Zbuffering retraction

[karron at MCIRPS2.MED.NYU.EDU]

>
>My feeling is that sgi could use a floaing point Zbuffer, as the
>dynamic range of a 16 bit z short int z buffer is to small to
>be of real use in many real time applications.
>

Please note that the zbuffering is a hardware dependent value generally > 16
bits, not 16 bit.

I was in error.

My experience with Zbuffering and Depthcuing/Shading is that in virtual
reality type modeling, where you don't really know how far or near away your
eye is from the object, you are forced to open up the Z clipping ranges to the
ranges that you do not have a reasonable value for Hardware depthcueing.

Your entire object can lie in a Z band of one count if your near is too
close or your far is too far. There is not enough dynamic range in a
int or long value. A double floating point option would seem more appropriate.
However, then you would have to do Zclipping and Zinterpolation with
floating point numbers, and all the overhead that that would entail.



+-----------------------------------------------------------------------------+
| karron at nyu.edu                          Dan Karron                          |
| . . . . . . . . . . . . . .             New York University Medical Center  |
| 560 First Avenue           \ \    Pager <1> (212) 397 9330                  |
| New York, New York 10016    \**\        <2> 10896   <3> <your-number-here>  |
| (212) 340 5210               \**\__________________________________________ |
| Please Note :Soon to move to dan at karron.med.nyu.edu 128.122.135.3  (Mid Oct)|
+-----------------------------------------------------------------------------+