Re: pfEarthSky with off-axis viewing frustra
Gan Wang (rock.csd.sgi.com!sgi.com!uunet.uu.net!ds9!cavalier!gan)
Mon, 14 Apr 1997 12:51:52 -0700
On Apr 11, 11:58am, Christopher R Volpe wrote:
> Subject: Re: pfEarthSky with off-axis viewing frustra
> Gan Wang wrote:
> >
> > On Apr 9, 1:14pm, Christopher R Volpe wrote:
> > > Subject: pfEarthSky with off-axis viewing frustra
> > > The man page for pfEarthSky says it doesn't work with off-axis viewing
> > > frusta, but that is what I need. Can someone tell me what the nature of
> > > the problem is and if there is a workaround?
> > >
> >
> > PfEarthSky only works with "Z-up" coordinates. So you have make sure that
your
> > rendering coordinates have Z-up. If you scene graph is built in a non-Z-up
> > coordinate system, you have to take care of the coordinate transformation
> > (rotation and maybe translation if the two origin do not coincide) from
your
> > coordinates to the rendering coordinates. This can be done by adding a DCS
on
> > the top of your scene graph. Hope this help.
>
> The database I'm using has Z being in the up direction. As for
> "rendering coordinates" having Z-up, I suspect this means (correct me if
> I'm wrong), that the plane of my physical display device must be
> vertical, and the view-up within that plane must also be vertical. I.e.,
> I can have a screen straight ahead of me, and another one off to the
> left or right, angled toward me or not, and I can have a screen above me
> that's not angled down towards me, but if I have a screen above me
> that's angled down towards me, I'm in trouble because the "up direction"
> in viewing coordinates is not parallel to the "up direction" in world
> coordinates. In fact, this would cause me trouble even if the frustum
> for that screen were completely on-axis (i.e. even if the vector from my
> eyepoint to the center of this screen up above me were perpendicular to
> the screen). Furthermore, if the display were "rolled" rather than
> "pitched", I'd still have a problem because even though the plane of the
> display device is vertical, the view up vector is not. So the disclaimer
> about off-axis viewing frusta is incorrect and should be a disclaimer
> about viewing planes with pitch and roll offsets.
>
By rendering coordinates, I mean the coordinates in which the channel eyepoints
are calculated and the database is loaded. pfEarthSky is rendered in this
coordinate and is always render with its horizon line perpendicular to the z
axis. However, the geometry in the database can be computed in another
coordinate system (i.e., geocentric) and that's when the coordinate
transformation is needed so that its horizon will be parallel to that of esky.
I don't think (I could be wrong) that a channel offset (it sounds like what
you do) will affect anything because that is only a view offset and doesn't
change the coordinates.
PfEarthSky would still not work well if your eye point is high up in the sky,
then you would need a true geocentric earth sky, something you have to do
yourself.
Hope it helps.
Cheers,
Gan
> Is this a correct description of the situation?
>
> thanks,
> Chris
> --
>
> Chris Volpe Phone: (518) 387-7766
> GE Corporate R&D Fax: (518) 387-6560
> PO Box 8 Email: volpecr++at++crd.ge.com
> Schenectady, NY 12301 Web: http://www.crd.ge.com/~volpecr
>-- End of excerpt from Christopher R Volpe
--
Gan Wang
Cambridge Research Associates Voice: 703-790-0505 ext.7210
1430 Spring Hill Road, Suite 200 Fax: 703-790-0370
McLean, VA 22102 E-mail: gan++at++cambridge.com
=======================================================================
List Archives, FAQ, FTP: http://www.sgi.com/Technology/Performer/
Submissions: info-performer++at++sgi.com
Admin. requests: info-performer-request++at++sgi.com
This archive was generated by hypermail 2.0b2
on Mon Aug 10 1998 - 17:55:03 PDT