> (you may decide to go to povray.off-topic, I read that also)

it's not off-topic to povray.general, just off-topic to "Boned mesh file 
formats?" - anyway, we can continue it here until someone complains ;-)

>> My problem is that a cubic bezier triangle can't be connected smoothly 
>> to cubic bezier patches 
> 
> Sorry, you lost me there. Exactly what is your problem?

the "standard" rectangular (tensor product) patches in povray are 
bicubic, thus they have 16 controlpoints.

A "standard" cubic bezier triangle 
http://www.google.at/search?q=bezier+triangle has got 10 controlpoints 
(the outlines are cubic bezier curves, and there is one center-point). 
As far as I know, you can't smoothly connect such a triangle to 
rectangular (cubic) patches. I think (but I'm not sure) that it would be 
possible for bezier-triangles of order 4 (quartic) - (by degree 
elevating http://www.google.at/search?degree+elevation the outlines) 
bacause it has three inner controlpoints...

The only three- and five-sided patches I know of, which can be smootly 
connected to the rectangular ones are Hosaka/Kimura (M. Hosaka and F. 
Kimura. Non-four-sided patch expressions with control points. Computer 
Aided Geometric Design 1(1): 75-86, 1984.) for the cubic case, and 
Zheng/Ball for the general case (arbitrary degree).

Thus, I'm looking for a conversion scheme for such patches, similar to 
that one for "classical" bezier-triangles into "standard" rectangular 
patches (as in my macro).

> I am quite sure that I downloaded and printed that article then.
> The paper will be in a stack somewhere and I probably have the PDF on
> a backup disk.
> There is a new elsevier policy that allows the authors to have a PDF
> of their article for download at a page of their institution.

I have a zip archive with the zheng/ball publication (it was available 
for download on Mr. Zhengs webpage some time age), but I don't know if 
it is legal to share it.

It's a little bit beyond my math skills, so I'd be interested in reading 
the Hosaka/Kimura version, but I'Ve never found a download link :-/

Elsevier: I'm not a student, do you know if universities have (free) 
access to such material?
Would it be legal to share such a paper within an open-source project 
(i.e. not for commercial use)?

 > Perhaps there should be a tutorial
 > on advanced bezier techniques, but I do not have the time to write
 > that.

We could start with a discussion about "advanced bezier techniques" to 
share expirence and discuss all that scary math stuff ;-)
It could eventually become something like a tutrial with tips on using 
bezier patches in povray.
Not sure what the best platform would be - a thread in p.advanced-users?
Alternatively I could add a forum on my webpage's phpbb board 
http://www.jpatch.com/forum_frameset.html
or a wiki...

-sascha

Post a reply to this message

Sascha Ledinsky wrote:

>  > (you may decide to go to povray.off-topic, I read that also)
> 
> it's not off-topic to povray.general, just off-topic to "Boned mesh file 
> formats?" - anyway, we can continue it here until someone complains ;-)
> 
>>> My problem is that a cubic bezier triangle can't be connected 
>>> smoothly to cubic bezier patches 
>>
>>
>> Sorry, you lost me there. Exactly what is your problem?
> 
> 
> the "standard" rectangular (tensor product) patches in povray are 
> bicubic, thus they have 16 controlpoints.
> 
> A "standard" cubic bezier triangle 
> http://www.google.at/search?q=bezier+triangle has got 10 controlpoints 
> (the outlines are cubic bezier curves, and there is one center-point). 
> As far as I know, you can't smoothly connect such a triangle to 
> rectangular (cubic) patches. I think (but I'm not sure) that it would be 
> possible for bezier-triangles of order 4 (quartic) - (by degree 
> elevating http://www.google.at/search?degree+elevation the outlines) 
> bacause it has three inner controlpoints...
> 
> The only three- and five-sided patches I know of, which can be smootly 
> connected to the rectangular ones are Hosaka/Kimura (M. Hosaka and F. 
> Kimura. Non-four-sided patch expressions with control points. Computer 
> Aided Geometric Design 1(1): 75-86, 1984.) for the cubic case, and 
> Zheng/Ball for the general case (arbitrary degree).
> 
(to be honest I did not know that the standard bezier triangles had
only 10 control points, so that explains my confusion)
IIRC the Zheng Ball triangles had 3 inner control points. That implied
that every edge had 4 neighbour control points in the triangle
(two inside and 2 at the other edges). The derivative perpendicular
to the edge was defined by these 4 control points, just as in quad
patches. connecting to quad patches was thus conceptively simple.
Therefore, I think that patches with 12 control points will be more
useful for POV users, and true bezier triangles are only of academic
interest. You already seem to have a bezier patch with 12 control
points and I assume(hope) that they have the same properties as
the Zheng Ball triangles. So for all practical purposes my question
is: why look further?
And (just to annoy Thorsten): can we get Zheng Ball triangles and
pentangles as primitives in the next version of POV?
> Thus, I'm looking for a conversion scheme for such patches, similar to 
> that one for "classical" bezier-triangles into "standard" rectangular 
> patches (as in my macro).
If you insist, you can always compute the real 3D coordinates of the
points you want to use as the control points of your three patches
and invert the bezier computation to get the actual control points
that are needed by POV.
> 
>> I am quite sure that I downloaded and printed that article then.
>> The paper will be in a stack somewhere and I probably have the PDF on
>> a backup disk.
>> There is a new elsevier policy that allows the authors to have a PDF
>> of their article for download at a page of their institution.
> 
> 
> I have a zip archive with the zheng/ball publication (it was available 
> for download on Mr. Zhengs webpage some time age), but I don't know if 
> it is legal to share it.
Ah, getting off-topic. Yes an interesting IP question. I downloaded
it also and at that time it was free to use and distribute. Now,
it appears that it is not anymore. Is my copy still free and free
to distribute or am I not allowed anymore to share it. Basically
the question is can you change the legal status of a document
later without consent and without informing the holder of that
document.
> 
> It's a little bit beyond my math skills, so I'd be interested in reading 
> the Hosaka/Kimura version, but I'Ve never found a download link :-/
> 
> Elsevier: I'm not a student, do you know if universities have (free) 
> access to such material?
No, unless they have a subscription to the paper.
> Would it be legal to share such a paper within an open-source project 
> (i.e. not for commercial use)?
Ask a lawyer :)
I guess it is, as the paper is not a part of the project, only
secondary literature.
>  > Perhaps there should be a tutorial
>  > on advanced bezier techniques, but I do not have the time to write
>  > that.
> 
> We could start with a discussion about "advanced bezier techniques" to 
> share expirence and discuss all that scary math stuff ;-)
> It could eventually become something like a tutrial with tips on using 
> bezier patches in povray.
> Not sure what the best platform would be - a thread in p.advanced-users?
> Alternatively I could add a forum on my webpage's phpbb board 
> http://www.jpatch.com/forum_frameset.html
> or a wiki...
I do not know either.
> 
> -sascha

Post a reply to this message

> My problem is that a cubic bezier triangle can't be connected smoothly 
> to cubic bezier patches - what I'd need is a similar conversion 
> algorithm for something like Zheng-Ball patches (J.J. Zheng, A.A. Ball.

It was my own inability to find a way to stitch the patches smoothly
togeher that led me to developing a hybrid suface subdivision scheme.
This makes POV-Ray parsing a lot slower, though.

Charles Loop has been working on S-patches, which can have any number
of sides and can be as smooth as the modeler needs.  The problem with
them is that they require a s***load of control points.

Regards,
John

Post a reply to this message

I did this **in povray** for this anim:
(6 MB download:  http://www.irtc.org/ftp/pub/anims/2004-04-15/ought.mpg )

I used Mike William's file for extruding a mesh along a spline.  I had
another walk cycle which generated the points of the skeleton, then simply
put a natural spline along them. Mike Williams then generates the mesh 2.
See also smaller anims at:

http://news.povray.org/povray.binaries.animations/thread/%3C4088703a%40news.povray.org%3E/
http://news.povray.org/povray.binaries.animations/thread/%3C4036deea%40news.povray.org%3E/?ttop=197897&toff=50

So I think it is tres cool to keep it all in povray.

The Next Big Thing in this arena would of course to go beyond the "walking
worm" to the "walking pair of pants."  That is, have some code which
provides for a smooth leg-hip-otherleg.



"Rune" <run### [at] runevisioncom> wrote in message
news:412d243a$1@news.povray.org...
> I'm not sure I have the required skills, but I'd like to write a program
in
> Java that can import a boned mesh in some format, move the bones around,
> deform the mesh according to the moved bones and then export the mesh as a
> mesh2 for example.
>
> Are there any widely used formats for boned meshes?
>
> It should contain the mesh (typically a character in some standard pose),
> the bones in the same standard pose, and data about which vertices are
> affected by which bones.
>
> However, in order to deform the mesh correctly, I'd probably also have to
> know the algorithm used for the deforming. Even with that available I
might
> not have the skills, but I can always hope.
>
> I think I have to base it all on a pre-existing format for boned meshes,
> because how else would people be able to actually create those boned
meshes?
> I certainly don't have the skills to create a graphical mesh and boning
> modeler/editor.
>
> Any ideas?
>
> Rune
> -- 
> 3D images and anims, include files, tutorials and more:
> rune|vision:  http://runevision.com
> POV-Ray Ring: http://webring.povray.co.uk
>
>

Post a reply to this message

Or even cooler, do it in a povray INC.
"andrel" <a_l### [at] hotmailcom> wrote in message
news:412### [at] hotmailcom...

>
> So much I understood. My point still is that if you make a utility that
> imports some fileformat and you distribute that utility the format
> could become a de facto standard, with all the consequences.
>

Post a reply to this message

Greg M. Johnson wrote:
> I did this **in povray** for this anim:
> (6 MB download:
> http://www.irtc.org/ftp/pub/anims/2004-04-15/ought.mpg )
>
> I used Mike William's file for extruding a mesh along a spline.  I had
> another walk cycle which generated the points of the skeleton, then
> simply put a natural spline along them. Mike Williams then generates
> the mesh 2.

Extruding a mesh along a spline which happens to go through some points in a
skeleton is not the same as deforming a mesh directly according to the bones
in a skeleton.

Your technique no doubt worked for your specific alien characters (that look
like they're made of sphere_sweeps) but are you claiming it can be used as a
general method to animate mesh characters such as your MIME Man or my AL the
Alien character (provided they were meshes to begin with)?

Rune
-- 
3D images and anims, include files, tutorials and more:
rune|vision:  http://runevision.com
POV-Ray Ring: http://webring.povray.co.uk

Post a reply to this message

"Rune" <run### [at] runevisioncom> wrote:
>
> Extruding a mesh along a spline which happens to go through some points in a
> skeleton is not the same as deforming a mesh directly according to the bones
> in a skeleton.
>
> Your technique no doubt worked for your specific alien characters (that look
> like they're made of sphere_sweeps) but are you claiming it can be used as a
> general method to animate mesh characters such as your MIME Man or my AL the
> Alien character (provided they were meshes to begin with)?
>



Good question. First of all, rather than "boning", I would use the term if
we're talkig povray of "transforms".  I have shin transforms, right eye
transforms, forearm transforms, even left index finger middle digit
transforms in my system.  If you set up transforms, then you can have
"anything" you want move to the same set of transforms. I primarily think
in transforms but I can have it spit out points if I choose.


There are several levels of difficulty to the problem you describe, perhaps
it's best to consider individually:

1).  Animation of a "single long stocking"-- say going from a foot to an
ankle to a knee up to the crotch.   This I believe is quite doable using
(hoping I'm not insulting him by invoking his name  and quality macro in
assocation with my mediocre work) the Mike Williams method.  I made
(perhaps unpublished?) a simple example like this and think it's quite
"workable" with a more professional-quality construction if one were to
choose.   If you don't care exactly how smoothly Al's arms connect to his
torso, if you'd settle for mere overlapping objects, then the answer is a
flat "yes".

2). Animation of "a pair of pants" or a "sweater"-- say two independently
moving legs which smoothly join to a "hip & but" or two independently
moving arms smoothly joining to a "torso". (A more complex version would be
multiple fingers joining a hand).  This is currently over my head, although
I'm currently skeptical that it's impossible with the Mike Williams method.
I think it's just a slightly bigger coding headache and this discussion is
almost about made me want to try it.

3). Animation of a moving jaw within a mouth with an opening.  This I think
is doable, too.

Post a reply to this message

gregjohn wrote:
> "Rune" <run### [at] runevisioncom> wrote:
>>
>> Extruding a mesh along a spline which happens to go through some
>> points in a skeleton is not the same as deforming a mesh directly
>> according to the bones in a skeleton.
>>
>> Your technique no doubt worked for your specific alien characters
>> (that look like they're made of sphere_sweeps) but are you claiming
>> it can be used as a general method to animate mesh characters such
>> as your MIME Man or my AL the Alien character (provided they were
>> meshes to begin with)?
>
> Good question. First of all, rather than "boning", I would use the
> term if we're talkig povray of "transforms".

It's basically the same thing. You need one or two transforms to specify the
alignment of a bone.

> If you don't care exactly how smoothly Al's arms connect to
> his torso, if you'd settle for mere overlapping objects, then
> the answer is a flat "yes".

But I do. That's kind of the whole point with the technique.

> 2). Animation of "a pair of pants" or a "sweater"-- say two
> independently moving legs which smoothly join to a "hip & but" or two
> independently moving arms smoothly joining to a "torso".

> I think it's just a slightly bigger coding headache and
> this discussion is almost about made me want to try it.

If you mean it'd be a new coding headache every time you have a new model or
skeleton, then you're missing the point. Deforming a skeleton according to
bones is a general technique that, once coded, don't require any additional
coding no matter how different the models and skeletons are. If on the other
hand you think you *do* have a general-purpose solution, then I wish you
good luck coding it.

> 3). Animation of a moving jaw within a mouth with an opening.  This I
> think is doable, too.

Again, this should be possible using the same code as for the pants and
sweater.

Rune
-- 
3D images and anims, include files, tutorials and more:
rune|vision:  http://runevision.com
POV-Ray Ring: http://webring.povray.co.uk

Post a reply to this message

"Rune" <run### [at] runevisioncom> wrote in message
news:41324cbb$1@news.povray.org...
> > 2). Animation of "a pair of pants" or a "sweater"-- say two
> > independently moving legs which smoothly join to a "hip & but" or two
> > independently moving arms smoothly joining to a "torso".
>
> > I think it's just a slightly bigger coding headache and
> > this discussion is almost about made me want to try it.
>
> If you mean it'd be a new coding headache every time you have a new model
or
> skeleton, then you're missing the point. Deforming a skeleton according to
> bones is a general technique that, once coded, don't require any
additional
> coding no matter how different the models and skeletons are. If on the
other
> hand you think you *do* have a general-purpose solution, then I wish you
> good luck coding it.
>


Consider the 'sweater' example.  I can envision a general model which works
of course for the first character you design with it. But if everything is
procedurally generated, then you can tweak skeletal points as well as the
splines that define the arm diameters as well as the spline that defines the
shape of the torso.  It shouldn't be too hard to go from robot to man to
woman to puppy dog once you've got the system.  You would however be locked
into designing the character within this system--or to put another spin on
it, blessed by having a totally procedurally generated, portable, 100%
povray, SDL-only character.  I intuitively appreciate this philosophy,
especially if it were politically incorrect with certain artsy-fartsy types.
To quote a recent winner of an IRTC Animation round, (from memory),  "Using
povray for character animation must be a complete waste of time".

Post a reply to this message

Greg M. Johnson wrote:
> It shouldn't be too hard to go from robot to man to woman to puppy
> dog once you've got the system.

Those all have four limbs, placed relatively the same way. How about bird,
dragon, three-armed alien, spider, worm?

> You would however be locked into designing the
> character within this system--or to put another spin on it, blessed
> by having a totally procedurally generated, portable, 100% povray,
> SDL-only character.

You just don't do detailed or realistic models that way. You haven't done
it, I haven't done it, nobody have yet. Meshes are the way to go for any
serious purposes (or possible bicubic patches). And I know you can do
procedural meshes, but those too are not of the detail level you can do with
a modeler. If you only need toons, you might do fine, but for serious work
it won't do.

> I intuitively appreciate this philosophy,
> especially if it were politically incorrect with certain artsy-fartsy
> types. To quote a recent winner of an IRTC Animation round, (from
> memory),  "Using povray for character animation must be a complete
> waste of time".

I've been very much a POV purist. My animation attempts with AL the Alien
proves that. Despite that I've come to the conclusion that doing character
animation *completely* within POV-Ray *is* a waste of time. That is, if
you're not satisfied with blobby toons, unsmooth joints etc.

Rune
-- 
3D images and anims, include files, tutorials and more:
rune|vision:  http://runevision.com
POV-Ray Ring: http://webring.povray.co.uk

Post a reply to this message