Chris,

You might want to get more detailed information from Sun.  I have a
section about z-buffer written by Doug (Sun Engineer) on my website at:

http://www.starfireresearch.com/services/java3d/supplementalDocumentation.html

I think I have a decent understanding of how z-buffer works but I won't
go so far as to know exactly how Java 3d -> OpenGL -> video card
communication works.

My understanding was that Java 3D maintains positions with doubles (even
though we mostly work with floats).  However there might be a conversion
(and loss of precision) when sent to the video card.  I don't know if
any of that is relevant to the z-buffer issue though.  From my
conversations with Sun the z-buffer problems stem from the fact that
most video cards use a 16 bit z-buffer and that if your video card does
use a larger z-buffer these issues will go become minor (ie unrelated to
Java 3D).

- John Wright
Starfire Research

ChrisThorne wrote:
>
> John,
>
> Q: what precision of numbers for z depth is passed between the graphics card and 
>Java3D: single precision floats?  The reason I
> ask is that if J3D is restricting itself to
> single precision then there is not enough precision to handle the range Thomas 
>Auinger
> required for his planets while at the same time having adequate precision close in 
>to a
> small satelite or on the surface of the planet.  I am more familiar with this sort 
>of mathematical limitation in a different
> type of problem which it manifests itself in jittery motion in VRML due to VRML 
>being constrained to single
> precision fp for field values.
>
> Chris
>
> John Wright wrote:
>
> > Chris,
> >
> > The precision of the Z-buffer is dependent upon the graphics card /
> > driver not Java 3D. For a typical video card we assume 16 bits of
> > Z-Buffer and the recommendation of Sun for a 3000 ratio.
> >
> > Note: the values that you set for front and back clip are NOT the actual
> > positions where the clipping will occur.  These values are scaled based
> > on other factors of your view platform.
> >
> > - John Wright
> > Starfire Research
> >
> > chris thorne wrote:
> > >
> > (snip)
> > > I do not know where he gets the 3000 from.
> > >
> > > Question to Doug and others:
> > >
> > > Q: what is the precision of the Z-buffer in J3D?
> > > Q: is the precision of the Z-buffer in J3D platform dependent?
> > > Q: what ratio of back/front is appropriate for typical platforms?
> > >
> > > regards,
> > >
> > > Chris
> >
> > ===========================================================================
> > To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
> > of the message "signoff JAVA3D-INTEREST".  For general help, send email to
> > [EMAIL PROTECTED] and include in the body of the message "help".
>
> --
>                   ,--_|\
>                  /  Oz  \
> Chris Thorne    :)_.--._/
> [EMAIL PROTECTED]     v
>
> ===========================================================================
> To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
> of the message "signoff JAVA3D-INTEREST".  For general help, send email to
> [EMAIL PROTECTED] and include in the body of the message "help".

===========================================================================
To unsubscribe, send email to [EMAIL PROTECTED] and include in the body
of the message "signoff JAVA3D-INTEREST".  For general help, send email to
[EMAIL PROTECTED] and include in the body of the message "help".

Reply via email to