Focus Question

[JacquesBalthazar]

Advertisement (gone after registration)

I agree with HC, Jaap and others on the fact that we are noticing issues now that we simply did not "focus" on in the film age, mainly due to the magnification ratio at which we inspect images and the ease with which such inspection can take place.

 

But, coming from the "middle of the road" DSLR world (Nikon D100 and KM 7D) and using similarly fast fixed focal length lenses as the ones I use with the M8, I must confess I never noticed this sort of focus shift, which immediately leads one to wonder if a rangefinder is a suitable focusing tool in this new age.

 

The coupled rangefinder in effect "simulates" the focusing process, while the SLR allows focusing the image actually transmitted by the lens (supposing mirror, AF sensors, captor and prism all in correct alignment). A rangefinder system cannot "see" focus shift, while a SLR can, and allows correction.

 

That is quite an existential question, taking into account our investments in this...

 

That said, have not tested my combo for focus shift, but do have the impression of a higher rate of slight focus errors than with my M6. Again this impression could very well be blamed on the observation method (the 1:1 mouse click) rather than on reality. And most of these "flawed" images translate into perfectly acceptable prints at usual print sizes.

[nscali]

HC,

Your comments make me feel at ease.

Also, because I am a bit nuts, I went to my Leica dealer today at Lunchtime.

He had 2 New M8's in stock and graciously allowed me to perform my tests on these bodies.

Well, guess what? Exactly the same focus shift issues as with my body - at f4.0 and beyond.

Nicky

[wattsy]

A rangefinder system cannot "see" focus shift, while a SLR can, and allows correction.

 

This is not the case. The SLR focussing system is just as susceptible to the focus shift problem as is the rangefinder. SLR lenses are usually focussed wide-open and then stopped-down automatically as the shot is taken. As far as I know there is no readjustment of focus as the lens is stopped down.

[wattsy]

Puts says the solution to this non-problem is to use the Leica lenses at or near full aperture.

 

Are you sure this is what Puts says? Restricting yourself to shooting at or near wide-open is a ludicrous solution - there are innumerable situations in which any serious photographer will want to stop down for creative reasons.

 

I think it is important to be clear about the difference between a minor (partly 'theoretical') focus shift that I am sure all lenses have to some extent and the large shifts that some users have been observing recently with some lenses. The latter surely cannot be explained away or excused by simply suggesting that such lenses are designed to be used only at the widest apertures.

 

I'm interested in what LFI will have to say on this matter. If the problem is essentially insuperable (existing because of the flat 'zero thickness' sensor plane compared with the traditionally slightly curved and thicker film plane) then Leica will need to rethink the whole idea that the M8 is compatible with almost all existing M lenses. I can live with IR filters but being expected to accept that £1.5K lenses can only be used reliably at F1.4 to F2 is a compromise too far.

[jaapv]

This is not the case. The SLR focussing system is just as susceptible to the focus shift problem as is the rangefinder. SLR lenses are usually focussed wide-open and then stopped-down automatically as the shot is taken. As far as I know there is no readjustment of focus as the lens is stopped down.

 

Focus shift on SLR lenses is less pronounced because of the longer register distance, I think.

[wattsy]

Focus shift on SLR lenses is less pronounced because of the longer register distance, I think.

 

Sounds like a reasonable theory.

 

One thing is for sure - I've never encountered focus shift with any of the SLR systems and lenses that I have owned.

[carstenw]

Advertisement (gone after registration)

I am wondering how much of the sharpness loss can actually be seen on A3 printouts, or even A2 printouts. I will test my lenses at some point and see if I can spot something, but given that many printers use 240, 300 or 360dpi, I cannot really imagine that this sharpness loss is even visible. Nor should it be visible in web-sized reductions. In other words, we may all be fretting about something which in the end amounts to nothing. I am interested in hearing counter-points.

[kobold]

Howard: excellent summary, thank you. And, wattsy, your points are well made too, IMHO.

 

Looking at images at 100% on a 30" screen has a lot to do with this concern, for sure. The printed image looks different, also for sure. Interesting thread. cheers, kl

[Woody Campbell]

This is not the case. The SLR focussing system is just as susceptible to the focus shift problem as is the rangefinder. SLR lenses are usually focussed wide-open and then stopped-down automatically as the shot is taken. As far as I know there is no readjustment of focus as the lens is stopped down.

 

You are quite right. Canon does focus with lenses wide open so focus shift happens when they stop down to shoot. I have followed endless threads to the effect that "My xx mm f xx backfocuses. This is ridiculous in a camera that costs $x,xxx." on the Canon-oriented boards (my other gear is a 1DsII, 5D and a closet full of lenses). The usual cause is focus shift and focus and recompose.

 

Wider manufacturing tolerenaces in the Canon world lead to greater sample variability among lenses - it's common to send your body and lenses to Canon USA to have focus adjusted so they work together properly (I've done so). Turn around times at Canon USA are quite good and of course they've chosen to address this issue in their new offering.

[ho_co]

Are you sure this is what Puts says? Restricting yourself to shooting at or near wide-open is a ludicrous solution - there are innumerable situations in which any serious photographer will want to stop down for creative reasons.

Ian--

Thanks for holding me in bounds. Again I overstated, but not by much. At the link given above, Erwin says "The other side of the medal is the fact that focus shift might be more noticeable when stopping down: another argument not to stop down too much when you need critical sharpness at the focus point."

 

One thing is for sure - I've never encountered focus shift with any of the SLR systems and lenses that I have owned.

That's what worried me most about the focus issue discoveries with the M8: If the same people designed the R lenses, why didn't we see the same outcry when the DMR appeared?

 

Jaap's answer is interesting because it would solve the problem.

 

Why don't we see the same problem with the R-D1? Perhaps because of smaller sensor; certainly because of thicker sensor cover glass with AA filter.

 

These are the same lenses we've been using for years; they haven't changed. Carsten is right: Make big prints and look at them from a 'normal' distance, not from 12 inches. Jacques (welcome to the forum, Jacques) said the same thing: "And most of these 'flawed' images translate into perfectly acceptable prints at usual print sizes."

 

We're literally bumping against the edges of the possible here. Remember, David Adamson says the M8 delivers the best images he has seen from any 35mm format camera. A lot of other photographers concur, but they're not making tests of focus shift; they're just taking pictures.

 

This isn't to denigrate those who have made us aware of the issue, particularly Tim, without whose persistence we would still be wondering what was up. I'm also a nit-picker. But let's face it, guys and gals: The M8 and these lenses are so good that they give us the nits we want to pick. Drop an anti-aliasing filter into the chain and you've got just another digital camera.

 

And remember, a few years ago Leica said you couldn't make a digital M. I think they've done a pretty good job for the impossible.

 

--HC

[tashley]

I shan't speak the unspeakable, but if I were to do so, here's what I would say:

 

1) We've always known that Leica lenses were amazingly good wide open. Other manufacturers' lenses don't come close.

 

2) We've always known that as you stopped down, the others' lenses improved and by about f/8 were comparable to the Leica lenses.

 

There's no question about it; those are simply facts.

 

3) Now with the M8 we discover a focus shift caused by spherical aberration that we hadn't noticed before.

 

Put the two together and you have one and the same thing. Zeiss says the same about their 50/1.5 C Sonnar's adjustment (http://www.leica-camera-user.com/digital-forum/21452-adjusting-zeiss-sonnar.html). Again:

 

Leica lenses are perfectly in focus wide open (our experience in point 1 and explained by Puts as cited above).

 

As we stop them down, focus shifts. It does the same with other manufacturers' lenses as well (see Puts again) BUT THE OTHERS HAVEN'T OPTIMIZED THE LENSES AS LEICA HAS. That is, the focus shift takes the Leica lenses slightly out of focus because they were perfect before. But with others' lenses, since they weren't adjusted for accuracy wide open, the focus moves INTO the rangefinder's set distance (see Puts again).

 

All lenses shift focus as you stop them down. But non-Leica lenses were adjusted with smaller apertures in mind, so we just see them improving as we stop them down. We aren't aware of the focus shift because the lens is getting better.

 

We didn't notice the focus shift with film because film isn't flat and because we couldn't blow film up to 400% at the click of a mouse. Ever wonder why it's so hard to get good scans from 35mm transparencies? It's because the film wasn't flat and we sat ten or fifteen feet from the screen, but now we're trying to extract everything and can't find it. Same thing we're seeing in digital.

 

Okay, that's my theory. It's made up out of my sheer brilliance and imagination and a little bit of whole cloth. In summary: Leica lenses work so well because they're in focus wide open, and as the focus shifts, it stays "within depth of field." But depth of field doesn't exist except in the brain, and we aren't looking at depth of field when we enlarge an image on a computer screen. The proper test is to make a print and evaluate it as an artist, not a technician.

 

(In the Zeiss thread I mentioned, the thread starter said he had asked Zeiss to adjust the focus of the 1950's style lens to work more like the Leica lenses, and they gave him exactly this explanation of what would happen with the focus adjustment. Puts says that Zeiss claims to have less focus shift but that they actually reach this claim by redefining the 'in-focus' plane to be thicker. [i suppose that has to do with how they locate the focus plane between the point of highest contrast and the point of highest resolution.])

 

Puts says the solution to this non-problem is to use the Leica lenses at or near full aperture. The "Form Follows Format" article in LFI a few months ago also said if we're not using the lenses wide open, we're missing a great artistic tool. Phrased completely differently, but a clear hint in the same direction.

 

I think we've got the world's best lenses and ought to go out and shoot them and forget about focusing on test targets. (Heaven forfend! No tests? What am I doing on the Leica Forum? )

 

Tim--I think that's all in Puts, but he puts it more genteelly. I'm also interested in LFI's approach; like Puts, they tend to tell it straight but a little hidden behind a veil.

 

--HC

 

I think you have the whole truth and nothing but here HC, as usual: but I would add a couple of small flies to the ointment:

 

All my Leica lenses perform better when stopped down a little, however awesome they are wide open. For most of the luxes and crons, stopping down one or two stops is something we'd all like to do from time to time and on most it is not a problem - but there are enough people out there for whom this creates an OOF situation that it kind of does matter.

 

Those people need to observe their lenses' behaviour carefully because as we now know, this OOF zone is around the centre of the file, around 1/3 of it. That means that recomposing to a third will get your subject in focus BUT it also means that front focussing on purpose will put the edges of the frame OOF! In other words, we need new techniques.

 

Also, the high base ISO of the M8, and the need for some DOF in certain shots, means that using these lenses wide open is just not an always solution, nor should it be.

 

To summarise: if a photographer is one of those with focus shift of a degree that outstrips increases of DOF, (s)he needs to learn to front-focus if the subject is central, or to recompose to a third, or to go straight to F8 and maybe higher ISO.

 

I do believe that this is practical, but it is a new technique and whilst it is now second nature to me, it takes a while to get it right!

 

Best

 

Tim

[jaapv]

Are you sure this is what Puts says? Restricting yourself to shooting at or near wide-open is a ludicrous solution - there are innumerable situations in which any serious photographer will want to stop down for creative reasons.

 

I think it is important to be clear about the difference between a minor (partly 'theoretical') focus shift that I am sure all lenses have to some extent and the large shifts that some users have been observing recently with some lenses. The latter surely cannot be explained away or excused by simply suggesting that such lenses are designed to be used only at the widest apertures.

 

I'm interested in what LFI will have to say on this matter. If the problem is essentially insuperable (existing because of the flat 'zero thickness' sensor plane compared with the traditionally slightly curved and thicker film plane) then Leica will need to rethink the whole idea that the M8 is compatible with almost all existing M lenses. I can live with IR filters but being expected to accept that £1.5K lenses can only be used reliably at F1.4 to F2 is a compromise too far.

 

 

I may be a bit of the devils advocate here, but when we look at it from a creative point of view, there is a lot to be said for regarding the middle apertures as less important. In creating a photograph one must decide for either shallow or deep DOF. It makes sense to shoot as wide open as possible for the one and on -or slightly beyond- the diffraction limit for the other. In film camera's the main use of the middle apertures was to adjust exposure to the film used. On the M8? Well, we can go down to 1/8000th and adjust ISO up to 2500/3200 so the need is not as pressing.

That kind of makes these apertures in the middle bland and not very interesting. Exact plane of focus wide open, and focus shift lost in DOF at f 11.0,not a mad idea maybe?

[kobold]

Jaap: precisely.

[JacquesBalthazar]

SLR lenses are usually focussed wide-open and then stopped-down automatically as the shot is taken. As far as I know there is no readjustment of focus as the lens is stopped down.

 

 

That is absolutely true, and that obvious fact escaped me completely. Guess you could focus the "real image" by stopping down using DoF preview, or R lenses on C bodies....

[ho_co]

Again, Jacques, rhe further you stop down the greater the DoF. My guess is that the autofocus designs make use of that to cover some of the error.

 

It would be possible to program a focus correction into the lens ROM in an AF camera, but I think we would have heard about it if someone were doing that.

 

And although stopping down to taking aperture in order to get the actual focus is a good idea, just remember how dark it gets in the finder when you do that!

 

Tim--that curvature is a bug-a-bear, as is the fact that a couple people seem to have lenses that don't exhibit the more general behavior. As for the first, and not to diminish its importance, remember the picture of Venice you showed with the question "What's wrong with this?" You had to ask twice before someone identified the problem; and another person said in his opinion the picture was already good enough for a double-page spread.

 

Also, because I am a bit nuts, I went to my Leica dealer today at Lunchtime.

Nicky, we're all a bit nuts with you. That's why we like this camera.

 

--HC

[gman]

 

In summary: Leica lenses work so well because they're in focus wide open, and as the focus shifts, it stays "within depth of field." But depth of field doesn't exist except in the brain, and we aren't looking at depth of field when we enlarge an image on a computer screen. The proper test is to make a print and evaluate it as an artist, not a technician.

--HC

 

Howard-

 

I don't disagree that we sometimes look to quantify everything about the camera or lenses without actually looking at photographs, but all the issues I've had with testing my 50/1.4 ASPH on a focus chart were at wide open aperture.

 

So far, this is the one lens I have giving me backfocus. I'm not too happy with that.

 

jeff

[ho_co]

Jeff--

According to all I've read, it should be properly in focus wide open. There's a problem; could be with the rangefinder, could be with the lens.

 

And you're right, it doesn't do any good for someone to tell you your back shouldn't hurt when you know you've got a backache.

 

--HC

[wattsy]

I may be a bit of the devils advocate here, but when we look at it from a creative point of view, there is a lot to be said for regarding the middle apertures as less important....

That kind of makes these apertures in the middle bland and not very interesting. Exact plane of focus wide open, and focus shift lost in DOF at f 11.0,not a mad idea maybe?

 

I understand your point but it is a pretty desperate argument for £2k lenses that are great at F1.4-2 and F8-11 but which might be average or poor in-between. I don't agree with the premise anyway - there are many situations where F4 or F5.6 might be the preferred creative choice. Going wide-open just for the sake of an often cheesy looking 'razor thin DOF' isn't always the right approach in my opinion (and suggests a primary interest in effects rather than subject).