POV-Ray : Newsgroups : povray.programming : A possible solution for coincident surfaces? : Re: A possible solution for coincident surfaces? Server Time
28 Apr 2024 13:08:48 EDT (-0400)
  Re: A possible solution for coincident surfaces?  
From: MessyBlob
Date: 16 Apr 2009 20:35:01
Message: <web.49e7ce3b149f0cf9addfbead0@news.povray.org>
"MessyBlob" <nomail@nomail> wrote:
> The only approach I think will work is to assign an 'outsideness order'
> to each object, and two global thresholds for precision: 'start'
> (larger number) and 'end' (smaller number). The 'end' would be larger
> than the rounding error of the FP calculations.
? [...]

For completeness, I should add that a more than two surfaces can be coincident.
One can use the same algorithm to interpolate between two of three ordered
surfaces, noting that the terms would need to be recalculated for each ray.

Also, you could keep the assignment of 'outsideness order' simple, by assigning
incremental numbers to each successive object added to the scene. You might
then want to allow a script to assign absolute or relative changes to the
value, for users who need to change this order.

p.s. I think this method is the answer for a frequent requirement, so would be
interested to hear if anyone has any notes on feasibility, and whether they'd
like this feature to be implemented without unacceptably affecting performance.


Post a reply to this message

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