>f-number has absolutely nothing to do with beams of
>light, axial or otherwise.  Nor does it have anything
>to do with light transmission, vignetting or filters.

The included ange of the cone of rays converging at a 
point on the focal plane determines the depth  of 
field.  This in turn is determined by the numerical
aperture (as well as the focal length and subject
distance).

>An f-stop is simply the ratio between the focal length
>of the lens and the diameter of the aperture.  Nothing
>more, nothing less.

But life is not that simple.  See below.

>It's not at all unusual for a zoom lens to transmit
>less light than an equivilent fixed lens at the same
>aperture.  All those extra glass surfaces increase
>internal reflection and reduce transmission.  With
>modern lens coatings the effect is much smaller than
>it used to be, but it's still there.

I have always assumed that the most important 
aspect of lens aperture was the following
desideratum:

"When a lens is set to a given aperture or reports
that aperture to the camera, it shall transmit
the same amount of light as a _theoretical lossless_
lens at that (geometrically defined) aperture."

An honest lens designer/marketer would design
his lens to behave this way.  It implies that he would
design his (real-world) lens with enough extra
_geometrical_ aperture to compensate for its 
internal transmission losses.  Thus, if he designed
a lens whose maximum aperture would be advertised
(both to the consumer and to the camera exposure
electronics) as an f/2.8, the geometry of the lens
might have to be f/2.4 to _transmit as much light_
as a theoretical f/2.8.

Sure.  This means that the lens actually exhibits less
depth of field than the theoretical lossless f/2.8 lens,
but it does something that is more important: _it
transmits as much light_.  An unavoidable tradeoff,
and that's the direction I would want it in.

If lens designers went around calling their lenses by the
geometrically defined aperture, and those lenses
reported the same to their camera bodies, how
could the designers of the exposure electronics know
how much transmission loss to factor in?
The information on a given lens would be unavailable; 
it would have to be assumed as some kind of brand 
or industry average.  Is that what they really do?

The only way out, if they really do all trade on
geometrical aperture, would be if the transmission
loss were negligible; I believe, as  you do at least
for zoom lenses, that it is not.

DGW
*
****
*******
***********************************************************
*  For list instructions, including unsubscribe, see:
*    http://www.a1.nl/phomepag/markerink/eos_list.htm
***********************************************************

Reply via email to