Is brilliance ignored by radiosity by design? When an object
is lit by ambient source the shading cannot be altered with
brilliance. I would seem to be illogical that otherwise
equivalent ambient source and light source shade an object
differently.


_____________
Kari Kivisalo

Post a reply to this message

> Is brilliance ignored by radiosity by design? When an object
> is lit by ambient source the shading cannot be altered with
> brilliance. I would seem to be illogical that otherwise
> equivalent ambient source and light source shade an object
> differently.

Hmm, I'm afraid a little to step up on your level.  ;o)  What info do I have
to offer? Only personal guessing. You're the light expert. It's true that
brilliance is ignored by radiosity, just as specular highlight is ignored,
and I consider brilliance to actually be a kind of specular reflection.
Better say, an artificial mix between specular (pure reflection) and diffuse
(most wide spread reflection).

So I imagine that brilliance relies on the same formula as specular
reflection, and that formula cannot be handled by radiosity because "rays"
from radiosity are not precise enough. Their direction is ... diffuse.

Please let me know if my answer is wrong.

Btw, with radiosity it's not possible to have an ambient source *equivalent*
to a light source as you said, because the ambient source has to be a lot
bigger in size. Otherwise they aren't "catched" well at all, by objects in
the scene.. But you probably knows this.

Regards,
Hugo

Post a reply to this message

Hugo wrote:
>
> and that formula cannot be handled by radiosity because "rays"
> from radiosity are not precise enough. Their direction is ... diffuse.

Just look at this http://luxlab.com/tmp/rad_brilliance.gif
The radiosity rays definitely have a direction. The algorithm may
be limited only to Lambertian surfaces because of difficult integrals
or something.


> Btw, with radiosity it's not possible to have an ambient source *equivalent*
> to a light source as you said, because the ambient source has to be a lot
> bigger in size.

Equivalent ambient source has the same dimensions as the light source.
Just look up "equivalent" in the images group. I guess I did the impossible.
I marked it as experimental as I knew this was pushing the limits of what
POV-Ray can do. Brilliance is the limit :)


_____________
Kari Kivisalo

Post a reply to this message

> Brilliance is the limit :)

I don't know. The math minded people here don't reply, and it's in their
hands to explain. I saw your link. I assume you used low quality radiosity
(it looked that way) and with the small size of the light, I'm satisfied.

From a human perspective it might seem easy to implement brilliance to pure
radiosity light. But the current radiosity feature typically gives strong
artifacts, even with very high settings and it can be soooo slow. To have
brilliance included might contribute to a messy looking picture, without
insane high settings.

I suppose when computers go beyond 10 Ghz, people will code better
simulations of light. And our images will still take hours to render. But
for now, patience is .... the limit.

Regards,
Hugo

Post a reply to this message

On Thu, 18 Apr 2002 15:42:36 -0500, Kari Kivisalo wrote:


> Is brilliance ignored by radiosity by design? When an object is lit by
> ambient source the shading cannot be altered with brilliance. I would
> seem to be illogical that otherwise equivalent ambient source and light
> source shade an object differently.

I think what is happening is that POV uses radiosity to compute the
ambient light at each point, and ambient light has no direction. POV
can't include the normal in the calculations, because the samples don't
all go in the same direction. And it can't store the directions of each
individual sample ray, that would slow calculations down a *lot* and 
use much more memory.


-- 
Christopher James Huff <chr### [at] maccom>
POV-Ray TAG e-mail: <chr### [at] tagpovrayorg>
WWW: http://homepage.mac.com/chrishuff/

Post a reply to this message

Christopher James Huff wrote:
> I think what is happening is that POV uses radiosity
> to compute the ambient light at each point, and ambient
> light has no direction.

No wait a minute. The incoming rays do have directions. It's only the
reflected rays that don't have directions. Thus radiosity samples should
be adjusted with brilliance depending on the direction of the incoming
ray. I see no reason whatsoever why it can't be done. Granted, my
understanding of the radiosity algorithm is limited, but it does already
shade dependent on the directions, or you would get no shading at all.
Since it radiosity already supports shading for brilliance 1, I don't
see why it wouldn't work for other brilliance values.

Rune
--
3D images and anims, include files, tutorials and more:
Rune's World:  http://rsj.mobilixnet.dk (updated Apr 14)
POV-Ray Users: http://rsj.mobilixnet.dk/povrayusers/
POV-Ray Ring:  http://webring.povray.co.uk

Post a reply to this message

On Sun, 21 Apr 2002 17:11:06 -0500, Rune wrote:

> No wait a minute. The incoming rays do have directions. It's only the
> reflected rays that don't have directions. Thus radiosity samples should
> be adjusted with brilliance depending on the direction of the incoming
> ray. I see no reason whatsoever why it can't be done. Granted, my
> understanding of the radiosity algorithm is limited, but it does already
> shade dependent on the directions, or you would get no shading at all.
> Since it radiosity already supports shading for brilliance 1, I don't
> see why it wouldn't work for other brilliance values.

The direction of the incoming ray doesn't matter. Brilliance only uses
the angle of the light ray to the surface normal, what you would suggest
would make it look like all radiosity lighting was coming from the camera.
You would have to do it for every radiosity ray, using the direction of
the radiosity sample...and I don't know how this would interact with the
interpolation POV uses.

-- 
Christopher James Huff <chr### [at] maccom>
POV-Ray TAG e-mail: <chr### [at] tagpovrayorg>
WWW: http://homepage.mac.com/chrishuff/

Post a reply to this message

Christopher James Huff wrote:
> The direction of the incoming ray doesn't matter.

I meant the incoming ray to the surface of the object, i.e. the rays
that hit an object. Naturally the direction of those rays matter, more
presicely the angle of the light ray to the surface normal.

> Brilliance only uses the angle of the light ray to
> the surface normal

Exactly, and radiosity should take into account the brilliance when
dealing with this angle. How much the surface is lightened for a given
point and a given angle depends on the brilliance value used.

> what you would suggest would make it look like all
> radiosity lighting was coming from the camera.

You must have misunderstood me.

> You would have to do it for every radiosity ray,
> using the direction of the radiosity sample...

Radiosity is already currently shaded dependent on the angle of the
light ray to the surface normal, so adding support for brilliance should
not require any new information, but simply raise the influence of the
light ray to the power specified by the brilliance value.

Rune
--
3D images and anims, include files, tutorials and more:
Rune's World:  http://rsj.mobilixnet.dk (updated Apr 14)
POV-Ray Users: http://rsj.mobilixnet.dk/povrayusers/
POV-Ray Ring:  http://webring.povray.co.uk

Post a reply to this message