Please define focus shift

[wizard]

Advertisement (gone after registration)

vor einer Stunde schrieb Al Brown:

I suggest to check and read the whole thread, it is exactly what I’ve been saying.

If so, apologies for not reading the whole thread, but your limitation to RF lenses did suggest otherwise.

[Advertisement]

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

[pgk]

2 hours ago, pegelli said:

I'm not sure either, maybe I should have said "some" instead of "many". Is your statement that autofocus lenses might be "more corrected" for for focus shift at the expense of correction of other abberations a theory or is there a source with real world data that supports it?

Not offhand - try searching. But if you are less concerned with say distortion and chromatic aberration (software can do a good job of dealing with these in pp), then it means that you are free to concentrate on better correction of other problmes such as spherical aberration. Many lenses used on AF cameras do show significant distortion if let unadjusted.

[pegelli]

6 minutes ago, pgk said:

Not offhand - try searching. But if you are less concerned with say distortion and chromatic aberration (software can do a good job of dealing with these in pp), then it means that you are free to concentrate on better correction of other problmes such as spherical aberration. Many lenses used on AF cameras do show significant distortion if let unadjusted.

But the same would be true for digital rangefinder cameras, the fact AF or MF lens/camera makes no difference in that regard.

Also many lenses that still stem from the film area show less distortion because that was harder to correct, I think "distorting lenses" corrected by (in camera or external) software is a more recent phenomena, and certainly didn't apply to earlier digital RF or DSLR cameras.

Edited September 28, 2022 by pegelli

[pgk]

13 minutes ago, pegelli said:

But the same would be true for digital rangefinder cameras, the fact AF or MF lens/camera makes no difference in that regard.

Also many lenses that still stem from the film area show less distortion because that was harder to correct, I think "distorting lenses" corrected by (in camera or external) software is a more recent phenomena, and certainly didn't apply to earlier digital RF or DSLR cameras.

To be correctable using software, lenses have to communicate with the camera. In the case of digital rangefinder camera, te M uses only 6-bit coding. This does not help the camera know which aperture is in use (although it can make an 'educated' guess) whilst fully electronic lenses can tell the camera suh information as aperture in use and focus distance. This could be used to add in correction for focus adjustment to compensate fr focus shift and/or adjustment for less well corrected aberrations which can be used in-camera or in post processing. Older lenses were often less well corrected than current ones although the greatest improvements in optics haveactually been in terms of ever increasing usable focal length, fast aperture lenses and so on.

Focus shift has been a problem in photography since photographic lenses were first made (try actinic focus). Today we can deal wth it in many ways but not, unfortunately, with mechanical rangefinder cameras and mechanical lenses. These require that lenses should not suffer from it. The fact that Leica introduced the 35mm FLE Summilux shows that there is a need for M lenses free from focus shift.

[pegelli]

@pgk, I'm not questioning your logic or theory, I agree with those. I however have my doubts that avoiding focus shift is a significant aspect of modern (AF) lens design used on SLR/mirrorless cameras. Yes, more and more lenses distort and get software corrected but I've always understood that the main purpose of that is to make lenses smaller, lighter and cheaper and I've never heard (until this thread) that avoiding focus shift was an important factor in that as well. It might be, but it's new to me.

[pippy]

Apologies if this is a silly question but if DSLR and Mirrorless cameras auto-focus at max. ap. will there not be the same back-focus issue if the lens is used stopped-down or does the electronic brain re-focus at actual shooting aperture?

Philip.

[pegelli]

Advertisement (gone after registration)

21 minutes ago, pippy said:

Apologies if this is a silly question but if DSLR and Mirrorless cameras auto-focus at max. ap. will there not be the same back-focus issue if the lens is used stopped-down or does the electronic brain re-focus at actual shooting aperture?

Philip.

I can only speak for my DSLR and mirrorless cameras (Sony).

My mirrorless cameras focus at the set aperture with their native and most 3rd party AF lenses, so no focus shift.

My DSLR's focus at max aperture and close to shooting aperture w/o adjusting the focus, so they're susceptible to focus shift, allthough you can set "micro focus adjust" for individual lenses and therefore compensate for focus shift to the most used aperture, however you can't set it so it depends on aperture used so other apertures may again not focus correctly depending on the amount of focus shift the lens has.

I don't know if the other brands are different, maybe others who know can chime in.

Edited September 28, 2022 by pegelli

[jsrockit]

22 hours ago, UliWer said:

No, it isn't. As i said, focus may be absolutely correct at f/2.8, and way off at f/1.4, if your lens is calibrated at f/2.8. Due to optical laws the focus shifts when you change the aperture. 

 

I took it to mean he was using the OP´s example of 2.8 and F4 and not literally only those F stops in all cases.  Of course, there is an optimum F stop and it shifts from there.

Edited September 28, 2022 by jsrockit

[wizard]

vor einer Stunde schrieb pgk:

The fact that Leica introduced the 35mm FLE Summilux shows that there is a need for M lenses free from focus shift.

This wasn't much of an issue in the film days, as the thickness of the film emulsion more or less covered all focus shift, but today's digital sensors are pretty much unforgiving here.

[pgk]

1 hour ago, pippy said:

Apologies if this is a silly question but if DSLR and Mirrorless cameras auto-focus at max. ap. will there not be the same back-focus issue if the lens is used stopped-down or does the electronic brain re-focus at actual shooting aperture?

Philip.

Having owned some very fast lenses (Canon 50/1.2, 85/1.2 MkI and 24/14 Canon and 20/1.8 and 35/1.8 Sony, all I can say is that I've never found focus shift a problem, and my guess is because it isn't significant, unlike my 35/1.4M pre-FLE which some seem to find problematic (I can't say that I do though). Mirrorless focus at working aperture, dSLRs wide-open. Conclusion: there is no significant focus shift with these lenses, or the makers have software which deals with it (unlikely). However distortion and chomatic aberration wre issues at times most d/w in post. 

[pippy]

28 minutes ago, pgk said:

Having owned some very fast lenses.........all I can say is that I've never found focus shift a problem, and my guess is because it isn't significant...

So might focus-shift be less of a problem with these systems because there are more generous parameters regarding lens design so correcting this issue can be achieved because the lens-body doesn't have to be as small as, for instance, M lenses? Very fast Leica M lenses tend to be considerably smaller/shorter than the equivalent in DSLR-World. By way of an example take a look at the v2 35mm Summilux and compare its size with the same spec lens from Canon...

Philip.

[pgk]

1 hour ago, pippy said:

So might focus-shift be less of a problem with these systems because there are more generous parameters regarding lens design so correcting this issue can be achieved because the lens-body doesn't have to be as small as, for instance, M lenses? Very fast Leica M lenses tend to be considerably smaller/shorter than the equivalent in DSLR-World. By way of an example take a look at the v2 35mm Summilux and compare its size with the same spec lens from Canon...

Philip.

Makes sense to me😁.

[adan]

2 hours ago, pippy said:

So might focus-shift be less of a problem with these systems because there are more generous parameters regarding lens design so correcting this issue can be achieved because the lens-body doesn't have to be as small as, for instance, M lenses?

I agree with pgk's response above. But in addition (re-capping some points already made):

1) Film tends to mask many sins. The emulsion is thicker, so that any focus shift still ends to fall within the imaging surface. And up until about 2003 (Canon 10D) film was it for 99.99999% of photography.

There was also less fixation on "nothing but bokeh" pictures shot at large apertures just for the heck of it - photographers tended to stop down to the "best" aperture of f/4.0-5.6 unless they really needed to work in very dim light, which also meant very grainy, unsharp film (and modest lens performance) anyway. Royal-X 120 at ISO 1250 or Tri-X pushed to ISO 1600-3200 was so fuzzy, you'd never notice a bit of blur ± cause by focus shift.

Who knows if 1962-2001 Nikon F f/1.4s had focus shift - no one would ever see it.

And there were fewer fast lenses in use overall - recall that the Nikon F system was introduced in 1959 with four lenses, of apertures f/2.8, f/2.0, f/2.5, and f/3.5. Only a few photographers with special needs bought lenses f/1.4 or faster (except 50s) for the next couple of decades. Even when they became available,(35 f/1.4 - 1971, 85mm f/1.4 AI-S - 1981) -  they were very expensive and large and heavy.

2) Tech capabilities of larger SLR lenses, probably. Nikon produced their first FLE lens for general use on SLRs in 1967. 24mm Nikkor-N f/2.8. Canon began a series of FLE lenses with the FD lenses for the F-1 in 1970.

Leica did not produce an FLE lens for M until 35-38 years later (75 APO-Summicron-M - 2005).

That assumes that FLE actually fixes focus shift, which I am not sure about - how does the optical design "know" which aperture is in use, in a mechanical M lens, and correct for it?

Edited September 28, 2022 by adan

[UliWer]

33 minutes ago, pgk said:

By way of an example take a look at the v2 35mm Summilux and compare its size with the same spec lens from Canon...

 

Though you won't see much focus shift with the pre-aspherical 35mm Summilux. The aspherical 35mm Summilux (non FLE) was much better corrected with much higher resolution fully opened. Its correction to achieve this high resolution resulted in focus shift. You may say that the performance of the aspherical Summilux was "overstretched". They had to correct it even more to avoid focus shift for the FLE.

The most "famous" lens for focus shift is the 1:1.5/50mm Sonnar for the  Zeiss-Ikon Contax. I have a pre-war version and a post-war one. When I compared both I was astonished that on longer distances the resolution of the pre-war version was better than the post-war version. But the pre-war version shows very strong focus shift, which is very much reduced with the post-war version. The first version was certainly "overstretched", and they reduced the performance just a little bit to avoid focus-shift. the modern 1:1.5/50 Sonnar is a simplified version of the original Sonnar design. It shows better resolution than the old design - but has very strong focus shift: "overstretched" again. 

[pgk]

7 minutes ago, adan said:

That assumes that FLE actually fixes focus shift, which I am not sure about - how does the optical design "know" which aperture is in use, in a mechanical M lens, and correct for it?

If I remember crrectly. Focus shift seems to be more evident at closer focus distances. FLE adjusts the position of rear group as the lens focusses closer, and this improves (reduces) spherical aberration at these closer distances and as a result also reduces focus shift. So I don't think that it needs to "know" the aperture.

[UliWer]

Yes, you can generally say that lenses which show focus shift, will show weaker performance on short distances. The 1:1.5/50 Sonnar is again a good example: the pre-war version with high resolution on long distances and strong focus shift gets weak on short distances. The post-war version with lower resolution on long distances and less focus shift was much better on short distances. Of course it had no floating elements. And I don't know if lens designers can achieve a reduction of focus shift by floating elements.

Edited September 28, 2022 by UliWer

[Jeff S]

28 minutes ago, adan said:

That assumes that FLE actually fixes focus shift, which I am not sure about - how does the optical design "know" which aperture is in use, in a mechanical M lens, and correct for it?

Discussed frequently here, as in the following discussion (comments from Rick, 01af and me… Karbe quote). Focus shift mitigation is a secondary benefit; the primary being improved performance at close distances, resulting from reduced spherical aberrations and better contrast.

 

 

Jeff

[TomB_tx]

I believe glass types for the elements also have a lot of influence on focus shift. The Voigtlander Nokton 35 f1.4 version 1 shows significant focus shift when stopping down. The version 2 has MUCH less shift without resorting to FLE, etc. 

I also think the concern with focus shift is overblown for most of us. I never worried about it for 50 years, and have been happy with my film-era M lenses on digital. I did find shift on my Nokton 1.4 ver 1 once I read about it and tested carefully for it, but used it happily with good results before then. (Moral: don't read forums - just use your equipment and enjoy it.)

I should also note that I prefer f2 lenses for their smaller size, lower cost, and fine performance, and shift is more of an issue with faster lenses.

[Advertisement]

Hallo,

Look here: Leica M

[01af]

Floating elements improve close-range performance. They don't address aperture-related focus shift.

But then, anything that reduces spherical aberrations also will reduce focus shift ... simply because focus shift is a consequence of less-than-perfect spherical correction. Hence, floating elements indirectly reduce focus shift ... at least for shorter focusing distances.

[Harpomatic]

On 9/28/2022 at 8:58 PM, adan said:

That assumes that FLE actually fixes focus shift, which I am not sure about - how does the optical design "know" which aperture is in use, in a mechanical M lens, and correct for it?

I think Adan, as usual, makes really good points in his post.

About the quoted text, FLE simply mitigates the issue: to have no focus shift you would need no spherical aberration at all, which is impossible. 
The FLE lenses like the Summilux-M 50mm ASPH, which is one I tested, has so little focus shift that it’s negligible for all intents and purposes because its focus plane always falls well within the DoF of the chosen aperture.

But focus shift is called so because that’s what the focus plane does shift forward when closing the aperture, backwards when opening the diaphragm. No matter what system or focusing mechanisms are used. 
 

On a RF the best bet is to get and FLE lens or a lens not renown for its focus shift.

Or you just get on with it, make pictures, and complain only if you find your subject out of focus because of it. Which is rare on most modern lenses.