I'm trying to make a branching lightning bolt using a series of cones with media
in them. The cones connect to each other, but when I make them transparent so I
can only see the media, there are gaps between them.

Post a reply to this message

Archpawn <nomail@nomail> wrote:
> I'm trying to make a branching lightning bolt using a series of cones with media
> in them. The cones connect to each other, but when I make them transparent so I
> can only see the media, there are gaps between them.

  Add the keyword 'open' to the cones to remove the endcaps.

  If there joints are still visible it means that there are holes in
your series-of-connected-cones. You have to either position the cones
better, of you have to use some other method. Look at the documentation
about sphere sweeps.

-- 
                                                          - Warp

Post a reply to this message

"Archpawn" <nomail@nomail> wrote:
> I'm trying to make a branching lightning bolt using a series of cones with media
> in them. The cones connect to each other, but when I make them transparent so I
> can only see the media, there are gaps between them.

This is only a *GUESS*, but here's what I would try:

1.) Make the radius of the small end of the cone
a little bigger than zero.

2.) Make the region of overlap between one cone
and the next a little bigger and "merge" the cones
into a single object to which the media can be
applied.

3.) Cheat. Sometimes a NEARLY transparent object
(say a tiny amount of reflection with no ambient
or diffuse light) can nicely fudge a result that
"correct" use of media is reluctant to deliver up.
If you try this, remeber to SPECIFY zero values for
ambient and diffuse,as the default vaues are not
zero.

Good luck and best regards,
Mike C.

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
>   Add the keyword 'open' to the cones to remove the endcaps.
Didn't work. I don't see how it could cause the problem, as the endcaps are just
as transparent as the rest of the cones

>   If there joints are still visible it means that there are holes in
> your series-of-connected-cones. You have to either position the cones
> better,...
When the cones aren't transparent, there are no gaps. With media, some gaps are
several pixels thick.

> ...or you have to use some other method. Look at the documentation
> about sphere sweeps.
That would solve that problem, but the media isn't constant. I currently have it
so its intensity is inversely proportional to the square of the distance from
the center. This would be virtually impossible without making the media in
separate pieces and translating it.

"Mike the Elder" <nomail@nomail> wrote:
> This is only a *GUESS*, but here's what I would try:
>
> 1.) Make the radius of the small end of the cone
> a little bigger than zero.
They're not cones in the geometric sense. Almost all of them have both ends
significantly bigger than zero.

> 2.) Make the region of overlap between one cone
> and the next a little bigger and "merge" the cones
> into a single object to which the media can be
> applied.
When I made an overlap, it made a black line at the end of each cone, so there
were pairs of black lines. This seems to provide evidence for the endcap
theory, but adding open didn't change anything. Changing it from a union to a
merge significantly decreased the thickness of the lines in this versions,
though it did little when the didn't overlap.

> 3.) Cheat. Sometimes a NEARLY transparent object
> (say a tiny amount of reflection with no ambient
> or diffuse light) can nicely fudge a result that
> "correct" use of media is reluctant to deliver up.
> If you try this, remember to SPECIFY zero values for
> ambient and diffuse,as the default values are not
> zero.
I assume you mean to use specular highlights. Didn't work. The brightness is
determined way to much by the angle of the cone. The ones parallel to a line
drawn between the camera and the light source end up reflecting the light
better and being way brighter than the ones that aren't. If I were to use
diffuse, it would have the same problem, as it's still solely dependent on the
angle of the surface. Ambient isn't, but it would just result in a solid color,
and I'd be better off to just merge the thing together and use a solid media, so
it's at least dependent on the thickness.

Should I just give up on branching lightning and go with a single bolt?

Post a reply to this message

"Archpawn" <nomail@nomail> wrote:
...
> Should I just give up on branching lightning and go with a single bolt?

Since I haven't seen the project and don't know what
your timetable and aesthetic priorities are I can only
say that it would be up to you.

One last try at helping... An OLD .inc file for
lightning generation by Jeffry Brickley can be
found here:

http://webpages.charter.net/omniverse/private/lightnin.zip

It was written using "halo", ... I did say OLD ;-)
but you may be able to adapt the idea to media.

Good luck.  If you feel inclined to share the image resulting
from your work, I, for one, would like to see it.

Best Regards,
Mike C.

Post a reply to this message

I think I found the problem. I've been intersecting off parts of the cones
after, rather than before, putting the media in them. When I took out the
intersections, and allowed the cones to intersect a little, the big gaps went
away. There's still smaller ones, but these might just be caused by the small
gaps where pieces of spheres are going to go.

By the way, this is for my own entertainment, so my time frame for this project
is forever, which is fitting, as I never consider any of my projects done.

Post a reply to this message

Never mind. I made the media be applied after the intersection and it still
doesn't work. By the way, is there any way for me to edit my old message rather
than replying to it?

Post a reply to this message

"Archpawn" <nomail@nomail> wrote in message 
news:web.47d430c944d655be9c37dbc40@news.povray.org...

> By the way, is there any way for me to edit my old message rather
> than replying to it?
>

I don't believe you can update a message in place. You should be able to 
cancel a message and submit a replacement response, though it would be easy 
to confuse people who already read your initial response, so I'd recommend 
using sparingly. Outlook Express provides an option to 'Cancel Message' on 
the 'Message' menu. There's a 'povray.test' newsgroup that you should be 
able use to practise with whatever newsgroup reader you have.

Regards,
Chris B.

Post a reply to this message