home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.graphics:12891 comp.sys.sgi:18016
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!saimiri.primate.wisc.edu!ames!purdue!bu.edu!bass.bu.edu!kiki
- From: kiki@bass.bu.edu (Keith Baccki)
- Newsgroups: comp.graphics,comp.sys.sgi
- Subject: Re: Polhemus Head Tracking / GL problem
- Message-ID: <104769@bu.edu>
- Date: 13 Dec 92 02:19:38 GMT
- References: <1992Dec8.144411.16003@hellgate.utah.edu> <1992Dec9.013147.2709@u.washington.edu>
- Sender: news@bu.edu
- Reply-To: kiki@bass.bu.edu (Keith Baccki)
- Followup-To: comp.graphics
- Organization: Boston University
- Lines: 59
-
- In article <1992Dec9.013147.2709@u.washington.edu> chuckb@stein.u.washington.edu (Charles Bass) writes:
- >I need more information to answer this question however I can
- >give the following advice:
- >
- >1) Don't use the lookat(..) function in GL there are
- >singularities that will cause "interesting" behaviour.
-
- Not sure what you mean by interesting behaviour.
- Might you be able to elaborate? Surely, you don't want
- to restrict the use of lookat() (a very intuitive function
- when combined with window() IMHO). What exactly are you
- doing with lookat(), and what interesting behaviour does
- it produce? Just curious.
-
-
- >2) Be careful on how you set up your near and far clipping
- >planes. This range is mapped into a 16 bit range (I'm pretty
- >sure it's 16 bit). Thus if you do something like setting your
- >near plane to .1 and your far plane to 10,000,000 you will get
- >"interesting" behaviour. You want to set the near and far planes
- >to be as close together as possible to maximize your resolution.
-
- Hmmm, not sure about the bit resolution (would guess it's
- a long, not a short). Whatever. Again, I must ask what you mean by
- interesting behaviour?!?!? Perhaps it's because you're viewing
- volume is squat and it distorts the dimensions of your image?
- (HINT: yes it is, and that's what it does ;) Far plane only sets
- your far clipping plane - i.e. has no effect on the image other
- than clipping beyond that distance. Near plane,however, defines
- the shape of your viewing volume (together with the dimensions of
- your view plane in window()) So your above statement is misleading.
- Any "interesting behaviour" is caused by your view plane dimensions
- being disproportionate (is that a word?) to your near plane distance.
- Think about it:
- . .
- . .
- . |
- .| |far
- eye. |near --> |
- .| |
- | |
- . .
- . .
- .
- .
-
- Your view volume is defined by near, far, and dimensions of your
- view plane. Note that the viewplane dimension is ~2 X distance to
- the near plane => big spread in perspective projection. Consider
- putting the near plane twice as far as it is from the eye => the
- spread of the volume becomes smaller. It's called "perspective",
- it's a wonderful thing. Now, when using zbuffer(), I've been told
- that it's accuracy depends on the far plane's distance - anyone
- care to comment on this?
-
-
-
- keith
-
-