focus shift - please explain it

[135format]

I repeatedly see talk of focus shift in some lenses but I am curious why adjusting aperture actually changes the focus point. Can someone explain the theory of why it happens and does it happen in all lenses to a lesser degree. And if not why does it happen in some and not in others.
Thanks

 

[john_s]

You've probably seen the simple diagrams showing a lens focusing an object at a distance, with the lines of light meeting at a point where the film is. The film has to be where the lines meet at a neat point.

The reality is that the lines don't neatly meet at a point. They don't meet exactly. They approach each other in a bundle called a caustic (probably an archaic term). The designer chooses a point for focus that gives the best compromise. The lines from the wider part of the lens (i.e. widest aperture) meet at a different spot, so that as the aperture is closed the optimum spot is different. The degree of this is different for different lens designs, and is probably most exaggerated at very close focusing distances which is why tests at minimum focus distance might not be the best indication of general focus correction that's needed. The effect is much more pronounced with very wide aperture lenses.

There was a diagram on Erwin Puts's site but I can't find it now. I'll have a look for a diagram.

Edit: http://toothwalker.org/optics/spherical.html

 

[BobYIL]

http://mansurovs.com/what-is-focus-shift

 

[f16sunshine]

Focus shift is caused by the internet.

😀

 

[135format]

http://mansurovs.com/what-is-focus-shift

That's cleared it up. So basically it's due to a poorly corrected/designed lens. Or a lens designed for one purpose which doesn't suit another. i.e. it wasn't designed to be used that wide open or that closed down but they stuck an aperture in there which could be used like that.
And basically I guess all lenses will suffer to an extent unless they perfect correction at all apertures and all distances which is unlikely.

 

[Richard G]

Focus shift is caused by the internet.

😀

Amen. I doubt I would have known of it without the internet, and the C Sonnar is my favourite lens.

 

[user237428934]

Focus shift is caused by the internet.

😀

😕 My Sonnar is unsharp at f1.5. I think it's not too bad to know the scientific reason for that.

 

[rxmd]

😕 My Sonnar is unsharp at f1.5. I think it's not too bad to know the scientific reason for that.

Knowing the scientific reasons for something is fine, I guess it's the exaggerated bogeyman attention that focus shift has been getting that is the problem.

 

[Gabriel M.A.]

It's also Leica's fault.

 

[Keith]

It's also Leica's fault.

No .... Nikon's fault surely! 😀

 

[rxmd]

No .... Nikon's fault surely! 😀

I thought it was Obama's fault.

Or San Andreas' fault.

 

[Sejanus.Aelianus]

Can I blame it on my Tomcat?

 

[BobYIL]

That's cleared it up. So basically it's due to a poorly corrected/designed lens. Or a lens designed for one purpose which doesn't suit another. i.e. it wasn't designed to be used that wide open or that closed down but they stuck an aperture in there which could be used like that.
And basically I guess all lenses will suffer to an extent unless they perfect correction at all apertures and all distances which is unlikely.

"poorly corrected/designed lens" to a certain extent would rather be unfair; as what we call as focus shift is actually caused by residual spherical aberrations; i.e. corrected for most of the aperture range except the first one or two. Such lenses usually exhibit quick drop in resolution toward edges at wide apertures. The C-Sonnar, 35 Summilux pre, Summicron 35 V.1 (8-elements) just to name few.

However focus shift can be observed also with some well corrected lenses exhibiting excellent characteristics wide open: The former version of the Summilux 35 asph. for example. Especially the use of hi-res sensors began to accentuate such faults which were overlooked in the film era as many lenses do suffer of focus shift -except the apochromatics- however in negligible amounts.

BTW, some of the lenses having residual spherical aberrations could be the ones having the most "character"; especially for B&W. Especially the ones I mentioned above have always been and still are among the sought-after lenses for many.

 

[jarski]

Roger and Frances site describes focus shift in Sonnar review.

been also cautious (for years) about this "horrid" problem, but reading the article realized how trivial it actually is, also on digital. Sonnars (except very early ones) are optimized for f1.5, where the DOF is narrowest. once you stop down, DOF increases and hides practical shift effect. I doubt meaningful focus issue like "nose in focus, eyes unfocused" will not happen in f2.8 and beyond.

 

[user237428934]

Sonnars (except very early ones) are optimized for f1.5, where the DOF is narrowest.

What is a very early one? Mine is not even 2 years old and it's optimized for 2.8

 

[jarski]

What is a very early one? Mine is not even 2 years old and it's optimized for 2.8

hmm strange, things have changed again it seems. R&F comment: "... Later lenses -- all after early 2007 -- were therefore calibrated for sharp focus at f/1.5"

 

[ssmc]

Digital made focus shift easier to recognize because few people enlarged their 35mm film images to the same sort of poster-size equivalents that 1:1 viewing on a monitor allows with nothing more than Ctrl+Alt+0.

Of course once you become aware of it and start shooting test scenes and such to see how bad (or otherwise) it is, you're trapped on that merry-go-round (I'm not saying it's not important - for some lenses focus shift across certain aperture ranges can be a show-stopper).

One lens I have (the MD 45/2 Rokkor) that is quite mediocre on the ol' "topo maps taped to a wall" sharpness/centering test has given me some of my absolute favourite landscape images in actual use (in this case not because of focus shift, but field curvature at close range)

IMO the important thing is to know how a particular lens behaves ahead of time and not be caught out by surprise (either on digital or film). If it's characteristics don't suit the way the you shoot, use something else!

Scott

 

[Freakscene]

been also cautious (for years) about this "horrid" problem, but reading the article realized how trivial it actually is, also on digital. Sonnars (except very early ones) are optimized for f1.5, where the DOF is narrowest. once you stop down, DOF increases and hides practical shift effect. I doubt meaningful focus issue like "nose in focus, eyes unfocused" will not happen in f2.8 and beyond.

The difficulty with focus shift in a lens like the 50/1.5 Sonnar is that if the lens is adjusted for accurate focus at f1.5, as you stop down the depth-of-field increases, but it shifts backwards sufficiently that the point you focus on is in front of the plane of acceptable sharpness - and therefore unsharp. If you focus on an eye you end up with a sharp ear. The increase in depth-of-field in the 50/1.5 Sonnar does not overcome the focus shift until ~f4.5 as this test shows quite clearly: http://photo.net/leica-rangefinders-forum/00aKSW

Focus shift is not imaginary, or an invention, but it's also not fatal. Understanding it can facilitate you being able to better use of lenses that display it, and make photos that look like you want them to, which is the point after all.

Marty

 

[JHutchins]

An interesting discussion of the relationship between focus shift, spherical undercorrection and bokeh here from Zeiss's director of lens design:
http://www.zeiss.com/c12567a8003b8b6f/embedtitelintern/cln_35_bokeh_en/$file/cln35_bokeh_en.pdf

 

[ampguy]

Anyone who has used a stop down (metering, DOF preview, etc.) SLR, and observed critical focus, will have seen it.

So those who haven't seen it, have either a) bad eyesight, b) wide open SLR or RF use only, or c) have seen it, but haven't been bothered by slightly out of focus subjects.