T3D Re: A Dull Double

[john bercovitz <bercov@xxxxxxxxxxx>]

> Measuring subject distances from the film plane, instead of lens
> center (or "nodal point") would seem to reduce a' to zero, making
> triangles T1 and T3 non-existant and reducing deviations on the film
> plane to zero.  What have I done to myself?  Is this a special case 
> or what? ;-) Obviously I am twisted in some fallacy resulting from 
> my profound mathematical ignorance.

Sorry for the delay - it took me a little while to find the article to 
which you refer.  Frankly, I'm unsure how to answer your question so 
let me flail about for a while and you can tell me when I'm getting 
warmer or colder.  8-)  

The reason camera lens manufacturers scale their lenses to measure 
from the film plane is that it's a fixed point on the camera body, 
not a point which is fixed on the lens and so moves with the lens 
as the lens is focussed.  Also, if the lens is a zoom, the point 
would move relative to the lens as you zoomed.

To answer your question, I think you probably need to think of the
lenses in the diagram as pinholes and then try your argument again.
I think you'll see you need to run rays through the pinhole or they
won't travel from object to film.  So picking a point on the film as
the origin won't work.

John B


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