From: toerting@kentvax.mcs.kent.edu (Tim Oerting)
Subject: Re: Position judgement at distance
Date: Mon, 8 Jul 1991 01:25:25 GMT
Organization: Kent State University


In article <1991Jun27.010103.10247@milton.u.washington.edu> mike@x.co.uk (Mike Moore) writes:
>
>
>My armchair virtual world program has a problem....
>
>I've picked up an object (a pyramid) and now I want to put it down so that
>it rests on top of another object (a cube).  I don't want point-of-view of
>the user to change as the operation is done. 

>I was thinking of using a shadow which appeared underneath the pyramid and
>deformation of the shadow as it travelled over the cube would inform the
>user of the pyramids position relative to the cube.  This implies that the
>light source generating the shadow must always be at a defined point,
>preferably immediately above each object#.

this sounds like the right approach to me.  The only way we can judge position
(depth) in reality is by its position in relation to other objects whose 
distance we know or by things like shadows.  Even more basic is that we see an
object when light reflects off its surface.  But, is this reflection of light 
(from one or many source(s)) enough? I don't think so.. take an object in space
 you see it because of light but there is nothing for the shadow to be cast on..
so you don't get a sense of it real size. You need to provide a differentiatable
(not likely a real word) setting.  If you have a floor or something that the 
cube is setting on then you should be able to use any single point source of
light and project it at the edges of the object to get a projection of the light
on the ground (or elsewhere).

>books which discuss various methods of giving distance and relative
>position feedback to the user?
 Sorry but I'm just winging it here...I have no real knowledge in the area and
have never done any graphics works...these are just some ideas.
 You said pick up the pyramid...so I think a neat way to show distance would be
to show the hand/arm moving outward and changing size (this is the only setup
I could envision in which the point of view doesn't change as the object moves
away from the operator).

>Another point of interest in relation to perspective:
>
>                        Infinity
>                          /\<----/_e
>                         /  \
>                        / a  \
>                       /~~~~~~\            -
>                      /        \           |
>                     /          \          |d
>                    /     b      \         |
>                   /~~~~~~~~~~~~~~\        -
>                  /                \       |c
>                 /                  \      -
>
>                        Observer
>
> o If an object has an actual width of 10 units then the apparent widths
>   a and b at distances c and d depend on the angle e.  What is angle e?
>   Note: This is a genuine question, not an exercise.
 It seems to me that you are misrepresenting the problem with your drawing.
The eyes don't exactly work like this I don't think (feel free to correct me
if I'm wrong here but after reading this I have thought about it for a while).
 I believe that although "mathematically correct" and fairly reasonable your 
picture doesn't capture the true operation of the eyes. The eyes must be focus-
ed on a singular point at all times (by this I mean that the eyes can't be 
focused on the entire 10cm object at one time unless it is far enough away to
be represented by a point source). Therefore you must know the point at which
the eyes are focusing.  By theoretically moving the object far enough away so
that it could be seen as a point you can solve for the angle:
 (If the object is directly in front of observer:)
                            / .5 distance bet. pupils of eye  \
   half angle of e =arctan < ---------------------------------->
                            \ perpendicular distance to object/
 (General case if not directly facing object then you must also know the dist
  right or left of the plane dividing your face and a few other calculations
  are necessary and I didn't solve that one)
  
> o Do the same rules apply for the height of an object?
  I would imagine so ... the eyes see in two cones of vision which I think are
of circular x-section (rather than possibly ellipses).. but again I add that 
this is all just hypothetical speculation and might be shot down by someone
else with more knowledge of vision and of graphics.

>Mike Moore
> Sprechen Sie Deutsch?
  Non! Est ce que tu parles Anglise ? (no flames its been a while since I tried
french)

Tim Oerting  -->toerting@mcs.kent.edu