P3D Re: Parallax tech, deviation

[abram klooswyk <abram.klooswyk@xxxxxx>]

Larry Berlin (P3D Digest 3623, 30 Nov 1999) wrote on
parallax in stereo, also touching upon deviation.

Gregory J. Wageman correctly answered (same ref.): (...)
>Parallax is an angular measurement, and can be expressed in
>any convenient angular unit (degrees, radians, etc.).(...)
and:
>Parallax angle becomes a distance in stereophotography
>when the two perspectives are projected onto the film plane.
>We can then measure the apparent shift of nearby objects
>with respect to the distant background.

Since "parallax" is often also used in the sense of the
linear shift on the stereopicture, it might be good to
stress that this should be avoided.

Chris Jones (1 Dec 1999 ):
>I think astronomers just liked drawing right-angled triangles ;)

Koo Ferwerda had a degree in astronomy and wrote a thesis on
an astronomical subject, it was not entirely on right-angled
triabgles :-). Indeed, parallax is
> not specific to astronomy
but it does originate from astronomers, and it seems good to
stick to their definition.

Ferwerda again points out in his book "The World of 3-D"
(see http://www.stereoscopy.com/3d-books) that the parallax
of a point is the angle, subtended at that point, of the "visual
lines" from two different (viewing) positions.

"Deviation" is the linear shift Greg mentions, by definition it is
the difference of the separation of a homologous pair of
points on the stereophotograph and the separation of a pair
of very distant or "infinity" homologues.

"Separation" again is the preferred word for the distance
between left and right "items", to avoid confusion with
distances in space, e.g. from the stereocamera to an object.

Since we have these three dimensions, it is convenient to
have accurate terms to avoid getting lost in space...

Abram Klooswyk