Accuracy of M8's rangefinder system

[diogenis]

Advertisement (gone after registration)

Ok, as promised, I continue from where I was left off:

Fortunately, Mr. Putts has provided us with a very nice article regarding rangefinders:

Accuracy is some 1.5% up to 10m and raises after that, provided you aim @ thin lines (vernier acuity). Therefore, provided that lens used are accurately set (no focusing errors), all that's needed is actually the correlation of %error with horizontal movement of the lens for focusing, and how this is affected for different types of lenses. Unfortunately this correlation is not given by this excellent article from Putts.

 

I am very happy with how this simple optic system works, and I am almost certain that 50 years of heritage guarantee that the system works great with lenses upto 90mm as Leica suggests, but I would really like to know what it actually means to operate within this error and how this is related to output.

 

I will begin a new post about this to avoid this thread's hijacking though, so we can continue this conversation...

So here is my suggestion:

How about we try to answer this, it will be a great help to those (like me) that tries rangefinding for first time or are afraid to switch because of this. In the mean time this will help us find real life limits and ways to work a lot faster, instead of endless chasing of overlapping the split image

[jaapv]

Interesting issue. There have been, over the years, many publications on the accuracy of rangefinder focussing vs. SLR focussing. For a synopsis I recommend either the "Leica M-advanced photo school" book by Günther Osterloh, which has tables and graphs, or Jonathan Eastland's "Leica M compendium".

To say anything sensible, we must, of course, compare manual SLR focussing with manual RF focussing. This can be expressed in a simple, linear graph. With a 0.72 viewfinder the RF is magnitudes more accurate at 21 mm and the SLR draws level at just under 135 mm, to go on to be more accurate as focal length increases. With the M8 rangefinder at 0,62 I would put the value where the graphs cross at 90 mm.

Nowadays there are comparisons between manual RF focussing and Autofocus SLR's. Personally, but one would have to work it out, I feel that the RF wins in accuracy, certainly up to and including 50 mm. If the theoretical 90 mm is really as easily obtained as theory suggests (and the same goes for the 75 Summilux), may be doubtful imo. The digital-mechanical coupling is probably more consistent that the human eye-brain-hand coordination though.

Erwin Puts' 1.5 % is clearly more close to the actual plane of focus than the official tolerance that Canon uses which is "within 1/2 DOF" (at the COC of the relevant format).

Much as I hate to shift the blame to the user, the factors of dexterity and eyesight, notably unnoticed astigmatism, are the main contributing factor to the problems that are seen with rangefinder systems. A matte screen is easier on the eye than the angle comparison of a rangefinder patch. The visibility of the DOF helps the eye as well in judging sharpness. I notice, especially with long lenses, that I "walk" the plane of focus through the projected image to get accurate.

Of course, a sensor's exactitude and instant review throw the spotlight on the placement of the plane of focus. It would be interesting to do a scientific comparison between an M8 and a manually focussed 1.3x crop camera. I think I am off with my DMR at least as often as I am with the M8,possibly more often, but I do not keep statistics.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[diogenis]

Nice one Jaapv, and thanks for the info.

I will try to find those books, if still available, as I am curious.

However, let's keep things easy. Let me ask you from your own personal experience, or say Guy's experience, or anyone else (I hardly know you lot guys) that is experienced with that system (RF), If you ever had any real problems, sharp focusing a 90mm lens.

 

As for accuracy, it is quite easy to raise it and cover all those issues Marc already mentioned by implementing a magnifier. But is there really needed?

Sean can help us here too. And he has already mentioned that he is shooting for his tests many shots pre, on and past focus point for his reviews.

 

So what is your personal feeling? A real issue or just something that lives in our minds?

[jaapv]

I'm not the norm - I've been using Leica M's since 1976....

With my ageing eyes, I am most comfortable with the Megaperls 1.15x magnifier on my cameras. I use it up to and including 135 mm, and the Summilux 75, and the Summicron 90 (and the 28 for that matter). When I miss focus, I know I am to blame....

[luigi bertolotti]

Jaap started with RF in '76, me in '75 (but was a Zorky... "student's camera":) , Leica from '79) so my exp is similar : the simple truth is that WHEN YOU GET ACCUSTOMED RF focusing is quick and precise in MOST of the standard situations an amateur gets into: I know i can be blamed for this, but, to speak honestly, lot of times i used my 1,4 lenses at 2-2,8 or 4 just "to be sure on focus" ... as a "dumb rule" I used my 135 f4 at f8 minimum... I enjoied a lot, in past years, an old Summilux 35 1,4... almost never wide-open...

With M8 the problem is growing... sensor and, maybe, manufacturing quality and tolerances on the M8 RF have generated an issue : you speak of using a 90 at f2... to be simple, I'd say thet in such a situation with M8, it's anyway better to get accustomed to focus bracketing... Sean is right... but let's also say that this is one of the many advantages of Digital: during my film life, I think I made focus bracketing no more than 10-12 times.. throw away film was annoying, while who cares of 2 or 3 files that you can remove soon after ?

[rosuna]

All digital cameras have focusing problems. Autofocus reflex cameras can adjust to these errors, by means of software (firmware), but a manual focus camera cannot:

 

The Online Photographer: Pentax K20D Report—Part I: Focus

 

So, at this moment, in 2008, for digital photography, the old comparisons are meaningless.

[diogenis]

Advertisement (gone after registration)

It is counter productive Luigi.

I want a system that is trustworthy enough, that I will take a shot the way I want it.

If I can't have this, then I need to know the limits.

So, for a 90mm @ f2 and low distances it must be very accurate. It even stops somewhere after 7 meters and goes to infinite after that. Given the fact that up to 10m, I get ±1cm accuracy, it feels like very accurate to me..

[jaapv]

I do not focus bracket, except for IR, and I find the 90/2.0AA extremely accurate at all distances. What I do do is put the focus plane in different places for the same shot. Not for bracketing, but for judging the different photographic effects on screen. Luigi is right, digital is ideal for burning "film"to get the best out of our lenses.

[diogenis]

All digital cameras have focusing problems. Autofocus reflex cameras can adjust to these errors, by means of software (firmware), but a manual focus camera cannot:

 

Why can't they not? It's a matter of quality lens and precise manufacturing and alignment with the rangefinder. It's already happened for more than 50 yrs

 

SLRs use a more direct approach, but don't forget that it took many years to perfect and the help of very modern electronics to do what RFs did for half a century...

[jaapv]

I think he means that on autofocus on a digital camera can be adjusted in electronics, whereas a rangefinder is a separate mechanical system, which will alway operate discretely from the sensor and electronics. More fine tuning and more fiddly all in all.I agree with what Mark Norton is always saying, that it is on the limit of what is feasable on an mechanical-electronic interface, but I find, that once set up correctly, it works excellently.

[wstotler]

I want a system that is trustworthy enough, that I will take a shot the way I want it. If I can't have this, then I need to know the limits. So, for a 90mm @ f2 and low distances it must be very accurate. It even stops somewhere after 7 meters and goes to infinite after that. Given the fact that up to 10m, I get ±1cm accuracy, it feels like very accurate to me.

 

As for accuracy, it is quite easy to raise it and cover all those issues Marc already mentioned by implementing a magnifier. But is there really needed?

 

To address the 90mm/f2 question directly? I can hit the 90mm with spot-on focus every time (haven't lost a 90mm photo yet because of unsharp focus) provided I'm using the 1.25x magnifier and concentrating a bit. The 90mm required more concentration and active thought when compared to the 50mm, which is no sweat at all to get spot on with the 1.25x magnifier. 35mm (no magnifier)? Focus is always right where I want it to be.

 

Quick comparison? Focusing on SLR and Rangefinder. . . . It's not a scientific test, but here's what I experienced recently.

 

I jumped repeatedly back and forth between the M8's rangefinder (at f/2) and a Nikon setup (at f/1.8) for about two straight weeks. Both systems were manually focused. I can say that with the M8 I could hit the focus I wanted with little effort. With the Nikon I was "walking" the focus a lot (like Jaapv mentioned) and there were times I was unsure if my focus was "spot on" or not. (Especially in lower light conditions.)

 

I did get good focus from the Nikon setup most of the time, but I always got excellent focus results from the M8.

 

I guess what I confirmed for myself during the split Nikon/M8 use is that the rangefinder patch is *absolutely invaluable* to get certainty in focus. I did miss the certainty the rangefinder patch provides when using the Nikon system.

 

The key issue for me when I was learning the M8's rangefinder focusing system (for the first thousand shots or so) was the time it required me to get focus. It used to take me a long time (4 to 8 seconds, depending) with the RF patch to get focus spot on. At 20k photos now, I can twiddle the focus ring, almost unconsciously and instantly, and nail it under almost all conditions. That was a practice thing.

 

Exception to this and a test of my patience? Shooting my kid on anything moving (carousels, swings, etc.) at night (with only environmental lighting) and at f/2 (to get better speed) is about as hard as it gets in terms of moving targets and focus. (Daytime motion? OK. Can usually get just what I want if I'm really concentrating.) Yes, I know an SLR with autofocus would be better for "action shots" like this. But I haven't "lost" photos I wanted otherwise because I didn't have one.

 

So, there you go. When measure scientifically, there are limits. I'm sure! But, in practice, the M8's rangefinder is very accurate and (in my experience) removes focusing uncertainty in a way the SLR doesn't.

 

Thanks,

Will

[diogenis]

Right, just what I was hoping to hear

So, it is safe to assume, that focusing is easy and accurate with lenses up to and including the 90mm apo, granted with a bit of practice and maybe with help of a magnification aid.

It is not certainly not difficult, and it gives you a rewarding feeling of accomplishment to every photo you shoot. It also operates under some strict rules (135mm difficult, no good for moving targets that move closer to 10-20 meters)

 

On the other hand a very good autofocus dSLR, is precise and fast, It still has its limitations -low light- even though these are being worked and they won't exist in the future (maybe), it can shoot moving targets and all this is done by the xyz processor.

[jaapv]

It also operates under some strict rules (135mm difficult, no good for moving targets that move closer to 10-20 meters)

 

 

Ummm..TE 135/4.0

 

 

On the other hand a very good autofocus dSLR, is precise and fast, It still has its limitations -low light- even though these are being worked and they won't exist in the future (maybe), it can shoot moving targets and all this is done by the xyz processor.

 

Not really, one needs to focus on eye/beak for birds for instance. Very hard with autofocus (as long as they don't offer beak recognition )

[diogenis]

Wow!

And you cursing your ageing eyes ???

Nice photo indeed.

Now how can you tell to that dumb xyz chip to focus the seagull?

But the eye can never fail nice one Jaapv

[dalippe]

To address the 90mm/f2 question directly? I can hit the 90mm with spot-on focus every time (haven't lost a 90mm photo yet because of unsharp focus) provided I'm using the 1.25x magnifier and concentrating a bit. The 90mm required more concentration and active thought when compared to the 50mm, which is no sweat at all to get spot on with the 1.25x magnifier. 35mm (no magnifier)? Focus is always right where I want it to be.

 

Quick comparison? Focusing on SLR and Rangefinder. . . . It's not a scientific test, but here's what I experienced recently.

 

I jumped repeatedly back and forth between the M8's rangefinder (at f/2) and a Nikon setup (at f/1.8) for about two straight weeks. Both systems were manually focused. I can say that with the M8 I could hit the focus I wanted with little effort. With the Nikon I was "walking" the focus a lot (like Jaapv mentioned) and there were times I was unsure if my focus was "spot on" or not. (Especially in lower light conditions.)

 

I did get good focus from the Nikon setup most of the time, but I always got excellent focus results from the M8.

 

I guess what I confirmed for myself during the split Nikon/M8 use is that the rangefinder patch is *absolutely invaluable* to get certainty in focus. I did miss the certainty the rangefinder patch provides when using the Nikon system.

 

The key issue for me when I was learning the M8's rangefinder focusing system (for the first thousand shots or so) was the time it required me to get focus. It used to take me a long time (4 to 8 seconds, depending) with the RF patch to get focus spot on. At 20k photos now, I can twiddle the focus ring, almost unconsciously and instantly, and nail it under almost all conditions. That was a practice thing.

 

Exception to this and a test of my patience? Shooting my kid on anything moving (carousels, swings, etc.) at night (with only environmental lighting) and at f/2 (to get better speed) is about as hard as it gets in terms of moving targets and focus. (Daytime motion? OK. Can usually get just what I want if I'm really concentrating.) Yes, I know an SLR with autofocus would be better for "action shots" like this. But I haven't "lost" photos I wanted otherwise because I didn't have one.

 

So, there you go. When measure scientifically, there are limits. I'm sure! But, in practice, the M8's rangefinder is very accurate and (in my experience) removes focusing uncertainty in a way the SLR doesn't.

 

Thanks,

Will

 

Was your Nikon with a simple matte screen or did it have the focusing aids typical of older manual focus SLRs? Perhaps it is because I learned it when I was younger, but I still find it faster to focus with my old high school Minolta SLR whose screen has two aids: a horizontal bar which causes a "break" in vertical lines when focus is off, and a centered ring whose interior goes from "bumpy" to transparent when focus is achieved.

 

Note that I haven't done any comparisons regarding *accuracy* of focus and I never scrutinized my negatives and slides the way I scrutinize my digital files. I just know that it takes me less time until I *believe* I've achieved focus with the old SLR focusing screens.

 

After returning to manual focus with the M8, I'm interested in trying some manual focus lenses on my DSLR. But I haven't yet tried because I think judging focus on a matte screen is very difficult. If I try this, I'll definitely replace my focusing screen with one that has some sort of aid.

 

David

[jaapv]

The focussing wedge works best for lenses up to 90 mm, from 90 to 180 the central microprism patch; for slow lenses, macro and long lenses only a full matte screen will work comfortably.

[lxlim]

Glad this topic came up.

 

I have been trying out several Leica 1.25X magnifiers but the viewfinder look out of focus to my eyes. I cannot see fine details in the viewfinder, not just the rangefinder patch.

 

Without the magnifier the M8's viewfinder is very bright and clear and easy to focus. I have good focus 95% of the time with my 90mm elmarit-M at f2.8 but with the magnifier I cannot even see details in the viewfinder let alone focus with the rangefinder.

 

My eyesight has been recently fully corrected (even for astigmatism)

 

What could be the problem? Its not just me, I checked it out with two other magnifiers and with two of the sales staff who have close to perfect eyesight.

 

This becomes an issue especially when I am shooting with the 90 for several days in a row. It gets really tiring without some form of magnification. I got the megaperls 1.35X but the diopter adjustment is not stable and is not accurate when I press against it.

 

Thanks to anyone who can help me out here.

[jaapv]

Just get one of the Megaperls magnifiers; they have an inbuilt diopter correction which solves the problem. I marked the 1.35 one with a scratch in the right place but I prefer the 1.15x one.

[dalippe]

The focussing wedge works best for lenses up to 90 mm, from 90 to 180 the central microprism patch, for slow lenses, macro and long lenses only a full matte screen will work comfortably.

 

Hi Jaap,

 

I'm guessing the focusing wedge is the horizontal line I referred to, and the microprism patch is the ring I referred. Thanks for the terminology-- there was no internet when I was shooting in high school so I never learned the proper names for these things ;-)

 

Anyway, I'm curious which you find faster and more accurate for fast lenses below 90mm, which are after all the typical rangefinder lenses, at least for most of us. SLR with aforementioned aids, or rangefinder? I know you often shoot very long lenses with quite impressive results, but certainly that isn't the typical rangefinder application these days.

 

David

[Advertisement]

Hallo,

Look here: Leica M

[jaapv]

I must confess, David, that I even have "misplaced" my universal focussing screen on the R9. I prefer the full matte screen in all circumstances on an SLR. It must be my RF background. I hate obstructions between myself and the photograph. But common knowledge, and common sense, I suppose:(- find the universal screen with central focussing wedge and microprism ring the most versatile for fast (say faster than f 4.0) and short (under 90) lenses. Which means that quite a few modern zoom lenses suffer from blackout on the focussing wedge and even on the microprisms, as they run into 5.6 as max. speed.