"POVMAN" <s### [at] acom> wrote:
> Looks like the same tech' I used in SOTW.  get enough boxes and the surface
> comes to life.  real cool image BTW.

Its a fun technique, nice SOTW too!

Post a reply to this message

"PM 2Ring" <nomail@nomail> wrote:
> Nice one, Janet. The harmonious colour balance complements the jaggedness
> created by the boxes.
Thanks!
> > Also just wanted to say all the greeble stuff looks so fun!
>
> Yeah! Me wanna play with greebles, too... but I've got a bunch of povray.org
> Links to fix first!
Me wanna play all the time. Hmm, since you're fixing links, do you have this
one? http://www.pims.math.ca/knotplot/

Post a reply to this message

"Janet" <par### [at] attnet> wrote:
> Well, its pretty much what the subject line says. The code can be found here
> http://www.deviantart.com/view/33552164/  It can be modified for the number
> of boxes (n) and the box sizes and size variance too. Also just wanted to
> say all the greeble stuff looks so fun!
> Janet

Nice image & great colours!
But could you have found a harder way to code it?
I've never used a matrix transformation.  I don't know how it works.

This code will do the same thing, just put the box in a good starting
position and let 'rotate' do all the work...

#version 3.6;
// --------------------------------------------------------------
#global_settings{
  assumed_gamma 1.0
  max_trace_level 5
  radiosity{count 35 error_bound 0.1}
}
// --------------------------------------------------------------
#macro RandLH(L,H)
  (rand(R0)*(H-L)+L)
#end
// --------------------------------------------------------------
#declare R0 = seed(1); // random seed
#declare SR = 10; // sphere radius
#declare Bmin = 0.3; // minimum box scale
#declare Bmax = 1.0; // maximum box scale
// --------------------------------------------------------------
union{
  #declare num = 3500;
  #local ca = 0;
  #while(ca<num)
    //box{-0.5, 0.5
    superellipsoid{<0.2, 0.2> scale 0.5
      scale<RandLH(Bmin, Bmax), RandLH(Bmin, Bmax), RandLH(Bmin, Bmax)>
      translate SR*y
      rotate<RandLH(0,180), RandLH(0,360), 0>
    }
    #local ca = ca+1;
  #end
  texture{pigment{color rgb 1} finish{ambient 0 diffuse 1}}
}
// --------------------------------------------------------------
light_source{<-2, 2, 1>*1000 color rgb <1.0, 1.0, 0.5>*2}
light_source{<0, 0, 0> color rgb <0.2, 0.5, 0.5>*2}
camera{
  perspective
  location<0, 0, -50>
  up y*(image_height/image_width)
  right x
  sky y
  look_at <0, 0, 0>
  angle 40
}
// -------------------------------------------------- end of file

Post a reply to this message

"Mark Birch" <las### [at] hotmailcom> wrote:
> Nice image & great colours!
Thanks
> But could you have found a harder way to code it?
Yes, I probably could have.
> This code will do the same thing, just put the box in a good starting
> position and let 'rotate' do all the work...
This is a very good point Mark, and now I feel like an idiot for not having
thought of it that way. My thinking started out as using the equation for
the surface of a sphere and it kind of went from there. The matrix
transform is kind of interesting, though. There's a good explanation of it
here: http://www.geocities.com/evilsnack/matrix.htm

Post a reply to this message

"Janet" <par### [at] attnet> wrote:
> The matrix
> transform is kind of interesting, though. There's a good explanation of it
> here: http://www.geocities.com/evilsnack/matrix.htm

That's much clearer.  Thanks for the link!

Post a reply to this message

"Mark Birch" <las### [at] hotmailcom> wrote:
I tried your code. Works great. Looks cool with the superellipsoids too.
Thanks, Janet

Post a reply to this message

"Mark Birch" <las### [at] hotmailcom> wrote:
> I've never used a matrix transformation.  I don't know how it works.

basically, it's 3 vectors that define the "new" x,y and z. if you make sure
that
the new x y and z are still perpendicular and have a length of one, it's
just a rotation. The remaining 3 numbers in a transformation matrix is just
a translation vector. Internally povray uses matrices also. The big
advantage is that it's very simple (and fast!) to combine and use
transformations.

the matrix transformation in povray is very useful if your data is already
in matrix form because it's exported form some other 3D program/simulation.
you can even position a povray camera using a matrix.
I used it to render some ODE simulations in povray:
http://news.povray.org/povray.binaries.animations/thread/%3Cweb.443593c62b28687ba8399d8d0%40news.povray.org%3E/?mtop=22
6348&moff=10
you will have make some of the values in the matrix negative, because povray
and ODE use different left/right-hand coordinate systems. also, everyone
seems to store and write the numbers of the matrix in a different way :-)
starting top-down or left-to-right.

Post a reply to this message

"Janet" <par### [at] attnet> wrote:
> "PM 2Ring" <nomail@nomail> wrote:
> > Nice one, Janet. The harmonious colour balance complements the jaggedness
> > created by the boxes.
> Thanks!
> > > Also just wanted to say all the greeble stuff looks so fun!
> >
> > Yeah! Me wanna play with greebles, too... but I've got a bunch of povray.org
> > Links to fix first!
> Me wanna play all the time. Hmm, since you're fixing links, do you have this
> one? http://www.pims.math.ca/knotplot/

There are a couple of links to knots, at least one of which is dead, but I
can't see that one in the list. At least, it's not in the top level,
although it may be linked to from one of the existing links. I don't know
if I'm actually allowed to add links to the list, but I'll certainly
recommend this one: it's excellent, Janet! I could spend many hours
browsing that page.

I tried to do a simple trefoil knot many years ago in DKBtrace, using a REXX
program to calculate the positions of the spheres with trig functions. I got
some pretty images, but I was never able to get the exact shape I wanted.

Post a reply to this message

"PM 2Ring" <nomail@nomail> wrote:
> There are a couple of links to knots, at least one of which is dead, but I
> can't see that one in the list. At least, it's not in the top level,
> although it may be linked to from one of the existing links. I don't know
> if I'm actually allowed to add links to the list, but I'll certainly
> recommend this one: it's excellent, Janet! I could spend many hours
> browsing that page.
Knotplot has been mentioned in the POV-Ray newsgroups. The creator does seem
to be updating it, so thats always a good thing.
> I tried to do a simple trefoil knot many years ago in DKBtrace, using a REXX
> program to calculate the positions of the spheres with trig functions. I got
> some pretty images, but I was never able to get the exact shape I wanted.
You are a brave, brave person. :)
A knotplot image in my gallery ==> http://www.deviantart.com/view/31865726/

Post a reply to this message

"Janet" <par### [at] attnet> wrote:
> "PM 2Ring" <nomail@nomail> wrote:
> > There are a couple of links to knots, at least one of which is dead, but I
> > can't see that one in the list. At least, it's not in the top level,
> > although it may be linked to from one of the existing links. I don't know
> > if I'm actually allowed to add links to the list, but I'll certainly
> > recommend this one: it's excellent, Janet! I could spend many hours
> > browsing that page.
> Knotplot has been mentioned in the POV-Ray newsgroups. The creator does seem
> to be updating it, so thats always a good thing.

Ok, it should already be linked then, but if not I'll try to get it on the
list.

> > I tried to do a simple trefoil knot many years ago in DKBtrace, using a REXX
> > program to calculate the positions of the spheres with trig functions. I got
> > some pretty images, but I was never able to get the exact shape I wanted.
> You are a brave, brave person. :)

I was just playing around with 3D Lissajous figures. DKBtrace was a pure
SDL, there was no programming abilities, so anything fancy had to be done
with external programs. REXX is really good for quick development.

> A knotplot image in my gallery ==> http://www.deviantart.com/view/31865726/

Nice! You should try it with some reflection mapping.

Post a reply to this message