 |
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Mon, 14 Jan 2002 14:31:59 -0800, Dearmad <dea### [at] applesnake net>
wrote:
>no, they'll intersect. I meant to say the light should be
>slightly *larger* than the cone so that the cone is sure to be
>hit by the matching light.
A cone is defined as the locus of all lines passing through a point
(vertex) and forming the same angle with a given line (the axis) also
passing through above point. Another way to define a cone is to rotate
a line (ruler) about another line (axis), whereas the two lines
intersect, and the surface formed by above rotation (ruled surface, in
this case also surface of revolution) is a cone.
You can easily see that if you place your cone so that the vertex is
at the location of the light source, be it point or spot, the rays of
light will be either inside the cone, outside the cone or coincident
with it. The latter is highly unlikely due to numeric inaccuracies
inherent to the way computers treat floating point numbers, so in
practice no way will intersect the cone's surface (except the base),
and that's what really counts in raytracing.
Peter Popov ICQ : 15002700
Personal e-mail : pet### [at] vipbg
TAG e-mail : pet### [at] tagpovrayorg
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
In article <k3q64usckpm6tm8lc135hcojuglr284i49@4ax.com>,
Peter Popov <pet### [at] vipbg> wrote:
> You can easily see that if you place your cone so that the vertex is
> at the location of the light source, be it point or spot, the rays of
> light will be either inside the cone, outside the cone or coincident
> with it. The latter is highly unlikely due to numeric inaccuracies
> inherent to the way computers treat floating point numbers, so in
> practice no way will intersect the cone's surface (except the base),
> and that's what really counts in raytracing.
However, this looks like just the type of situation where a point will
sometimes be lit, sometimes in shadow, etc. exactly because of precision
errors, though it probably wouldn't cause any difference in illumination
because of the steep angle. This is why I recommended using ambient only
and a diffuse of 0. Similarly, the specular highlights will only show up
if the cone is hit by another light, and I can't think of any reason
they would be desireable.
--
--
Christopher James Huff <chr### [at] maccom>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Peter, you're reading waaaay too much into it. The point of
origins aren't the same so all you're doing is lighting up a cone
with a spotlight from the top and making the cone slightly
transpaerent so it mimics the appearence of a "cone of light."
all your proofs are great, but ultimately inapplicable because:
I've done it. :o) I understand what you've written below, it
really doesn't apply here, because you don't line them up
perfectly to match, as I said (about 4x now) make the cone
*slightly* smaller than the spot light.
-peter
Peter Popov wrote:
>
> On Mon, 14 Jan 2002 14:31:59 -0800, Dearmad <dea### [at] applesnakenet>
> wrote:
>
> >no, they'll intersect. I meant to say the light should be
> >slightly *larger* than the cone so that the cone is sure to be
> >hit by the matching light.
>
> A cone is defined as the locus of all lines passing through a point
> (vertex) and forming the same angle with a given line (the axis) also
> passing through above point. Another way to define a cone is to rotate
> a line (ruler) about another line (axis), whereas the two lines
> intersect, and the surface formed by above rotation (ruled surface, in
> this case also surface of revolution) is a cone.
>
> You can easily see that if you place your cone so that the vertex is
> at the location of the light source, be it point or spot, the rays of
> light will be either inside the cone, outside the cone or coincident
> with it. The latter is highly unlikely due to numeric inaccuracies
> inherent to the way computers treat floating point numbers, so in
> practice no way will intersect the cone's surface (except the base),
> and that's what really counts in raytracing.
>
> Peter Popov ICQ : 15002700
> Personal e-mail : pet### [at] vipbg
> TAG e-mail : pet### [at] tagpovrayorg
--
Current obsession: "Ballet pour ma fille."
http://www.applesnake.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
On Tue, 15 Jan 2002 00:07:12 -0800, Dearmad <dea### [at] applesnakenet>
wrote:
>Peter, you're reading waaaay too much into it.
Yep. Trying to look smart after I made a fool of myself by not reading
:( I apologize.
At least what I posted wasn't all wrong :)
Peter Popov ICQ : 15002700
Personal e-mail : pet### [at] vipbg
TAG e-mail : pet### [at] tagpovrayorg
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Peter Popov wrote:
>
> On Tue, 15 Jan 2002 00:07:12 -0800, Dearmad <dea### [at] applesnakenet>
> wrote:
>
> >Peter, you're reading waaaay too much into it.
>
> Yep. Trying to look smart after I made a fool of myself by not reading
> :( I apologize.
>
> At least what I posted wasn't all wrong :)
Yup, in fact what you posted is *exactly* what I encountered when
I *first* tried the technique... loads of black artifacts along
the surface... then I figured it out. :o)
-peter
--
Current obsession: "Ballet pour ma fille."
http://www.applesnake.net
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Bob have you tried this? I get completely black when I run it in 3.5. When I
run the prior one I do get a spot light but it has a strange anomoly where
there seems to be a change in shade at about one third down the cone.
"bob h" <omn### [at] charternet> wrote in message
news:3c424849$1@news.povray.org...
> Here it is with scattering so shadows show up in the light beam.
> Admittedly, I don't know geometry well enough to keep the distance and
width
> of light beam in sync. Must be a formula for keeping angle relative to
the
> cone apex height and base diameter ratio.
> BTW, this was fairly quick for me to render.
>
> #declare F=1.333; // fudge factor
> #declare D=10*F; // distance
> #declare S=5*F; // scale
>
> #declare ConeOfLite=
> cone {
> 0,0,y,1
> pigment {rgbt 1}
> interior {
> media {
> emission 1/3/S // intensity of air glow
> scattering {1,0.2}
> density {
> cylindrical
> }
> }
> }
> scale <S,D,S> // x and z are width, y is length or distance
> hollow
> }
>
> light_source {
> 0, 1
> spotlight
> point_at y
> radius 6*S
> falloff 12*S
> looks_like {ConeOfLite}
> rotate <120,90,0> translate <-3,2,0>
> }
>
> plane {y,-1
> pigment {rgb <.9,.6,.3>}
> }
>
> sphere {0,1 pigment {rgb .5}}
>
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Was straight from a render to posting it. Not that it means anything the
way things are going at the moment while I try to track down a weird
transparency thing {bug} :-)
Oh, make sure the area remains 'hollow' for that cone media of course, if
you add things into your scene you don't want a plane or surrounding sphere
to cut off the media. And there was a ground plane so that will be seen to
cut through the cone of light, however out beyond it.
bob h
"Mitchell Waite" <mit### [at] dnaicom> wrote in message
news:3c4533c4$1@news.povray.org...
> Bob have you tried this? I get completely black when I run it in 3.5. When
I
> run the prior one I do get a spot light but it has a strange anomoly where
> there seems to be a change in shade at about one third down the cone.
>
> "bob h" <omn### [at] charternet> wrote in message
> news:3c424849$1@news.povray.org...
> > Here it is with scattering so shadows show up in the light beam.
> > Admittedly, I don't know geometry well enough to keep the distance and
> width
> > of light beam in sync. Must be a formula for keeping angle relative to
> the
> > cone apex height and base diameter ratio.
> > BTW, this was fairly quick for me to render.
> >
> > #declare F=1.333; // fudge factor
> > #declare D=10*F; // distance
> > #declare S=5*F; // scale
> >
> > #declare ConeOfLite=
> > cone {
> > 0,0,y,1
> > pigment {rgbt 1}
> > interior {
> > media {
> > emission 1/3/S // intensity of air glow
> > scattering {1,0.2}
> > density {
> > cylindrical
> > }
> > }
> > }
> > scale <S,D,S> // x and z are width, y is length or distance
> > hollow
> > }
> >
> > light_source {
> > 0, 1
> > spotlight
> > point_at y
> > radius 6*S
> > falloff 12*S
> > looks_like {ConeOfLite}
> > rotate <120,90,0> translate <-3,2,0>
> > }
> >
> > plane {y,-1
> > pigment {rgb <.9,.6,.3>}
> > }
> >
> > sphere {0,1 pigment {rgb .5}}
> >
> >
> >
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|
