Alain <ele### [at] netscapenet> wrote:
> With the no_shadow, the lights will illuminate the shadowed area just as if the 
> object was not there at all, making the shadow very faint.

  What does that last part mean?

-- 
                                                          - Warp

Post a reply to this message

Warp nous apporta ses lumieres en ce 23-01-2007 04:24:
> Alain <ele### [at] netscapenet> wrote:
>> With the no_shadow, the lights will illuminate the shadowed area just as if the 
>> object was not there at all, making the shadow very faint.
> 
>   What does that last part mean?
> 
Kenneth suggested having two copies of an object. One with no_image and lighted 
by an area_light, the second with no_shadow lighted by an aray of point lights.

-- 
Alain
-------------------------------------------------

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
> Kenneth <kdw### [at] earthlinknet> wrote:
> > What
> > strikes me immediately is that lightgrid.jpg (which I take to be an array
> > of real point lights) produces *much* smoother illumination on the wall and
> > sphere than does the area light.  Surprisingly so.
>
>   Why surprisingly?

Ah, I see what you mean (and you're right, it's not so surprising after
all): The area light's *illumination* is just a point--producing the same
direct illumination on objects as a single point light would (disregarding
shadows.)  While the light grid of real point lights is lighting the
objects from multiple sources and directions.  Thus automatically smoother.

What *is* surprising though, is that the large highlight on the sphere is so
smooth, when using the light grid; I would have expected multiple, small and
discreet highlights. (Perhaps there are, but they appear to blend together
quite well.) An interesting and useful discovery, for me.
>
> > And the shadows from the
> > sphere look smoother than I would have imagined.
>
>   Well, there's the same amount of point lights than in the area light.
> The shadows should be pretty similar.

But given that the light grid is casting multiple *distinct* shadows, while
the area light is creating "jittered" shadows from all(?) of its lights,
shouldn't the area light produce much smoother shadowing?  (Although, I
admit, your experiment seems to prove me wrong there.)

KW

Post a reply to this message

Alain <ele### [at] netscapenet> wrote:

> With the no_shadow object, the lights will illuminate the shadowed area just as if
the
> object was not there at all, making the shadow very faint.

You mean, the shadow will be partially "lit," no longer black or
dense.That's an interesting point, and would seem to be true. But in
experiments I've done, using two simple spheres (one with no_shadow, one
with no_image) and a single point light just for simplicity, what I see is
that the final shadow, cast onto a white ground plane, is still quite dark.
*Almost* black...even when compared to a pure black object
next to it. (And this is using assumed_gamma of
1.0)  Plugging in assumed_gamma of 2.0--which I usually do, for my own
reasons--the shadow becomes even darker.

I'm a bit mystified as to why the shadow *is* so dark. As you say, the
darkness in the shadow area *should* be a mix of "shadow" and "no shadow"
..... that is, 50% illuminated. But that isn't so. Not that I'm complaining!
The trick seems to work.  But I haven't yet tried it with an area light and
a corresponding light grid, where there could be a subtle but distinct
"shift" in shadow darkness between the fully shadowed area and the soft
shadow from the area light. That's just a guess, though.

KW

Post a reply to this message

Kenneth <kdw### [at] earthlinknet> wrote:
> What *is* surprising though, is that the large highlight on the sphere is so
> smooth, when using the light grid; I would have expected multiple, small and
> discreet highlights. (Perhaps there are, but they appear to blend together
> quite well.) An interesting and useful discovery, for me.

  POV-Ray uses the phong reflection model
(http://en.wikipedia.org/wiki/Phong_reflection_model) for highlights,
and the whole idea of this model is that it "blurs" the reflection of
the light source on the surface of the object.
  If you have many such blurred reflections close together, of course
they will form a uniform mass.

> But given that the light grid is casting multiple *distinct* shadows, while
> the area light is creating "jittered" shadows from all(?) of its lights,
> shouldn't the area light produce much smoother shadowing?  (Although, I
> admit, your experiment seems to prove me wrong there.)

  The area_light is not jittered in my example scene. Both examples
should produce identical shadows (except for the tiny differences
produced by 'adaptive').

  There's no visible banding because there are so many point lights
in both cases (10x20). If the camera was moved close enough of one of
the shadow borders then some banding could become visible in both cases.

-- 
                                                          - Warp

Post a reply to this message

"Kenneth" <kdw### [at] earthlinknet> wrote:
>...in
> experiments I've done, using two simple spheres (one with no_shadow, one
> with no_image) and a single point light just for simplicity, what I see is
> that the final shadow, cast onto a white ground plane, is still quite dark.
> *Almost* black...

I took a more critical look at my own test image, and I have to conclude
that the final shadow isn't "almost black," it IS black. The shadow (from
the no_image sphere) is somehow taking complete precedence. There's no
"brightness mixing" at all. Odd, but interesting. Perhaps that's just the
logical operation of POV's shadow calculations.

Meanwhile, my little test has presented another useful trick, something I've
been pondering for awhile but couldn't figure out how to do, until now:
creating a variable-density shadow. By making the pigment of the no_image
sphere something like rgbt <1,0,0,.5> (the rgb components don't really
matter), its shadow can be made more or less dense/dark. So that an object
that appears to be solid (the no_shadow sphere, in my case) can cast a
variable-darkness shadow.  May not be physically realistic, but could be a
useful artistic tool. Likewise, using rgbf <1,0,0,.5> would cast a reddish,
variable-density shadow.

Ken

Post a reply to this message

"Kenneth" <kdw### [at] earthlinknet> wrote:

>>Likewise, using rgbf <1,0,0,.5> would cast a reddish,
> variable-density shadow.
>

Actually, using rgbft is better; varying f and t produces any shade or
density of shadow.

SO... after *finally* getting around to rendering my "two identical objects"
experiment using an area light, I discovered something interesting: The
order of the two objects in the scene file -- which one comes first --
becomes important. My original test used just a single point light, to
prove the concept; and the order didn't matter.  But with an area light,
the order DOES matter.  The no_image sphere--the one casting the
shadow--has to come first in the code, otherwise NO shadow is produced.
I'll leave it up to more brilliant minds than mine to figure out why that
is so.  :-) But the trick works quite well, even using rgbft as above, with
no shadow problems that I can see.

Ken

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:

>   POV-Ray uses the phong reflection model
> (http://en.wikipedia.org/wiki/Phong_reflection_model) for highlights...

Thanks. (The page also contains a link to an explanation of "dot product,"
another useful read for me.)

KW

Post a reply to this message

Kenneth nous apporta ses lumieres en ce 29-01-2007 04:46:
> "Kenneth" <kdw### [at] earthlinknet> wrote:
>> ...in
>> experiments I've done, using two simple spheres (one with no_shadow, one
>> with no_image) and a single point light just for simplicity, what I see is
>> that the final shadow, cast onto a white ground plane, is still quite dark.
>> *Almost* black...

> I took a more critical look at my own test image, and I have to conclude
> that the final shadow isn't "almost black," it IS black. The shadow (from
> the no_image sphere) is somehow taking complete precedence. There's no
> "brightness mixing" at all. Odd, but interesting. Perhaps that's just the
> logical operation of POV's shadow calculations.

> Meanwhile, my little test has presented another useful trick, something I've
> been pondering for awhile but couldn't figure out how to do, until now:
> creating a variable-density shadow. By making the pigment of the no_image
> sphere something like rgbt <1,0,0,.5> (the rgb components don't really
> matter), its shadow can be made more or less dense/dark. So that an object
> that appears to be solid (the no_shadow sphere, in my case) can cast a
> variable-darkness shadow.  May not be physically realistic, but could be a
> useful artistic tool. Likewise, using rgbf <1,0,0,.5> would cast a reddish,
> variable-density shadow.

> Ken


Are you using light_group? In that case, if the no_shadow object is in a group 
with a given light source, that light will only illuminate that object, unless I 
missunderstood the working of a light_group.

-- 
Alain
-------------------------------------------------
Don't try so hard, the best things come when you least expect them to.

Post a reply to this message

Alain <ele### [at] netscapenet> wrote:

> Are you using light_group? In that case, if the no_shadow object is in a group
> with a given light source, that light will only illuminate that object, unless I
> missunderstood the working of a light_group.
>

Nope, no light_group.  But I think your understanding is correct.

You're question gives me some more thoughts. I'll try these tricks with a
light_group, to see if anything odd or interesting pops up.

KW

Post a reply to this message