🚀 go-pugleaf

dchapman@ncsa.uiuc.edu (Dee A. Chapman) writes:

>I've noticed some "distortion" in viewing 3D graphics objects which
>were rendered using an SGI SkyWriter (GL Library).  For example, if I 
>create two divergent rectangular boxes, and view them on the flat screen
>of the computer console, they will appear to converge or be parallel:
>
>I believe this effect is due to the use of the "perspective" GL
>library call.  (void perspective(Angle fovy,aspect, znear, zfar)) I
>realize that the objects should appear to converge as they get farther
>away from the viewer, however the effect I'm referring to is that
>them seem to converge excessively. I can lessen this "converging" effect 
>>by varying the fovy parameter.  However, this can introduce other 
>inaccuracies. 

The effect is caused by an excesive FOV for your viewing situation.
The FOV csnnot be too wide: the angle should be 2*atan(w/2/d), where d
is the distance from your eye to the monitor, and w is the width of
the monitor screen (or window).  You can often go up to twice this FOV
without adding excessive distortion.  Still, it's restrictive.

>>From a simulation point of view, I'd like to make this graphics 
>representation 
>as close to "real" life representation as possible.


>
>In addition, I've noticed when I view the 3D graphics objects using
>a Flight Helmet head mounted display, the "converging" effect is
>lessened.  I'm thinking that is probably due to the lensing in the

No, this is because the FOV of the HMD matches the FOV parameter in
your software better.  I see this all the time with wide-angle
displays for REND386--it's not a bug, it's a feature! (B-{))


--------------------------------------------------------------------------
| My life is Hardware,                    |         Dave Stampe          | 
| my destiny is Software,                 | dstampe@psych.toronto.edu    |
| my CPU is Wetware...                    | dstampe@sunee.uwaterloo.ca   | 
| Am I a techno-psychologist, or just an engineer dabbling in psychology?|
---------------------------------------------------------------------------