2007/5/4, Jean-Sébastien Guay <[EMAIL PROTECTED]>:
Hi Vladimir,
> I've just uploaded preliminary version of osgCal to sourceforge svn.
Great! I was looking forward to this as I was planning a "free-time"
project soon which would need character skinning/animation. Thanks!
Seems you have some makefiles to compile under linux and possibly
cygwin; did you try compiling this under Windows?
Yes. It actually was developed under windows. But I use mingw/msys for
compilation. I have some plans to use cmake (as in current OSG). I
personally never used cmake before, but I think I can use OSG for
reference (btw, current makefiles are slightly modified old OSG ones).
Also, this is the new version of the LGPL osgCal right? Not the GPL
one? (maybe another name would be more appropriate, instead of mixing
up osgCal/osgCal2 all the time...)
Yes. It is LGPL version. Another name... Maybe it will be renamed to
osgCal2 or osgCal3d. Don't know what name is better (osgCal2 v0.3.0 --
to many digits :).
> Some preliminary version features:
> * Uses GLSL hardware skinning, yet supporting OSG picking.
> * Can be swiched to fixed function implementation.
> * Supports normal mapped, two-sided & transparent meshes.
> * Uses different shaders (with minimum of instructions) for different
> materials.
> * Calculates deformations only when bone positions are changed.
> * Uses non-skinning shader for fast drawing of non-deformed meshes.
Very nice! Have you tested the speed compared to old versions of
osgCal? I think in particular there was some discussion a few weeks
ago about speed when there were multiple characters onscreen at
once... Has that improved? Do you have comparative numbers?
It's hard to say about comparative numbers. It certainly faster than
previous osgCal when using fixed function implementation. When using
GLSL imlpementation it maybe slower due to fillrate problems (since
per-pixel lighting, normal mapping and transparency can slowdown
fragment rendering). Fillrate problems usually occurs on high
resolution and/or when mesh with complex material occupy large portion
of screen.
As for multiple characters -- these are usually geometry bound
problems (vertex shader). These are handled very well. Our models are
pretty hi-poly and library was optimized for simultaneous handling of
many hi-poly models.
Also remark, that vertex positions calculation is duplicated on CPU.
It used for correct picking and bounding box caclulation. It uses
additional CPU time, but less that for fixed function (normals are
discarded). Also calculated postions are not transferred to video
board (since it calculates them faster). So library has some load on
CPU. But still it is smaller than in software rendering.
_______________________________________________
osg-users mailing list
[email protected]
http://openscenegraph.net/mailman/listinfo/osg-users
http://www.openscenegraph.org/
