|
|
|
|
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ard" <ard### [at] waikatoacnz> wrote in message
news:web.42bbf552bd7481d397d402360@news.povray.org...
>
>> web.42b69f52a6ab8cf34f17dd8d0@news.povray.org
>
> Thanks for that. However I'm not sure that my problem is the same as
> Burki's, or at least the solution isn't the same. He had artefacts in
> conical spotlights; I'm getting them at the origin of cylindrical
> spotlights, the beams of which travel only a couple of units before
> striking an object. It looks like the light is emanating from a black
> saucer.
>
> I have limited the media thus:
>
> box {-2 2
> hollow pigment{rgbt 1}
> interior {media { scattering { 2, rgb 1/2 extinction 0} samples 50}}
> }
>
> ....with no improvement.
>
> Thanks for your help so far. I wonder if I could trouble you for some
> more
> :-)
Hello,
After futzing around with it here, unfortunately I haven't come to a
solution either. I have found however that it does appear to be a
media-related problem, but as you say, is not the same as Burki's.
I think I must have done much the same experiement with media as you --
encase the scene and media inside a hollow object to eliminate infinite rays
as a cause. Bumping up samples and intervals also made no difference to the
black circles.
I added looks_like {disc {}} clauses to the light sources to give them
visibility as physical objects. The discs are oriented in the direction of
the beams, representing the front lens/filter of the light (see code
below).
What leads me to still suspect it is media, is this: When the interior
{media{...}} line is commented out, the cylinder lights now properly look
like red and blue circles, due to the discs. They were invisible before.
Woohoo, progress, etc. That multiplier of 0.001 I put on on the disc
location in the pointing direction, is merely to avoid coincident surface
speckle with the light origin, otherwise its value would have been <0,0,0>.
BUT when the interior {media{...}} line is re-enabled, the lights become
black holes again. It seems as though media is not calculated at all for
rays traced from the camera that hit a light source object(?) But I'm not
sure if that really explains why the discs (which are very slightly in front
of the light origins) also disappear when media is active. Look closely,
and you'll see that the part of the disc which is WIDER than the radius of
the light, is still visible around the black part.
I've pushed your "furniture" around a bit - everything is not where you last
left it! :-)
Sorry I haven't been more help than just finding out "more" about the
problem!
Cheers,
Brian
---- code ----
camera {
// location <0, 0.5, -1> look_at y/3
location <0, 0.5, -2> look_at y/3
}
plane {-z, -1 pigment {rgb 1}} // Backdrop
plane { y, 0 pigment {rgb 1}} // Floor
sphere
{ 0, 10 hollow
// interior {media { scattering { 2, rgb 1/2 extinction 0} intervals 1
samples 50 }}
}
#declare Light_Location_Red = <0.1, 0.6, 0.5>;
#declare Light_PointAt_Red = <-0.05, 0.15, -0.75>; // <-0.05, 0.35, -1>;
#declare Light_Direction_Red = vnormalize (Light_PointAt_Red -
Light_Location_Red);
#declare Light_Location_Blue = <-0.5, 0.4, 0>;
#declare Light_PointAt_Blue = <0.3, 0, -0.1>;
#declare Light_Direction_Blue = vnormalize (Light_PointAt_Blue -
Light_Location_Blue);
// Illuminate the background so that the black discs at the ends
// off the cylindrical lights are more obvious
light_source {<10, 10 -10> rgb 0.3}
// Blue light on left, aimed right, striking the floor:
light_source {
Light_Location_Blue
color rgb <0, 0, 5>
cylinder
parallel
radius 0.05
falloff 0.10
point_at Light_PointAt_Blue
looks_like
{ disc
{ Light_Direction_Blue*0.001, Light_Direction_Blue, 0.11
pigment {rgb 1} double_illuminate // finish {ambient 1}
}
}
}
// Red light aimed nearly at the camera
light_source {
Light_Location_Red
color rgb <5, 0, 0>
cylinder
radius 0.03
falloff 0.07
point_at Light_PointAt_Red
looks_like
{ disc
{ Light_Direction_Red*0.001, Light_Direction_Red, 0.11
pigment {rgb 1} double_illuminate //finish {ambient 1}
}
}
}
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
Ard nous apporta ses lumieres en ce 2005-06-24 00:55:
> I'm trying to render stage lighting using cylindrical light sources and
> environmental media. Spotlights above a stage in a smoky bar, kind of
> thing.
>
> The light sources are in the image, surrounded by objects that make them
> look like conventional spotlights, and the beams being emitted look
> fantastic as do the objects (musicians) they illuminate. However the light
> sources look awful.
>
> Here is a very simple example that exaggerates the problem. The ends of the
> cylindrical lights are rendered as stark black discs.
>
> camera {
> location <0, 0.5, -1> look_at y/3
> }
> plane {-z, -1 pigment {rgb 1}} // Backdrop
> plane { y, 0 pigment {rgb 1}} // Floor
> // Illuminate the background so that the black discs at the ends
> // off the cylindrical lights are more obvious
> light_source {<10, 10 -10> rgb 0.3}
> // Blue light on left, aimed right, striking the floor:
> light_source {
> <-0.5, 0.4, 0>
> color rgb <0, 0, 5>
> cylinder
> parallel
> radius 0.05
> falloff 0.10
> point_at <0.3, 0, -0.1>
> }
> // Red light aimed nearly at the camera
> light_source {
> <0.1, 0.6, 0.5>
> color rgb <5, 0, 0>
> cylinder
> radius 0.03
> falloff 0.07
> point_at <-0.05, 0.35, -1>
> }
> media { scattering { 2, rgb 1/2 extinction 0} }
>
> I've various ways to replace the disks with illuminating objects, but they
> all look awful and I get the feeling I'm just missing something obvious.
>
> Please set me right. Is there a simple way to remove those black discs?
>
> Many, many thanks in advance.
>
>
>
A few solutions that I think of:
- use spotlight instead of cylinder. You get conical beams, no black disks. Increase
the radius
and falloff angles: 0.05 become 5
- As you already have spotlight objects, just put a regular light_source _inside_
those objects.
Don't use look_like. Make sure that the front end is actualy open. An advantage is
that you can have
working bafles just like on real spotlights, enabeling you to have a shaped beam, not
just a
circular one. If you want fuzzy edges, use an area_light with circular and orient (try
17*17 aray
for smooth results) and keep the vectors short. Add adaptive 0 and increase if you get
artefacts
(light spot in shadow or dark spot in lighted area)
You can combine those 2 suggestions: area_light spotlight.
If you get artefacts IN the beams, increase the samples value, DON'T increase
intervals!
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Alain" <ele### [at] netscapenet> wrote in message
news:42bc1900$1@news.povray.org...
> Ard nous apporta ses lumieres en ce 2005-06-24 00:55:
> A few solutions that I think of:
> - use spotlight instead of cylinder. You get conical beams, no black
> disks. Increase the radius and falloff angles: 0.05 become 5
> - As you already have spotlight objects, just put a regular light_source
> _inside_ those objects. Don't use look_like. Make sure that the front end
> is actualy open. An advantage is that you can have working bafles just
> like on real spotlights, enabeling you to have a shaped beam, not just a
> circular one. If you want fuzzy edges, use an area_light with circular and
> orient (try 17*17 aray for smooth results) and keep the vectors short. Add
> adaptive 0 and increase if you get artefacts (light spot in shadow or dark
> spot in lighted area)
> You can combine those 2 suggestions: area_light spotlight.
> If you get artefacts IN the beams, increase the samples value, DON'T
> increase intervals!
That is such a strange behaviour -- and a scene-breaker. I've never used
cylindrical lights before, let alone use them in media and placed within the
viewport, and would never have expected this.
I've just experimented a bit more placing objects in Ard's scene. No colour
is calculated for pixels which map over the aperture of a cylindrical light
when media is on. Even when a nearer opaque object obscures the light
source from view, that part of it still becomes black, and ONLY when media
is on.
Is it known if this issue is tested for in the Pov-Ray 3.7 beta?
Brian
---- code ----
camera {
// location <0, 0.5, -1> look_at y/3
location <0, 0.5, -2> look_at y/3
}
plane {-z, -1 pigment {rgb 1}} // Backdrop
plane { y, 0 pigment {rgb 1}} // Floor
sphere
{ 0, 10 hollow
interior {media { scattering { 2, rgb 1/2 extinction 0} intervals 1
samples 20 }}
}
#declare Light_Location_Red = <0.1, 0.6, 0.5>;
#declare Light_PointAt_Red = <-0.05, 0.15, -0.75>; // <-0.05, 0.35, -1>;
#declare Light_Direction_Red = vnormalize (Light_PointAt_Red -
Light_Location_Red);
#declare Light_Location_Blue = <-0.5, 0.4, 0>;
#declare Light_PointAt_Blue = <0.3, 0, -0.1>;
#declare Light_Direction_Blue = vnormalize (Light_PointAt_Blue -
Light_Location_Blue);
// Illuminate the background so that the black discs at the ends
// off the cylindrical lights are more obvious
light_source {<10, 10 -10> rgb 0.3}
// Blue light on left, aimed right, striking the floor:
light_source {
Light_Location_Blue
color rgb <0, 0, 5>
spotlight //cylinder
parallel
radius 5 // 0.05
falloff 10 // 0.10
point_at Light_PointAt_Blue
/* looks_like
{ disc
{ Light_Direction_Blue*0.001, Light_Direction_Blue, 0.11
pigment {rgb 1} double_illuminate // finish {ambient 1}
}
} */
}
// Red light aimed nearly at the camera
light_source {
Light_Location_Red
color rgb <5, 0, 0>
/*spotlight*/ cylinder
radius 0.03 // 0.03
falloff 0.07 //0.07
point_at Light_PointAt_Red
looks_like
{ disc
{ Light_Direction_Red*0.001, Light_Direction_Red, 0.11
pigment {rgb 1} double_illuminate //finish {ambient 1}
}
}
}
disc
{ Light_Location_Blue+Light_Direction_Blue*0.2, Light_Direction_Blue, 0.11
pigment {rgb 1} double_illuminate no_shadow // finish {ambient 1}
}
/*
disc
{ Light_Location_Red+Light_Direction_Red*0.1, Light_Direction_Red, 0.11
pigment {rgb 1} double_illuminate no_shadow //finish {ambient 1}
}
*/
box
{ Light_Location_Red + Light_Direction_Red*0.3 - 0.1,
Light_Location_Red + Light_Direction_Red*0.3 + 0.1
texture
{ pigment {rgb <0.3, 1, 0.3>}
finish {ambient 0.2}
}
}
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Brian Elliott" <NotForSpam@AskIfUWant> wrote in message
news:42bc285c$1@news.povray.org...
>
> That is such a strange behaviour -- and a scene-breaker. I've never used
> cylindrical lights before, let alone use them in media and placed within
> the viewport, and would never have expected this.
>
> I've just experimented a bit more placing objects in Ard's scene. No
> colour is calculated for pixels which map over the aperture of a
> cylindrical light when media is on. Even when a nearer opaque object
> obscures the light source from view, that part of it still becomes black,
> and ONLY when media is on.
>
> Is it known if this issue is tested for in the Pov-Ray 3.7 beta?
>
> Brian
>
> ---- code ----
> [snip]
What I should have said: In that scene, the green box is "burnt right
through" by the cylindrical light behind it when the "interior {media}" line
is used, but is correct with media commented out.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
> That is such a strange behaviour -- and a scene-breaker. I've never used
[...]
> Is it known if this issue is tested for in the Pov-Ray 3.7 beta?
[...]
> What I should have said: In that scene, the green box is "burnt right
> through" by the cylindrical light behind it when the "interior {media}" line
> is used, but is correct with media commented out.
In the more complicated scene that first displayed this behaviour, one
object *was* mapped on top of the black disc. It was about 0.1u behind the
aperture, in a light_group lit by a source about 0.05u away - the
extremely-brightly lit insides of the real-world spotlight behind the
origin of the POV cylindrical light. So some things do show through. Not
that that makes any difference, but it's another observation to throw on
the pile.
Anyway, how can I bring this to the attention of the developers (my
apologies for not recognising you, if you are one) and make sure it is
tracked?
Thanks again.
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ard" <ard### [at] waikatoacnz> wrote:
> I'm not sure that my problem is the same as
> Burki's, or at least the solution isn't the same.
Oops... i apologize, my answer was definitely not relevant. Clearly not the
same issue. I see that others played with your scene today, and i can't
quite come up with a better idea... :(
I even tried to constrain the media inside the cylinder you wish to see and
locate the light_source slightly outside... to no avail. It seems that the
location of the light_source does NOT have to be IN the media for this
artifact to show up.
JYR
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ard" <ard### [at] waikatoacnz> wrote in message
news:web.42bca198bd7481d39a0bf9720@news.povray.org...
>> That is such a strange behaviour -- and a scene-breaker. I've never used
> [...]
>> Is it known if this issue is tested for in the Pov-Ray 3.7 beta?
> [...]
>> What I should have said: In that scene, the green box is "burnt right
>> through" by the cylindrical light behind it when the "interior {media}"
>> line
>> is used, but is correct with media commented out.
>
> In the more complicated scene that first displayed this behaviour, one
> object *was* mapped on top of the black disc. It was about 0.1u behind
> the
> aperture, in a light_group lit by a source about 0.05u away - the
> extremely-brightly lit insides of the real-world spotlight behind the
> origin of the POV cylindrical light. So some things do show through. Not
> that that makes any difference, but it's another observation to throw on
> the pile.
>
> Anyway, how can I bring this to the attention of the developers (my
> apologies for not recognising you, if you are one) and make sure it is
> tracked?
Thanks, but I'm only another ordinary user.
All the same, I was curious about the question of will this be fixed in 3.7,
so today I checked in povray.beta-test and on www.povray.org/beta and in the
release and change notes. There was quite a bit of work going on in media
code in April, with many change entries particularly by Thorsten and some by
Nathan and Chris. No explicit description of this cylinder lights in media
problem, but a possible candidate among others is a comment by Thorsten in
change 3587: "Some cutting on light tests in media".
So I downloaded the Win32 3.7 beta6 version to check it, and I'm happy to
report the problem has been fixed in 3.7.
Out of interest, I tried this scene out in PovRay 3.5, and it gave the black
circles like 3.6.1 and 3.6.1a do. So the bug has been around a while. Of
course, don't expect to see a final release of 3.7 for some months, it's
still in beta with some features disabled, and once Beta finally passes,
there's still the series of Release Candidates to get through!
But ONE DAY we can have cylinder spots in our virtual discos!
:-)
Cheers,
Brian
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
JYR nous apporta ses lumieres en ce 2005-06-24 22:12:
> "Ard" <ard### [at] waikatoacnz> wrote:
>
>>I'm not sure that my problem is the same as
>>Burki's, or at least the solution isn't the same.
>
>
> Oops... i apologize, my answer was definitely not relevant. Clearly not the
> same issue. I see that others played with your scene today, and i can't
> quite come up with a better idea... :(
> I even tried to constrain the media inside the cylinder you wish to see and
> locate the light_source slightly outside... to no avail. It seems that the
> location of the light_source does NOT have to be IN the media for this
> artifact to show up.
>
> JYR
>
>
>
It's definitely NOT a case of infinite ray! There is a plane at Z*1. the longets ray
are less than 3
units long, at the corners.
Alain
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Brian Elliott" <NotForSpam@AskIfUWant> wrote:
> I'm happy to report the problem has been fixed in 3.7.
Are you sure? When I render my original demo, or yours without the
looks_like objects, objects behind or in front of the aperture render
correctly, but the media mapped on top of the aperture does not. It's as
though the light is being emitted from the inside surface of the cylinder,
rather that from the outside surface. Instead of black discs I'm getting
grey ones.
> don't expect to see a final release of 3.7 for some months
I started this project 15 years ago on an Amiga so a few months more won't
kill it :-}
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
| |
|
|
"Ard" <ard### [at] waikatoacnz> wrote in message
news:web.42bde0abbd7481d39bd64a230@news.povray.org...
> "Brian Elliott" <NotForSpam@AskIfUWant> wrote:
>
>> I'm happy to report the problem has been fixed in 3.7.
>
> Are you sure? When I render my original demo, or yours without the
> looks_like objects, objects behind or in front of the aperture render
> correctly, but the media mapped on top of the aperture does not. It's as
> though the light is being emitted from the inside surface of the cylinder,
> rather that from the outside surface. Instead of black discs I'm getting
> grey ones.
Oops you are right. I only paid attention to the black hole "burn-through",
not the media effect. So I made a minimal scene out of it and posted a
report in povray.beta-test -- hope you don't mind me taking that liberty.
>> don't expect to see a final release of 3.7 for some months
>
> I started this project 15 years ago on an Amiga so a few months more won't
> kill it :-}
>
>
>
Post a reply to this message
|
|
| |
| |
|
|
|
|
| |
|
|