Focus shift

[Reeray]

Advertisement (gone after registration)

I hope this isn't a daft question but if so please bear with me.

 

When considering a lens infamous for focus shift my understanding is that such a problem only occurs when focusing wide open and subsequently stopping down. Conversely, if one focuses at the desired aperture and adjust shutter speed for correct exposure then this issue is a non event. Am I correct in this assumption?

 

As an aside, is there any benefit in actually focusing wide open prior to stopping down?

 

I'm a bit naive on these points and would appreciate your advices.

 

Thank you in anticipation of any responses.

 

Apologies, should be in lens thread. Please move.

Edited September 2, 2016 by Reeray

[edwardkaraa]

I hope this isn't a daft question but if so please bear with me.

 

When considering a lens infamous for focus shift my understanding is that such a problem only occurs when focusing wide open and subsequently stopping down. Conversely, if one focuses at the desired aperture and adjust shutter speed for correct exposure then this issue is a non event. Am I correct in this assumption?

 

As an aside, is there any benefit in actually focusing wide open prior to stopping down?

 

I'm a bit naive on these points and would appreciate your advices.

 

Thank you in anticipation of any responses.

 

Apologies, should be in lens thread. Please move.

 

If you focus with the rangefinder, the lens is normally calibrated for a certain aperture. In the case of the ZM 50/1.5 I have seen two varieties, one calibrated for wide open and one calibrated for f/2.8.

 

If you shoot at a different aperture then you must practice how to compensate.

 

If you focus with the EVF, you will need to do it at working aperture. Focusing wide open and then stopping down will give you back focus.

[Advertisement]

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

[Reeray]

I see, so assuming the lens is prone to focus shift and one focuses with the rangefinder at something other than its optimised aperture, then focus shift may still be evident?

[edwardkaraa]

I see, so assuming the lens is prone to focus shift and one focuses with the rangefinder at something other than its optimised aperture, then focus shift may still be evident?

 

Yes, until the focus shift is covered by the depth of field. It is easy to compensate though. I normally focus normally with the rangefinder patch then move the focusing ring by a certain amount using muscle memory and the distance between the double images in the patch.

 

If the lens is optimized for 1.4, you will experience shift at every other aperture and is barely covered by depth of field at small apertures, so unless you shoot most often wide open, this is not an ideal setting.

 

If the lens is optimized for 2.8, you will get the front focus shift at wide open, but the advantage is that the shift is not noticeable at smaller apertures.

[luigi bertolotti]

I see, so assuming the lens is prone to focus shift and one focuses with the rangefinder at something other than its optimised aperture, then focus shift may still be evident?

As you say, MAY be evident... is a subtle question because if, for instance, the lens is optimised at full aperture , you focus through RF and then close a pair of stops, CAN be that the shift is compensated by the depth of focus that has became deeper stopping down... and with lenses in the 35/50 range this is rather easy to occur. Fact is that the focus shift issue has became an issue (in 24x36 photography) mainly with digital, and in the Leica world mainly with the Summilux 35 asph... personally, I noticed it only on certain lenses (Summicron 90 un-asph , Tele Elmarit 135) and only at close distances and enlarging significantly the image... at film times, I knew of it only from books on operations on Large Format cameras,,, and never noticed it in my 35mm negatives...

Edited September 2, 2016 by luigi bertolotti

[willeica]

I agree with Luigi. This is largely a digital phenomenon in my experience. An EVF eliminates this altogether as the focussing is TTL.

 

William

[edwardkaraa]

Advertisement (gone after registration)

Depth of field in digital is noticeably shallower than film. I agree that shift becomes much more of a problem with digital, especially high mp count.

 

I think if the lens is optimized for 1.4, the depth of field won't catch up with shift until f/5.6-8, depending on how picky the user is. Certainly couple of stops won't be enough on the M240/262.

Edited September 2, 2016 by edwardkaraa

[Reeray]

Mmm - I still don't get it. If I focus, via RF and a given selected aperture and nail it, why it it so then possibly subject to focus shift? I can accept that subsequently changing aperture may be influential but in the absence of such change I'm lost. What am I missing?

 

Sorry if I'm coming across as dense but this is really confusing me me.

[edwardkaraa]

Obviously, when you focus with the RF, the lens aperture is not taken into account. It depends on for which aperture the focusing mechanism has been optimized for. Only at that aperture will you get accurate results, assuming of course a lens with focus shift.

[LocalHero1953]

Rangefinder focusing is based on an exact match between the position of the lens in relation to the sensor, and the overlap between two images in the viewfinder: when the images overlap at the point you want to focus on, THEN the position of the lens in relation to the sensor means that that object is in focus.

 

The connection between the image overlap and the position of the lens in relation to the sensor is a mechanical calibration: it is determined by the shape of that angled surface on the focusing mount on the inner end of the lens, and how it engages with the roller inside the body.

 

That calibration is made on the assumption that the lens is perfect i.e. at ALL apertures, the focus of the lens stays the same. In a lens with focus shift, that calibration is only true for one aperture. If you try to focus at another aperture the lens will be placed at an out of focus position in relation to the sensor.

 

So if your lens is calibrated to the Leica RF at f/2.8, and you try to focus at f/8, you may think it is in focus, because the two images have overlapped perfectly; but the calibration is no longer correct, and the lens is out of focus.

 

I expect I have confused matters even more!

[adan]

Bottom line with a rangefinder camera using the rangefinder (as opposed to live view or EVF, since we are in the M240 category):

 

The lens - i.e. the glass and aperture - play NO ROLE AT ALL IN THE PROCESS OF FOCUSING!

 

One can stick an empty lens barrel (no glass, no aperture blades, no image being created) on an M camera, and turn the metal focus ring, and the rangefinder patch will move perfectly normally, and appear to indicate focused or not-focused conditions (RF image aligned, or not aligned).

 

(and you can even get "empty lens barrels" to try that out - many older lenses were intentionally designed to separate the glass and aperture from the M-mount focusing barrel, so that they could be used on the SLR-like Visoflex accessory - https://www.cameraquest.com/jpg3/leicaM_50dr1_04.jpg )

 

So - Reeray - when you say that you focus perfectly "via a given selected aperture" - really, you do not. You simply focus via the lens focus ring. The "selected aperture" is not involved of the process, and may not even be present.

 

The focus cam in the lens is a solid piece of metal. It produces the same displacement of the focusing images all the time. It does not "adjust itself" according to the aperture the photographer chooses.

 

In theory, in lenses known to have focus shift at different apertures, an additional "aperture cam" connected to the aperture ring could probably be designed in, that nudges the whole focus cam in or out a bit as the aperture is changed. In addition to the movement the focusing ring itself produces. But this have never been done in rangefinder lenses**, and without such communication of aperture in use, the focus cam alone cannot correct for focus shift across apertures. In the absence of such a "variable mechanism," the best that can be done is factory calibration for one focus point or the other or a reasonable compromise (as Ed mentions), depending on which works best for an individual photographer's purposes.

 

As a point of historical interest - the Nikon 45mm "GN" lens for flash use on SLRs did connect aperture to focusing - the reverse - so that when using a flash on the camera, the aperture would stop down as one focused closer, to keep the flash exposure correct at any distance (before there were self-adjusting flash units). You set the guide number for the flash in use, on the lens (thus the "GN" name), slid a tab to lock the aperture ring to the focus ring, and turning the second automatically moved the first.

 

As another aside, it important to understand why and how focus shift occurs. It is mostly due to a lens having some residual spherical aberration. When used wide open, the light rays from the extreme edges of the large aperture ("peripheral rays") do not focus in exactly the same place as the rays coming through the center of the lens. The point of focus is smeared out over a tangible distance, with a point of "best focus"  - or "circle of least confusion" - somewhere near, but not necessarily at the middle of, the smear.

 

When the lens is stopped down, the "aberrant" rays from the edges are cut off, leaving just the paraxial rays to form the image. And those focus more tightly in one place (the "paraxial focus"). Which is not the same place as the "best overall focus" at the larger aperture - i.e. the "best focus" point shifts from the first point to the second when stopping down. It is a function of which light rays you are using and which you are discarding, not the physical movement of the lens or light as such.

 

http://www.olympusmicro.com/primer/images/aberrations/sphericalfig1.jpg

_________________________________

**Probably because the extra mechanism would make the lenses more complex, larger and heavier - which goes against the whole ethos of the compact rangefinder idea. As well as adding cost. And if cost and complexity and size are no problem - you can just design the lens to be really big and with many aspherical elements, and get rid of the spherical aberration that causes the shift in the first place.

Edited September 2, 2016 by adan

[Bill Livingston]

Thank you Adan!

 

I was getting increasingly frustrated until your post... Nice clear explanation for the OP.

 

I think this is hard to 'get' if you come from a background of SLR's where the lens is actually being focussed via the mirror onto the focussing screen (which has its own potential for error) which is where I guess the OP comes from, so can well understand his confusion...

 

An EVF and reading off the sensor is about as direct a focus as you can ever get because you are looking at the result of the focus of the lens you are actually using (as opposed to the rangefinder lenses and the mechanical link) and can't have either of these errors. (rangefinder calibration issues or, in the case of an SLR or DSLR, mirror/focussing screen/film/sensor plane misalignment). Which is why an EVF is always the best option for truly accurate focus regardless of lens chosen. Frankly, rangefinder focussing is susceptible to so many potential errors it's a wonder it works at all... and the fact it does so well on the Leica M is a testament to the engineering and build quality of the camera and justifies the expense... not the principle at all.

 

Which is why I agree with Erik so often (and am afraid to say!). I love my M because of its built, its look, it's engineering and it's feel... I wouldn't worry if there was an EVF M or electronic rangefinder in the future... it would still be an M to me...

 

As far as your explanation of 'focus shift' is concerned, that's the best explanation I have ever read it simplified the rather convoluted 'explanation I had in my head and pointed out obvious points I hadn't considered before. I learned something there... thank you.

Edited September 2, 2016 by Bill Livingston

[jaapv]

Err.. the normal (D)SLR is as prone to focus shift as a rangefinder, as it is focused wide open and stopped down by the auto-diaphragm at exposure.

[jdlaing]

As jaap sez.

Actually more so.

[Reeray]

Thank for all the responses. I now realise that the mistake I was making and causing my confusion was overlooking that I'm not focusing through the lens! Duh!

 

Now it all makes sense.

 

Great forum, very helpful people

 

Regards

[edwardkaraa]

Err.. the normal (D)SLR is as prone to focus shift as a rangefinder, as it is focused wide open and stopped down by the auto-diaphragm at exposure.

Not only that, but with AF cameras, what you see on the screen is equivalent to f/2.8, and with AF lenses that have focus shift, like the canon 50 1.0, the AF sensors see the effect at f/5.6. Effectively this causes severe front focus.

[Bill Livingston]

Err.. the normal (D)SLR is as prone to focus shift as a rangefinder, as it is focused wide open and stopped down by the auto-diaphragm at exposure.

Err... isn't it the lens that's prone to 'focus shift'..?

 

Nothing to do with whether its a rangefinder or an SLR.

 

But if you are focussing through the lens, then pretty much every time you are taking that out of the equation, such as on an SLR... unless the focus shift only occurs at a smaller aperture(s)... which isn't exactly common! And on those extremely rare examples. you could focus with the lens stopped down and focus at the shooting aperture... which is pretty much what you are doing on an EVF.

 

On the other hand, there isn't much you can do with a rangefinder apart from compensate from experience.

 

The advantage of the M for me isn't the 'rangefinder'', it's the viewfinder. You get a field of view that approximates that of a 28mm, with frame lines showing an approximation of the mounted lens, (which is cool as you see outside the frame without having to search), but the main advantage is you are looking through and seeing exactly the ambient light... and that's a massive advantage... and obviously the whole mechanical process of taking the photo has absolutely nothing whatever to do with what you see through the viewfinder. You are recording your vision, you first creative choice, then further creative element is then added in what choices you have made in aperture and shutter speed, what your point of focus is and what you do in PP... whether that's film or digital... it's just the same.

 

That's what I enjoy about M photography. The rangefinder is simply a focussing aid, and not always the best choice of focussing in certain circumstances... But in terms of creative photography, it's the most natural and instinctive camera I'm aware of... especially if you are like me and have grown up understanding aperture and shutter speed and focus.

 

Anyway, the OP's question was answered perfectly by Adan... and as I said in my first post, it was obvious that the OP was thinking he was focussing TTL from the points he was making. It was the only possible explanation for his confusion.

Edited September 3, 2016 by Bill Livingston

[Steve Ricoh]

Great thread and thanks to Adan for the excellent explanation.

 

All I need to know now is the optimised aperture for each lens I own, the 50/1.4 ASPH and 35/f2 ASPH, and the 28/f2.8 on my list.

 

I'm guessing it's maximum aperture as that would make more sense.

[edwardkaraa]

Great thread and thanks to Adan for the excellent explanation.

 

All I need to know now is the optimised aperture for each lens I own, the 50/1.4 ASPH and 35/f2 ASPH, and the 28/f2.8 on my list.

 

I'm guessing it's maximum aperture as that would make more sense.

None of your lenses should have focus shift but it's easy to verify. Put the camera on tripod and shoot a ruler at 45 degrees at apertures from wide open to f/8, focusing of course with the RF. Not all lenses are optimized for the same aperture.

Edited September 3, 2016 by edwardkaraa

[Advertisement]

Hallo,

Look here: Leica M

[TomB_tx]

I have a few lenses with significant focus shift when stopping down, notably both the Voigtlander 35 Nokton 1.4 and Nickel Heliar 50 f2. On film RF I've never noticed a problem with them, but on digital M the sharpest details at mid-stops are often behind the subject.

I use a Sony A7 (EVF) for my R (and other legacy SLR lenses), and find the 35 1.4 Nokton gives sharper subjects with it, since you focus TTL at actual aperture. (Of course, don't look too close at the smeared corners with the A7 and an RF 35 lens. Need to get the kolari sensor modification.)

I also realized my M9 RF was slightly off at middle distances, making the focus shift effect worse. I'll get it back next week from DAG and see how these lenses do then.