Subject:
|
Re: 1-8sphe.dat tesselations compaire image
|
Newsgroups:
|
lugnet.cad
|
Date:
|
Tue, 3 Oct 2000 21:56:44 GMT
|
Viewed:
|
1620 times
|
| |
|
|
I agree with Manfred's conclusions.
I'd also like to add that if Travis had made his submission at a different
time, so we didn't have the alternative, I would have been very happy to
accept his version of the primitive.
Another thought: what about the 48\1-8sphe.dat? The current file does not
follow James' approach to the original 1-8sphe.dat, and it also does not
have a particularly pleasant arrangement of polygons.
Steve
In lugnet.cad, Manfred Moolhuysen wrote:
> 4 days ago I've posted that I've made a compaire image showing the two
> possible alternatives for surface tesselations for the 1-8sphe.dat primitive.
> I've waited a while before giving my own opinion, but here it is:
>
> For me, the proposed alternative for the tesselation, using 16 triangles per
> sector, has no substantial benifit over the tesselation as used in the
> existing official primitive, using 9 quads and 4 triangles per sector.
> My arguments:
>
> The inaccurate coordinate values that the old 1-8sphe.dat primitive still had,
> have been resolved with a fix submitted by Axel Poqué and I in August this
> year. (This fix is part of the recent 2000-02 update)
>
> I think that face distribution with 9 quads and 4 triangles isn't a
> disadvantage compaired with Travis Cobbs his distribution with 16 triangles.
> In the compaire image http://members.ams.chello.nl/mmoolhuy/compaire.jpg
> it seems to me that the existing 9 quads and 4 triangles tesselation is even
> better looking that the proposed 16 triangle alternative. In my opinion, the
> dome on the right looks a little smoother than the dome on the left.
>
> Calculation of intersection points with other faces is easyer for the existing
> primitive than for the proposed alternartive.
>
>
> However, I appreciate that Travis did made an effort, and hope I haven't
> discouraged him to make other contributions in the future. After all, he was
> right in saying that (at least) the accuracy of the old primitive had to be
> improved.
>
>
> References for those of you who want to do some research of their own:
>
> My compaire image followed a proposal by Travis Cobbs to replace the
> 1-8shpe.dat primitive. See this mail and its follow-ups:
> http://news.lugnet.com/cad/dat/parts/primitives/?n=70
>
> My mail containing the link to the compaire image:
> > Recently Travis Cobbs proposed to replace the existing 1-8sphe.dat primitive
> > with an alternative that uses a different tesselation of the sphere surface.
> > (tesselation is the way in wich a curved surface is represented by dividing it
> > into plane quads and/or triangles)
> >
> > For the benefit of those who want to have a say in this choice, I've
> > constucted an image where whan can compaire the effects of the proposed
> > tesselation agains the current one. This image can be viewed at:
> >
> > http://members.ams.chello.nl/mmoolhuy/compaire.jpg
> >
> > On the left is the proposed "Travis Cobbs"-method, which uses
> > 16 triangles per sector.
> > On the right is te existing "James Jessiman"-method, which uses
> > 9 quads and 4 triangles per sector.
> >
> > Tecnical data of the picture:
> > The objects you see are two identical 4-4cyli.dat primitives (radius 100),
> > each topped with a dome consisting of four 1-8sphe.dat primitives of the
> > appropriate tesselation type.
> > I've rendered the picture with L3Lab,
> > Shading: on,
> > and Color Scheme: 3 Ldlite+Mlcad
> >
> > Because Travis Cobbs proposed primitive hasn't type 5 lines yet, I've also
> > turned Show Line Type 5 off, to have a fair comparison.
> > The background has been manualy changed into white.
> >
> > Greetings, M. Moolhuysen
|
|
Message has 1 Reply:
 | | Re: 1-8sphe.dat tesselations compaire image
|
| "Steve Bliss" <steve.bliss@home.com> wrote in message news:6alktscro14u6dr...4ax.com... (...) Quite frankly, I have no idea what a 48 version would look like that followed James' approach. Would there be a single row of quads with the rest being (...) (26 years ago, 4-Oct-00, to lugnet.cad)
|
Message is in Reply To:
| | Re: 1-8sphe.dat tesselations compaire image
|
| 4 days ago I've posted that I've made a compaire image showing the two possible alternatives for surface tesselations for the 1-8sphe.dat primitive. I've waited a while before giving my own opinion, but here it is: For me, the proposed alternative (...) (26 years ago, 3-Oct-00, to lugnet.cad)
|
19 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
|
|
|
|
