POV-Ray : Newsgroups : povray.general : Infinite Objects Server Time
27 Nov 2024 16:32:19 EST (-0500)
  Infinite Objects (Message 1 to 10 of 15)  
Goto Latest 10 Messages Next 5 Messages >>>
From: Stephen
Subject: Infinite Objects
Date: 21 May 2008 12:24:16
Message: <f4j834991908jrf4arraldfilngp0ro8tc@4ax.com>
I've just noticed that if I scale a sphere by 10000 it becomes an
infinite object. Are there any implications to this?
-- 

Regards
     Stephen


Post a reply to this message

From: Jan Dvorak
Subject: Re: Infinite Objects
Date: 22 May 2008 02:28:58
Message: <483512aa@news.povray.org>
Stephen napsal(a):
> I've just noticed that if I scale a sphere by 10000 it becomes an
> infinite object. Are there any implications to this?
-since double precision is only 16 digits you cannot have objects with 
scales differing by more than 8 digits.
-if the camera is inside the sphere you must make it hollow or inverse.
-such object will probably be traced by every ray. This has performance 
implications of using complicated objects (an unlikely need)

this method is often used to create a background because a sky_sphere 
doesn't cooperate with radiosity very well.
-- 
the ultimate time-killer:
+a0.0 +am2 +r9

Johnny D


Post a reply to this message

From: Warp
Subject: Re: Infinite Objects
Date: 22 May 2008 03:21:51
Message: <48351f0f@news.povray.org>
Jan Dvorak <jan### [at] centrumcz> wrote:
> -if the camera is inside the sphere you must make it hollow or inverse.

  It's not a must.

> this method is often used to create a background because a sky_sphere 
> doesn't cooperate with radiosity very well.

  What do you mean? I have never noticed any problems.

-- 
                                                          - Warp


Post a reply to this message

From: Jan Dvorak
Subject: Re: Infinite Objects
Date: 22 May 2008 04:17:50
Message: <48352c2e$1@news.povray.org>
Warp napsal(a):
> Jan Dvorak <jan### [at] centrumcz> wrote:
>> -if the camera is inside the sphere you must make it hollow or inverse.
> 
>   It's not a must.
>

I'm sorry for this.
This is only needed if one uses media.

>> this method is often used to create a background because a sky_sphere 
>> doesn't cooperate with radiosity very well.
> 
>   What do you mean? I have never noticed any problems.
> 

Some people have reported black spots. I have noticed some too, but I 
thought that it was caused by sharing samples between parallel surfaces.


Post a reply to this message

From: Stephen
Subject: Re: Infinite Objects
Date: 23 May 2008 11:14:03
Message: <hond34hug8d4nrln91npietsj7a04vgbcf@4ax.com>
On Thu, 22 May 2008 08:28:52 +0200, Jan Dvorak
<jan### [at] centrumcz> wrote:

>-since double precision is only 16 digits you cannot have objects with 
>scales differing by more than 8 digits.
>-if the camera is inside the sphere you must make it hollow or inverse.
>-such object will probably be traced by every ray. This has performance 
>implications of using complicated objects (an unlikely need)

Thanks Jan,
	I was just wondering. I am using a large sphere for a
skysphere as I want a planet to be obscured slightly by clouds and a
skysphere will not do this. Incidentally I had to add the hollow
keyword to the sphere and the union it was in to get rid of the
warning.
-- 

Regards
     Stephen


Post a reply to this message

From: Warp
Subject: Re: Infinite Objects
Date: 23 May 2008 11:28:34
Message: <4836e2a2@news.povray.org>
Stephen <mcavoysAT@aoldotcom> wrote:
>         I was just wondering. I am using a large sphere for a
> skysphere as I want a planet to be obscured slightly by clouds and a
> skysphere will not do this.

  Btw, a layer of clouds is never a sphere centered at the camera.
If you are making a huge sphere to work as a sky sphere centered at
the origin (or the camera), that won't really work as realistic clouds.

  The sky_sphere is not even intended to be used for clouds (never mind
what the documentation and example scenes claim). It's intended for the
change in coloration of the sky which depends on the angle of incidence
(sky usually looks whiter closer to the horizon). Clouds should usually
be added as a plane or similar object.

-- 
                                                          - Warp


Post a reply to this message

From: Stephen
Subject: Re: Infinite Objects
Date: 23 May 2008 15:14:25
Message: <0e5e34hfm9davr1r5ote0322im2o4utcv7@4ax.com>
On 23 May 2008 11:28:34 -0400, Warp <war### [at] tagpovrayorg> wrote:

>Stephen <mcavoysAT@aoldotcom> wrote:
>>         I was just wondering. I am using a large sphere for a
>> skysphere as I want a planet to be obscured slightly by clouds and a
>> skysphere will not do this.
>
>  Btw, a layer of clouds is never a sphere centered at the camera.
>If you are making a huge sphere to work as a sky sphere centered at
>the origin (or the camera), that won't really work as realistic clouds.

Drat! Would you expand a little, please?
At the moment I'm just using the sphere as a means of having a sky for
reflections on the sea and I will start working on the textures soon.
So any help will be appreciated.

>  The sky_sphere is not even intended to be used for clouds (never mind
>what the documentation and example scenes claim). It's intended for the
>change in coloration of the sky which depends on the angle of incidence
>(sky usually looks whiter closer to the horizon). Clouds should usually
>be added as a plane or similar object.

Interesting, would you suggest a sky_sphere for the air with a "cloud
plane" in front of my background planet?
-- 

Regards
     Stephen


Post a reply to this message

From: Nicolas Alvarez
Subject: Re: Infinite Objects
Date: 23 May 2008 15:34:33
Message: <48371c49@news.povray.org>
Stephen wrote:
> On 23 May 2008 11:28:34 -0400, Warp <war### [at] tagpovrayorg> wrote:
> 
>>Stephen <mcavoysAT@aoldotcom> wrote:
>>>         I was just wondering. I am using a large sphere for a
>>> skysphere as I want a planet to be obscured slightly by clouds and a
>>> skysphere will not do this.
>>
>>  Btw, a layer of clouds is never a sphere centered at the camera.
>>If you are making a huge sphere to work as a sky sphere centered at
>>the origin (or the camera), that won't really work as realistic clouds.
> 
> Drat! Would you expand a little, please?
> At the moment I'm just using the sphere as a means of having a sky for
> reflections on the sea and I will start working on the textures soon.
> So any help will be appreciated.

http://f-lohmueller.de/pov_tut/backgrnd/p_sky1.htm


Post a reply to this message

From: Alain
Subject: Re: Infinite Objects
Date: 23 May 2008 15:38:17
Message: <48371d29$1@news.povray.org>
Stephen nous illumina en ce 2008-05-23 15:14 -->
> On 23 May 2008 11:28:34 -0400, Warp <war### [at] tagpovrayorg> wrote:
> 
>> Stephen <mcavoysAT@aoldotcom> wrote:
>>>         I was just wondering. I am using a large sphere for a
>>> skysphere as I want a planet to be obscured slightly by clouds and a
>>> skysphere will not do this.
>>  Btw, a layer of clouds is never a sphere centered at the camera.
Unless you are at the center of a free floating bubble of air.
>> If you are making a huge sphere to work as a sky sphere centered at
>> the origin (or the camera), that won't really work as realistic clouds.
> 
> Drat! Would you expand a little, please?
> At the moment I'm just using the sphere as a means of having a sky for
> reflections on the sea and I will start working on the textures soon.
> So any help will be appreciated.
You can flatten the sphere with an uneven scaling and shift it down 
(scale<1,0.5,1> translate -Radius/4*y). You'd still get the reflections.
The sky_sphere will also get reflected. Good for the background of the sky.
Remember that a sky_sphere never interact with any light and never cast any 
shadow. It's a unit sphere centered on the camera, whose pigment get returned 
whenever a ray is not stopped by an object.
> 
>>  The sky_sphere is not even intended to be used for clouds (never mind
>> what the documentation and example scenes claim). It's intended for the
>> change in coloration of the sky which depends on the angle of incidence
>> (sky usually looks whiter closer to the horizon). Clouds should usually
>> be added as a plane or similar object.
> 
> Interesting, would you suggest a sky_sphere for the air with a "cloud
> plane" in front of my background planet?
That's the general idea. The cloud plane is transparent exept for the clouds 
that will have various levels of opacity.
The sky_sphere hold the background pigment that is beyong everything else.
You can then add some fog or media.

-- 
Alain
-------------------------------------------------
Please hassle me, I thrive on stress.


Post a reply to this message

From: Warp
Subject: Re: Infinite Objects
Date: 23 May 2008 15:41:43
Message: <48371df6@news.povray.org>
Stephen <mcavoysAT@aoldotcom> wrote:
> Interesting, would you suggest a sky_sphere for the air with a "cloud
> plane" in front of my background planet?

  Maybe something along the lines of:

camera { location <0, 0, -1> look_at z angle 60 }
light_source { <-5, 10, 0>*1000, 1 }

sky_sphere
{ pigment
  { gradient y color_map
    { [0 rgb <.8, .9, 1>]
      [.4 rgb z]
      [1 rgb z*.7]
    }
  }
}

plane // ground
{ y, -10
  pigment
  { granite color_map
    { [0 rgb <1, .75, .5>]
      [1 rgb <.9, .6, .4>]
    }
    scale 10
  }
}

plane // clouds
{ -y, -100
  pigment
  { bozo color_map
    { [0 rgb .7]
      [.3 rgb 1 filter .4]
      [.5 rgbf 1]
    }
    turbulence 1
    scale 100
  }
  finish { ambient 1 }
}

sphere
{ <0, -2, 70>, 8
  finish { reflection { .75, 1 } }
}

-- 
                                                          - Warp


Post a reply to this message

Goto Latest 10 Messages Next 5 Messages >>>

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.