P3D Re: the value of toe-in

[Andrew Woods <andrew@xxxxxxxxxxxxxxxxxxxxx>]

George Themelis wrote:
> Dave Spacey wrote:
> >My method was to zoom in close, and use the status bar of my fave graphics
> >editor to read the pixel positions directly.  I figured that keystoning
> >would show up most where a foreground object approached the four corners.
> 
> I would say the "two edges", not four corners.  Keystone distortion has 
> an axial symmetry:  It is zero in a vertical axis that passes from the 
> center of the image and it increases as we move away from this axis.  This 
> makes the two edges the worse possible places for distortion.  Masking down
> (vertically) a keystoned pair will help make things better.

Dave is right - the four corners are the best place to find keystone
distortion.  Keystone distortion is most easily identified by the
vertical parallax it introduces to the image.  Keystone distortion also
introduces horizontal parallaxes but it's difficult to identify this in
the presence of the stereoscopic horizontal parallaxes.

The nature of the vertical parallax introduced by Keystone distortion
is symetrical in both the vertical axis (passing through the centre of the
image) and the horizontal axis (passing through the centre fo the image).

> Perhaps Gabriel can make a good ASCII drawing?

There's no need for an ASCII version.  There's a perfectly good GIF
image on my web site which illustrates Keystone distortion very clearly.

  http://info.curtin.edu.au/~iwoodsa/keyston.gif

which is from the paper "Image Distortions in Stereoscopic Video Systems"
  http://info.curtin.edu.au/~iwoodsa/spie93pa.html

Andrew Woods.              http://info.curtin.edu.au/~iwoodsa


------------------------------