[Leica] RD-1 users: Light loss with fast lenses?

[henningw at archiphoto.com (Henning Wulff)]

At 8:50 PM -0800 3/10/05, Peter Klein wrote:
>Folks: I've noticed an interesting phenomenon with my E-1 and fast 
>OM Zuiko lenses, and I'm curious if something similar happens with 
>RF lenses on the RD-1.
>
>The widest apertures don't give me as much light on the E-1 as they 
>do on film. My 50/1.4 is more like a 50/1.8 wide open.  My 50/1.8 is 
>more like a 50/2.
>
>Note that I'm not talking about metering error here, nor am I 
>talking about vignetting at the edges.  I'm talking about using my 
>50/1.4 to take a bunch of bracketed shots of a blank wall with 
>manual exposure. If I get a pixel level of 128 near the image center 
>at 1/30 at f/2.8, then I would expect to get the same level at 1/125 
>at f/1.4.  But I don't.  I need to slow the shutter to 1/80 to get 
>the same shade of grey.  This is 2/3 of a stop more exposure than 
>expected.  The same lens shows less than 1/3 stop loss with film.
>
>I'm curious if the RD-1 has a similar effect with f/2 and especially 
>f/1.4 lenses.  Could some of you RD-1 owners who have Summiluxes and 
>Noktons check this out and let us know?
>
>I'm sure all this has to do with sensor angle of acceptance vs. lens 
>characteristics like exit pupil size and angle of the cone of light. 
>There has been some talk of this on digicam forums, with the usual 
>indistinguishable combination of heat and light.  I personally 
>believe that something is indeed going on.  I've seen too many 
>examples of weird DSLR behavior with film lenses at maximum and 
>minimum apertures.
>
>DSLR owners who have used the same fast lens on both a DSLR and film 
>body, feel free to chime in, too.
>
>Thanks!
>--Peter

Good to hear you're feeling better!

Happens on film, too. If you ever try to be systematic on film, and 
measure your exposures with a densitometer, you'll note the same 
thing. The suspicion might come up that the manufacturers have 
inflated their maximum f-stop numbers for the sake of marketing, but 
due to the less than optimal diameters of lens elements, especially 
fast ones (we're neither willing to pay for nor carry lenses with 
optimal sizes) there is a lessening of effective, or T-stop for very 
fast lenses. An f/1.4 lens will still be faster than an f/1.8, but 
neither actually produces twice as much light on the film as the same 
lens stopped down to f/2.8 resp. 3.5. Note that an f/1.4 lens stopped 
down to f/2 will produce a denser neg than an f/2 lens wide open; but 
you are carrying a much larger and more expensive 'f/2' lens in the 
f/1.4 stopped down.

The aperture marked on the lens is a geometric aperture, not a 
transmission aperture, from whence comes the 'T-stop'.
---------------------------------------------------

I should add that older lenses especially, and then rangefinder 
lenses on the R-D1 in particular, will have more fall off at the 
corners as discussed here and at various other places, _and_ will 
also have somewhat lower exposure levels at the center with fast 
lenses due to the edge rays necessarily striking the center more 
obliquely when the lens is used wide open, and thus exhibit some of 
the same 'vignetting' at the center due to non-perpendicular rays.

It might be that the E-1 is more sensitive to this as it was designed 
with 'digitally optimized' lenses in mind, and less compromised for 
the sake of older lenses which did not have their ray bundles as 
perpendicular to the sensor array as the E-1's own lenses.

This isn't very noticeable on my 20D and might not be that noticeable 
on the R-D1 as the latter is definitely intended for 'non-optimized' 
lenses.

-- 
    *            Henning J. Wulff
   /|\      Wulff Photography & Design
  /###\   mailto:henningw@archiphoto.com
  |[ ]|     http://www.archiphoto.com