Mark Dubovoy's Leica S2 Review

[David Grover / Capture One]

I have started this thread as I wish to point out Mark's innacurate description of Leica's Autofocus system and also of ours.

Ill start with this quote...

An important thing that few photographers are aware of is that autofocus systems are not continuous; they work in zones. The Leica S2 has many more zones of focus from close up to infinity versus the competition. In other words, the S2 can focus more accurately at a specific distance.

This is absolutely not the case with Leica, Hasselblad, Canon or Nikon for that matter.

With a Phase Detection Auto Focus system we can measure any distance from near to infinity.  There is a tolerance on the last measured focus error but this has still nothing to do with zones.

The only 'zoned' cameras were small compact cameras which used a different system to measure the approximate distance and then set the lens to one of a certain number of zones.  Either a couple of zones or a few more.

Digital Compact cameras and I would guess cameras such as the Canon 5DMKII (in Live View mode) use contrast detection to determine focus.  NOT zones!

If Mark had intended to mean Leica's last measured value was a higher tolerance than Hasselblad, Nikon and Canon then I would assume Mark would know our tolerance, Nikon / Canon's tolerance and finally Leica's tolerance.   Please Mark, tell me your sources for these values - I have never seen these published anywhere?.  I have not be asked by you for any information on the Hasselblad Auto Focus system.


Next....

Suppose I want to focus on a tree at about 200-300 feet. Using autofocus with a Medium Format camera, the system often goes into the infinity focusing zone. But, the tree is not at infinity. Add to this a high resolution MF back like a P65+, and the tree is guaranteed to look out of focus. Finally, focus shift could make matters worse if it goes the wrong way.


To be absolutely clear, this is not the way the Phase One 645DF works, or the Hasselblad H System.


Furthermore...

Most lenses suffer from focus shift as you change the aperture. This is a problem, since autofocus systems focus with the lens wide open. Hasselblad makes a correction (in the H4 series) with a software model. The camera figures out the lens, the focusing distance and aperture, and just before the exposure it applies an algorithm that re-focuses the lens.  The problem is that the algorithm is based on one "ideal sample" lens. Unfortunately, each real lens differs somewhat from the “model lens”, so the system is not totally accurate. The PhaseOne lenses in my experience seem to have less focus shift. As far as I know, PhaseOne does not make any corrections to compensate for focus shift; PhaseOne users simply have to live with it.

We have been making this correction since the H1 was produced.  Again, please check your sources, or actually request the information.

How do you know by how far the model lens deviates from a produced lens?  Again this is not published information and for what it is worth the model lens is actually very close to the produced lens, close enough for the focus correction to work and to be of value.

Have you made specific focus shift tests on the Phase One and Hasselblad lenses?  It would not be possible to test the Hasselblad lenses anyway as the focus correction cannot be switched off.  So please tell me how you measured this compared to Phase One?

In conclusion, as usual a review on Luminous Landscape which mentions Hasselblad contains inaccurate information.  I am more than happy to answer questions, supply technical data or even proofread articles for technical aspects.  You only have to ask.  

Best Regards,



David

[tho_mas]

who really gives anything on Mark D.'s pointless essays (only containing claims based on half knowledge)...?

[eronald]

Re. AF, the question is not only what can the camera measure, it is also what can the lens be set to by the camera's AF. In other words the design of the actuator on each lens comes into play.

I do think that Mark's review is a paper which the referees would like revised ....when I was organizing conferences, we would simply ask an author to take the referres's views into account, and the paper would usually get improved.


Edmund

[dfarkas]

I can't speak to the accuracy of the writer's technical reasons WHY he feels that the autofocus on the Leica S2 is faster and more accurate than other camera systems he has owned or currently owns, but the point is that he still observed that the AF on the S2 IS faster and more accurate.

David

[PaulSchneider]

... it appears that Lula sometimes is quite easy to understand nowadays:

first: one has just to bear in mind that "all is subjective", including autofocus mechanisms and dynamic range ...

and remember a few rules of thumb so that one can find one's way around here more easily:

Phase One > Hasselblad (the color is so awful)
Leica > Japanase 35mm (eaten alive)
Alpa > other technical cameras (sharper, better)
and not to forget: dont believe Puts, he should stick to his leica glass observations ...

but, all is subjective ...

And if you agree you can join and participate in a unique workshop where Phase One sponsors camera gear,
where you can try out some Alpas and Leicas ...

but don't listen to me, I think I'm being subjective  

[eronald]

I can't speak to the accuracy of the writer's technical reasons WHY he feels that the autofocus on the Leica S2 is faster and more accurate than other camera systems he has owned or currently owns, but the point is that he still observed that the AF on the S2 IS faster and more accurate.

David

My feeling of said speed is glacial, as *my* emotions seem to have bonded with my Nikon. Oh, and by the way, so far I've only met one MF camera which could focus almost decently and it's a plastic fantastic Hasslemebad.  They have this model which does some vodoo and somehow changes focus as you wave it around. I even once tried using a camera called flexoreil or something, but I never got it to focus, and then suddenly !woosh! it was not for sale, I guess Leaves drop when autumn comes. Another camera I've used , the Mummykam, kind of wakes up focuses a bit and goes to sleep again and does so over and over - I think I can hear it snore, cough, stutter, snore. And then there is the Nikon! It says Yes! and it's sharp.

My opinion of this whole fiasco is that editorial discretion should have prevailed, someone should have told Mark to polish his piece *before* it went out, as he's a very smart guy he would have done so and controversy ,broken eggshells and ruined suits would have been avoided. However, a rewrite now might still make sense. And A nice brick wall photograph with brands P65, M9, S2, D3x and 5D2 would make my day. Please stick the $100 50mm/1.8 on the Nikon and Canon for the test.


Edmund
Edmund

[michael]

The article is in fact in error on the matter of autofocus.

I accept responsibility for not catching it. I prepared it for publication but did not read it carefully because I was in transit yesterday.

I have contacted Mark to see how he wants to handle editing the piece.

I apologize for the error.

Michael

[ndevlin]

...that said, Mark's 'misstatement' correlates precisely with my real-world experience of the systems in question.  

And yes, it is all subjective. Anyone who claims to be able to reduce anything of true significance in photography to fixed mathematical certainty is either a liar or a dullard, or both.  So much is dependent upon the user's abilities and methods that any claims to 'universal' measures are essentially meaningless. To be honest, that's what makes it interesting.  There's room for them all: the fanboys, the measurebators and the real photographers who take the time to share their subjective experiences.    

- N.

[rogan]

On top of that I have never seen a mf camera with more CA than the Leica. It's as bad as the canon when using one of their cheaper lenses. And if you are going to review something, review it. Mention the tethering is 3 hours and 9 minutes between images until you hit the buffer, then 6 hours each. Sorry kids the S2 had soooo much potential. I know of not one pro who makes his money shooting(not counting trust fund kids or seal) that has purchased one.
And he really used those pictures to test it?

[alastairbird]

On top of that I have never seen a mf camera with more CA than the Leica. It's as bad as the canon when using one of their cheaper lenses. And if you are going to review something, review it. Mention the tethering is 3 hours and 9 minutes between images until you hit the buffer, then 6 hours each. Sorry kids the S2 had soooo much potential. I know of not one pro who makes his money shooting(not counting trust fund kids or seal) that has purchased one.
And he really used those pictures to test it?

I'm with Rogan. I shot a few frames with a production S2 in late March and I was alarmed at the CA.  We're talking CA on the highlight on a black wristwatch placed in the middle of a frame shot in open shade - somewhere I've never seen CA before - and there wasn't any in the frame from the GF1 we were shooting with as well.  The AF was fast and incredibly accurate, however...

[Ben Rubinstein]

Question: How can you have accurate AF when the moment you recompose you've just thrown off the plane of focus from the point it was focused on? Incredibly accurate AF in all light and all lenses is what my 5D has. But like these systems it's only the center point. Moment you recompose, the plane of focus is shifted and you are relying on DOF to hold the focus for you. If there was ever anything that digital allowed us to see it was how false the perception of DOF is, the plane of focus is sharp, everything else is just less unsharp at any fstop until you hit diffraction and it evens out a bit.

I've never even seen the new 'blad and my 1Ds mkIII after 3 trips to canon still won't focus accurately with the off center points, I've no horse in this race. I do know however that focus is only ever going to be as accurate as the amount of recompose you are forced to do, especially the more megapixels you have to show focus differences.

A single point AF camera can only ever be described as 'incredibly accurate' if you always frame with the subject in the middle or stop down to the point that the accuracy is moot anyway.

[David Grover / Capture One]

I can't speak to the accuracy of the writer's technical reasons WHY he feels that the autofocus on the Leica S2 is faster and more accurate than other camera systems he has owned or currently owns, but the point is that he still observed that the AF on the S2 IS faster and more accurate.

David

If that is mark's conclusion then of course I am not qualified to argue against it.

What I am more interested in is the accurate portrayal of AF systems in general and the misinformation regarding Hasselblad.

[David Grover / Capture One]

The article is in fact in error on the matter of autofocus.

I accept responsibility for not catching it. I prepared it for publication but did not read it carefully because I was in transit yesterday.

I have contacted Mark to see how he wants to handle editing the piece.

I apologize for the error.

Michael

Thanks Michael. If Mark needs the correct information on Hasselblad he can always ask.

[David Grover / Capture One]

Question: How can you have accurate AF when the moment you recompose you've just thrown off the plane of focus from the point it was focused on? Incredibly accurate AF in all light and all lenses is what my 5D has. But like these systems it's only the center point. Moment you recompose, the plane of focus is shifted and you are relying on DOF to hold the focus for you. If there was ever anything that digital allowed us to see it was how false the perception of DOF is, the plane of focus is sharp, everything else is just less unsharp at any fstop until you hit diffraction and it evens out a bit.

I've never even seen the new 'blad and my 1Ds mkIII after 3 trips to canon still won't focus accurately with the off center points, I've no horse in this race. I do know however that focus is only ever going to be as accurate as the amount of recompose you are forced to do, especially the more megapixels you have to show focus differences.

A single point AF camera can only ever be described as 'incredibly accurate' if you always frame with the subject in the middle or stop down to the point that the accuracy is moot anyway.

Excellent post.

Placing AF sensors out wide on an MF camera to a useful position would make them pretty hopeless, especially with shallow DOF.

[ErikKaffehr]

Hi,

Just because Mark was mistaken about zones in autofocusing it is quite possible that S2 autofocusing is simply better. It seems that Hasselblad has spent a lot of effort, behind the scenes, on improving real world AF-performance. They may deserve some credit.

Best regards
Erik

The article is in fact in error on the matter of autofocus.

I accept responsibility for not catching it. I prepared it for publication but did not read it carefully because I was in transit yesterday.

I have contacted Mark to see how he wants to handle editing the piece.

I apologize for the error.

Michael

[ErikKaffehr]

Hi,

Photography, technically seen, is applied physics, with some mathematics blended in. So scientific approach does apply to photography.

The art of photography has a lot to do with perception less with "science", but science is still there.

Best regards
Erik

...that said, Mark's 'misstatement' correlates precisely with my real-world experience of the systems in question.  

And yes, it is all subjective. Anyone who claims to be able to reduce anything of true significance in photography to fixed mathematical certainty is either a liar or a dullard, or both.  So much is dependent upon the user's abilities and methods that any claims to 'universal' measures are essentially meaningless. To be honest, that's what makes it interesting.  There's room for them all: the fanboys, the measurebators and the real photographers who take the time to share their subjective experiences.    

- N.

[fredjeang]

Question: How can you have accurate AF when the moment you recompose you've just thrown off the plane of focus from the point it was focused on? Incredibly accurate AF in all light and all lenses is what my 5D has. But like these systems it's only the center point. Moment you recompose, the plane of focus is shifted and you are relying on DOF to hold the focus for you. If there was ever anything that digital allowed us to see it was how false the perception of DOF is, the plane of focus is sharp, everything else is just less unsharp at any fstop until you hit diffraction and it evens out a bit.

I've never even seen the new 'blad and my 1Ds mkIII after 3 trips to canon still won't focus accurately with the off center points, I've no horse in this race. I do know however that focus is only ever going to be as accurate as the amount of recompose you are forced to do, especially the more megapixels you have to show focus differences.

A single point AF camera can only ever be described as 'incredibly accurate' if you always frame with the subject in the middle or stop down to the point that the accuracy is moot anyway.

Exactly!
Thanks for this precision Ben.

-----------

Some thoughts however: it is curious how in those days photographers have focussing problems with MF and have to relay on DSLR ultra high tech to get things in focus...

It is like those new models who can not handle a 2K HMI and start to cry...

I wonder what would think Avedon of all that in his tumb.

But those where "real men" with the correct proportion of testosterons to handle any kind of situation and therefore knew how to focus properlly with unperfect gear, didn't they?

Do you have focus issues and need a dslr to fix them?

mmm...wired times really, wired times.

Ps precisions: my use of "real men" has nothing sexist but an image, as I've also noticed that women's Lu-La forum never complain about gear specs, or editorial errors, or
have focusing issues with their devices. They take pictures and generally pretty well. Who wears the trousers? But I knew that we still have a lot to learn from the ladies.

[guyharrison]

I had a very nice demo at a Capture Integration event where I shot the S-2 for an hour or so, with a model and also landscape/cityscape scenes. A Phase One DF was also present and I used it, more than I expected.

There was a lot to love about the Leica (weather sealing, quality, handling, simplicity in shooting--regarding all of which Mark was dead on), but we need to get real.

The Leica is not faster or more accurate at focus than the Phase One, period. They are very similar in focus speed. From my demo, I would say that you would never get a shot with the Leica that you would miss with the Phase One due to either speed or accuracy issues. Both focus very well and very quickly in dim light. I was surprised at this because Leica was claiming that their focus was significantly and substantially better. It is better than the older Mamiya bodies, and better than my Contax 645 (in speed but not in accuracy), but not so as against the newest-generation Phase.

This does not even account for what happens when you recompose, as has been thoroughly covered in other posts. I can't comment on Hassy because one was not there for my use. However, I believe that, in theory, their method for compensating for re-composition might actually be the best focus system yet for a format limited to one-center-point focusing systems.

That being said, I do not see where focus-and-recompose has ever lead me, with any system, to have chronically unsharp or unacceptable pictures. This is true even with 35mm systems using 85mm 1.4 lenses wide open. Of course, no comparison at all between Canikon AF and any medium format system which are still completely primitive by comparison.

[feppe]

Some thoughts however: it is curious how in those days photographers have focussing problems with MF and have to relay on DSLR ultra high tech to get things in focus...

It is like those new models who can not handle a 2K HMI and start to cry...

I wonder what would think Avedon of all that in his tumb.

But those where "real men" with the correct proportion of testosterons to handle any kind of situation and therefore knew how to focus properlly with unperfect gear, didn't they?

Do you have focus issues and need a dslr to fix them?

mmm...wired times really, wired times.

[political correctness snipped]

Here's why focusing is a whole different ballgame these days. Nothing to do with testosterone or lack of focusing skills in modern photographers.

[fredjeang]

I had a very nice demo at a Capture Integration event where I shot the S-2 for an hour or so, with a model and also landscape/cityscape scenes. A Phase One DF was also present and I used it, more than I expected.

There was a lot to love about the Leica (weather sealing, quality, handling, simplicity in shooting--regarding all of which Mark was dead on), but we need to get real.

The Leica is not faster or more accurate at focus than the Phase One, period. They are very similar in focus speed. From my demo, I would say that you would never get a shot with the Leica that you would miss with the Phase One due to either speed or accuracy issues. Both focus very well and very quickly in dim light. I was surprised at this because Leica was claiming that their focus was significantly and substantially better. It is better than the older Mamiya bodies, and better than my Contax 645 (in speed but not in accuracy), but not so as against the newest-generation Phase.

This does not even account for what happens when you recompose, as has been thoroughly covered in other posts. I can't comment on Hassy because one was not there for my use. However, I believe that, in theory, their method for compensating for re-composition might actually be the best focus system yet for a format limited to one-center-point focusing systems.

That being said, I do not see where focus-and-recompose has ever lead me, with any system, to have chronically unsharp or unacceptable pictures. This is true even with 35mm systems using 85mm 1.4 lenses wide open. Of course, no comparison at all between Canikon AF and any medium format system which are still completely primitive by comparison.

Yes,

But the reality with those 35mm AF lenses is that the focus ring is so bad that they must be accurate with their AF system, and they are.

Then when one uses a proper manual focus prime, the precision of the mechanism allow the couple brain+hand to be extremely precise, and with training the natural correction when recomposing is done with a very high level of accuracy.
All you need are lenses that are built such a way that become if I might say, part of the body.

Then, the precision is IMO superior to any automatised system in most situations, and speed and accuracy can be very high. If one experiments another reality is for lack of practise in manual focusing.
A 6th sense as to be trained, but with the right lenses, not with AF lenses.

Also, it is true that with the digital resolutions we have today, focusing (and focusing fast) has more importance than in film age,

but I think that all these focusing issues are indeed exacerbated in a world where we are used to get everything pre-coocked. No surprise.

I still like to go to the market and cook my meal by myself. (don't need mum for that)

IMO.