Wouldn't it be simpler to explain it as something like "round off
error"? (perhaps not technically accurate but it should be easy for most
people to understand).

Imagine "compressing" doubles to floats, the fine precision is lost but
the rough number is the same (and uses half the storage).

The important aspect is that a vertex is not lost but it's location may
"drift".

- John Wright
Starfire Research

Mark Hood wrote:
>
> > Date:         Mon, 11 Dec 2000 14:56:08 -0500
> > From: Joe Boccanfuso <[EMAIL PROTECTED]>
> >
> > I am particularly concerned with the words "... with very little loss in
> > object quality".  My concern/question is as follows.  Does this refer to
> > (possibly) losing the actual co-ordinate (i.e. x, y, z) and color data or
> > does this refer to losing Appearance attributes above and beyond color
(e.g.
> > Polygon fill mode, Texturing, Transparency, etc.)?  My hope is that the
> > actual geometric/co-ordinate data is NOT lost.  What data may be lost?
>
> Geometry compression is lossy.  Positions, colors, normals (and in Level
II,
> texture coordinates) are quantized.
>
> The intent is to effectively produce perceptually lossless compression, as
in
> JPEG and MP3.  Quality is controlled by setting quantization levels (up to
16
> bits per component for positions, colors, and textures, and up to a total
of 18
> bits for normals).
>
> Although it works well for many applications, it's not appropriate for
those
> that must retain the exact bit-for-bit representation after decompression.
> You can get an idea of the quality by playing around with the example
programs
> provided for compressing .obj files.
>
> -- Mark Hood
>
> ===========================================================================
> 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