POV-Ray : Newsgroups : povray.programming : gradient : Re: gradient Server Time
6 Oct 2024 13:47:06 EDT (-0400)
  Re: gradient  
From: Thomas Willhalm
Date: 3 Dec 2002 02:57:34
Message: <3dec63ed@news.povray.org>
Christoph Hormann wrote:
> 
> Thomas Willhalm wrote:
>> 
>> Hello,
>> 
>> if I remember correctly, I came across some code in povray or one of its
>> patches that calculates the gradient of a pattern or function. By this
>> I mean a function that numerically finds the direction of a real valued
>> function f at some point p in space in which f increases most. This was
>> done by evaluating the function slightly off p. To improve numerical
>> stability, the code in question used the four vertices of a tetrahedron
>> and evaluated the function f there. So, if v1,...,v4 denote the vertices
>> of a tetrahedron centered at the origin, the function f was evaluated at
>> p+v1, p+v2, p+v3 and p+v4.
>> 
>> Unfortunately I cannot find this code anymore. Does somebody know what
>> I'm writing about and the file to look at?
> 
> You have some POV-SDL based macros for gradient calculation in
> 'math.inc'.  Apart from that the isosurface normal calculation code does
> something similar too, function 'IsoSurface_Normal()' in 'isosurf.cpp'.

Thanks, this function does what I want, although it's not the function
I vaguely remember. Perhaps it was in a different program, but I have 
no clue what program this should be.

Thanks again
Thomas


Post a reply to this message

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