Subject:
|
Re: Some Words To BFC
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Wed, 5 Apr 2000 16:09:11 GMT
|
Reply-To:
|
RUI.MARTINS@LINK.ihatespamPT
|
Viewed:
|
2714 times
|
| |
|
|
On Wed, 5 Apr 2000, Don Heyse wrote:
> In lugnet.cad.dev, Rui Manuel Silva Martins writes:
> >
> > Which method are you using.
> >
> > I have researched (Odd/Even), (BSP) and (Winged Edge). There is also the brute
> > force method of supersampling raycasting, but this one is a NO GO.
>
> Perhaps this approach is a bit naive, but couldn't you use something similar
> to a point in polygon test?
>
> Cast a ray along the normal from the center of each polygon. If it intersects
> an odd number of surfaces then the normal points toward the inside of the
> solid and should be reversed for BFC compliancy. Is this the (Odd/even) test?
That's is the (Odd/Even) test, also with Raycasting, but no supersampling, in
the supersampling case, you just subdivide the polys further and further, or
shoot pseudo-random rays until you find the correct normals for all faces.
this does NOT work with our files!
Why you ask ?
Because we have intersecting faces and non explicit connected
faces (ex: studs in parts)
look at this cut of a 3005.dat
_____
| |
-----------
| ------- |
|_| |_|
^
|
|
Now if you shoot a ray along the vertical that passes in the midel of the
stud (point 0,0,0 ), how many faces do you have there ?
3, so it's Odd, and you don't know if the part/piece is incorrect or just if
you should assume that all volume between the 3 faces should be considered
interior volume, and what about if you have 5 faces ?
_____
| |
-----------
| ---------
| |
| |
| |
| |_________
|___________|
^
|
|
How will the program now which ones are in/out volume ?
> If so, what are the stumbling blocks?
Did I give you any hint iet ? (see above)
The intersecting face just make it even harder.
P.S.
When will the lugnet web server show the messages with a fixed size font, so
that ascii art is shown correctly.
See ya
Rui Martins
|
|
Message has 1 Reply:
 | | Re: Some Words To BFC
|
| (...) This points out that ray traced renderings of the transparent version of these pieces will not render the refraction correctly. In the best of all worlds, the interior surface would be removed. For ray tracing, the model could be replaced with (...) (25 years ago, 5-Apr-00, to lugnet.cad.dev)
|
Message is in Reply To:
| | Re: Some Words To BFC
|
| (...) Perhaps this approach is a bit naive, but couldn't you use something similar to a point in polygon test? Cast a ray along the normal from the center of each polygon. If it intersects an odd number of surfaces then the normal points toward the (...) (25 years ago, 5-Apr-00, to lugnet.cad.dev)
|
61 Messages in This Thread:
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
This Message and its Replies on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
