Subject:
|
Re: The use of type 5 lines for smoothing of objects.
|
Newsgroups:
|
lugnet.cad.dev
|
Date:
|
Tue, 21 May 2002 17:00:33 GMT
|
Viewed:
|
516 times
|
| |
|
|
In lugnet.cad.dev, Don Heyse writes:
> If you ignore the type 5 control points, what points do you use to
> generate the cross products? I would think to generate "fair" lengthy
> vectors you'd have to use the center of gravity point for the adjacent
> quads or triangles. Otherwise you won't get good results if say
> one point of an adjacent quad is near, and the other is far.
I only use the actual (un-normalized) normals of the faces adjacent to the
type 5 line. It's a little more complicated, because when a single point
has multiple type 5 lines radiating out from it, the surface normal I
compute at that point takes all the adjacent faces into account. I saw one
model which had a very shallow cone with at least 16 segments, and the
surface normal was right. On the other hand, the algorigthm won't properly
handle a steep cone, since the surface normal at the point would point
either straight up or straight down. I have to think about how to solve
that problem.
As for my weighting scheme; I agree that it probably won't produce
mathematically accurate results. However, with surface normals, you have an
amazing amount of play available without producing noticably incorrect
results. All the visuals I've looked at so far seem reasonable. I also
think that my mechanism is producing results that are more accurate than you
would think on first glance. Cross products have some really nice
properties that I believe make up for much of the seeming problems in my scheme.
I'll investigate a little further, though. Incidentally, I did get it sped
up considerably last night. I still have to look into using a real Map
class, though.
--Travis Cobbs (tcobbs@REMOVE.halibut.com)
|
|
Message is in Reply To:
 | | Re: The use of type 5 lines for smoothing of objects.
|
| (...) If you ignore the type 5 control points, what points do you use to generate the cross products? I would think to generate "fair" lengthy vectors you'd have to use the center of gravity point for the adjacent quads or triangles. Otherwise you (...) (22 years ago, 21-May-02, to lugnet.cad.dev)
|
22 Messages in This Thread:
 
 
- Entire Thread on One Page:
- Nested:
All | Brief | Compact | Dots
Linear:
All | Brief | Compact
|
|
|
|
