Nikodemus Siivola wrote:
> 
> Remco de Korte <rem### [at] xs4allnl> wrote in <3A9AFF6D.BC7DD671@xs4all.nl>:
> 
> >I did a similar thing some time ago.
> >I've attached some small examples.
> 
> >If the codec is a problem I could
> >post the larger (file) versions .
> 
> Dunno if it is the codec or my computer, but unfortunately I was unable to
> view these...

I have already posted an mpeg version.
I didn't really remember what codec I used. 
I created the AVI with Fast Movie Processor 
and I just found it also supports the (a?)
MPEG4-codec. It could just be I used this, 
but FMP gave the file an AVI extension just 
as well. Perhaps you can view the files if 
you change the extension?

> 
> >Since I needed to have images on both
> >sides I couldn't use a mesh (or at
> 
> I think that with MegaPOV you could, since all the triangles should have a
> normal vector - and therefore you could use the inside texture thingie. I
> think. Haven't tried.

Me neither. At the time I made this I didn't 
have MegaPov yet.

> 
> Actually, it could possibly be done with meshes even in normal POV-Ray, but
> it would have been a bit messy: You can apply textures to individual
> triangles, just make a "hollow" paper, where the top and bottom surfaces
> meet only at folding points, and are something like 0.000001 apart at
> maximum.

The problem with this approach is that you can 
have folds crossing each other so it would be 
rather difficult to calculate the triangles 
(at least to me).

> 
> >thin box with imagemaps, cutting and
> >folding. This is a recursive process
> >(isn't it?) and it took ages to render.
> 
> Darn I hope I could see this. Before I decided to go the mesh-route I was
> considering something similar, along with a plenty of other options. How
> did you handle the points where the paper is folded several times over the
> same spot?
> 

That's the problem. What I did was a more or 
less recursive process. You start out with a 
thin box, cut it in half and fold them on top.
Then you take this result, cut it in two parts
and fold those together, etc. Obviously with
only a few folds this will result in a whole
clutter of CSG for POV-Ray to wade through.

Anyway, I did get the imagemaps right and that
was the whole point.

There is something else, which I saw you also
handled really well: the thickness of the 
paper and the room between folds. I don't know 
how you did it but it took me quite some 
trial and error to get it right.

> Yours,
> 
>   -- Nikodemus

Kind regards,

Remco
http://www.onwijs.com/

Post a reply to this message

In article <Xns90567D065894tsiivolacchutfi@204.213.191.228>, 
tsi### [at] cchutfi (Nikodemus Siivola) wrote:

> I think that with MegaPOV you could, since all the triangles should 
> have a normal vector - and therefore you could use the inside texture 
> thingie. I think. Haven't tried.

The interior_texture feature should work fine with triangles, the side 
that is "outside" or "up" is determined by the order of the vertices. As 
I recall, "up" is the direction from which the points wind clockwise.

-- 
Christopher James Huff
Personal: chr### [at] maccom, http://homepage.mac.com/chrishuff/
TAG: chr### [at] tagpovrayorg, http://tag.povray.org/

<><

Post a reply to this message

Remco de Korte <rem### [at] xs4allnl> wrote in

>I have already posted an mpeg version.
>I didn't really remember what codec I used. 
>I created the AVI with Fast Movie Processor 

I cannot really claim to know this stuff - like I mentioned, the paperplane 
was my first animation (discounting the tutorial) - but try VideoMach. It 
is a successor of the Fast Movie Processor, and it supports .png as input 
format, and mpeg1 (among others) as output. And most importantly for me 
it's easy to use...

>> Actually, it could possibly be done with meshes even in normal
>> POV-Ray, but it would have been a bit messy: You can apply textures to
>> individual triangles, just make a "hollow" paper, where the top and
>> bottom surfaces meet only at folding points, and are something like
>> 0.000001 apart at maximum.

>The problem with this approach is that you can 
>have folds crossing each other so it would be 
>rather difficult to calculate the triangles 
>(at least to me).

What I did was first fold the a real paperplane, unfold it and draw the 
creases on graph paper. Then it was time for some head-banging and 
trigonometry to find out the points where the creases cross each other or 
the edge of the paper. This process was much aided by rendering each 
relevant set of triangles (one set per area delimeted by creases) in a 
different color while having the "paper" in it's unfolded state. Easy 
visual verification for having the points right.

>There is something else, which I saw you also
>handled really well: the thickness of the 
>paper and the room between folds. I don't know 
>how you did it but it took me quite some 
>trial and error to get it right.

Thanks! It wasn't the hard part, really. Once I had the points (vectors) 
declared I used a simple macro to rotate the relevant points around the 
fold-lines. The hard part was the topology of the paper: which fold folds 
when, where and around what.

Also, I cheated:

A) only the "right" side of the plane is really modelled, the left side is 
just a copy of the right scaled <-1, 1, 1>. B) a couple of points along the 
center-line are checked to see if they have crossed to the wrong side. If 
they have, they are returned to the center line. If I didn't do this the 
paper would "tear" on a couple of points.

Yours,

 -- Nikodemus

Post a reply to this message

Chris Huff <chr### [at] maccom> wrote:

>The interior_texture feature should work fine with triangles, the side 
>that is "outside" or "up" is determined by the order of the vertices. As 
>I recall, "up" is the direction from which the points wind clockwise.

Thanks! Will have to try. 

 - Nikodemus

Post a reply to this message

Remco de Korte <rem### [at] xs4allnl> wrote in
<3A9C3389.D6DF4BE7@xs4all.nl>: 

>I have already posted an mpeg version.

Was finally able to view it: very nice! I can see why you wanted it done 
with pigments...

 - Nikodemus

Post a reply to this message

Nikodemus Siivola wrote:
> 

> Also, I cheated:
> 
> A) only the "right" side of the plane is really modelled, the left side is
> just a copy of the right scaled <-1, 1, 1>. B) a couple of points along the
> center-line are checked to see if they have crossed to the wrong side. If
> they have, they are returned to the center line. If I didn't do this the
> paper would "tear" on a couple of points.
> 
> Yours,
> 
>  -- Nikodemus

Cheater! 8)
It's a good trick, though. I was glad I figured out how to do it the hard way
and after that I could leave the rest up to my system (which needed a couple of
days to render the whole bunch).

BTW the method I used wasn't really very hard once I figured out how to do it.
Just find the line on which to fold, cut the model and rotate the parts around
this line. I also figured out how to more or less automate the correction for
the thickness of the paper. It's the long render times that were a problem
during the tests.


Remco

Post a reply to this message