From: dstamp@watserv1.waterloo.edu (Dave Stampe-Psy+Eng)
Subject: Head motion and resolution (Was Re: sci.virtual-worlds)
Date: Thu, 14 Nov 1991 05:55:10 GMT
Message-ID: <1991Nov14.055510.23895@watserv1.waterloo.edu>
Organization: University of Waterloo


jraymond@BBN.COM (Jayson Raymond) writes:

>>Problem with variable-resolution is that it's unnatural if done
>>indisciminantly.  Does the world "decay" when you move your head, 
>>then slowly recover?  Of course not.  You'll see it too, since
>>your eyes can see very well during head movements, due to nystagmus
>>stablization.  
>
>While the world may not actually decay as you turn your head,
>your ability to detect details in moving objects does. This is in fact
>being exploited by HDTV/video compression researchers trying to get the
>most bang (aesthetic appeal) for the byte. Even the Mac and other
>windowing systems recognize that while something in is in motion, you
>don't have to draw the whole thing to clue the user, such as in the acts
>of moving or closing windows. 

Problem is that the loss of resolution in moving objects on video
displays is partly the fault of the display technology.  In the
real, continuous world, the eyes can track objects moving at less
than 20 visual degrees/sec with NO loss of acuity.  But because video
transmits images as a series of "stills", the edges "smudge" on
the retina due to the eye's continuous tracking motion and the
images intermittant motion.

>What Craig describes is an approach I have seen before, where a very
>course image is rendered, and continually refined as time goes on (if I
>recall correctly, a UW professor was working on just such a system for
>medical imaging). It has very powerful potential, and I would caution
>against it's quick dismissal. 

The key point I was making was INDISCRIMINANT.  I know this is a 
useful technique, but the possibilities for abuse are high.  I
believe that caution is important if realism is a goal.  A simple
but stable object is much more real than a detailed object that 
metamorphises when moved.

Of course, if all you're doing is showing an object on a screen,
this is OK as you can sit back and let it stabilize before touching
it again.  But with eyephone-based systems, it's impossible for the
user to hold still enough to prevent near-constant updates.  This
would lead to a constant swirl of changes to nearby objects.  Better
to use a relatively constant parameter to determine detail in such a
case, such as distance, lighting level (if the system supports 
radiosity) or even how much of an object is obsured from the viewpoint.

With constant updates in an eyephone system, motion is not liable to
be a good parameter for economization during the rendering phase.  Of
course, the database processes (i.e. BSP generation) might force a
limit on motion rates. 
 
>Again, not only are Level-of-Detail changes useful (simplifying distant
>objects), but so are near ground moving objects, if you can't spare the
>CPU's to render them. Another instance of variable resolution worlds is
>the variable acuity systems by the large simulator companies (GE/E&S),
>which track the eye and give high resolution insets where one is looking
>(though perhaps this has already been covered, before I caught on to
>this thread).

Yes, and I beleive that this is the BEST way to use variable-resolution
objects (especially for eyephone-based systems).  But currently the high 
cost and difficulty of calibration of eye-tracking systems limit them
in usefulness.  Watch for this to change in the future, unless 100 Mpix
displays become common (as discussed a few weks ago).

To summarize, my main reservations in the use of variable-resolution
objects is for eyephone based systems where world-constancy and
realism is an important consideration.  For viewing of CAD images or
"desktop VR" on a monitor (with or without 3D glasses) this is not
as critical, as the sense of "presence" is not required.  Too bad,
really, as it's eyephone-based VR that really needs that extra
rendering boost. 

--------------------------------------------------------------------------
| My life is Hardware,                    |                              | 
| my destiny is Software,                 |         Dave Stampe          |
| my CPU is Wetware...                    |                              | 
| Anybody got a SDB I can borrow?         | dstamp@watserv1.uwaterloo.ca |
__________________________________________________________________________