[x3d-public] Fw: Re: [x3d] Minutes from Nov 12 X3D Blender suppport call

Fri Nov 22 13:20:10 PST 2024

Note that I upgraded my Blender import code to latest, and the skin and
skeleton seem to be in different coordinate systems.  I have not tried
negating Z yet, but it appears to be more than just that.  Principled
testing, design and knowledge of Blender may be required.

I’ve tried various things, but I don’t have the right recipe yet.

John

On Thu, Nov 14, 2024 at 3:04 PM John Carlson <yottzumm at gmail.com> wrote:

> Joe, I acknowledge that problems rotations or joints centers may be key in
> getting animations working properly.  More than just swapping bone head and
> tail.  I acknowledge that I may not know “The Blender way.”  I also
> acknowledge that animations weren’t working at all with .x3d imports before
> my feeble attempts.
>
> Mostly, I’ve been adapting existing code that might have been incorrect.
> AFAICT, it only possibly worked with some version of .wrl ROUTEs I am
> unaware of.
>
> Specifically, I’ve written code that does animation for Blender bones,
> particularly bone orientation, based on existing OrientationInterpolator
> code.  To me, the armature animation looks a lot better than what the skin
> does, so I’ve been focusing on something to do with skin weights.  I am not
> saying the bone animations are correct yet, particularly for Jin and
> segment and site geometry.  I am not focusing on segment and site geometry
> animation presently.  That has been particularly challenging.  I am trying
> for low hanging fruit.
>
> Yes, my import Python looks a bit funny with its strange parameters.
> They’re the best I have so far.
>
> I am fully open to criticism, I’ve done the best with the knowledge I
> have.  I don’t know how HAnim will work with existing scenes.
>
> I would hope that people can review my Python code here:
>
> https://projects.blender.org/yottzumm/io_scene_x3d
>
> Thanks,
>
> John
>
> On Wed, Nov 13, 2024 at 10:18 PM Joe D Williams via x3d-public <
> x3d-public at web3d.org> wrote:
>
>>
>> -----Forwarded Message-----
>> From: Joe D Williams <joedwil at earthlink.net>
>> Sent: Nov 13, 2024 7:38 PM
>> To: Vincent Marchetti <vmarchetti at kshell.com>
>> Cc: X3D Ecosystem public discussion <x3d-ecosystem at web3d.org>
>> Subject: Re: [x3d-public] [x3d] Minutes from Nov 12 X3D Blender suppport
>> call
>>
>> Hi Vince,
>> The basic idea is that x3d is cast in the GL coordinate system where +y
>> is up, model faces +z is toward the viewer, and +x is to the model left,
>> viewer right. Axis rotations follow right hand rule. + is ccw. .
>> So for the ideal HAnim source pose before animation all joints and points
>> are 0 0 1 0.
>> The host world coord sys is same as hanim model coord sys.
>>
>> From what I seem to see I think the Blender world coordinate system is +z
>> up and left hand rule for rotations. That is why the gltf needs the +Yup
>> flag, because gltf is cast in GL coord sys.
>> I am not sure what that gltf export +Yup flag does to the data from
>> blender. Does it just do that x-axis rotation, or more?
>>
>> So, this may be called the problem of using a Dx model in a GL system.
>> Converting seems not so easy to export by just rotation of x-axis as shown
>> in the example because joint and point coordinates may not be as expected
>> GL and x3d 0 0 1 0 and animations may be a further problem.
>>
>> This is opportunity that needs some work.to understand. Right,we need a
>> simple example that shows default x3d model coordinate system and a simple
>> animation in that system,, import to Blender and see what happens, make
>> some simple mods and animations then see what it exports and how it plays
>> in x3d. Turns out that Vulcan is the only khronos that uses what think is
>> the Dx as a basis
>>
>> John actually has some experience in this in his work with the kicker.
>>
>> Good Luck, it is Important, since may authoring and run times have chosen
>> the Dx and others something different from either.
>>
>> All Best,
>> Joe
>>
>>
>> -----Original Message-----
>> From: Vincent Marchetti
>> Sent: Nov 13, 2024 1:24 PM
>> To: Joe D Williams
>> Cc: X3D Ecosystem public discussion
>> Subject: Re: [x3d-public] [x3d] Minutes from Nov 12 X3D Blender suppport
>> call
>>
>> I have seen this issue brought up before. Is the problem particular to
>> exporting HAnim type models, or is it generic to all Blender to X3D exports
>>
>> What would help understand and potentially create a solution would be if
>> someone created a .blend file that had
>> -- a single shape, with low poly mesh but with a visually obvious front
>> and up direction
>> -- that shape oriented in Blender space with the up direction along
>> Blender +Z and the forward direction along Blender -Y, and this achieved
>> without any rotation of the mesh defined in Blender.
>>
>> Then we can examine X3D exports with various options to see if there's a
>> problem to fix in the code.
>>
>> Vince Marchetti
>>
>>
>> > On Nov 13, 2024, at 4:03 PM, Joe D Williams wrote:
>> >
>> >
>> https://github.com/andreasplesch/x3dom/wiki/RiggedSimple-glTF-conversion
>> >
>> > X3D
>> > experimental xml
>> > ...
>> >
>> > Transform DEF='node_0' rotation='1 0 0 1.57
>> >
>> > This is not good. It shows the thing was drawn +zup.
>> > This should be fixed in blender either by fixing blender to do the GL
>> default, not depending on the Transform of the model to put the thing into
>> GL space.
>> > Thanks,
>> > Joe
>> > -----Original Message-----
>> > From: Extensible 3D (X3D) Graphics public discussion
>> > Sent: Nov 13, 2024 8:00 AM
>> > To: Vincent Marchetti , X3D Ecosystem public discussion , ,
>> > Cc: Joe D Williams
>> > Subject: Re: [x3d-public] [x3d] Minutes from Nov 12 X3D Blender
>> suppport call
>> >
>> > So, to me for hanim it seems like we must fix Blender at the source.
>> > Just about the idea that Blender considers all Joints can affect all
>> points, seems like that is a non-starter. Doesn't this liimit the
>> skeleton-driven skin animation to less than 16 Joint nodes?
>> > If so, then Blender needs to be fixed to operate more like x3d hanim,
>> and every other tool, to allow for more skeleton joints(?) Gltf has a
>> suggested? limit of, four joints to affect any point.
>> > What am I missing?
>> > Thanks for all the great work on this import/export with Blender.
>> > Joe
>> >
>> >
>> > -----Original Message-----
>> > From: Vincent Marchetti
>> > Sent: Nov 13, 2024 6:36 AM
>> > To: X3D Ecosystem public discussion ,
>> > Subject: [x3d] Minutes from Nov 12 X3D Blender suppport call
>> >
>> > Minutes from the Nov 12 2024 meeting on Blender 4.2 support for X3D are
>> available for view and comment
>> >
>> >
>> https://docs.google.com/document/d/1wvjvDpJTMFDMpS53ImuoRHNOCt9zrCiq/edit
>> >
>> > Vince Marchetti
>> >
>> >
>> > _______________________________________________
>> > x3d mailing list
>> > x3d at web3d.org
>> > http://web3d.org/mailman/listinfo/x3d_web3d.org
>> >
>> >
>> >
>> > _______________________________________________
>> > x3d-public mailing list
>> > x3d-public at web3d.org
>> > http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>> >
>> >
>>
>>
>>
>>
>>
>>
>> _______________________________________________
>> x3d-public mailing list
>> x3d-public at web3d.org
>> http://web3d.org/mailman/listinfo/x3d-public_web3d.org
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://web3d.org/pipermail/x3d-public_web3d.org/attachments/20241122/f883f4f5/attachment.html>