Followup on animation of VRML
Greetings ...
1 ) Is the Mpeg4 standard suited to your task ?
If not, how is it lacking ?
2 ) Did these work on your system ?
http://ligwww.epfl.ch/~babski/StandardBody/mpeg4/
http://www.frontiernet.net/~imaging/HAnimPoser.html
3 ) Having equations that approximate the
joint angles has advantages :
- it's compact
- it's simpler for the program
- it reflects the analog nature of the motion
- it smoothes & verifies the data
- equations imply a higher level understanding ...
In the 3d imaging world, in an analogous
fashion, curvy 3d objects are approximated
with 2nd degrees equations such that each
patch blends smoothly into the next ( Nurbs ).
This is a bit frivolous, and grossly inaccurate :
http://www.frontiernet.net/~imaging/character_animation.html
but with a few guesses and a few sine functions I have a walking motion.
4 )
> ... told me it was far too big to use over the net.
>
> ... standard data files ... I guess this might need
> some sort of CGI interface to get the data from a URL.
Java can read streams of data with no real limit,
and reads regardless of its origin ( excepting
security concerns ), CGI seams unnecessary ...
HAnimPoser accepts data local and remote.
Is the issue slow modems ?
. . . .
I'll look at the animation data you mentioned.
The H-Anim mailing list archive has details on what people are doing :
http://www.vrml.org/WorkingGroups/h-anim/hypermail/1999/date.html
Ballreich is alive and well and is either in hiding or is very busy.
Look for a separate e-mail.
-- Paul
------------------------------------------------------------------------
"Imaging the Imagined : Modeling with Math & a Keyboard"
http://www.frontiernet.net/~imaging/contents.html