So I've got a scene with radiosity settings like this:

global_settings
 {radiosity
   {count 250
    nearest_count 10
    error_bound .5
    minimum_reuse .005
    brightness 1}}

Yet the scene, a green room with a few green objects lit by an outer cyan
sphere with ambient 1, renders like this:
http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_rad_b1.png
If I increase the brightness to 2, I get this:
http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_rad_b2.png
But that just artificially brightens the scene. I shouldn't have to do that
to get the scene to be decently lit, or so I gather from the documentation.
Here's what the scene looks like with just a point light and no radiosity:
http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_conv.png
These images were generated with Display_Gamma left unset. I've tried
playing around with the gamma settings, both assumed_gamma and
Display_Gamma, but anything that makes this scene bright enough causes
other scenes to be too bright. This skeleton of a scene, for instance:

sphere
 {3*z, 1
  pigment {rgb <.75, .75, .75>}
  finish {diffuse 0 ambient 1}}

renders a circle that is more than 75% gray when I use the same gamma
correction that makes the green room look right. Saving POV-Ray's output to
a different format from PNG doesn't seem to help.

So I guess my question is, what am I doing wrong? Radiosity isn't supposed
to be this dark, is it? And I should be able to use the same gamma settings
for conventional lighting and radiosity, right?

Post a reply to this message

Sudo schrieb:
> 
> So I guess my question is, what am I doing wrong? Radiosity isn't supposed
> to be this dark, is it?

It is.

When you have a room with a window for example the sky outside is easily 
10 times brighter than the wall of the room.  Your materials have to 
account for this to lead to realistic results.  Radiosity brightness has 
nothing to do with this.

-- Christoph

Post a reply to this message

There's nothing strange about your images, this is what i would expect from
my experiences with radiosity. Rather than raising brightness, it might be
a better idea to make your sky brighter instead, by raising either the
colour or the ambient well above 1. This better approximates reality - when
indoors, the outside environment is incredibly bright compared to any
possible interior lights. This should light the room better while keeping
the interior objects' interaction realistic at the same time.

Also, as I'm sure others here will soon confirm, as a rule it's a good idea
to use assumed_gamma 1. Hopefully somebody will also explain why, because I
can't remember the reasoning! :)

Bill

Post a reply to this message

Sudo nous apporta ses lumieres en ce 2007/06/14 09:03:
> So I've got a scene with radiosity settings like this:
> 
> global_settings
>  {radiosity
>    {count 250
>     nearest_count 10
>     error_bound .5
>     minimum_reuse .005
>     brightness 1}}
> 
> Yet the scene, a green room with a few green objects lit by an outer cyan
> sphere with ambient 1, renders like this:
> http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_rad_b1.png
> If I increase the brightness to 2, I get this:
> http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_rad_b2.png
> But that just artificially brightens the scene. I shouldn't have to do that
> to get the scene to be decently lit, or so I gather from the documentation.
> Here's what the scene looks like with just a point light and no radiosity:
> http://i31.photobucket.com/albums/c395/mimir227/povray/greenroom_conv.png
> These images were generated with Display_Gamma left unset. I've tried
> playing around with the gamma settings, both assumed_gamma and
> Display_Gamma, but anything that makes this scene bright enough causes
> other scenes to be too bright. This skeleton of a scene, for instance:
> 
> sphere
>  {3*z, 1
>   pigment {rgb <.75, .75, .75>}
>   finish {diffuse 0 ambient 1}}
> 
> renders a circle that is more than 75% gray when I use the same gamma
> correction that makes the green room look right. Saving POV-Ray's output to
> a different format from PNG doesn't seem to help.
> 
> So I guess my question is, what am I doing wrong? Radiosity isn't supposed
> to be this dark, is it? And I should be able to use the same gamma settings
> for conventional lighting and radiosity, right?
> 
> 
> 
The cyan sphere is not bright enough. Some possible solutions:
- As you tried, add an actual light_source. Make it an area_light.
- Increase the ambient around 10~20.
- Usr rgb <.5,10,10> for your sphere.
- Place a cyan box with a high ambient and/or rgb value in front of the window, 
but out of sight from the camera location.


-- 
Alain
-------------------------------------------------
You know you've been raytracing too long when you can describe in perfect, 
legal, pov syntax, how to re-create everything in your computer room using 
primitives and csg operations.
fish-head

Post a reply to this message