T3D Re: stereo math

[john bercovitz <bercov@xxxxxxxxxxx>]

>>And is that why the 1.2 mm is actually f/30?<

>Not sure what you mean here.  The ISO standard specifies lenses of 35mm
>focal length, and 35/30=1.166666...

And 1.166.. rounds off to 1.2 which is plenty accurate for our purposes.
What we're looking at is the maximum on-film deviation approach to keeping 
stereo depth within viewable bounds.  For 35 mm cameras, the max dev is 
about 1.2 mm.  For medium format, it is about 2.7 mm.  For large format, it 
is about 4.5 mm.  All these figures are 1/30th of the focal length of the 
taking lens.

If you're unfamiliar with the on-film deviation approach, I would suggest
reviewing Bob Mannle's posts on P3D.  He spent a lot of time devising good
explanations for the method he uses in his vocation.

Bob's basic approach was to lay the left transparency on top of the right 
transparency and superimpose the pair of homologues representing the farthest 
point in the scene and then measure the distance between the homologues 
representing the nearest point in the scene.  This distance is the on-film 
deviation and should not exceed the figures given above for easy viewability.

Since the allowable on-film deviation is a certain fraction of the focal 
length, it really represents an angle, in this case about 2 degrees.  That 
seems to be the amount we can easily tolerate without the scene coming apart.

Of course some scenes are easier to fuse than others and so the f/30 rule
for on-film deviation is really just a safe number.

John B


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