[Leica] Full frame again!

[henningw at archiphoto.com (Henning Wulff)]

I hate to be a spoil sport when interesting ideas 
are floated, but all of those solutions have 
serious drawbacks.

The last, which equates to a scanning back, is 
basically unacceptable due to the inertia of the 
system. Scanning backs are useful for large 
formats where one shot sensors are not 
(economically) available, but for general 
photography we're moving away from them. Also, a 
scanning solution doesn't address the steep angle 
of incidence, except possibly to some degree by 
rotating the sensor strip as it moves across the 
frame. The mechanics of this would be a 
nightmare, and to have this go evenly across the 
sensor plane in 1/250 second and also not 
increase the thickness of the M cameras seems 
extremely unlikely in any practical sense.

The fibre optic solution that Marty Forscher 
used, as far as I know, was only direct 
transmission, not reduction. He made a polaroid 
back for cameras where he couldn't actually fit 
the polaroid material at the true film plane, so 
he 'extended' the film plane far enough behind 
the camera to be able to fit the bulky polaroid 
back. This solution also doesn't really solve the 
angle of incidence problem. The proposed solution 
addresses a problem that isn't there, namely the 
sensor size. FF sensors are available, and prices 
by now are reasonable. The problem is fitting a 
sensor at the image plane of the lenses. Nikon 
had polaroid backs which used a folded optical 
path to expand and extend the image to fill a 
polaroid frame, giving a large enough image to 
actually judge composition. The Forscher backs 
produced 35mm sized images on polaroids, which 
were hard to do anything with.

The intermediate lens solution, as the previous 
one, would make a camera that is significantly 
thicker than an M, which is unacceptable. I would 
not use a camera that is 20-30mm thicker; or I 
would certainly not prefer a FF M camera of that 
size  over the current M8. The M8 is a much more 
workable solution. The intermediate lens solution 
introduces horrendous optical issues, as does the 
fibre optic solution. Again, the angle of 
incidence issue is passed onto another part of 
the camera that can't deal with it either.

The problem with the Leica M lenses and the 
sensor isn't just that the angle of incidence is 
steep, it's that the angle of incidence in the 
corners covers such a wide range of angles, from 
almost 90? for some of the tele lenses to less 
than 30? for some of the wides. That's why 
microlenses over the photosites is a relatively 
good solution as these increase the acceptable 
angle of incidence. A further optimization and 
more precise configuring of these microlenses 
might well be the best solution at this time. 
Software unfortunately cannot really fix things 
completely because in the end, all software can 
do is throw image information away, and in the 
corners with reduced photon impingement and 
variable spectral response dynamic range will 
suffer.



At 4:57 PM -0400 4/1/09, Lawrence Zeitlin wrote:
>Chris' wonderful April Fool joke gave me a few 
>seconds hope that Leica had finally returned to 
>its senses. It is not impossible to make a full 
>frame Leica which can use all the legacy lenses, 
>just difficult. It is perhaps more of a 
>technical tour de force than a financially 
>strapped company can support.
>
>Before the M8 was introduced I had lunch and a 
>few beers with a couple of photographers who, 
>incidentally, were optics gurus at IBM's 
>Yorktown technical center. The topic of a full 
>frame digital Leica came up and a number of 
>ideas were floated to solve the problem of the 
>short back focus between the lenses and the 
>digital sensor. The first idea was offset 
>prismatic micro lenses to turn the angular light 
>rays in a more vertical direction, an idea Leica 
>partially implemented in the M8. But this was 
>not the only solution.
>
>Another idea was to use an intermediate lens to 
>capture the real full frame image of the prime 
>lens and reimage it on a smaller size sensor. 
>This is a common technique in instrument optics. 
>Admittedly the tube length would be longer than 
>desired and the intermediate field lens would 
>have to be of the highest optical quality, but a 
>few million dollars of engineering time should 
>be able to solve both problems and reduce the 
>size of the package to Leica dimensions.
>
>A third solution was to use an optical fiber 
>reducing bundle, full 35 mm frame on one end, 
>sensor size at the other. Marty Forscher used 
>this method 40 years ago to produce 35 mm images 
>from a Hassleblad. The bundle of fibers, each 
>fiber smaller than a pixel, would be compressed 
>on one end to the dimensions of the digital 
>sensor. The compression need not be linear since 
>software could correct any geometric distortion 
>(these guys worked for IBM). Indeed, a long thin 
>mechanically moved sensor could scan the bundle. 
>An electronic shutter slot, so to speak. That's 
>the way film is exposed in a conventional Leica.
>
>The final solution was to keep the current 
>format of the present Leica and move the digital 
>sensor across the frame as in the previous 
>example. If the sensor could move across the 
>frame in 1/30 second, a digital full frame image 
>could be recorded.  This is a miniature version 
>of the old Hassleblad Leaf system. Of course 
>there would be difficulties with slow speed 
>exposures but I'm sure the Elves of Solms could 
>come up with an answer.
>
>Any other ideas?
>
>Larry Z

-- 

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