Geoff Wedig wrote:

> I think a sphere sweep might be your perfect solution.  Assuming you can
> compute the points that the cylinder begins and ends and intermediate
> points, this will give exactly the effect you want if you use one of the
> splines.  Even if you don't use a spline function, it should be cleaner than
> using a lot of cone segments as the edges won't be visible.
>

For the record, here is a croissant-making isosurface I posted in p.advanced-users
some time ago.

#declare major_axis=5;
#declare minor_axis=1;
#declare a=function{sqrt(x^2+z^2)-major_axis}
isosurface{
    function{sqrt(y^2+a^2)-minor_axis*(0.9+z*0.2)}
    contained_by{sphere{0,10}}
    eval
    threshold 0
    pigment{Red}
    rotate y*45
}

G.

Post a reply to this message