Yep, the size of different primitives cause differing effects. Though
the scaling testing I was carrying out was based on the other epsilon
value for Windows and DOS, that being 10,000,000 if I remember right.
And the scaling test was actually created by another person, and I
forgot the name, but believe it was at c.g.r.r. Ingeniously he scaled
both camera and objects together. I just reformatted the script some.
Here it is again:

// this shows several basic primitives at various sizes
// for epsilon value 10,000,000 in POV-Ray for Windows 3.0 version
// this has since been changed for 3.1d, 3.1e
// Note: Watch For Line Wrap!

#include "colors.inc"

//#declare SCALE = 1
 // at 1 normal.

#declare SCALE0 = 0.000000001
 // at .000000001 the cone remains normal! (gone in version 3.*?)
#declare SCALE1 = 0.00000666
 // at .0000066 the torii are chopped! no spheres or blobs.
#declare SCALE2 = 0.0031624
 // at .0031623 the spheres shrink out of sight (640x480)!
#declare SCALE3 = 44444
 // above 36000 the box distorts!
#declare SCALE4 = 666666
 // at 666666 the torii are cut in (opposite) half! blobs need sturm. no
box.
#declare SCALE5 = 999999999
 // at 999999999 the cone is indestructable (not in 3.*?)! spheres gone
again. no box, blobs or torii.

#declare SCALES = SCALE0  // replace # with 0-5

#declare OBJ = union{
cylinder {-y,y,1 rotate <30,45,30> pigment {Red}}
sphere {0,1 translate <0,-2.5,0> pigment {Green}}
sphere {0,1 translate <0,2.5,0> pigment {Yellow}}
cone {-y,1,y,0 rotate <30,30,30> translate <-3,0,0> pigment {Red}}
torus {1,.25 rotate <45,15,30> translate <3,-2.5,0> pigment {White}}
torus {1,.25 rotate <-45,-15,-30> translate <3,2.5,0> pigment {Black}}
box {<-.8,-.8,-.8>,<.8,.8,.8> rotate <45,45,45> translate <3,0,0>
 pigment {Red}}
blob {threshold .5
 sphere	{-x,1.25,.75 pigment {Green}}
 sphere	{x,1.25,.75 pigment {Yellow}}
 cylinder {-y,y,1.25,.75 pigment {DarkPurple}}
  rotate <-40,-30,-20> translate <-3,-2.5,0> sturm}
blob {threshold .5
 sphere	{-x,1.25,.75 pigment {Blue}}
 sphere	{x,1.25,.75 pigment {Orange}}
 cylinder {-y,y,1.25,.75 pigment {Pink}}
  rotate <40,30,20> translate <-3,2.5,0> sturm}
}
camera		{ location <0,0,-15> look_at <0,0,0> angle 40 scale SCALES }
light_source{ <+100,+100,-1000> color rgb<1.5,1.5,1.5> scale SCALES }
background	{ color Gray}
object		{ OBJ scale SCALES }

//END


Ken wrote:
> 
> Phil Clute wrote:
> >
> > Maybe someone else has tried this, or maybe I'm alone in my stupidity
> > but I tried to render an earth to scale 1pov unit = 1 meter which of
> > course comes
> > out to a mean value of 12,756 km in diameter or 6,378,000m as a sphere
> > radius.
> > Also, I put my light source(the sun) at <0,0,0> and then moved my earth
> > sphere
> > to <0,0,149600000000>(mean distance from sun).
> >
> > 1.What is the limit? How many places can I carry out my numbers around
> > the decimal point?
> 
>  Whatever your system's epsilon will support. I forget what the numbers
> are on a system by system basis. There was a mention that the Pov team made
> a change increasing the value it now supports in the release of Pov v3.15e
> They had choosen a lower valu previously and many users complained so they
> extended it's capability some. Check in the announcements group and see
> if the release notes give the upper value for that.
> 
> > 2.Why does the earth appear to have a grainy texture when it is only has
> > a blue
> > pigment? Scale is beyond pov's capabilities?
> 
> Posibility.
> 
> > 3.I was surprised that the earth was actually lit up at that distance.
> > Is this because
> > there is no falloff in the point lightsource?
> 
> Yes.
> 
> > 4.I also wanted to be able to rotate the camera around the earth but
> > couldn't
> > seem to make it work. I tried rearanging the order of the code with
> > rotate
> > before or after translate to no avail. How would this be accomplished?
> 
> camera { location < 0, 0, -120,000,000> look_at 0 rotate y*360 }
> 
> > Note: I know doing it at this size really isn't necessary. I just wanted
> > to give it a shot!
> 
> Bob Hughs did a rather extensive study on this issue of distance within
> the last 3-4 months or so. I forget where he psted his results but
> a little searching should be able answer your questions in more detail.
> 
> It was probably posted to this group. If not look in
> .tutorials
> .windows
> .scene-files
> 
> --
> Ken Tyler
> 
> mailto://tylereng@pacbell.net

-- 
 omniVERSE: beyond the universe
  http://members.aol.com/inversez/homepage.htm
 mailto:inv### [at] aolcom?Subject=PoV-News

Post a reply to this message