Focus shift question

[Guest #12]

Advertisement (gone after registration)

Try this book:

 

The art and science of optical design by Robert Rennie Shannon

 

Thanks, I'll check that out.

[Advertisement]

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

[algrove]

OK, so i've read all this, but to help a novice what is the correct way to focus a 35 or 50mm lens at 3 feet, what happens at 6 feet, 30 feet?

 

How do you compensate for this in day to day use? Do I change my focus forward (in front of the main subject) or backwards (or behind the main subject)?

 

Maybe a general comment is not easy, but try.

 

Since I have around 15 lenses I would like to know how I make them work great without having them sent to Leica. Sending lenses and/or camera to Leica seems to me contrary to buying quality product. Never had to send an old Nikon into Nikon. Never had to send an old Hasselblad into Hasselblad. Never has to send a recent Panasonic into Panasonic.

What the heck is going on here? Why do you all put up with this crap!

[ho_co]

Al, you say you have 15 lenses. Have you used them? Do they exhibit focus shift? If not, you don't do anything.

 

Some lenses make use of focus shift to give improved performance overall. Since Leica designed them that way, sending them to Leica isn't going to do you much good.

 

How do you focus at 3 ft, 6 ft, 30 ft? Use the rangefinder, focus, shoot. If your lens exhibits focus shift, you'll see how to compensate.

 

All Leica lenses (not true of all brands) should focus accurately (as determined by the rangefinder) when wide open. Any residual focus shift will be visible as you stop down a couple stops from there.

 

With the arrival of digital sensors, some of the design adjustments were found to be less beneficial, and Leica is redesigning those lenses. See for example the 35/1.4 asph, which on film was considered one of Leica's best lenses. It was redesigned to make use of floating elements, and the new version is better both on film and on the sensor. Few people had problems with the earlier version (even on digital cameras), and studying it enlightened a lot of us as to how Leica makes their fast lenses so good. You can retrace our collective discovery in Tim Ashley's thread, http://www.l-camera-forum.com/leica-forum/digital-forum/17699-very-interesting-answer-leica-35mm-1-a.html.

 

 

What the heck is going on here? Why do you all put up with this crap!

Whoa! Get a handle! Leica didn't write the laws of optics. All high-speed lenses with variable aperture suffer from some focus shift. Leica just did the math better than others and improved the overall performance by making practical use of it.

[algrove]

I own the new 35 you talk about. No problems. I also have older Leica lenses and perhaps I have problems. If I do have problems, I am asking how one overcomes focus shift problems when using a lens without having to send them in to Leica.

 

I would still like to have practical knowledge on how to overcome focus shift-a practical explanation.

[ho_co]

In its discussion of focus shift, LFI suggested with the predecessor of the current 35 Summilux that:

• if one was working in the aperture range and distance range where focus shift was noticeable;
  
• one should focus with the rangefinder, then nudge the focus a slight amount closer.
  

 

The question you're raising is analogous to, "If I lose control of my car on the highway, what do I do?"

 

Until you know whether you've got understeer or oversteer, or whether there's ice on the road, or generally, why you lost control in the first place, it's impossible to say what the best way out of the situation may be.

 

The question, "If I have some kind of problem with some lens in future, how do I correct for it?" is too general. I'd wait till the problem manifests itself, then see if LFI's recommendation doesn't work.

 

Most people working with the lenses with undercorrected spherical aberration never notice a problem with the lenses. Wait till the problem arises before trying to figure out how to correct it.

 

You keep saying you don't want to send your lenses to Leica. Remember, if the lens was designed to exhibit focus shift--see Tim Ashley's thread--there's nothing wrong with the lens, so sending it to Leica won't help.

[lars_bergquist]

There are two pieces of advice circulating about how to deal with focus shift:

 

(1) Lean a little backwards after focusing with the rangefinder.

(2) Nudge the focus ring a bit towards closer focus, after focusing with the rangefinder.

 

Now as the actual amount of back focusing (in centimeters or inches or whatever) depends on the operating distance, leaning backwards is no solution. And adjusting focus by guesswork is imprecise, and slow.

 

So my solution is very simple: If a lens causes problems because of focus shift, dump it and replace it. The whole point of Leica M photography is that you know what you are doing. If you don't, you can just as well switch to a plastic wonder with compulsory autofocus, which will be far cheaper.

 

The old man from the Simple Age of Chromatic Aberration

[asmith]

Advertisement (gone after registration)

With Leica lenses, focus shift is a problem in practical photography only with those with a very wide maximum aperture - such as the earlier Noctilux and 35 mm Summilux. With these lenses the practical remedy is to use either the maximum aperture or a much smaller one where depth of focus will cover the error. Most users have been quite happy with these lenses used in this way.

More modern designs have reduced the effect so that it has little or no practical consequence.

Alwyn

[algrove]

There are two pieces of advice circulating about how to deal with focus shift:

 

(1) Lean a little backwards after focusing with the rangefinder.

(2) Nudge the focus ring a bit towards closer focus, after focusing with the rangefinder.

 

Now as the actual amount of back focusing (in centimeters or inches or whatever) depends on the operating distance, leaning backwards is no solution. And adjusting focus by guesswork is imprecise, and slow.

 

So my solution is very simple: If a lens causes problems because of focus shift, dump it and replace it. The whole point of Leica M photography is that you know what you are doing. If you don't, you can just as well switch to a plastic wonder with compulsory autofocus, which will be far cheaper.

 

The old man from the Simple Age of Chromatic Aberration

 

Your comment has bite.

 

I do not agree that "The whole point of Leica M photography is that you know what you are doing". My point in having a Leica M is enjoying it to its fullest and have images I can be proud of.

 

Since I am new to Leica and have embraced the system, I want to know about issues I keep reading about. I guess it negatively affects some people owning Leicas to know their are some negatives to an otherwise very good system. I want to learn what the negatives are and then how to overcome them in an efficient manner-not sell my Leica kit for a plastic kit!

 

Learning does not mean I do not know what I am doing. It's just that-- learning.

[jaapv]

Focus shift is not a specific Leica problem - it is a general optics problem with fast lenses.

[lars_bergquist]

Focus shift is not a specific Leica problem - it is a general optics problem with fast lenses.

 

True. In most cases film use has helped to hide the problem – as the thickness of the emulsion ensured that the sharp image was somewhere inside it, and not least because you did not see the result until later – in the case of Kodachrome, days later. And then you had often forgotten where you tried to put that focus ...

 

Also, in the days when Tri-X was rated at ASA 200, you did not quibble about image quality if a f:1.4 lens got you any kind of useable picture.

 

The old man from the days when ...

[mjh]

Since I am new to Leica and have embraced the system, I want to know about issues I keep reading about.

Focus shift has been an issue long before Leitz built the first Leica and it continues to be an issue with some lenses even today – Leica lenses, Zeiss lenses, Hasselblad lenses, you name it. Not with all lenses, mind you; with most modern lenses the effect is negligible.

[TomB_tx]

Looking at both my '62 and '69 Summicrons, I suspect that Leica (like Zeiss) has changed the way they calibrate focus because of the current rage to use very wide-aperture lenses, and usually at full aperture. Both these Summicrons show some shift, and both were calibrated for best focus at 2-3 full stops closed; so that is more the middle of the common f-stop range. Wide open, they both show the indicated focus point is at the back edge of the "sharp" band of focus; with the '62 lens a bit farther back. As the lenses are stopped down, the middle of the sharp band shifts farther towards the indicated focus point, so that the '69 is at best focus (sharpest at the indicated point) at about f4.0, and the '62 is at about f5.6. I doubt if anyone would have noticed such a shift working with film cameras.

When these lenses were made I don't think people used their lenses wide open as much as many do today, so the lenses likely were used more at the stops where they are in best focus.

Zeiss acknowledged on the Sonnar that they recently changed their procedures to set best focus wide open. I wonder if Leica has done the same, but just hasn't publicized it.

[250swb]

 

Learning does not mean I do not know what I am doing. It's just that-- learning.

 

Mmmm, so you asked your teacher at school why they 'put up with crap' if you didn't understand them? That is why you find people having a snipe at you.

 

Focus shift is often over played, simply because as you stop the lens down the increasing DOF compensates for it. The problems appear most when you have a fast lens and there is more ground to cover between wide open and where any appreciable DOF kicks in. As has been said, with film there was much more leeway. But with digital there are many more pixel peepers as well, who actually search for faults rather than try to understand lens dynamics. On top of all that nowadays you have naive people who buy lenses thinking they should always be using them wide open or nearly wide open. With people thinking they are not getting value for money because they don't understand what is going on Leica are powerless to respond, they give you the best and it still isn't good enough.

 

Steve

[ho_co]

...

Zeiss acknowledged on the Sonnar that they recently changed their procedures to set best focus wide open. I wonder if Leica has done the same, but just hasn't publicized it.

 

Tom, I don't know that Leica has changed the way they adjust lenses, but as long as I've been aware of their designs, they've designed the lenses to show accurate focus at full aperture.

 

See Puts on the topic. (I don't have the reference at the moment, but it's been since the introduction of the M8.)

 

I don't doubt that your lenses behave as you say, but Leica normally don't leave the factory that way because that's contrary to the way Leica has designed the camera.

[Guest #12]

...Robert Rennie Shannon

 

Did you read this book?

 

I read a nice memoir by him a while back.

[jaapv]

You wanted an in-depth reference - you got an in-depthe reference.

[Guest #12]

It can be generalized to all lenses with under-corrected spherical aberration; these are the lenses exhibiting focus shift. This is standard optics stuff, nothing fancy or new. Once you’ve understood what is happening...

 

Still want to know how a diagram for one spherical element explains focus shift in a lens with more than one element.

 

Let me ask it this way...say I wanted to know the focus shift (along the axis) for each aperture...maybe I want to compare this to the depth-of-field. Say I give you the corrected doublet in Jaap's book (it has all the parameters for the two lenses). If you "understand what is happening," you should be able to get a number for the focus shift at each aperture, right? How would you do it?

[mjh]

Still want to know how a diagram for one spherical element explains focus shift in a lens with more than one element.

In just the same way. You generally use single element lenses in diagrams not because you want to explain something unique to single element lenses, but to keep it simple. Think of the single element as a stand-in for any actual lens which will generally consist of several elements.

 

In this case the diagram is about a lens with under-corrected spherical aberration. The lens design doesn’t matter as long as there is some spherical aberration.

 

Let me ask it this way...say I wanted to know the focus shift (along the axis) for each aperture...maybe I want to compare this to the depth-of-field. Say I give you the corrected doublet in Jaap's book (it has all the parameters for the two lenses). If you "understand what is happening," you should be able to get a number for the focus shift at each aperture, right?

In principle, yes.

 

How would you do it?

Realistically I would buy some software for simulating lenses; it would answer this and many more questions about lenses. Of course one would need all the data about the lens which the vendor may not disclose (in sufficient detail).

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[Guest #12]

...Think of the single element as a stand-in for any actual lens which will generally consist of several elements.

 

In this case the diagram is about a lens with under-corrected spherical aberration. The lens design doesn’t matter as long as there is some spherical aberration.

...

 

My understanding is that a multi-element lens usually ends up with spherical aberration under-corrected for some zones and over-corrected for others, depending on the design. The only lens I know that behaves like your example is a Petzval portrait. But I can't find much in the way of a discussion of this anywhere.

 

Anyway, thanks

[mjh]

My understanding is that a multi-element lens usually ends up with spherical aberration under-corrected for some zones and over-corrected for others, depending on the design.

What do you mean by “zones”?