Archived posting to the Leica Users Group, 2006/06/15

[Author Prev] [Author Next] [Thread Prev] [Thread Next] [Author Index] [Topic Index] [Home] [Search]

Subject: [Leica] 24x36mm sensor for the RFDM
From: henningw at archiphoto.com (Henning Wulff)
Date: Thu Jun 15 16:58:42 2006
References: <00d801c690cf$d68c8080$6401a8c0@FrankDell2>

At 4:02 PM -0700 6/15/06, Frank Filippone wrote:
>Maybe I missed something, but I have not read an argument as to the 
>need to make digital lenses bigger for the smaller sensor than
>24x36 film.  There may be an issue that I am not aware of, but 
>marketing issues must be put aside to make sense technically.  The
>availability of some Olympus lens of size X is a marketing issue. 
>Technical details would be helpful. 
>
>Comparing the size of true WA to Retrofocus designs is apples to 
>oranges.  But if this is the comparison issue, then yes, the
>retrofocus design will be bigger and yes it will have higher angle ( 
>less oblique) of attack to the sensor, and therefore "better"
>for digital sensors.  But I already stated this point.  My comment 
>about using R lens designs on a M mount lens.  Makes sense to me.
>
>What I said is that for a smaller target sensor ( digital or film) 
>the lens can be designed smaller because the image circle is
>smaller. This is of course for the same F stop and Fl lens.  It is 
>the same argument for Retrofocus lenses as it is for true WA
>lenses.  Smaller target = smaller optics. Always true given the same 
>lens design.
>
>Why is the Olympus lens larger?  Beats me.  I have no idea why the 
>designers made it that way.  Is there a way we can ask them?
>
>I am open to continued discussion on this subject.
>
>Frank Filippone
>red735i@earthlink.net

Frank,

The issue is to what degree the lenses are 'telecentric'. Think of 
this as a kind of 'Ueber-Retrofocus'.

 From pure geometry, you can deduce that if you want _all_ rays to 
have an angle of incidence of 90 degrees, ie, a fully parallel 
bundle, the exit pupil of the lens has to have a diameter at least as 
large as the diagonal of the sensor. This is the optimal 
configuration, and is what causes the lenses to be huge. This isn't 
possible for 24x36mm sensors in an M body, because the opening in the 
throat of the lens mount, after taking the lens mount into 
consideration is less than 43.3mm. So fully telecentric lenses aren't 
possible under these conditions.

However, if you want to optimize a lens for use with digital sensors, 
you want to get as close to telecentric on the image side as 
possible, and that dictates a large exit pupil and a large rear 
element, and some very large elements inside the lens, and then 
larger elements on the subject side of the lens. That's the direction 
that Olympus has gone, and that is why even with a 4/3 sensor, which 
is a lot smaller than that planned for the D-M, the lenses are so 
large. It comes down to basic geometric and optical constraints, not 
all of which can be overcome through better designs with better 
glass, techniques etc.

Due to these basic physical constraints, the Olympus designers had to 
make the lenses as big as they did. That's partly why they perform as 
well as they do.

If you're not as concerned with some vignetting and inefficient use 
of the sensors, you can make them less telecentric, and just make 
sure that most of the other parameters of good lens design are 
handled well. With that sort of attitude we wind up with lenses like 
the 50, 75 and 90 ASPH lenses. Outstanding optical achievements which 
should display little if any CA or other aberrations that are 
especially annoying to digital sensors. Shorter lenses, such as the 
21 and 24 are not so good, because the steep angle of incidence not 
only creates vignetting and causes inefficiencies in the dynamic 
range of the sensors, but are also more sensitive to CA and caused by 
residual aberrations like coma and astigmatism as only a very slight 
smearing of the light rays will cause different photosites to pick up 
conflicting information. So correction algorithms, and later newer, 
more telecentric (and bigger) designs will overcome these problems. 
The current M lenses are still likely to produce much higher quality 
than lenses like the Canon wideangles, but there are ultimately 
better designs out there.

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

Replies: Reply from red735i at earthlink.net (Frank Filippone) ([Leica] 24x36mm sensor for the RFDM)
In reply to: Message from red735i at earthlink.net (Frank Filippone) ([Leica] 24x36mm sensor for the RFDM)