> Unfortunately I don't think it's possible.  Area lights only effect shadows
> as they fall onto other objects not the object creating the shadow.  I'm
> sure many people have tried many times to get a area light to do this and
> found the same thing.  Seems they will only affect the shadow which is cast
> away from the object.  And the highlights don't change whether from point
> sources or area lights either, same goes for the diffuse too I guess.  So
> any such attempt is probably futile.

i do remember that limitation of area lights, but that makes them faster
than a bunch of points right?  but how about a small array of points for
this?  or maybe spots?  can an angle of close to 180 be specified?  are
there any calculation speed ups with spots since light obviously can't
go behind them?  or are light calculations done another way?  and can
spot lights have their penumbra (is that the right word?) set to zero so
all light from the spot light is of the same intensity (ie no lateral
falloff)?  let's say i make a planet 70000 units in radius, how should i
go about finding the minimum acceptable distance (given the
size/distance ratio of 1/100 in christopher's post)?  would i use
different distances for spots than i would for points?  and what about
all those new keywords in megapov like circular and parallel?  might
they be of any use here?  has anyone ever done a looks_like to make a
sun?

Post a reply to this message

Basically, I agree with you.  The highlight position or shape won't
change and it won't get bigger, even though the light source is now
bigger.

	But, the block_area_light function basically figures out how much of
the area light is visible from any point and multiplies the phong or
specular or whatever highlight by that percentage.  This is totally
visible in this post.

	However, you can still see that POV cheats!  The sharp shadow line on
the second sphere wouldn't be there in reality, because part of the
light is still visible over the horizon.  There should be a smooth
transition.

	Well the idea works, but it looks like a 1:100 sun won't do anything
for you.

In Him,
	Ben

Post a reply to this message

Attachments:
Download 'orient_area_light.jpg' (9 KB)

Preview of image 'orient_area_light.jpg'

"ryan constantine" <rco### [at] yahoocom> wrote in message
news:39C08E5D.682CD947@yahoo.com...
| > Unfortunately I don't think it's possible.  Area lights only effect
shadows
| > as they fall onto other objects not the object creating the shadow.
|
| i do remember that limitation of area lights, but that makes them faster
| than a bunch of points right?  but how about a small array of points for
| this?  or maybe spots?  can an angle of close to 180 be specified?  are
| there any calculation speed ups with spots since light obviously can't
| go behind them?  or are light calculations done another way?  and can
| spot lights have their penumbra (is that the right word?) set to zero so
| all light from the spot light is of the same intensity (ie no lateral
| falloff)?  let's say i make a planet 70000 units in radius, how should i
| go about finding the minimum acceptable distance (given the
| size/distance ratio of 1/100 in christopher's post)?  would i use
| different distances for spots than i would for points?  and what about
| all those new keywords in megapov like circular and parallel?  might
| they be of any use here?  has anyone ever done a looks_like to make a
| sun?

Wow, so many questions :-)
I just noticed you said before that you thought the Sun is 2 degrees across.
Wrong.  It's closer to 0.5 degree, but then there is the surrounding
brightness in the air.  Still probably not 2 degrees wide of a source of
light enough to contribute to the main source, I wouldn't think anyway.
So the Sun is 1/360th of the arc of the whole sky.
Not sure what is meant by that 1:100 ratio.  The Sun being about 864,000
miles across and 93 million miles away and the Earth being about 8,000 miles
across makes for about a 1:11,625 ratio of Earth diameter to Sun distance,
with about a 1:107 ratio for Sun diameter to Earth distance.  Must be the
Sun then eh? :-)
70,000 unit planet isn't the thing to figure from anyway, just the sun(s)
size(s) and that's still solar system-like.
All moot if Tatooine and it's system is totally different, but judging from
the movie the ratios are probably going to be similar.  Simply add the
second sun.
Anyhow, back to the spotlight question(s).
You can go up to but not quite 90 for the both radius and falloff (same
value each), that's what you'd want to do for the non-penumbra form (not
certain of use of that word concerning lights, umbra should mean the unlit
area though and penumbra the partially lit).  But you really wouldn't need
it that wide I'd think.  If everything were to scale then a 1 degree light
ought to be fine for the distance you would have.
Not sure if it matters at all as long as no other object is in the scene
away from the planet, concerning the speed of render.  Also 'tightness 1'
(or zero?) will close the penumbral part of a spotlight if you were to use a
falloff larger than the radius.
You do know of course that area_light using spotlight will be the same way
as a point light too.  That array of spotlights could be a real slowdown but
maybe it wouldn't be too slow.

Bob

Post a reply to this message

Bob said:
>
> Not sure what is meant by that 1:100 ratio.

I was the one to come up with this ratio. See, I thought it would
be easy to remember: "Put the sun area_light at a distance of 100 times
the area_light size." :)

______________________________________________________________________
Kari Kivisalo                                  http://www.kivisalo.net

Post a reply to this message

> > > exterior should be (the current one is WAY too small), and makes the
> > > much needed inside layout changes.  i really like the way he did it,
> > > even if it isn't EXACTLY the way the real interior set was.  if you
> > > finish the exterior and size it correctly (did you know the ertl/amt
> > I will, but it's not very likely I make the interior.
> > What's ertl/amt?  I thought I knew most of the SW-shorts.
>
> ertl/amt makes plastic model kits.  i used their model as the basis for
> my xwing, using dial calipers to measure the parts, but comparing to
> pics from the movies to make up for problems in the model.
At B squared they said you couldn't trust these kits, so I didn't even think
about looking at one.
Another reason I didn't might be that I don't have one...

> >  But It shouldn't be too
> > difficult to scale it to 1unit=1 meter.
>
> my xwing was roughly a meter=unit, but had to scale it a little in the
> end to make it longer.  i knew the overall length, and i knew my
> proportions were correct, so i just made a box the right length and then
> made sure it barely fit inside length wise.

What size should I use for the Falcon?  25m is too small, but 35m is too big
I think.

> > I think I have about everything I need to finish it.  Maybe a more
detailed
> > picture of the bottom (for the armor-plating)
>
> i have a huge one of the bottom straight on.  want me to email it to
> you?
It's possible that's the one I already have, but If it's not to big (-500KB)
you can email it.
unless someone else wants it too, then it might be a better idea to post it
here.

> > If you're using MegaPov (which I believe you are) I recommend you use
> > image_pattern.  I think it's more flexible.
>
> i'll look at that route.  thanks.
>
> > Oh, in that case material_maps might be more useful.
>
> still, it's good to have options.
Right, but you have to make a choice :-)

> > I wish you luck.
>
> thanks again.
>
> > What's your next project?  I believe you once mentioned the DeathStar?
>
> next project is a far model of the death star (i'll eventually do 3
> models; far (low detail), medium (a chunk of it where you can see
> detail, but also the curvature), and a close model (for running through
> the trench)).  right after i do the far model, i'm going to blow it up
> using a combination of several explosions others have posted that i feel
> i can modify and put together to get EXACTLY the same result as the
> movie.  that'll be exciting.  thanks for asking!

Good luck!  It'll be very difficult to make the exact same explosion!
I think after the Falcon is finished (next year?) I'll finish my
StarDestroyer.  Have you already seen it?

ZK
http://www.povplace.be.tf

Post a reply to this message

> Wow, so many questions :-)
> I just noticed you said before that you thought the Sun is 2 degrees across.
> Wrong.  It's closer to 0.5 degree, but then there is the surrounding
> brightness in the air.  Still probably not 2 degrees wide of a source of
> light enough to contribute to the main source, I wouldn't think anyway.
> So the Sun is 1/360th of the arc of the whole sky.
> Not sure what is meant by that 1:100 ratio.  The Sun being about 864,000
> miles across and 93 million miles away and the Earth being about 8,000 miles
> across makes for about a 1:11,625 ratio of Earth diameter to Sun distance,
> with about a 1:107 ratio for Sun diameter to Earth distance.  Must be the
> Sun then eh? :-)
> 70,000 unit planet isn't the thing to figure from anyway, just the sun(s)
> size(s) and that's still solar system-like.
> All moot if Tatooine and it's system is totally different, but judging from
> the movie the ratios are probably going to be similar.  Simply add the
> second sun.

someone else mentioned the 1/100, and since i don't know about those
things, i figured it was right.  the reason i mentioned the planet size
is because i don't want to make the sun the right size at the right
distance.  what i want is to bring the sun closer, and make it smaller
because pov freaks out when you have both really big and really small
numbers in the same scene.  so if i could make the sun proportionately
closer so that it were still 1% and illuminate the planet as if it were
at the correct size and distance, that would be great.  and what about
the parallel keyword?

> Anyhow, back to the spotlight question(s).
> You can go up to but not quite 90 for the both radius and falloff (same
> value each), that's what you'd want to do for the non-penumbra form (not
> certain of use of that word concerning lights, umbra should mean the unlit
> area though and penumbra the partially lit).  But you really wouldn't need
> it that wide I'd think.  If everything were to scale then a 1 degree light
> ought to be fine for the distance you would have.
> Not sure if it matters at all as long as no other object is in the scene
> away from the planet, concerning the speed of render.  Also 'tightness 1'
> (or zero?) will close the penumbral part of a spotlight if you were to use a
> falloff larger than the radius.
> You do know of course that area_light using spotlight will be the same way
> as a point light too.  That array of spotlights could be a real slowdown but
> maybe it wouldn't be too slow.
> 
> Bob

Post a reply to this message

> > ertl/amt makes plastic model kits.  i used their model as the basis for
> > my xwing, using dial calipers to measure the parts, but comparing to
> > pics from the movies to make up for problems in the model.
> At B squared they said you couldn't trust these kits, so I didn't even think
> about looking at one.
> Another reason I didn't might be that I don't have one...

right, they can't be trusted, but there are enough accurization websites
out there to tell you exactly what is wrong and how to fix them.  the
xwing is probably one of their least screwed up.  the falcon is really
close, except for a few big mistakes.
 
> > >  But It shouldn't be too
> > > difficult to scale it to 1unit=1 meter.
> >
> > my xwing was roughly a meter=unit, but had to scale it a little in the
> > end to make it longer.  i knew the overall length, and i knew my
> > proportions were correct, so i just made a box the right length and then
> > made sure it barely fit inside length wise.
> 
> What size should I use for the Falcon?  25m is too small, but 35m is too big
> I think.

if 35m is what b squared says, i'd do it!  i think the overal multiplier
he gave was 40% larger, which is 35m right?
 
> > > I think I have about everything I need to finish it.  Maybe a more
> detailed
> > > picture of the bottom (for the armor-plating)
> >
> > i have a huge one of the bottom straight on.  want me to email it to
> > you?
> It's possible that's the one I already have, but If it's not to big (-500KB)
> you can email it.
> unless someone else wants it too, then it might be a better idea to post it
> here.

> Good luck!  It'll be very difficult to make the exact same explosion!
> I think after the Falcon is finished (next year?) I'll finish my
> StarDestroyer.  Have you already seen it?

nope.  haven't seen it.

Post a reply to this message

>
> if 35m is what b squared says, i'd do it!  i think the overal multiplier
> he gave was 40% larger, which is 35m right?
I don't remember if he said 35m, but 35m is indeed 40% more than 25m, so
that had to be it :-)

> > I think after the Falcon is finished (next year?) I'll finish my
> > StarDestroyer.  Have you already seen it?
>
> nope.  haven't seen it.
It's on my site.  Well, some images are.

ZK
http://www.povplace.be.tf

Post a reply to this message

i dont know?, the second image look perfect (real perfect!) untill you
notice the shadow covered highlight..

Rick

Post a reply to this message

Pov, I don't use Moray anymore :) Good program, I just don't use it.

Grim
"ryan constantine" <rco### [at] yahoocom> wrote in message
news:39C08B45.654599ED@yahoo.com...
> flavor: moray or povray?
>
> GrimDude wrote:
> >
> > By all means, take your time. I've got plenty to toy with :)
> >
> > Flavor?
> >
> > Grim

Post a reply to this message