[Leica] Noctilux DOF Film vs M8

[henningw at archiphoto.com (Henning Wulff)]

>Steve: It has to do with the range of the RF roller adjustment that 
>will result in acceptable focus with each lens.  The lenses are 
>calibrated to a standard, like the body.  But in the real world, 
>things may not be quite right, and the M8 is less forgiving than 
>with film--in other words, the tolerances before things go awry are 
>less.
>
>Note: The diagrams below will only look correct if you are using a 
>Monospace font like Courier to view your email.
>
>Think of a yardstick and a 1-foot ruler placed on a table, with the 
>ruler below the yardstick. Now imagine that the yardstick is the 
>adjustment range that work with the 50/2 Summicron, and the ruler is 
>the range that will work for the Noctilux at f/1. I've pictured it 
>below. The asterisk represents where the RF roller is adjusted. Here 
>is how your wide-open Noct and M8 may been set, before you adjusted 
>the camera:
>
>
>(Fig. 1)
>|SummicronSummicronSummicron|
>        |Noctilux|
>      *
>
>
>As adjusted above, the camera will focus properly with the 
>Summicron, but not with the wide-open Noct.  However, we could 
>adjust it so that the focus is OK with the Noct, and still be OK 
>with the Summicron. Below is what you probably did:
>
>
>(Fig. 2.)
>|SummicronSummicronSummicron|
>        |Noctilux|
>            *
>
>
>The problem with the Noct is that if you stop it down 2 or 3 stops, 
>you get something like this:
>
>
>(Fig. 3)
>|SummicronSummicronSummicron|
>                                       |NoctiluxNoct|
>            *
>
>
>Now there is no way to get the Noct exactly right without messing up 
>your other lenses like the Summicron.
>
>The Summicron also focus shifts a little bit when you stop down 2 or 
>3 stops, but nowhere near as much as the Noct.  So you might get the 
>typical M8 "all depth of field in back of the point of focus," but 
>things are still usable with the Summicron. Not with the Noct.
>
>
>(Fig. 4)
>           |SummicronSummicronSummicronSummicron|
>                                       |NoctiluxNoct|
>            *
>
>
>
>On film, the shift is less than with either lens.  The problem still 
>exists, but it is within tolerance for the Summicron, and still 
>visible (though less) for the Noct:
>
>
>(Fig. 5)
>    |SummicronSummicronSummicronSummicron|
>                               |NoctiluxNoct|
>                 *
>
>
>Does this make sense?  In real life, the initial position of the RF 
>roller adjustment may vary, (asterisk) as may the calibration or the 
>individual lenses (sideways position of the "ruler" or "yardstick."
>
>The point is that there is no absolute "perfect" point of focus. 
>There is only a range of acceptable tolerances.  With the M8, those 
>tolerances are noticably smaller.  It also appears that in order to 
>get some lenses to focus correctly wide-open, the RF and lens are 
>adjusted so that most of the DOF is behind the point of focus 
>(Summicron in Fig. 4), rather than the classic 1/3:2/3 distribution 
>most of us learned in the film era (Fig. 5).
>
>I'm sure the folks at Leica knew all this for a long time. It's just 
>that in the film era, most of us didn't notice.  The M8, with its 
>instant feedback and stricter requirements, opened our eyes.  The 
>confusion exists because the above is not intuitive--most of us 
>thing as focus as being either correct or not, and we don't think 
>about focus shift.  The situation has been made worse by quality 
>control problems, used lenses that were out of adjustment, and 
>perhaps by Leica initially adjusting lenses to a standard that was 
>viable for film but insufficient for the M8.
>
>I don't understand if or how focus shift also varies with respect to 
>subject distance. No one I've ever asked seems to understand it, 
>either. Sometimes DOF seems to cover focus shift at longer 
>distances. Some people have observed that focus shift is always the 
>same amount of offset "twist" from what the RF tells us, so it may 
>be in proportion to the subject distance, ie. linear with respect to 
>the focus cam.  This may break down at sufficiently large stops or 
>long focal lengths.

Focus shifts vary with distance as well as with colours; the latter 
is really only of interest when taking pictures under incandescent 
light vs. daylight. I have noted that some lenses that need only a 
slight tweaking at close distances for focus shifts need a lot more 
at 20' distances.

The problems is that in practical terms, for photographic lenses that 
have to be used at various reproduction ratios and under various 
aperture and lighting cvonditions, preferably at large apertures 
perfection is still a distant dream at any price point.

To get really high performance, say for electronic circuit 
production, lenses are required to only perform at one wavelength, at 
one reproduction ration and one f-stop. And still they aren't 
perfect, even though they might cost $250,000 or more. And they're 
huge.

With all the variables that camera lenses have to deal with plus the 
constraints of size and cost, it's amazing what they can do.

If you get apo correction, you come closer to eliminating the colour 
issues, but the 'apo' correction also changes over focussing distance 
and f-stop, so floating elements at least provide some help. That's 
why the Apo 75/2 is one of the best lenses.

The main problem is that most cameras are not perfectly adjusted when 
they leave the factory. It may not seem like it, but production costs 
and eventually price to the consumer are important even to Leica, and 
adjusting the rangefinder better by spending another $250 worth of 
factory time/facilities on it, relating to maybe another $500 in the 
purchase price might not be acceptable. So it's adjusted to a 98% 
standard instead of a 99% standard. Considering how demanding the M8 
is of excellent adjustment, maybe it would have been a reasonable 
decision to up the standard.

Personally, I would have liked it if the rangefinder had been 
constructed to be easier to fully adjust by the end user. Yes, there 
would be some seriously messed up rangefinder and complaints, but I 
believe in some personal responsibility. Nikon (or is it Canon?) now 
allows this individual adjustment through electronics on it's high 
end bodies; I wish Leica allowed it through reasonably accessible 
mechanics, and without expensive specialized tools.

As it is, my 8 is slightly mis-adjusted, and I will send it in for 
adjustment for my most critical lenses. Meanwhile, I know how to 
focus to get sharp pictures by 'misfocussing' slightly at times. This 
has become second nature.

With my film M's there are some lenses that also require this, but 
the knowledge was a lot harder to gather, so I don't know my lenses 
quite as well on those bodies.


>--Peter
>
>Steve Barbour wrote:
>
>>  a wonderful review and summary Peter...
>
>>  what I really don't understand is the following two points and their
>>  (apparant) incongruity...
>>  re fixing "back focus"....
>
>>  we either change the set screw in the body with an Allen wrench, OR we
>>  send our lenses somewhere, WITHOUT the body, to be adjusted...
>
>>  if you can fix the problem by changing something in the body, how can
>>  fixing the lenses with NO regard to the settings/status of the body
>>  possibly work?
>
>
>
>_______________________________________________
>Leica Users Group.
>See http://leica-users.org/mailman/listinfo/lug for more information

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