Path: utzoo!attcan!uunet!lll-winken!uwm.edu!rpi!masscomp!calvin!mark
From: mark@calvin..westford.ccur.com (Mark Thompson)
Newsgroups: comp.graphics
Subject: Re: A-Buffer source (and plotting lots of triangles)?
Message-ID: <31673@masscomp.ccur.com>
Date: 20 Jul 90 15:22:24 GMT
References: <1990Jul16.225553.2623@maths.tcd.ie> <1264@idunno.Princeton.EDU> <104 <31667@masscomp.ccur.com> <10588@odin.corp.sgi.com>
Sender: news@masscomp.ccur.com
Reply-To: mark@calvin.westford.ccur.com (Mark Thompson)
Organization: Concurrent Computer Corp. Westford MA.
Lines: 28

In article <10588@odin.corp.sgi.com> robert@sgi.com writes:
>
>In article <31667@masscomp.ccur.com>, mark@masscomp.ccur.com (Mark
>Thompson) writes:
>|> I'll bet those images weren't identical when rendering piles of
>|> overlaping partially transparent objects!
>
>right, I didn't bother to implement transparency, because I didn't 
>need it, and I didn't know if I would end up using the Z buffer. 
>
>The Z buffer algorithm can't do transparency.  
>I stopped just short of stating (what I thought) was obvious.

Not necessarily true. If you generate an alpha value for each pixel
as well as rgbZ, crude transparency can be implemented still using
a standard Z buffer algorithm. This of course begins to fall apart
when multiple transparent object overlap (because no linked list
of pixel contributions have been maintained). The main reason we used
an A buffer in the system I last worked on was to remove edge artifacts
on abutting antialiased polygons (using alpha lets the background
color seep through the 'cracks'). I'm still surprised that your Z buffer
was slower.
+--------------------------------------------------------------------------+
|  Mark Thompson                                                           |
|  mark@westford.ccur.com                                                  |
|  ...!{decvax,uunet}!masscomp!mark   Designing high performance graphics  |
|  (508)392-2480                      engines today for a better tomorrow. |
+------------------------------------------------------------------------- +