home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.sys.sgi
- Path: sparky!uunet!gumby!destroyer!ncar!virga.rap.ucar.edu!bill
- From: bill@virga.rap.ucar.edu (Bill Myers)
- Subject: z-buffering question
- Message-ID: <1992Aug14.175702.22720@ncar.ucar.edu>
- Sender: news@ncar.ucar.edu (USENET Maintenance)
- Reply-To: bill@virga.rap.ucar.edu (Bill Myers)
- Organization: Research Applications Program/NCAR, Boulder, CO
- Distribution: usa
- Date: Fri, 14 Aug 1992 17:57:02 GMT
- Lines: 30
-
- I have a question concerning z-buffering on an Indigo Elan.
-
- I am involved in a 3-d application in which I am trying to draw
- a line just "above" a polygon. I expect that I should be able to
- place the line just a few z-buffer notchs above the polygon and be
- able to see it overlaid on the polygon. So if I calculate what I believe
- to be the granularity in the z-buffer in world coordinates by
- (far - near)/ (2^#zbits), I should be able to translate the line "up"
- by a few times that amount and see that the line has passed the default
- z-buffer test. How one defines "up" is debatable, but the result I see
- seems to make that irrelevant -- Instead of having to be just a few
- notchs up, it seems to need to be many thousand notchs above to appear
- consistently.
-
-
- I wonder if I am really getting 24 bits of z, or just 12 as performance
- might suggest. That is, is my z running from 0x007fffff down to 0xff800000
- or only down to 0. I am using glcompat() and lsetdepth(), so I would
- expect to see the full range used.
-
- Other notes, I am double buffering, but don't expect that this should halve
- my effective z-buffer size. Also, I have gone from using a perspective
- projection to an orthographic projection to eliminate a level of complication.
-
- Also, the man page for lsetdepth suggests that there should be 32 bits of
- z-buffer available though getgdesc() seems to indicate otherwise.
-
- If anyone has any ideas, I would appreciate the input.
- Bill
-
-