"Bald Eagle" <cre### [at] netscapenet> wrote:
> "And" <49341109@ntnu.edu.tw> wrote:
>
> > I'm looking for patterns which have random characteristic(like bump), but I need
> > apply it to associate 0~360 degree(angular), because that is a not completed yet
> > challenge for me.
>
> OK, I ran a few more experiments and came up with the attached.
>
> #declare Rad = function {sqrt (x*x+y*y)}
> #declare S = 0.5;
> #declare W5 = function {Ang (x, y, z) * sin (Rad(x,y,z)*tau/RadFreq)*f_noise3d
> (x/S, y/S, z)}
>
> I hope that's more like what you want.
>
> > In fact, I know your W2 function have a seam... I have tried your (W2) method
> > when I started coding the macro for my wood generator program last year. So I
> > exchange it by a array value - function that time.
>
> Yeah - I've noticed a radial asymmetry and a seam as well - I'll try to think
> about a way to not have any discontinuities.
>
> - BE

I'm thinking and trying yours. In fact, that is interesting, too. But the seam
is still there. I have known that if you angular direct, (your Ang(x,y) ), you
unavoidable get the seam, or you mirror the pattern, but if so, you will produce
a symmetric pattern, not a randomize pattern.

I have found (just this thread I post) some combination to generate
angular-association or radial-association pattern to naturally avoid the seam,
they just don't have that seam. If you notice? I divide the (x, y) by its vector
length (your Rad(x,y,z) here) to f_bumps input, or just input a function which
used your Rad(x,y,0) to z. It is simple, I just never noticed the method before.

Post a reply to this message