The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Obsessions

scott kirkpatrick

Well-known member
I just got a second M8 body (shortly before the recent price rises), and was testing it for focus accuracy before making it finally mine by registering.

With a 75/2.0 installed, on a tripod, shooting down at 30-45 degrees angle onto a sheet of square-ruled paper, it focuses 1.5 cm in front of the desired point at 27 and 37". It's 1 cm in front at 48". Is this the lens or the body? I went to the other body and this lens, focussed on that body focussed 1 cm in front. I switched to a 35/2.0 asph and found that focus was about 1 cm off at these distances on both bodies. Finally, I got out my Olympus E-3 with its 50/2.0 macro lens and found that this one, using single point center-only (there's a "region around" setting as well) AF, it was consistently 0.5 cm in front of the desired distance.

A surprising observation was that with the older M8 body, moving my eye from side to side makes the crossing point in the rangefinder shift from front to back by as much as 2-3 cm. This didn't happen in the newer body.

Do I care? Send them all back to Solms? Definitely not! I've heard of much worse experiences, and these all seem to be within the range of a little focus-bracketing if I really want to do macro work. Don't know what I can do about that @#$% AF, though.

scott
 

Jack

Sr. Administrator
Staff member
My .02 on this subject...

First off, it is near impossible to design any focus system, AF or manual that can consistently focus to less than a 1% +/- error of a target. (That is equal to +/- 1cm at 1 meter.) In fact, the Leica M is the only camera to ever achieve those levels of accuracy with any regularity. (Though it is generally regarded the M RF system is capable of maintaining less than 1% total error, or under +/- 0.5% when properly calibrated...)

Note also, that focus accuracy is dependent on many factors, the primary one is the geometry of "triangulation". Without getting overly technical, accuracy increases with increased RF base-length which is directly related to finder magnification, and RELATIVE (not absolute) accuracy improves as distance increases. This is a complicated way of saying 1) the further away a subject is, the more likely it will be in focus, and 2) the higher magnification your viewfinder is, the more accurate focus will be.

Finally, in "real" shooting situations with the Leica M, it is unlikely you or the subject can remain static enough to keep within that relative error amount. (Note that as subject distance increases, the relative impact of camera and subject movements on focus accuracy decrease exponentially.)

Finally, digital sensors are less forgiving on focus accuracy than film ever was -- the reason for this is two-fold: 1) we can easily inspect a digital at 100% view, while it required a microscope to do so with film, and 2) digital sensors render their image on a 2-D plane while film had some "depth" to it's sensor, mitigating some measure of focus error...

Bottom line is since you are within 1 cm, I would shoot with what you got and forget about focus error --- you'll never see it in a print. If you're in a situation that requires greater accuracy, you'll be photographing a static subject with your M8 on a tripod, so you'll plenty of opportunity to chimp and make sure you nailed focus :)

More on the subject can be found in this technical article by Erwin Putts: http://www.imx.nl/photo/leica/rangefinder_accuracy.html

Cheers,
 

scott kirkpatrick

Well-known member
Jack, I agree that this is not an issue, just wanted to see how much not an issue it would be. Puts' article calculates the theoretical accuracy of a mechanically perfect rangefinder with the M8's dimensions, at a distance of 1m, as ranging from 1 to 6 mm depending on how sharply the two images presented can be aligned. Focusing on a line that passes thru the camera center is nearly ideal since the crossing point moves from top to bottom of the rangefinder rectangle as you pull focus. So in a studio setting, with lots of time and static objects, or for a resolution test, I should get within this window after chimping a few times. Otherwise biasing my error to one side (now that I know which side) should be sufficient.

Have you ever noticed the effect of eye position on apparent best focus?

scott
 

Jack

Sr. Administrator
Staff member
Have you ever noticed the effect of eye position on apparent best focus?

scott
Yes I have --- extreme side to side or up and down can make a difference in the perception of alignment... Interesting you only have it one body. I've noticed it M bodies since my film days and never thought more about it since I tend to keep my eye pretty well centered...

So it sounds like Leica may hove corrected something with the most recent boduies since your newest one does not exhibit this?

Cheers,
 

Peter Klein

New member
Scott, every lens I own of f/2 or faster has detectable focus shift on the M8. Now notice I didn't say "photographically significant" shift, just "detectable." Some lenses do it more than others.

There are two basic ways to deal with it. You can have the lens optimized for wide open, and it will slightly back focus at medium apertures. If it's photographically significant, then try to focus on the farthest thing you want in focus at medium apertures.

The other way: Optimize the lens at medium apertures, and deal with a little front focus at the widest stop or two.

My fastest lenses are already optimum wide open, which is fine, because that's how I use them.

My 35/2 Ver. IV Summicron front focuses a couple of cm close up at f/2. The RF focus point is at the back of the sharp zone at f/2.8. From f/4 on, no adjustment necessary. Since I use this lens more at medium apertures, that's best, and I asked DAG to optimize it this way when it got a CLA.

I never noticed this with film. But I thought the 35/2 'Cron was soft at f/2. Now I realize it was just that the Cron has sufficient shift that it was detectable on film. When I compensate, the lens is a lot better than I thought wide open. I never noticed my other lenses' shift with film. When I got the M8, it was a real eye-opener.

The good news is that our ability to measure now often exceeds photographic significance. If you can't see the focus issues at 1:2 on the screen (50%), then it probably isn't going to matter at print sizes below 11x14.

--Peter
 
Last edited:

scott kirkpatrick

Well-known member
Yes I have --- extreme side to side or up and down can make a difference in the perception of alignment... Interesting you only have it one body. I've noticed it M bodies since my film days and never thought more about it since I tend to keep my eye pretty well centered...
The new one shifts a little at the left and right extreme eye position, but not in the middle. The older one shifts steadily as I move my eye position across. With a little practice it isn't a problem. My M2 doesn't seem to do it at all.

scott
 
Top