Smalltalk › Squeak › Croquet › Croquet - Dev

avatar motion

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

2 messages Options

Howard Stearns

avatar motion

The avatar motion stuff from the Minnesota folks looks pretty darn cool.
See http://croquet-bento.blogspot.com/2006/06/better-avatars.html

The URL has a movie showing how to get some .obj (mesh) and .bvh
(motion) files from Second Life, and another showing how to run them
through an application called Poser to "zero" them, generating
new .obj and .bvh files.

I didn't quite follow what "zeroing" is. Is it that the SL avatars
are standing in a T (arms spread wide), but the motion files are
relative to some other initial position?
- Is this right? It seems odd that SL would distribute files that
don't "match" each other.
- I don't have Poser handy. (I'm cheap.) What other options do I
have?
- Is it possible to create an .obj file that matches the extensive
collection of SL .bvh files, or am I misunderstanding the issue?
Would someone like to post one or two .obj files that match the SL .bvh?

I did try just using the SL .obj/.bvh files as is, but ran into
several problems which may or may not be related to my confusion
about the above.

1. Using SL_avatar_OBJ.obj with avatar_stand.bvh and avatar_walk.bvh,
I get a null materialsDictionary within
TMeshMotion>>preRenderFaceGroup:currentFrame:using: and also in
#renderFaceGroup:currentFrame:using:.

- Are these bugs, or is SL_avatar_OBJ.obj not "the right kind"
of .obj file? (What is the right kind? How do I tell?)

"Fixing" these left the body exploded when brought into Croquet.

2. Using SL_Male.obj instead, I run into (1) and also that
TMeshMotion>>createTVertexArrays has a guard for null faceGroups, but
not for null vertexIndices within a faceGroup.

- Same question: bug or pilot error?

"Fixing" this left the figure imported, but standing in a T and
without any walking motion.

Anthony Zampogna

Re: avatar motion

Hi Howard,

Right now, we have only tested it with exported .obj and .bvh files
from Poser. Unfortunately, there are still some bugs with the parser,
so not all .obj files will import, and if they don't they won't
necessarily import with motions attached correctly.

Another thing is that the importer doesn't support any automatic
scaling of the .bvh motion to the .obj file yet.

Hopefully, we'll get around to fixing these things soon. But, for
now, the best thing to do is get a copy of Poser and use that.

Tony

On Jun 17, 2006, at 4:17 PM, Howard Stearns wrote:

> The avatar motion stuff from the Minnesota folks looks pretty darn
> cool.
> See http://croquet-bento.blogspot.com/2006/06/better-avatars.html
>
> The URL has a movie showing how to get some .obj (mesh) and .bvh
> (motion) files from Second Life, and another showing how to run
> them through an application called Poser to "zero" them, generating
> new .obj and .bvh files.
>
> I didn't quite follow what "zeroing" is. Is it that the SL avatars
> are standing in a T (arms spread wide), but the motion files are
> relative to some other initial position?
> - Is this right? It seems odd that SL would distribute files that
> don't "match" each other.
> - I don't have Poser handy. (I'm cheap.) What other options do I
> have?
> - Is it possible to create an .obj file that matches the
> extensive collection of SL .bvh files, or am I misunderstanding the
> issue? Would someone like to post one or two .obj files that match
> the SL .bvh?
>
> I did try just using the SL .obj/.bvh files as is, but ran into
> several problems which may or may not be related to my confusion
> about the above.
>
> 1. Using SL_avatar_OBJ.obj with avatar_stand.bvh and
> avatar_walk.bvh, I get a null materialsDictionary within
> TMeshMotion>>preRenderFaceGroup:currentFrame:using: and also in
> #renderFaceGroup:currentFrame:using:.
>
> - Are these bugs, or is SL_avatar_OBJ.obj not "the right kind"
> of .obj file? (What is the right kind? How do I tell?)
>
> "Fixing" these left the body exploded when brought into Croquet.
>
> 2. Using SL_Male.obj instead, I run into (1) and also that
> TMeshMotion>>createTVertexArrays has a guard for null faceGroups,
> but not for null vertexIndices within a faceGroup.
>
> - Same question: bug or pilot error?
>
> "Fixing" this left the figure imported, but standing in a T and
> without any walking motion.
>
>