I'm rendering a artist's studio that has sunlight coming in from the glass
roof.  Leaving that as the only light source and cranking up the ambiant on
the contents of the room doesn't look real at all.

Putting a fill light off to one side makes shapes look better (not "flat")
but adds shadows that are wrong.

Putting a large area light to one side looks great, for the major elements
of the scene.  Objects don't cast distinct shadows from this light source
but do show the vague darkening on the floor and close walls that we see in
a real room full of ambient light.  The =illumination= is still a point, not
a large surface, but it looks fine in this case.

However, I'm still getting stray shadows.  I got a sharp shadow from a small
object onto a nearby flat surface, which I presumed was because the
"adaptive" fails when the object is smaller than the light.  Hmm, on second
thought that's not right at all, as this would make =no= shadow (all four
corners of the light are unblocked when the object is in the middle), not a
sharp shadow with no prenumbra.

I tried turning off adaptive, and I still get a sharp shadow, even though
there are 25 sublights.  I'm trying a much larger number of sublights now,
but it's taking a long time to render.  If that shows no change, I don't
understand what's happening.

Anyway, I found that the "darkening" effect does a great deal toward making
a room look real.  If this isn't the best way to acheive that in an ambient
light situation, what is?  Would radiosity be the answer here, or does it
not help with this effect?

--John

Post a reply to this message

I've never encountered what you mention is wrong, and it sounds
peculiar. Did you use more than 'adaptive 1'?  I'm sure you must be
using 'jitter' in that too, aren't you? But what it almost sounds like
is a 'area_light' whose plane is aligned with the small object
somehow.

Bob

John M. Dlugosz <joh### [at] dlugoszcom> wrote in message
news:37cc8e2e@news.povray.org...
> I'm rendering a artist's studio that has sunlight coming in from the
glass
> roof.  Leaving that as the only light source and cranking up the
ambiant on
> the contents of the room doesn't look real at all.
>
> Putting a fill light off to one side makes shapes look better (not
"flat")
> but adds shadows that are wrong.
>
> Putting a large area light to one side looks great, for the major
elements
> of the scene.  Objects don't cast distinct shadows from this light
source
> but do show the vague darkening on the floor and close walls that we
see in
> a real room full of ambient light.  The =illumination= is still a
point, not
> a large surface, but it looks fine in this case.
>
> However, I'm still getting stray shadows.  I got a sharp shadow from
a small
> object onto a nearby flat surface, which I presumed was because the
> "adaptive" fails when the object is smaller than the light.  Hmm, on
second
> thought that's not right at all, as this would make =no= shadow (all
four
> corners of the light are unblocked when the object is in the
middle), not a
> sharp shadow with no prenumbra.
>
> I tried turning off adaptive, and I still get a sharp shadow, even
though
> there are 25 sublights.  I'm trying a much larger number of
sublights now,
> but it's taking a long time to render.  If that shows no change, I
don't
> understand what's happening.
>
> Anyway, I found that the "darkening" effect does a great deal toward
making
> a room look real.  If this isn't the best way to acheive that in an
ambient
> light situation, what is?  Would radiosity be the answer here, or
does it
> not help with this effect?
>
> --John
>
>
>

Post a reply to this message

Bob Hughes <inv### [at] aolcom> wrote in message
news:37cc9234@news.povray.org...
> I've never encountered what you mention is wrong, and it sounds
> peculiar. Did you use more than 'adaptive 1'?  I'm sure you must be
> using 'jitter' in that too, aren't you? But what it almost sounds like
> is a 'area_light' whose plane is aligned with the small object
> somehow.


It turns out that the shadow in question is correct, not generated by this
area light.  When I tested the image without the light to see if that shadow
went away, I must have used the wrong options, as the shadow went away <g>.

I'm still interested in general ideas for the "room edge" effect, or the
transition from ambiant to directional light.  I think my idea is OK for
some scenes, but because the illumination isn't an area (it only affects the
shadows) it can cause unwanted highlights too.

--John

Post a reply to this message

John M. Dlugosz wrote in message <37cc9f4f@news.povray.org>...
>
>Bob Hughes <inv### [at] aolcom> wrote in message
>news:37cc9234@news.povray.org...
>> I've never encountered what you mention is wrong, and it sounds
>> peculiar. Did you use more than 'adaptive 1'?  I'm sure you must be
>> using 'jitter' in that too, aren't you? But what it almost sounds like
>> is a 'area_light' whose plane is aligned with the small object
>> somehow.
>
>
>It turns out that the shadow in question is correct, not generated by this
>area light.  When I tested the image without the light to see if that
shadow
>went away, I must have used the wrong options, as the shadow went away <g>.
>
>I'm still interested in general ideas for the "room edge" effect, or the
>transition from ambiant to directional light.  I think my idea is OK for
>some scenes, but because the illumination isn't an area (it only affects
the
>shadows) it can cause unwanted highlights too.


Try using radiosity -- it has a much greater effect in this situation than
you'd expect.

Mark

Post a reply to this message

John M. Dlugosz <joh### [at] dlugoszcom> wrote:
: I'm rendering a artist's studio that has sunlight coming in from the glass
: roof.  Leaving that as the only light source and cranking up the ambiant on
: the contents of the room doesn't look real at all.

  If you want real interreflection of light, you need to use radiosity.
  See http://iki.fi/warp/pics/Radiosity_test2/

-- 
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):5;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/

Post a reply to this message

Mark Wagner <mar### [at] gtenet> wrote in message
news:37ccb22d@news.povray.org...
> Try using radiosity -- it has a much greater effect in this situation than
> you'd expect.


I did try it, briefly.  I'm sure I could twiddle the parameters a LOT, but
off hand I saw that it didn't illuminate metal the way I wanted.  When
backlit by the key light, it looked just like a high ambiant -- no curves or
shape visible.

Physically, the other end of the room is a light source, and lambert shading
applies.  And in the case of metal, more spectral highlights are noticable
as well.  Turning up ambiant light levels didn't give the objects any
=shape=, and my try at using radiosity was no different.  Would adjusting
the parameters help with that specific effect?

Also, I may be on the "painting" rather than "photo" side of things, making
it look like the way I see it in my mind, which is not necessarily the way a
camera sees it.  So I need some kind of "modeling light" in localized areas,
as well as the room-edge effect.

More than once I've wished that it was possible to declare that specific
light sources interact only with specific objects.

--John

Post a reply to this message

John M. Dlugosz wrote:
> 
> Mark Wagner <mar### [at] gtenet> wrote in message
> news:37ccb22d@news.povray.org...
> > Try using radiosity -- it has a much greater effect in this situation than
> > you'd expect.
> 
> I did try it, briefly.  I'm sure I could twiddle the parameters a LOT, but
> off hand I saw that it didn't illuminate metal the way I wanted.  When
> backlit by the key light, it looked just like a high ambiant -- no curves or
> shape visible.

Some 3D effects look best when animated (though you wouldn't necessarily
think so offhand); your brain uses the movement to see how an object's
appearance reacts to the objects around it. Metal and glass are probably
the best examples of this, as movement helps you perceive reflection and
refraction. (The ability to see different sides of an object, or to
judge its distance relative to other objects, also require movement, but
are based on perspective rather than interactivity.)

Of course, you don't sound like you're planning to animate this.

The next thing, then, would be to take a careful look at exactly what
you're illuminating, and how. With some thought and tweaking, you can
probably get it to look how you want. (I can't offer any more specific
suggestions, since I have almost no idea what the scene's supposed to
look like or what's in it.. I must have missed the first post.)

> Physically, the other end of the room is a light source, and lambert shading
> applies.  And in the case of metal, more spectral highlights are noticable
> as well.  Turning up ambiant light levels didn't give the objects any
> =shape=, and my try at using radiosity was no different.  Would adjusting
> the parameters help with that specific effect?

Ambient lighting (effectively) hits an object from every angle, so no,
obviously it wouldn't produce shadows that would help define the shape
of the object. In fact, it would do the opposite, by fading or
eliminating those shadows.

> More than once I've wished that it was possible to declare that specific
> light sources interact only with specific objects.

I don't think you can do this.. but in some cases, a carefully-placed
spotlight will do the trick, similar to the way that photographers use
reflectors to subtly illuminate a model's face (if you've seen an
outdoor photo shoot, you know what I'm getting at).

-Xplo

Post a reply to this message

"John M. Dlugosz" wrote:

> Mark Wagner <mar### [at] gtenet> wrote in message
> news:37ccb22d@news.povray.org...

<snip>

>
> More than once I've wished that it was possible to declare that specific
> light sources interact only with specific objects.
>
> --John

If you don't mind using a patch, the SuperPatch will allow light sources that
only interact with specific objects. It might be worth a try.
Mr.Art

Post a reply to this message

Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
news:37C### [at] unforgettablecom...
> Of course, you don't sound like you're planning to animate this.
No, it's meant to be a Giclée print, to hang on a wall.

> but in some cases, a carefully-placed
> spotlight will do the trick, similar to the way that photographers use
> reflectors to subtly illuminate a model's face (if you've seen an
> outdoor photo shoot, you know what I'm getting at).

That's basically what my 1m-square area "fill light" off-stage is doing to
the main figure, in addition to providing edge-effect radiosity-like
shadows.

> I have almost no idea what the scene's supposed to
> look like or what's in it..

See http://www.dlugosz.com/POV/dryad_dreams.html for my notes.

--John

Post a reply to this message

Am am indeed using the SuperPatch!  Thanks for the tip.

... reading txt file ...

So, are the lights nominated by the light_group for a specific object the
only lights to contribute to the illumination of that object?

I suppose I can make a modeling light by excluding it from every other
object (looks rather intrusive, since I can't change the meaning of "all"
... maybe can change the defaults) and also making it have no_shadow for
that object.

I'll have to experiment with that feature, or find a better
description/tutorial somewhere.  Any pointers?

--John


Mr. Art <mr.### [at] gcinet> wrote in message news:37CDF0EC.8CEA2118@gci.net...
>
>
> "John M. Dlugosz" wrote:
>
> > Mark Wagner <mar### [at] gtenet> wrote in message
> > news:37ccb22d@news.povray.org...
>
> <snip>
>
> >
> > More than once I've wished that it was possible to declare that specific
> > light sources interact only with specific objects.
> >
> > --John
>
> If you don't mind using a patch, the SuperPatch will allow light sources
that
> only interact with specific objects. It might be worth a try.
> Mr.Art
>

Post a reply to this message

Thanks for the sample images.  From that, I see that it can indeed give
"shape" to objects not directly illuminated, like your yellow sphere.

Nieminen Juha <war### [at] cctutfi> wrote in message
news:37cce0a9@news.povray.org...
>   If you want real interreflection of light, you need to use radiosity.
>   See http://iki.fi/warp/pics/Radiosity_test2/

Post a reply to this message

John M. Dlugosz wrote:
> 
> Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
> 
> > but in some cases, a carefully-placed
> > spotlight will do the trick, similar to the way that photographers use
> > reflectors to subtly illuminate a model's face (if you've seen an
> > outdoor photo shoot, you know what I'm getting at).
> 
> That's basically what my 1m-square area "fill light" off-stage is doing to
> the main figure, in addition to providing edge-effect radiosity-like
> shadows.
> 
> > I have almost no idea what the scene's supposed to
> > look like or what's in it..
> 
> See http://www.dlugosz.com/POV/dryad_dreams.html for my notes.

Lemme see if I can say something useful.

Well, your notes call for a gleaming 50s sort of robot. Such a gleaming
robot would probably be chromed.. in which case you want more reflection
and a much tighter specular highlight. I should point out, however, that
the more reflective your robot is, the less "shape" it will have, since
it'll be more and more a warped copy of what's around it and less
affected by traditional shading.

The robot you have now has a pretty good "soft metallic" finish, but
that may or may not be what you intended.

If the robot is highly reflective, it would be an excellent idea to
finish the room by adding walls behind the camera to reflect off of the
robot's skin (if you've done this already, they must be really boring
walls). The walls themselves will never appear, but their effect on the
scene will. Also note that adding a couple of walls may change the
lighting in the room, especially if you're using a lot of radiosity.

As for the lighting itself, hmm...

Well, it's supposed to be an artist's studio, yes? AFAIK, those tend to
be pretty well lit, and not with point light either. You probably want
(a) several wide-beam spotlights, similar to track lighting, or a lot of
sunlight streaming in through windows. (I can't see it real well, but
you appear to have something like a skylight, which isn't going to do
the trick.. unless of course you WANTED that.) With that much light, it
would be safe to turn down ambient light for the whole scene to 0 and
crank up the radiosity, letting what lighting exists flood the room in a
more natural manner.

You're looking for an artificial appearance, so I'll stop here. Hope you
got some ideas.

-Xplo

Post a reply to this message

Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
news:37C### [at] unforgettablecom...

> Lemme see if I can say something useful.

Thanks for the comments.

> The robot you have now has a pretty good "soft metallic" finish, but
> that may or may not be what you intended.

Whatever I intended, I kind of like what I wound up with.  I want to run
with the idea of the robot being very simple in shape, like a post, but the
arms and the attentive posture makes it alive.  The more I can increase this
juxtaposition...

>
> If the robot is highly reflective, it would be an excellent idea to
> finish the room by adding walls behind the camera to reflect off of the
> robot's skin (if you've done this already, they must be really boring
> walls). The walls themselves will never appear, but their effect on the
> scene will. Also note that adding a couple of walls may change the
> lighting in the room, especially if you're using a lot of radiosity.

There are indeed walls on all four sides.  The walls are blank, and I know
it affects the appearance of metal.  That's why I started with an empty room
rather than a black stage.

> As for the lighting itself, hmm...
>
> Well, it's supposed to be an artist's studio, yes? AFAIK, those tend to
> be pretty well lit, and not with point light either. You probably want
> (a) several wide-beam spotlights, similar to track lighting, or a lot of
> sunlight streaming in through windows. (I can't see it real well, but
> you appear to have something like a skylight, which isn't going to do
> the trick.. unless of course you WANTED that.) With that much light, it
> would be safe to turn down ambient light for the whole scene to 0 and
> crank up the radiosity, letting what lighting exists flood the room in a
> more natural manner.

Hmm, the scene is indeed lit from a skylight.  Actually, there is no ceiling
at all in the model <g>.  I'd put in grids to cast shadows if it was
visible.

So, given a "natural light" studio, how can it be well-lit other than a
single point source?  Lots of reflectors to make fill light?  If so, then
the large area light I already put in is the right idea -- models the large
white card or whatever he'll have set up.  Would I get the same effect by
using actual white reflectors and using a radiosity model?

> You're looking for an artificial appearance, so I'll stop here. Hope you
> got some ideas.

Yes, thanks.

--John

Post a reply to this message

Surely you must know by now that area lights have only a singular
point source for all purposes besides shadows.  Your statement seemed
to suggest otherwise.  But yes, I would think a white reflector
(ambient) should be very similar with the added plus of true
'area_light' appearance when used with 'radiosity'.  I'm thinking you
may have to go with both however in order to get good shadows too,
really not sure.

Bob

John M. Dlugosz <joh### [at] dlugoszcom> wrote in message
news:37cf302a@news.povray.org...
>
> So, given a "natural light" studio, how can it be well-lit other
than a
> single point source?  Lots of reflectors to make fill light?  If so,
then
> the large area light I already put in is the right idea -- models
the large
> white card or whatever he'll have set up.  Would I get the same
effect by
> using actual white reflectors and using a radiosity model?

Post a reply to this message

John M. Dlugosz wrote:
> 
> Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
> news:37C### [at] unforgettablecom...
> 
> > Well, it's supposed to be an artist's studio, yes? AFAIK, those tend to
> > be pretty well lit, and not with point light either. You probably want
> > (a) several wide-beam spotlights, similar to track lighting, or a lot of
> > sunlight streaming in through windows. (I can't see it real well, but
> > you appear to have something like a skylight, which isn't going to do
> > the trick.. unless of course you WANTED that.) With that much light, it
> > would be safe to turn down ambient light for the whole scene to 0 and
> > crank up the radiosity, letting what lighting exists flood the room in a
> > more natural manner.
> 
> Hmm, the scene is indeed lit from a skylight.  Actually, there is no ceiling
> at all in the model <g>.  I'd put in grids to cast shadows if it was
> visible.
> 
> So, given a "natural light" studio, how can it be well-lit other than a
> single point source?  Lots of reflectors to make fill light?  If so, then
> the large area light I already put in is the right idea -- models the large
> white card or whatever he'll have set up.  Would I get the same effect by
> using actual white reflectors and using a radiosity model?

If you're having light coming in through a large aperture, such as a
window or a skylight, using an area light roughly the size of the
aperture (preferably a fairly complex one, if you have the power to
render it) is probably the right way to go. Then, just render it with
the radiosity and see what happens.. it's entirely possible that the
objects in the room (most notably the walls.. real walls reflect a LOT
of light) will provide all the fill light you need/want. Then, if you
still need spot lighting, you can go and add that back in without
changing the overall lighting model very much.

If you wanted to play with white reflectors, you could do it that way as
well; the principle is basically the same, but the numbers won't be,
since your lighting will be different.

-Xplo

Post a reply to this message

Bob Hughes <inv### [at] aolcom> wrote in message
news:37cf3a85@news.povray.org...
> Surely you must know by now that area lights have only a singular
> point source for all purposes besides shadows.  Your statement seemed
> to suggest otherwise.

I know that.  soft shadows, but hard illumination.  In general, it can't
work because the lighting is still hard and a real array of lights will give
multiple phongs not soft illumination.  But for this scene, from this
viewing position, the fill light works OK.


> But yes, I would think a white reflector
> (ambient) should be very similar with the added plus of true
> 'area_light' appearance when used with 'radiosity'.  I'm thinking you
> may have to go with both however in order to get good shadows too,
> really not sure.

Post a reply to this message

Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
news:37D### [at] unforgettablecom...

> If you're having light coming in through a large aperture, such as a
> window or a skylight, using an area light roughly the size of the
> aperture (preferably a fairly complex one, if you have the power to
> render it) is probably the right way to go. Then, just render it with
> the radiosity and see what happens.. it's entirely possible that the
> objects in the room (most notably the walls.. real walls reflect a LOT
> of light) will provide all the fill light you need/want. Then, if you
> still need spot lighting, you can go and add that back in without
> changing the overall lighting model very much.


I don't think so -- the sun casts sharp shadows.  Ah! at least that's what
it looks like in Dallas, with a clear sky and no clouds at all.  On an
overcast day, you get even illumination from the sky.

As for using radiosity, I'll have to play with the numbers to see if I can
get it to "work".  I suppose varying the distance_maximum is enough to get
it working, or will I have to explore more degrees of freedom?  My earlier
try basically didn't work.

--John

Post a reply to this message

"John M. Dlugosz" wrote:

> As for using radiosity, I'll have to play with the numbers to see if I can
> get it to "work".  I suppose varying the distance_maximum is enough to get
> it working, or will I have to explore more degrees of freedom?  My earlier
> try basically didn't work.
> 
> --John

I am far from any kind of expert on the subject but as I understand it when
you use radiosity be sure to keep the distance of your objects from each
other down to a minimum. If your room is hundreds of units wide and it that
far from objects to walls the feature does not work very well. Small rooms
and scales are almost mandatory for radosity to work right. This I have heard
but have spent little time with the feature to date.

-- 
Ken Tyler

See my 850+ Povray and 3D Rendering and Raytracing Links at:
http://home.pacbell.net/tylereng/index.html

Post a reply to this message

John M. Dlugosz wrote:
> 
> Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
> news:37D### [at] unforgettablecom...
> 
> > If you're having light coming in through a large aperture, such as a
> > window or a skylight, using an area light roughly the size of the
> > aperture (preferably a fairly complex one, if you have the power to
> > render it) is probably the right way to go. Then, just render it with
> > the radiosity and see what happens.. it's entirely possible that the
> > objects in the room (most notably the walls.. real walls reflect a LOT
> > of light) will provide all the fill light you need/want. Then, if you
> > still need spot lighting, you can go and add that back in without
> > changing the overall lighting model very much.
> 
> I don't think so -- the sun casts sharp shadows.  Ah! at least that's what
> it looks like in Dallas, with a clear sky and no clouds at all.  On an
> overcast day, you get even illumination from the sky.

Oh, hmm. What on earth was I thinking? Oh right.. I was trying to
duplicate a "square" of light. Silly me. ^_^;;;

Yes, you WOULD be better off with a single light set back some distance
from the window.

-Xplo

Post a reply to this message

Lowering 'gray_threshold' nearer to zero helps to blend more color
from objects than the default of 0.5, that can sometimes help to show
more radiosity than would otherwise be seen (doesn't brighten it
though).  Also reducing 'distance_maximum' will tighten the blending
closer and thus seems to make it brighter if a lower 'count' is used.

Bob

John M. Dlugosz <joh### [at] dlugoszcom> wrote in message
news:37d091a5@news.povray.org...
>
> Xplo Eristotle <inq### [at] unforgettablecom> wrote in message
> news:37D### [at] unforgettablecom...
>
> > If you're having light coming in through a large aperture, such as
a
> > window or a skylight, using an area light roughly the size of the
> > aperture (preferably a fairly complex one, if you have the power
to
> > render it) is probably the right way to go. Then, just render it
with
> > the radiosity and see what happens.. it's entirely possible that
the
> > objects in the room (most notably the walls.. real walls reflect a
LOT
> > of light) will provide all the fill light you need/want. Then, if
you
> > still need spot lighting, you can go and add that back in without
> > changing the overall lighting model very much.
>
>
> I don't think so -- the sun casts sharp shadows.  Ah! at least
that's what
> it looks like in Dallas, with a clear sky and no clouds at all.  On
an
> overcast day, you get even illumination from the sky.
>
> As for using radiosity, I'll have to play with the numbers to see if
I can
> get it to "work".  I suppose varying the distance_maximum is enough
to get
> it working, or will I have to explore more degrees of freedom?  My
earlier
> try basically didn't work.
>
> --John
>
>

Post a reply to this message

John M. Dlugosz <joh### [at] dlugoszcom> wrote:
: As for using radiosity, I'll have to play with the numbers to see if I can
: get it to "work".  I suppose varying the distance_maximum is enough to get
: it working, or will I have to explore more degrees of freedom?  My earlier
: try basically didn't work.

  For tips about radiosity see:
http://iki.fi/warp/pics/Radiosity_test2/comments.html

-- 
main(i,_){for(_?--i,main(i+2,"FhhQHFIJD|FQTITFN]zRFHhhTBFHhhTBFysdB"[i]
):5;i&&_>1;printf("%s",_-70?_&1?"[]":" ":(_=0,"\n")),_/=2);} /*- Warp -*/

Post a reply to this message