The limits of the Optical Rangefinder

[scsambrook]

Advertisement (gone after registration)

My thanks to those who read my opening post carefully and made considered comments on it.

 

Let me touch quickly on a number of points raised by those responding, including what I fear are some misapprehensions about what I was trying to say:

 

To jaapv, post #3: I don’t recall mentioning anything about automating or idiot-proofing the rangefinding mechanism – in fact I referred specifically to “precise manual range measurement”.

 

To Steve Ash, post #7: I’m not trying to imply anything – let alone falsely – I asked a question which, in essence, was “Do the problems which were known to exist in military and naval rangefinders also exist in the Leica M rangefinder”.

 

To Xmas, post #8: As I’ve already pointed out above to jaapv, I made no suggestion of autofocus.

 

To lct, post #9: The laser rangefinder replaced the optical rangefinder in military use largely, though not wholly, because it was both more accurate and avoided possible human errors in aligning two images. I wonder – let me repeat “wonder” – if that principle might be a viable way to improve the efficiency of the camera rangefinder.

 

To Paul Verrips, post #10: I bought my first screw Leica in 1960, my first M Leica in 1964, and apart from a spell of poverty have owned one ever since. I have high regard for the M camera and am keen to see it evolve further.

 

To pico, post #12: I am not an engineer or a research or design person but I am perfectly comfortable to be in the company of those who are. I have no need to ‘be careful’ because I’m asking questions and putting forward questions and ideas for discussion. And although you and I may not seek radical innovation, any business such as Leica Camera AG needs – at the very least – to consider the market presence of those who do.

 

To adan, post #17: Thanks for your comments. I did begin by saying that “in general, focusing errors result from . . . “. Indeed there are others such as those you list, but I was writing in the context of the rangefinder itself and its ability to provide a sufficiently accurate range reading. As you say, the demands are greater in some circumstances than others. But calibration problems are not the same as a broken instrument that can’t work, they are to do with an instrument that appears to be working fine but is actually giving unreliable results.

 

I do not have any statistically worthwhile data on the reliability of the Leica M rangefinder, but it indisputable that errors in focusing accuracy do sometimes occur. They do not always occur, and I don’t think that I implied that the M rangefinder is chronically unreliable – that was certainly not my intention. I specifically referred to the discussions on focussing accuracy which regularly appear in this forum and raised the question of whether there might exist “temporary errors” such as those known in military and naval rangefinders – I wrote “I wonder to what extent might such operational inconsistencies lead to periodic variables in focusing accuracy with the M8 and M9 which are, perhaps wrongly, ascribed to other specific faults with camera body or lens?” That is a question, not an assertion.

 

The variable performance of service rangefinders was well known to armies and navies from the early 20th century when they began to be introduced into general service. For a useful primer, read Range and Vision by Michael Moss and Iain Russell published Edinburgh in 1988 to celebrate the centenary of Barr & Stroud who pioneered the development of a practicable rangefinder for military use. Your city library in Denver should be able to get a copy on loan for you to read. One can find a great deal of material in publicly available documents in the National Archives, Kew, London, mostly within the ADM series of files but also scattered elsewhere. The company records of Barr & Stroud held at the University of Glasgow (collection reference UGD 295) provide a very great deal of detail on the problems of maintaining consistent accuracy in service. I doubt that you’ll want to travel to check for yourself, so please feel welcome to contact me privately if you wish more details on this interesting, if arcane, topic. I’ve spent several years researching the evolution of the optical munitions industry in the UK and am happy to correspond with you if you wish. It’s good to know you have an open mind – please remember I’m not a politician running for office or a fifth columnist seeking to destabilise the Leica. I make a proposal for discussion, that’s all – !

 

To everyone, my thanks again for reading and responding.

[Advertisement]

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

[lct]

To lct, post #9: The laser rangefinder replaced the optical rangefinder in military use largely, though not wholly, because it was both more accurate and avoided possible human errors in aligning two images. I wonder – let me repeat “wonder” – if that principle might be a viable way to improve the efficiency of the camera rangefinder.

Problem is the subject matter receives the laser ray in photography. My Sony DSC-V1 uses a laser device to focus in the dark. Makes red spots which are quite visible actually.

[earleygallery]

How does the device I linked to work?

 

Clearly Leica have alternative technology capability, is there a reason why they haven't experimented with it further?

[jaapv]

I have a bit of a problem with transposing difficulties shown by naval rangefinders to a small camera thing - somewhat like referring to the Hubble telescope when discussing focus shift on a 35 mm lens.

[Xmas]

I have a bit of a problem with transposing difficulties shown by naval rangefinders to a small camera thing - somewhat like referring to the Hubble telescope when discussing focus shift on a 35 mm lens.

 

The Leica rangefinders at 3m is very sensitive (i.e. accurate) & some might say optimal gunnery at 30km is at the other extreme, so the gunnery analogue is poor. Close in is where the short focus lens needs accurate focus so in that sense the Leica optical co-incident rangefinder is optimal...

 

With a M8 or M9 it is easy to do a full check before every session.

 

Noel

[giordano]

One of these help? Leica Camera AG - Sport Optics - CRF 1600

 

We've been here before. Those devices are useless for general photography (accuracy ±1m at a minimum distance of 10m). Laser rangefinders that are accurate enough (e.g. Leica DISTO™ D5 - Your powerful partner for outdoors - Leica Geosystems - Leica DISTO) use beams that are powerful enough to damage the eyes of a person or animal that stares into them*, and only achieve their maximum range in subdued light with a nicely reflective target.

 

*maybe not a primary concern in military applications.

[scsambrook]

Advertisement (gone after registration)

This in particular to Xmas, re. post #25, but it's of general interest as well.

 

It's correct that the problems of rangefinder accuracy are inverted with regard to distance concerning cameras and gunnery. The permissable measuring inaccuracy for the camera diminishes as distance - and hence depth of field - decreases. For artillery, the permissable error decreases as range increases and the projectile's downward trajectory becomes steeper.

 

I've just emailed Leica to ask what is the intended measuring accuracy of the M9 rangefinder - I will find the information useful for another application besides this one. My assumption is that the factory sets minimum standards when making adjustments in assembling or repairing the camera. If anyone already has such an answer, I'm sure we will all be pleased to learn of it.

 

Looking at the depth of field tables for the 135/3.4 lens (seemingly the most demanding lens in this context), I note that the depth of field at full aperture and two meters range is from 1.978 to 2.023 metres, that is 45mm or 2.25% of the distance. As we know from photographic experience that the sharpest focus point will lie inside those extremes, the rangefinder will have to measure with an accuracy of better than plus or minus one percent. Of course, at closer distances that tolerance will have to be even less, but it need not be so small as distance increases. It will be interesting to see what figure(s) the factory uses.

 

Becoming more curious, I see that the Leica CRF 1600 digital rangefinder measures to within plus/minus 0.25% at 400 yards and to 0.5 percent at ranges over 800 yards. I have no idea whether accuracy alters as the range diminishes below 400 yards.

 

I've just seen Giordano's post - it looks as though that instrument measures to plus/minus one millimeter . . . I think a far smaller tolerance than needed for photography. The point about the harmful effects to vision is certainly important and I see there are warnings about careful use.

[lct]

The laser device of the Sony DSC-V1 works fine up to 2.5 meters. Here at 30 centimeters.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Xmas]

I'm sure we will all be pleased to learn of it.

No

Looking at the depth of field tables for the 135/3.4 lens (seemingly the most demanding lens in this context), I note that the depth of field at full aperture and two meters range is from 1.978 to 2.023 metres, that is 45mm or 2.25% of the distance. As we know from photographic experience that the sharpest focus point will lie inside those extremes, the rangefinder will have to measure with an accuracy of better than plus or minus one percent. Of course, at closer distances that tolerance will have to be even less, but it need not be so small as distance increases. It will be interesting to see what figure(s) the factory uses.

No, to the last sentence.

But every one is different.

 

note although the baselength of each M is the same (sensibly) the effective baselength is dependent on magnification of the finder, and hence accuracy.

 

I think you will find the big problem with the 135mm is the framelines, rather then the rfdr accuracy.

 

But before you become too happy if you have a e.g. eye safe laser rfdr, then the feedback of what you have focused on is not easy to arrange, e.g. Zeiss for the Contax G series, managed a totally impossible scheme.

 

For the M I used to focus half way across the nearest eye, low light cinima verity shots of children, 45 degrees to face, 1m away 9cm cron with M4... One needs to know that one hit or missed the middle eye lash. Used to miss a lot, not the M fault the little mites could move and eat at the same time... Eating was when they moved he least.

 

How are you going to tell where the laser focused on? Lots of people won't buy a Contax G clone...

 

Fixing something that is not broken is how to get fired...

 

The problem with the M8 and M9 and the M focusing is the lenses were not designed to work with a digital sensor, witness the new 3.5 cm lux,... There was a similar problem with the f/1 Noct 5cm in '75-'79, when people had to send cameras and lenses back, it dissappeared, for decades...

 

Suggest it is a depth of focus problem not depth of feld...

 

Noel

[jaapv]

I disagree. shooting laser beams all over the place is just the thing for street photography:rolleyes:

 

Well, maybe on Oxford Street in December...

[AlanG]

My opinion is that the rangefinder is probably accurate enough for the intended usage providing it is adjusted correctly. But the mechanical parts could probably be redesigned to allow for easy user adjustable fine tuning for each lens.

[pico]

To lct, post #9: The laser rangefinder replaced the optical rangefinder in military use largely, though not wholly, because it was both more accurate and avoided possible human errors in aligning two images. I wonder – let me repeat “wonder” – if that principle might be a viable way to improve the efficiency of the camera rangefinder.

 

The laser rangefinders with which I am familiar work by measuring the time it takes for the reflected beam to return to the sending unit. That makes them exceptionally accurate for long distances, and impractical for the ranges we use. Cameras that do use IR use it with autofocus to illuminate a subject, usually when there is not enough light for the autofocus to work properly. It's still blind-sighted autofocusing.

 

To pico, post #12: I am not an engineer or a research or design person but I am perfectly comfortable to be in the company of those who are. I have no need to ‘be careful’ because I’m asking questions and putting forward questions and ideas for discussion. And although you and I may not seek radical innovation, any business such as Leica Camera AG needs – at the very least – to consider the market presence of those who do.

 

Leica is not isolated. It has Panasonic as one partner, Kodak as another, and gosh only knows how many other high-powered marketing and engineering consultants.

 

I’ve spent several years researching the evolution of the optical munitions industry in the UK

 

Then perhaps you can propose an alternative to the current rangefinder, other than autofocus. Perhaps begin with focus confirmation.

 

I'm wondering - it's a long shot - but did we work together at Upper Heyford in the Sixties? None of the technology we had then is still classified.

[jaapv]

I think the best thing would be a focussing aid. Just something like two mirrors fitted to the current RF mechanism and a little sensor could activate a LED to indicate optimum focus. Nothing more precise than the current system, but very useful for those with less-than-perfect eyesight.

[pico]

We've been here before. Those devices are useless for general photography (accuracy ±1m at a minimum distance of 10m). Laser rangefinders that are accurate enough (e.g. Leica DISTO™ D5 - Your powerful partner for outdoors - Leica Geosystems - Leica DISTO) use beams that are powerful enough to damage the eyes of a person or animal that stares into them*, and only achieve their maximum range in subdued light with a nicely reflective target.

 

*maybe not a primary concern in military applications.

 

The military laser range-finder uses an exceedingly short exposure, lest one become a target.

[pico]

[...]Looking at the depth of field tables for the 135/3.4 lens (seemingly the most demanding lens in this context), I note that the depth of field at full aperture and two meters range is from 1.978 to 2.023 metres, that is 45mm or 2.25% of the distance. As we know from photographic experience that the sharpest focus point will lie inside those extremes, [...]

 

Not exactly between those extremes.

 

The sharpest focus point is that which is in focus. The acceptable depth-of-field depends upon what once considers an acceptable circle-of-confusion, which is qualitatively determined by degree of enlargement and viewing distance.

[adan]

@ lct - regarding your item #7. True.

 

@ scsambrook - Thank you for providing sources. I might add to our growing list "subject distance" -because there comes a point where any triangulating rangefinder can't distinguish infinity from "almost infinity".

 

A problem I had with the Contax G cameras (which did have a decent IR triangulating "electronic rangefinder") But it only measured distance out to about 60 ft (20 meters) - anything beyond that was "infinity". Which was within the DoF for a 35 or 45 f/2 lens, but not a 90 @ f/2.8. Things over 200 feet were in focus at "infinity", it could focus accurately on things closer than 60 feet, but there was a "black hole" between 60-200 feet where it just couldn't focus a 90.

 

Leica RFs have a similar limit at some point. Eyesight plays a role, but a basic check with my eyes (corrected) and the M9 .68x viewfinder is that the RF, focused at infinity, only begins to separate the images of a subject at about 100 meters (330 ft) or closer. In other words, to the M9 RF and my eye, anything beyond 100 meters is "infinity."

 

....or about 5x better than the Contax Gs

 

I guess the main reason I don't actually run into problems is that I don't take that many pictures in the "black hole" between 100 meters and infinity with lenses where DoF won't cover the error. Even the massive 75 f/1.4 almost has enough DoF (focused AT infinity, it will be "in focus" down to 118 meters - focused at a hyperfocal of 135 meters, DoF will cover from infinity to ~70 meters - coming in well short of the "black hole" limit of 100m.)

 

Calculations per this doo-hickey here: Online Depth of Field Calculator

____________

EDIT: and yes Pico is right - the doohickey uses 0.03mm for the CoC

[giordano]

The laser rangefinders with which I am familiar work by measuring the time it takes for the reflected beam to return to the sending unit. That makes them exceptionally accurate for long distances, and impractical for the ranges we use.

 

I'm not sure what method the Leica Disto and other "laser tape measures" use, but they have a much shorter range than military or hunting/golfing rangefinders, a very much shorter minimum range (<175mm in the case of my two-generations-old Disto A2), and millimetric accuracy at short to moderate distances.

[Xmas]

A problem I had with the Contax G cameras ...

The G1 had two rangfinder schemes, but the focus confirmation on the G1 was a digital display of the range, a display was rather short of any utility. It was not that there were inaccurate instead they

 

could decide that could not focus and flash a sorry cant display or

tell you they were focused at a distance which you had no way of validating, short of using a tape measure or a hot shoe rangefnder

 

They took some time to do either of these...

 

Superb lenses and rest of the camera was nice, and with the Hologon lens the camera was perfect cause the Hologon was manual scale focus.

 

Noel

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[IkarusJohn]

Well, given that my rangefinder on the M9 is consistently correct, only needs adjustment after a hefty bang and that all my errors are operator error ...

 

So if the rangefinder is consistently correct, how does focus shift fit?

 

I thought the issue was that you may focus a lens (say the previous Summilux-M 35 Asph) perfectly at f/1.4, and find that you needed to stop down. The focus remained perfect through the rangefinder, but the end result on your M9 was out of focus.

 

I seem to recall that this is a known problem, which SLR solves as you focus through the lens. I know I over simplify, but the point remains that the rangefinder focusing cannot manage focus shift. Isn't that really the only issue?

 

Cheers

John

[Jager]

So if the rangefinder is consistently correct, how does focus shift fit?

 

 

Focus shift is the phenomenon where the point of actual optical focus drifts rearward as the lens is stopped down from wide-open. It is a characteristic of all lenses - and most apparent in fast lenses. Fortunately, it is usually slight enough to not be noticeable in actual practice.

 

The rangefinder focusing mechanism is independent of aperture - the manual, physical focus so achieved is exactly the same whether you have the lens set at f1.4 or f16.