Re: [Algorithms] I hate mirrors...

[T.M. E. <tm...@ti...>]

Has anyone had any success with this? User clip planes are not properly =
supported on NVIDIA hardware (screws up depth test and is slooow), I'm =
desperately looking for a solution and this sounds like the best by far. =
I've spent some time trying to get it to work but didn't succeed, what =
would I need to do?

Cheers,
Madoc

  ----- Original Message -----=20
  From: John Pollard=20
  To: gda...@li... ; Charles Bloom=20
  Sent: Wednesday, June 26, 2002 2:47 AM
  Subject: Re: [Algorithms] I hate mirrors...


  Ok, you've got me all pumped up.  Gonna go write the code now.. :)

  ----- Original Message -----
  From: "Charles Bloom" <cb...@cb...>
  To: <gda...@li...>
  Sent: Tuesday, June 25, 2002 7:37 PM
  Subject: Re: [Algorithms] I hate mirrors...


  >
  > At 07:20 PM 6/25/2002 -0500, John Pollard wrote:
  > >5. Clear z's to 1 (assuming z space is 0..1).  Render the mirror =
poly, z
  > >writes on, no ztest.  Then, while rendering the mirror scene, you =
set the
  > >zbuffer to do a greater/equal test.  This will only accept objects =
in the
  > >reqion of the mirror, as well as clip to the mirror plane itself.  =
The
  > >problem is that you'd have to turn your objects *inside* out, =
because of
  > >the special ztest, or they wouldn't render properly.  Has anyone =
tried
  > >this?  I haven't yet, but it sounds promising.
  > >
  >
  > Yep, we haven't tried this, but we're seriously considering it; it =
sounds
  > very attractive, and it even works
  > for non-planar mirrors !  It's pretty easy to set up your projection
  matrix
  > to make your Z value come out
  > backwards.
  >
  >
  > ----------------------------------------------------
  > Charles Bloom    cb...@cb...    www.cbloom.com
  >