves the problem).
> Any advice?

Try not to worry about it.  Unless you're modeling or animating at
astronomical or nuclear scales (the film Powers of Ten comes to mind),
you aren't very likely to come across real animation situations in
which limited precision is going to be a big issue.

If you want to see limited precision in action, try this.  In Modeler,
create the planet Pluto as a sphere with a radius of 1.55 meters and a
center at (5870000, 0, 0) meters.  Use a large number of polygons, say
48 sides and 24 segments, just to convince yourself that you're not
looking at a low polygon count kind of ugliness.  After creating Pluto,
hit "a" to center it in all three views.

How do you avoid this?  Well, just don't do it.  You can divide the
biggest distance in your scene by the smallest distance, and if the
result is bigger than, oh, a million, you might get in trouble.  In the
case of Pluto, the coordinates need to be accurate to at most 50 mm in
order for it to look spherical, so the spread is

   5,870,000 / 0.05 = 117,400,000.

The obvious solution is to move Pluto a lot closer to the sun.  Your
audience certainly won't be able to tell the difference.

- Ernie

--
Ernie Wright <[e-mail]> sent this message.
To Post a Message           : [e-mail]
Un/Subscription Requests To : [e-mail]
(DIGEST)                 or : [e-mail]
Administrative Items To     : [e-mail]