Trevor G Quayle nous illumina en ce 2009-02-21 09:02 -->
> Warp <war### [at] tagpovrayorg> wrote:
>> Trevor G Quayle <Tin### [at] hotmailcom> wrote:
>>> Try using an area light but with an open cylinder around it to limit the light
>>> to a spot.  You can effect the amount of soft shadowing by the size of the
>>> light/cylinder, etc.  A while back I posted a method for making a square
>>> spotlight using area lights.  It would probably provide a good starting point.
>>> I'll dig around and see if I can find it.
>>   Out of curiosity: Didn't you even think of trying to define a light_source
>> as being a spotlight and an area light *at the same time*?
>>
>>   Because, as others noted, povray perfectly supports that.
>>
>> --
>>                                                           - Warp
> 
> No I didn't.  I guess I must've assumed that the user had maybe tried and it
> didn't work.
> 
> This trick does work for recatngular spotlights though...
> 
> -tgq
> 
> 
You can make it circular.

-- 
Alain
-------------------------------------------------
   I know I'm not sexy. When I put my underwear on I can hear the Fruit- 
of-the-Loom guys giggling.
	Rodney Dangerfield

Post a reply to this message

Alain wrote:
> You can...
> - add area_light to your spot_light. This will make the shadows it cast 
> blury.
> - using photons and your filter, average a few normals. Keep the normals 
> small.
> 
> Sample of the averaged normals:
> 
> cylinder{z, z+0.01, Radius pigment{rgbt 1}
>     #local Smpl = 5; #local R = seed(0);
>     normal{average normal_map{
>         #while(Smpl >0)
>         [1 bumps 0.1 scale 0.01 translate<rand(R),rand(R),rand(R)>*0.01]
>         #local Smpl = Smpl -1;
>         #end
>             }
>         }
>     interior{ior 1.5}
>     photons{target refraction on}
>     }
> 
> The value of "Smpl" controlls the smoothness of the bluring. Don't use a 
> large value, as it affect the rays on two surfaces, and squares the 
> number of samples.
> The value after bumps controlls the bluring's intensity.
> 
> 

The averaged normals idea seems to work fairly well, thanks!

When I place the filter close to the light source (so the projected 
light is a large radius) it requires a ridiculous amount of photons to 
look anywhere near smooth, but such is the price of the effect it seems.

I notice that I can't use any filtering objects between the light source 
and the diffuser without using pass_through, in which case the filtering 
effect is completely ignored. Is there a way to fix that? I'm using a 
simple image_map filter to effectively make an image projector, and if I 
use pass_through, it's gone, and if I don't the diffuser cylinder just 
makes a black shadow. Is this just something I have to live with?

Thanks again for the normals idea though!

Post a reply to this message

CShake wrote:

> I notice that I can't use any filtering objects between the light source 
> and the diffuser without using pass_through, in which case the filtering 
> effect is completely ignored. Is there a way to fix that? 

Did you make the filtering object a photon target?

Post a reply to this message

I've been giving this some more thought. It seems to me that whatever method you
come up with for your 'diffusion gel', the spotlight/area-light will need actual
multiple light sources (multiple point lights)--in order to cast 'diffused'
light onto objects themselves (which a regular area_light doesn't do, as it's
'really' just a point light.) Even with your good diffusion-gel object in the
beam, there's still that limitation. Perhaps using *multiple*
spotlight/area_lights would work, randomly placed in a small area or volume to
represent your single spotlight. Sounds render-intensive, though.

You might also want to consider an area-light's optional 'area_illumination'
feature (new to the 3.7 betas) that Warp came up with.

http://news.povray.org/povray.binaries.images/thread/%3C478e748d%40news.povray.org%3E/

I haven't worked with it yet, but his sample images look quite nice, with smooth
diffuse lighting on objects. And it renders faster than an equivalent group of
point lights.

It may just be that you can get the effect you want without even resorting to a
'diffusion gel.'

KW

Post a reply to this message