POV-Ray : Newsgroups : povray.beta-test : Torus artifacts with distant camera and small angle Server Time
30 Jul 2024 00:26:23 EDT (-0400)
  Torus artifacts with distant camera and small angle (Message 1 to 4 of 4)  
From: Scott Gammans
Subject: Torus artifacts with distant camera and small angle
Date: 21 Feb 2002 11:36:14
Message: <3c7521fe$1@news.povray.org>
I've been having no luck getting orthographic cameras to work with the new
syntax introduced a few versions ago, so while waiting for the documentation
to catch up on this feature I've been trying to simulate orthographic views
by placing the camera very far away and using a very small angle to zoom in.
That's how I stumbled on this bug:

/***** START *****/
torus {10,5 texture {pigment {color rgb 0.8}}}


//camera {location <0,100,0> look_at <0,0,0> angle 75}
camera {location <0,5000,0> look_at <0,0,0> angle 1.5}

light_source {<0,10000,0> color <1,1,1>}
/****** END ******/

If you use the first camera, the scene renders correctly. If you use the
second far-away camera, all kinds of surface artifacts appear on the torus.
None of the other primitives that I use in my scenes (box, sphere, cylinder,
plane, cone, prism, text) show these artifacts, but there are several that I
don't use so I haven't checked them all.

This looks exactly like what happens with a sphere_sweep object if the
tolerance value isn't set correctly. Is POV-Ray using sphere_sweep under the
covers to implement torus? Or is this bug part of the ongoing changes in the
camera object?

Thanks,
Scott Gammans


Post a reply to this message

From:
Subject: Re: Torus artifacts with distant camera and small angle
Date: 21 Feb 2002 11:49:15
Message: <8o8a7usj59euf00jog36knrlcdhni2071v@4ax.com>
On Thu, 21 Feb 2002 11:36:14 -0500, "Scott Gammans"
<nos### [at] yahoocom> wrote:
> This looks exactly like what happens with a sphere_sweep object if the
> tolerance value isn't set correctly.

Do you remember why tolerance parameter was added to sphere_sweep ?
From MegaPOV doc: "the object casts shadows onto itself at some points because
of calculation errors". This is connected with floating point accuracy
(precision problem). There is no way to remove this problem from POV-Ray. IMO
this is also connected with your (and mine) torus problem. I've reported it
before. Compare your problem with
http://news.povray.org/rh44rtoeu5ifj145bv3iu3uirf6jb7ckrj%404ax.com

ABX


Post a reply to this message

From: Slime
Subject: Re: Torus artifacts with distant camera and small angle
Date: 21 Feb 2002 21:05:16
Message: <3c75a75c@news.povray.org>
Yeah, I had this a long time ago... it's because torii require higher order
(quartic) equations, which become innacurate when seen from thousands of
units away.

My solution was to use an orthographic camera. Better figure out how to do
that =)

(It would be nice, however, if the engine noticed what was going to happen
and then tested the ray as though it were shot from just outside the
object's bounding box.)

- Slime
[ http://www.slimeland.com/ ]
[ http://www.slimeland.com/images/ ]


Post a reply to this message

From: Thorsten Froehlich
Subject: Re: Torus artifacts with distant camera and small angle
Date: 22 Feb 2002 03:55:21
Message: <3c760779@news.povray.org>
"Scott Gammans" <nos### [at] yahoocom> schrieb im Newsbeitrag
news:3c7521fe$1@news.povray.org...
> I've been having no luck getting orthographic cameras to work with the new
> syntax introduced a few versions ago

Whatever problems you are having, they are not caused by any camera changes
as those are not activte by default (as noted in revision.txt and pointed
out in this group before).  So what you see is an error on your part.

    Thorsten


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.