This was done in just a few hours time last night.  before.jpg is the
hull mesh and after.jpg is with three levels of subdivision plus normal
smoothing.

My modeler can now apply texturing to individual faces.  All I need now
is to add uv-mapping.

Regards,
John

Post a reply to this message

Attachments:
Download 'before.jpg' (27 KB) Download 'after.jpg' (27 KB)

Preview of image 'before.jpg'

Preview of image 'after.jpg'

John VanSickle wrote:
> This was done in just a few hours time last night.  before.jpg is the
> hull mesh and after.jpg is with three levels of subdivision plus normal
> smoothing.
> 
> My modeler can now apply texturing to individual faces.  All I need now
> is to add uv-mapping.
> 
> Regards,
> John
> 
> ------------------------------------------------------------------------
> 
> 
> ------------------------------------------------------------------------
> 
Very, very impressive!

Post a reply to this message

John VanSickle wrote:
> 
> This was done in just a few hours time last night.  before.jpg is the
> hull mesh and after.jpg is with three levels of subdivision plus normal
> smoothing.
> 
> My modeler can now apply texturing to individual faces.  All I need now
> is to add uv-mapping.
> 
> Regards,
> John
> 
>   ------------------------------------------------------------------------
>  [Image]  [Image]

Impressive.
Is there any predictable relationship between the original mesh and the
final result?
It now looks a bit like what you have when working with blobs: it needs
a lot of experimenting to get everything the right size.
Can you for instance be sure about the length or width of a paw or ear?

Again, I think it's very impressive and it looks very, very useful,
predictable or not.

Remco

Post a reply to this message

I agree.  Great job!

Post a reply to this message

John VanSickle wrote:
> This was done in just a few hours time last night.  before.jpg is the
> hull mesh and after.jpg is with three levels of subdivision plus normal
> smoothing.
That looks nice. Can you see the control mesh after subdivisions and 
modify it on the fly in your editor? It would make editing a whole lot 
easier (but needs a lot of computational muscles).
What scheme are you using for subdivisions? You might find this paper 
interesting:  http://tom.cs.byu.edu/~tom/papers/nurss.pdf. It's one way 
of preserving sharpe features. I don't know if it's the best option, but 
not too much work if you have a working Catmull-Clark.

  Ari-Matti

Post a reply to this message

Ari-Matti Leppaenen wrote:

> John VanSickle wrote:
> 
>> This was done in just a few hours time last night.  before.jpg is the
>> hull mesh and after.jpg is with three levels of subdivision plus normal
>> smoothing.
> 
> That looks nice. Can you see the control mesh after subdivisions and 
> modify it on the fly in your editor? It would make editing a whole lot 
> easier (but needs a lot of computational muscles).


No.  The modeler doesn't do the subdivisions.  The modeler exports the
basic mesh to a POV script file.  I have a set of macros here

   http://enphilistor.users4.50megs.com/nsss.htm

which take the polygon hull and turn it into the smooth form.

> What scheme are you using for subdivisions? You might find this paper 
> interesting:  http://tom.cs.byu.edu/~tom/papers/nurss.pdf. It's one way 
> of preserving sharpe features. I don't know if it's the best option, but 
> not too much work if you have a working Catmull-Clark.

Right now I use a modified Loop scheme (extended to include
quadrilateral faces).

Regards,
John

Post a reply to this message

Remco de Korte wrote:

> Impressive.


Thanks.

> Is there any predictable relationship between the original mesh and the
> final result?


The final mesh is always confined to the convex hull of the original
vertices.  I know that isn't saying much.

However, if I make a box and subdivide it, the resulting ball (close to
spherical) will be just slightly smaller than the box.  Another part of
this thread brought up the question of interactive display of the
refined mesh; I might add that later.

> It now looks a bit like what you have when working with blobs: it needs
> a lot of experimenting to get everything the right size.
> Can you for instance be sure about the length or width of a paw or ear?


The amount of shrinkage away from the mesh depends on how sharp the
angles are between the original mesh faces.  A tetrahedron shrinks a
lot, a cube shrinks less, an icosahedron shrinks a lot less.

> Again, I think it's very impressive and it looks very, very useful,
> predictable or not.

Yep.  The people who claim to be tired of Rusty may get their wish.

Regards,
John

Post a reply to this message

Very interesting.
Could one do some "sloppy" modelling first with povray primitives and then
have it go to mesh and then to ssd mesh?

Post a reply to this message

> Could one do some "sloppy" modelling first with povray primitives and then
> have it go to mesh and then to ssd mesh?
I just wrote some macros which output POV-Primitives as Mesh. Using 
those you could propably do it. I will post them when I have found 
time to improve them a little and to write some short documentation. 
But your idea is great and I will check if it's possible right now :)

Greetings,
Florian
-- 
//=================[web: http://www.torfbold.com]==================\\
#local a=-5;#while(a<5)sphere{<sin(a*pi)*5a*10pow(a,5)*.01>sin(a*a*a*
.1)+1pigment{rgb 9*z}}#local a=a+.01;#end camera{look_at-y*10location
<8,-3,-8>*10}// [www.povray.org]     [www.imp.org]     [www.irtc.org]

Post a reply to this message

John VanSickle wrote:
> 

Coorect me if I'm wrong, but what I understand of it is that you have
made a modeller to create a mesh which you feed to POV-macros to get the
subdivided mesh. If so, I understand how you came to this (the macros
came first) but wouldn't it be far more efficient to do the subdividing
in the modeller as well? That way you can get (sort of) a preview and
it's probably a lot faster then the POV parsing.

Remco

Post a reply to this message