Like they say, that *little* grows when doing long renders.
Btw, I've seen an average 8% difference on this 233MMX 32Meg SDRAM (now
64) but haven't done any DOS renders since).
POVpro, the dos "speed-optimized" version of POV-Ray 3, was usually 20
to 40 percent faster, depending on the scene.

Spider wrote:
> 
> I've made a little comparsion on my own PC, on a fireworks scene.
> Hardware :
> P166mmx overclocked to 200Mhz
> 32Mb EDO
> 1 Mb Cache.
> 
> win95 :
>   Litestep as shell, no unnecessary programs run.
>   no editor, no menubar, no animation.
> 
> Test scene can be posted in .text.scene-files on demand.
> Same .pov file and .ini file in both tests.
> 
> Scene info :
> two infinite spheres
> sky_sphere
> a HF, 512*334*16
> 4 lights of diff. colour
> a fog.
> 3207 objects total
> Resolution 640 x 480
> Antialiasing Off
> Radiosity Off
> 
> Dos-Stats
> Memory : 6 194 964
> Parse  : 20.0 seconds
> Trace  : 8 minutes   3.0 seconds
> Time   : 8 minutes  23.0 seconds
> 
> Win-Stats
> Memory : 6 191 008
> Parse  : 27 seconds
> Trace  : 8 minutes   56.0 seconds
> Time   : 9 minutes   23.0 seconds
> 
> As you can see there is not much difference between the two.
> 
> //Spider

-- 
 omniVERSE: beyond the universe
  http://members.aol.com/inversez/POVring.htm
=Bob

Post a reply to this message

Bob Hughes wrote:
> 
> Like they say, that *little* grows when doing long renders.
> Btw, I've seen an average 8% difference on this 233MMX 32Meg SDRAM (now
> 64) but haven't done any DOS renders since).
> POVpro, the dos "speed-optimized" version of POV-Ray 3, was usually 20
> to 40 percent faster, depending on the scene.
yeah, but I can't say taht this was fair to windoze.
If I shall make a true compariosn I'll boot with poray as shell, using
no editor and at highest priority.

But then... Perhaps not.

//Spider

Post a reply to this message

Spider wrote:
> 
> Ken wrote:
> >
> > Nieminen Mika wrote:
> > >
> > > Spider <spi### [at] bahnhofse> wrote:
> > > : two infinite spheres
> > >
> > >   What's an infinite sphere?
> >
> > An infinite sphere is one that your camera is inside of, such
> > as when you are using a sphere for the sky, and forget to use
> > the inverse keyword to flip the sphere's surface normal.
> > Why it is considered an infine sphere, I know not, but it has
> > puzzled me for years now.
> 
> So THAT's why. I thought it was enough to have it hollow.
> does the invere give me any other advantages except that the infinite
> goes away?
> 
> //Spider

It will flip the direction of the surface normal. I suppose
an example would be if you were using a bumps normal you
would be seeing dents without the inverse.

-- 
Ken Tyler

tyl### [at] pacbellnet

Post a reply to this message

Spider <spi### [at] bahnhofse> wrote:
: So THAT's why. I thought it was enough to have it hollow.
: does the invere give me any other advantages except that the infinite
: goes away?

  I think it's more a conceptual think. I personally prefer to look at objects
from _outside_. Although povray only models surfaces I like to think about
(non-transparent) objects as solid. If I put the camera inside an object then
it conflicts against how I imagine the world since it should be solid but it
isn't. So I personally don't feel comfortable if I put the camera inside
a sphere which should be solid. That's why I always make sky spheres, walls,
etc, which 'inside' part is away from the camera (ie. the camera is outside
the object). That way I'm also always sure that I'm really looking at the
outer surface of the wall so pigments, normals, etc will work as expected.
  This is even more important if I use things like fog. What I think is:
"There is a wall there; a wall is solid material and I'm looking it from
outside; I'm not inside the wall but outside; so the fog will be seen outside
the wall (where I am) but not inside (so if the wall was transparent, there
should not be fog inside there)".
  Many people use walls (ie. planes) the wrong way, ie. with the camera inside
the wall instead of outside and then just define the wall as hollow so that
fog etc will work. This conflicts entirely with my notion of the 3D world I
am modelling. I would feel very uncomfortable at the wrong side of walls...

-- 
main(i){char*_="BdsyFBThhHFBThhHFRz]NFTITQF|DJIFHQhhF";while(i=
*_++)for(;i>1;printf("%s",i-70?i&1?"[]":" ":(i=0,"\n")),i/=2);} /*- Warp. -*/

Post a reply to this message

I won't argue with your view of the world. But my definition of a plane,
that I got from my algebra classes is that it has no 3D definition, it
is 2D, and therefor shouldn't have inside or outside. It took some time
before I oculd understand why and how the thing had it in pov. now I
know(csg and so on.) but before that. Why the HECK does it think I'm
INSIDE it, it HAS no inside, it CAN'T, it's 2D !!! *smile*
Welcome to the world of pov.

As your thoughts of the other things, i'll meditate on it :-)
//Spider

Nieminen Mika wrote:
> 
> Spider <spi### [at] bahnhofse> wrote:
> : So THAT's why. I thought it was enough to have it hollow.
> : does the invere give me any other advantages except that the infinite
> : goes away?
> 
>   I think it's more a conceptual think. I personally prefer to look at objects
> from _outside_. Although povray only models surfaces I like to think about
> (non-transparent) objects as solid. If I put the camera inside an object then
> it conflicts against how I imagine the world since it should be solid but it
> isn't. So I personally don't feel comfortable if I put the camera inside
> a sphere which should be solid. That's why I always make sky spheres, walls,
> etc, which 'inside' part is away from the camera (ie. the camera is outside
> the object). That way I'm also always sure that I'm really looking at the
> outer surface of the wall so pigments, normals, etc will work as expected.
>   This is even more important if I use things like fog. What I think is:
> "There is a wall there; a wall is solid material and I'm looking it from
> outside; I'm not inside the wall but outside; so the fog will be seen outside
> the wall (where I am) but not inside (so if the wall was transparent, there
> should not be fog inside there)".
>   Many people use walls (ie. planes) the wrong way, ie. with the camera inside
> the wall instead of outside and then just define the wall as hollow so that
> fog etc will work. This conflicts entirely with my notion of the 3D world I
> am modelling. I would feel very uncomfortable at the wrong side of walls...
> 
> --
> main(i){char*_="BdsyFBThhHFBThhHFRz]NFTITQF|DJIFHQhhF";while(i=
> *_++)for(;i>1;printf("%s",i-70?i&1?"[]":" ":(i=0,"\n")),i/=2);} /*- Warp. -*/

Post a reply to this message

Hmm, i tried the inverse and it din't work.
Any idea why?
<code>
sphere {
  <0,0,0>, 1
  texture {T_Clouds scale 1/2}
  scale 100000
  inverse
}
sphere {
  <0,0,0>, 1
  texture { T_Clouds scale 1/3.3 }
  rotate x*20
  scale 200000
  inverse
}
</code>

This still is called 2 infinite. Why ?
//Spider


Ken wrote:
> 
> Nieminen Mika wrote:
> >
> > Spider <spi### [at] bahnhofse> wrote:
> > : two infinite spheres
> >
> >   What's an infinite sphere?
> 
> An infinite sphere is one that your camera is inside of, such
> as when you are using a sphere for the sky, and forget to use
> the inverse keyword to flip the sphere's surface normal.
> Why it is considered an infine sphere, I know not, but it has
> puzzled me for years now.
> 
> --
> Ken Tyler
> 
> tyl### [at] pacbellnet

Post a reply to this message

Spider wrote:
> 
> Hmm, i tried the inverse and it din't work.
> Any idea why?
<snip>
> This still is called 2 infinite. Why ?
> //Spider

Pov is still going to consider this an infinte object, any
time your camera is inside a sphere, even with the inverse
word attached.

The gain from doing so is:

One - you won't get a warning that your camera is inside
a non-hollow object and

Two - it will allow you to use interior restricted items, like
fog, without consequence. Essentialy Pov will look at an object
that uses the inverse keyword as an object that has been turned
inside out.

The reason pov calls it an infinite object in all likelyhood
is because one of the developers decided there was a reason
for treating it as infinite and it's now deply embedded in the
program. Even if it's physicaly misleading it may not be
conceptually and really it doesn't matter any way does it ?

-- 
Ken Tyler

tyl### [at] pacbellnet

Post a reply to this message

Ken <tyl### [at] pacbellnet> wrote:
: The reason pov calls it an infinite object in all likelyhood
: is because one of the developers decided there was a reason
: for treating it as infinite and it's now deply embedded in the
: program.

  I have an idea of what could this reason possibly be: Since the camera is
inside the sphere, povray has to calculate ray-sphere intersections for
each pixel, just like with ordinary infinite shapes. Perhaps it's faster
to treat it as an infinite surface when you are sure that you really will
need to test it for every pixel (perhaps it has something to do with
vista/light buffers or whatever).

-- 
main(i){char*_="BdsyFBThhHFBThhHFRz]NFTITQF|DJIFHQhhF";while(i=
*_++)for(;i>1;printf("%s",i-70?i&1?"[]":" ":(i=0,"\n")),i/=2);} /*- Warp. -*/

Post a reply to this message

The soul of the group wrote:
> Pov is still going to consider this an infinte object, any
> time your camera is inside a sphere, even with the inverse
> word attached.
Oki, I thought the inverse would kill the infinite :-) Sorry for
misundrerstandings.
 
> The gain from doing so is:
> 
> One - you won't get a warning that your camera is inside
> a non-hollow object and
Ahh, I used hollow, so... :-)
 
> Two - it will allow you to use interior restricted items, like
> fog, without consequence. Essentialy Pov will look at an object
> that uses the inverse keyword as an object that has been turned
> inside out.
That might be interestings, once I do media.
 
> The reason pov calls it an infinite object in all likelyhood
> is because one of the developers decided there was a reason
> for treating it as infinite and it's now deply embedded in the
> program. Even if it's physicaly misleading it may not be
> conceptually and really it doesn't matter any way does it ?
Nope, it doesn't matter, but it is an interesting fact to know.

//Spider

Post a reply to this message

Bob Hughes wrote:
> 
> Like they say, that *little* grows when doing long renders.
> Btw, I've seen an average 8% difference on this 233MMX 32Meg SDRAM (now
> 64) but haven't done any DOS renders since).
> POVpro, the dos "speed-optimized" version of POV-Ray 3, was usually 20
> to 40 percent faster, depending on the scene.

This is only true with "usual" scenes. POVPro is great but uses a lower
precision calculation (to increase speed), but this introduces noise in
the results and makes it more difficult to have AA converge correctly.
With strong AA, it can be far slower than the official version.

-- 

http://www.geocities.com/vienna/7709
_  _  _  _   ______  _____  _____      _____  _____  _
|| ||// //  //   // /___ / /___ /     / ___/ / ___/ //
|| |// //  //   //     //     //     / /__  / /    //__
|| // //  //   //     //     //     / ___/ / /    //|__|
||//|//  //   //     //___  //___  / /__  / /__  // ||
|//|//  //___//     /____/ /____/ /____/ /____/ //  ||

Post a reply to this message