Hi Mattijs,

Thank you for your reply. You seems have a better way to do the
triangulation. I also believe there must have some better ways to do that.
I tried AutoSEA, a software for acoustics modelling, which always shows a
better triangulation than my Java3D application with the same model.

Your idea looks good, but I guess the quality of the triangulation also
depends on which starting point you are choosing. Just in my case of the
2 polygons, if you don't choose the right starting point, you'll aslo end
up with some bad triangulation.

Regards,

white

On Thu, 21 Nov 2002 10:43:57 +0000, Mattijs Janssens <[EMAIL PROTECTED]> wrote:

>Hello,
>
>I had some problems with long-small triangles using some triangulation
>routine as well. I got some good ideas from the paper "Efficient
>Triangulation of Simple Polygons" (Godfried Toussaint, McGill
>University) which should be somewhere on the web. The idea is that you
>take any vertex, find another one which is 'visible' from that vertex
>and split the polygon into two along that line (and repeat the splitting
>on both polygons etc.). A good way of obtaining better quality
>triangulations is to first find all the visible vertices and then choose
>that one which best bisects the angle between the edges at your starting
>vertex.
>This is just something I hacked together (and not in Java unfortunately)
>and I am sure there are much better routines out there.
>
>Regards,
>
>Mattijs
>
>White Morph wrote:
>> Hi,
>>
>> I have 2 polygons each has 7 points, they share 3 points, so the total
>> unique points are 11. I tried 2 ways to render the 2 polygons.
>>
>> 1. Manually triangulate the 2 polygons, send the array of triangles to
>>    GeometryInfo class to calculate normals and render the shape. The result
>>    shape is shading smoothly across the 2 polygons, no obvious edge between
>>    the 2 polygons.
>>
>>    The printing result shows there are 11 unique points and 11 normals,
>>    so each point just has one normal.
>>
>>
>> 2. Send each polygon to a GeometryInfo class to do triangluation. Get
>>    the triangles from each GeometryInfo class and send them together to a
>>    third GeometryInfo class to calculate normals and then render the shape.
>>    The 2 ploygons are separted by an obvious edge.
>>
>>    The printing result shows there are 11 unique points and 16 normals, so
>>    some shared points have more than 1 normal.
>>
>>
>> The triangulation did by GeometryInfo generates some long-small triangles.
>> In my case, they generate a triangle using the 3 shared points for each
>> polygon but with opposite direction, so the 2 triangles differ 180 degree.
>> That causes the problem.
>>
>> I am wondering what the triangluation algorithm that Java3D uses, can it
>> be optimized to generate more elegant meshes???
>>
>> Thank you for your comments and suggestion----white.
>>
>> ===========================================================================
>> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>> of the message "signoff JAVA3D-INTEREST".  For general help, send email to
>> [EMAIL PROTECTED] and include in the body of the message "help".
>>
>> .
>>
>
>
>
>--
>/*---------------------------------------------------------*\
>| ===========           Mattijs Janssens                    |
>| \\        /           Development Engineer                |
>|  \\      /                                                |
>|   \\    /             Nabla Ltd.                          |
>|    \\  /              The Mews, Picketts Lodge            |
>|     \\/               Picketts Lane, Salfords,            |
>|     F ield            Surrey RH1 5RG.                     |
>|     O peration        Tel: +44 (0)1293 821272             |
>|     A nd              Email: [EMAIL PROTECTED]       |
>|     M anipulation     URL: http://www.Nabla.co.uk         |
>\*---------------------------------------------------------*/
>
> ==========================================================================
>To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
>of the message "signoff JAVA3D-INTEREST".  For general help, send email to
>[EMAIL PROTECTED] and include in the body of the message "help".

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to