Ack!  Neglected to mention I changed the OBS (observer) to be SUN instead of 
SHADOW, or else there wouldn't be anything to see.

Post a reply to this message

Thanks, Bob, I'm glad you found it useful.

I'd be careful about changing the size of the sun, moon and earth, but
changing the position of the camera can't do any harm.

I looked at your picture ... if you chose SUN as the viewpoint then (during
the eclipse) the moon should be in front of the earth.

I think I wrote somewhere in the code that a light source can't see the
shadows it makes.

I haven't tried running the code through POVRay in a while. I used rev 3.5
Does it still work with the latest revision?

Post a reply to this message

"pavium" <jrc### [at] tpgcomau> wrote in message 
news:web.44165789181997a38398f4b40@news.povray.org...
>
> I'd be careful about changing the size of the sun, moon and earth

I'm guessing you never meant for this to be about lunar eclipses, only solar 
eclipses.

> I looked at your picture ... if you chose SUN as the viewpoint then 
> (during
> the eclipse) the moon should be in front of the earth.

 Hmmm, well, it wasn't.  :)

> I think I wrote somewhere in the code that a light source can't see the
> shadows it makes.

Since I've raised the camera up out of the ecliptic it would have to see the 
shadow cast onto the moon from the earth... but if you meant shadowless, 
yeah the looks_like Sun object precludes any self-made shadows. And if that 
isn't what you're saying then I'm lost.

> I haven't tried running the code through POVRay in a while. I used rev 3.5
> Does it still work with the latest revision?

At first I thought it looked okay in 3.6.1b using the original date and time 
you had for the 1999 eclipse, but it seems like the second sun and moon you 
put in there don't belong. I can't really say what is happening, I jumped 
into changing things before I bothered to check on it. The area_light I 
added for the sunlight really changes the shadow a lot, too. I have that at 
the SUNRAD (radius) at the moment and it looks good, along with the moon 
being scale 1 not 2.

Something especially confusing me is how the day and night sides of the 
earth don't seem to correspond to the sun (when I changed to this upcoming 
lunar eclipse), which is rather strange considering the earth's shadow still 
shows up on the moon as expected. I'm sure I'd keep discovering new things 
if I kept reading through the script. The math itself I dare not touch!

Bob

Post a reply to this message

This is a perfect example of how it helps to read a file thoroughly. I 
wasn't paying any attention to most of it as I went in and began changing 
things just so I could get a POV-rendered look at that lunar penumbral 
eclipse before tonight. I didn't even try the commented out line for look_at 
<XMOON,YMOON,ZMOON> until a little while ago! Of course, I wanted the Earth 
in the view anyway, so... I've probably made a real mess of your original 
file, which I'm keeping separate.  ;)

I think maybe I know now what you were saying about the light and shadow, 
how the camera viewpoint is found along the moon and sun line then placed 
near the earth on the sunlit side. At least I guess I do, this kind of thing 
makes my head spin. The part immediately after the Earth and Moon object 
creation is the calculations for that. Not sure if the comments have an 
error where it says: x^2 + y2 + z^2 = EARTHRAD^2^ since I'd have thought it 
would say x^2 + y^2 + z^2 = EARTHRAD^2 but I'm not the one to second guess 
you on that. Ha ha!

Fun to mess around with it, regardless of my ineptitude.

Bob

Post a reply to this message

Well I only tried to model a lunar eclipse once so I could get the image on
the 'home' page. It wasn't ideal, but it gave the impression of the Moon in
shadow so I used it.

Eclipse.pov models the Sun, Earth and Moon at any given time and date.

Solar Eclipses seem more interesting because you can recognise the places
the Moon's shadow falls on, but recreating the circumstances of a lunar
eclipse should be just a matter of finding the date and time when it
happens (what astronomers call the central eclipse, I think) and then
pointing the camera at the Moon's co-ordinates instead of the Earth's. The
Earth's shadow is much bigger than the Moon and you might see only a
completely darkened Moon. On the other hand a penumbral eclipse might not be
the spectacle you'd like it to be.

My comment about the light source not being able to see the shadows it makes
follows from the idea that a shadow is a place where a light source can't
reach. If your eye was a light source, it couldn't see the shadows it makes
because they're in places where your eye (as light source) isn't visible.

In a way, it reminds me of the signs we see on trucks here in Australia - If
you can't see my mirrors, I can't see you.

And I'm talking in general terms about a simple system with one real light
source and not about POVRay constructs like shadowless light sources.

It's very strange if the night and day sides of the Earth don't correspond
to the Sun because eclipse.pov (in spite of the complexity of the
astronomical calculations) essentially creates a scene with a light source
and two bodies. If POVRay can't render a simple scene like that, there's
something seriously wrong ;-)

The only fiction I introduced to my simple scene was the additional Sun and
Moon so that I could get a distinct umbra and penumbra. These extra
components are arranged inline with the real Sun and Moon so the shadows
they create line up too.

An area light sounds like a more realistic approach, provided it's
dimensioned properly, but the edges of the Moon's shadow will be indistinct
and
you won't be able to see the path of totality. I once found on the Internet
a short movie of a solar eclipse taken by Shuttle astronauts. The shadow
was precisely as I've just described.

It looks like I'll have to invest in some hardware which will run POVRay
once more. I have actually tried running POVRay on a 75MHz Pentium under
OpenBSD with no graphics capability. It actually generated 25x80 ASCII text
images onscreen while saving GIF files to disk. It was fascinating to watch
but only marginally faster than watching grass grow.

Post a reply to this message

Bob,

It's been a while since I wrote eclipse.pov -- some of the details are still
coming back to me.

I had a closer look at your picture 'eclipse.jpg'.

It looks good to me - a penumbral eclipse of the moon might pass unnoticed,
depending on how far into the penumbra the Moon travels.

I can see a little bit of shadow on the lower edge of the Moon. Maybe this
is all we're supposed to see.

And of course for a lunar eclipse it should be opposite what I said before.
The Earth should be between the Sun and Moon, but if you raised the camera
out of the ecliptic you can see both with the Moon in the distance.

You could try commenting out sun2 and the tiny moon (if you haven't already
done so) as they only help in solar eclipses. Removing sun2 will make the
shadow on the Moon a bit darker.

Sorry about the bum steer ...

Post a reply to this message

"pavium" <jrc### [at] tpgcomau> wrote in message 
news:web.4416b350181997a3e8faad540@news.povray.org...
>
> a penumbral eclipse of the moon might pass unnoticed
> I can see a little bit of shadow on the lower edge of the Moon.

Yep. This one tonight, for the eastern US where I live, isn't going to look 
like much. Should just be dimmed on the righthand side judging by this 
eclipse.pov rendering. I see Australia is opposite the Moon at the time it 
happens, unlucky for you maybe, but you won't be missing this I'm sure.

> And of course for a lunar eclipse it should be opposite what I said 
> before.
> The Earth should be between the Sun and Moon, but if you raised the camera
> out of the ecliptic you can see both with the Moon in the distance.

Adding a second satellite (along with the few necessary parameters) seems to 
work okay, making it look past the Earth. But for now it's just there for 
today's date so I'm sure it would need merging into the other calculating 
somehow.

> You could try commenting out sun2 and the tiny moon (if you haven't 
> already
> done so) as they only help in solar eclipses. Removing sun2 will make the
> shadow on the Moon a bit darker.

Yeah, since that rendering I did comment out the secondary tiny moon, mainly 
because I wasn't sure how it fit into the scene and didn't want to bother 
figuring it out while I was having that peculiar lighting of the Earth. 
Found out I was probably using incompatible image maps so I removed the 
clouds altogether and put in a somewhat large jpeg of the Earth instead. 
After that I changed sun2 into a spotlight pointing at the Earth and reduced 
its brightness considerably so it wouldn't illuminate the Moon. That helped 
darken the shadow.

Most interesting to me is that stuff about locating a place along the 
alignments, so unless I give up trying to understand it I *might* learn 
something.

Post a reply to this message

pavium wrote:
> It looks like I'll have to invest in some hardware which will run POVRay
> once more. I have actually tried running POVRay on a 75MHz Pentium under
> OpenBSD with no graphics capability. It actually generated 25x80 ASCII text
> images onscreen while saving GIF files to disk. It was fascinating to watch
> but only marginally faster than watching grass grow.


Oh come on, I originally ran POV on a 486 with NO math coprocessor! 
Think of that - all that double precision math being emulated by the 
integer units!  IIRC, it would draw an RSOCP at ~20-30 pps.  Imagine how 
excited I was when I got my first 486 DX, and saw POV running at speeds 
of 100pps or more! :)

...Chambers

Post a reply to this message

Chambers <bdc### [at] yahoocom> wrote:
> Oh come on, I originally ran POV on a 486 with NO math coprocessor!
> Think of that - all that double precision math being emulated by the
> integer units!  IIRC, it would draw an RSOCP at ~20-30 pps.  Imagine how
> excited I was when I got my first 486 DX, and saw POV running at speeds
> of 100pps or more! :)

486? You were lucky to have a 486! We used to do the calculations by
counting on our fingers, with one arm tied behind our backs and no toes.
Half the fingers were missing, with someone playing old Monty Python
sketches to distract us!

But you try and tell the young people today that... and they won't believe
ya.

Post a reply to this message

Chambers <bdc### [at] yahoocom> wrote:
> Oh come on, I originally ran POV on a 486 with NO math coprocessor!
> Think of that - all that double precision math being emulated by the
> integer units!  IIRC, it would draw an RSOCP at ~20-30 pps.  Imagine how
> excited I was when I got my first 486 DX, and saw POV running at speeds
> of 100pps or more! :)

I can beat that. I first ran POV on an 8 MHz ARM2 processor in an old Acorn
Archimedes... (which many of you have probably never heard of!). My first
speed upgrade would have been the 33MHz ARM6 in the (slightly) newer Acorn
RiscPC. Neither machine had a co-processor, floating-point calculations
were by software emulation only. Test renders ran like continental drift...
:)

Bill

Post a reply to this message