If user-created meshes had skeletons, doesn't that mean that llRequestPermissions(lllGetKey(),PERMISSION_ANIMATION); would allow people to use BVH to animate a mesh ?

~ Marv.

Dahlia Trimble wrote:
There must be some support for skeleton or morph animation of meshes in the viewer as the avatars use those techniques. I have no idea if they could be applied to other meshes.

On Thu, Mar 19, 2009 at 5:54 PM, SignpostMarv Martin <[email protected] <mailto:[email protected]>> wrote:

    Didn't I read something about the possibility for sculpties to be
    animated/manipulated with shader language code ?

    How would arbitrary meshes be animated/manipulated ?


    ~ Marv.

    Dahlia Trimble wrote:

        One advantage sculpts have is they require little or no
        modifications to the SL asset storage/distribution model. I'm
        not sure how other mesh storage methods would fit into this
        model, but I suspect some modifications would be required.

        One problem I have with sculpties is the lack of precision for
        the vertex elements as a result of quantifying them to 8 bit
        integers to convert them to a color. Another is the need for
        square or rectangular arrays of vertices which makes some
        techniques of free form mesh modeling difficult or impossible
        as arbitrary polygons cannot be added or deleted.

        Still, I am impressed with the cleverness of sculpted prims
        and although I do share your desire for more flexible means of
        designing meshes for SL, I am grateful that sculpties exist
        and are being improved upon.

        On Thu, Mar 19, 2009 at 4:53 PM, Dale Mahalko
        <[email protected] <mailto:[email protected]>
        <mailto:[email protected] <mailto:[email protected]>>> wrote:

           It occurred to me recently that LL's reasoning for not
        supporting
           raw 3D meshes as a prim type is basically invalid, what
        with how
           SL functions right now.
               The whole premise behind sculpted primitives was the
        idea that
           wavelet/JPEG lossy compression could be used to store "organic"
           shapes that don't have a very definite shape or form. The lossy
           wavelet compression of the RGB map allows the data size to stay
           very small on the asset server, and that this could not work if
           applied to storing and rendering mesh data.
               However, that whole argument was thrown out the window
        after
           people kept demanding precision in sculpted shapes, and
        some time
           ago someone at LL said "Okay, FINE, we will allow the option of
           lossless compression for RGB sculpt-map uploads" and
        support for
           this lossless uploading has been added into the viewer.
                       So, uh, if lossless compression of sculpted
        primitives is
           permitted, doesn't that mean that there could equally be
        support
           for lossless compressed 3D mesh uploads, like traditional 3D
           editors use?
               There apparently isn't really any reason to prevent
        mesh uploads
           anymore, at least as long as the meshes are very small and
        simple,
           and the largest uploaded mesh byte size is equal to or smaller
           than the most randomized, incompressible compressed sculpt-map.
                       Allowing use of small meshes would be more
        efficient for the 3D
           renderer anyway. Since sculpts have a fixed number of
        triangles,
           if your model does not need the triangles for detail then they
           need to be "bunched up" into a pile to get them out of the way,
           possibly tucked inside the model so they don't show. But these
           unnecessary bunched up triangles are still being processed
        by the
           3D renderer whether or not they are needed.
               LL's own sample sculpt maps demonstrate this problem,
        like with
           the S-shaped sculpt that has a ton of prims scrunched up on
           it, and which could drop 50% or more of the triangle faces
        on it
           and still look good.
               So sculpted prims actually add to 3D scene rendering
        lag and
           contribute to a lower framerate and poor performance
        because they
           do not and cannot "turn off" unneeded triangle faces, vs a mesh
           which just simply would not include any more detail
        triangles than
           the mesh actually requires.
                       But there seems to be major LL heel-dragging
        against implementing
           actual 3D mesh support and making models a heck of a lot
        easier to
           build in external editors, vs the arcanary of sculpts that very
           few people understand or can use effectively to reduce prim
        usage.
               So, I am not expecting this post will change anything.
        But oh
           well, I thought I would bring up this small detail for
        discussion.
           The sldev list has been pretty quiet lately anyway.. ;-)
               - Scalar Tardis / Dale Mahalko
_______________________________________________
           Policies and (un)subscribe information available here:
           http://wiki.secondlife.com/wiki/SLDev
           Please read the policies before posting to keep unmoderated
           posting privileges


        ------------------------------------------------------------------------



        _______________________________________________
        Policies and (un)subscribe information available here:
        http://wiki.secondlife.com/wiki/SLDev
        Please read the policies before posting to keep unmoderated
        posting privileges


Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

_______________________________________________
Policies and (un)subscribe information available here:
http://wiki.secondlife.com/wiki/SLDev
Please read the policies before posting to keep unmoderated posting privileges

Reply via email to