Mark Dubovoy's Leica S2 Review

[David Grover / Capture One]

Ok, I have found myself with a bit of airport time so will respond to Mark and Michael in one post.


From Mark...

My understanding from the technical papers and my direct conversations with at least one MF manufacturer is that the issue is not measuring distances with a phase detection system. You are correct that this can be done continuously. The issue is the mechanics of the lens itself. You need quite a bit of torque to get the focusing system to start moving. The system has inertia, as well as gears and other moving parts. There is a limit as to how "continuous" these parts can move. There is also a limit as to how fast you can apply the brakes and stop the movement. When you put all that together, the mechanisms themselves are not continuous and cannot be moved in infinitesimally small increments. They are "jumpy" and they focus in discrete steps or zones.


I expect the manufacturer you are talking about does not use motors inside the lens to drive the focussing optics but rather a motor inside the camera body?  Hasselblad uses motors inside the lens, so a larger lens can use a larger motor etc etc.

The advantage of the motor inside the lens is that there is no mechanical connection needed to be made when attaching the lens to the camera body.  The small armature that protrudes from the lens mount to drive the mechanics I would expect to have some degree of mechanical slop, albeit a small amount.

Therefore as you get to the far range of focus and smaller amounts of movement on the optics need to be applied then the mechanical intolerance (if thats a word) would render it difficult to make precise steps and therefore have a jumpy effect with the result be a loss in finite focussing.  So I assume this is the conclusion from your conversations with Mamiya?  Please correct me if I am wrong.

The stepper motors in the H lenses do not suffer from this issue.  If you connect an H camera to our Phocus software you can see this yourself.  Using the camera control tool, you can manipulate the focus using the + and - buttons.  Hold down the Cmd key whilst clicking these buttons and this will give you an extremely fine focus.  ie they can be moved in infinitesimally small increments.

I hope that clears it up from my end.

If you wanted to research further, what you could investigate is when the AF drive goes back at forth at some point the camera will decide “Im in focus within my factory set tolerance”.  I cannot find any data from Canon and Nikon which describes this tolerance.  Perhaps it is mentioned in the Leica articles?

The tolerance for us does not change at either end of the focussing scale.

From Michael...

What I've learned so far is that while phase detection AF systems do not have "zones" the way that contrast detection systems do, they do indeed focus their lenses in discrete steps, and this is an understood issue in medium format AF design due to such systems' more shallow DOF than 35mm or smaller systems.

I hope with my words above I have tried to describe that this is not the case with the Hasselblad AF system.

From my work eons ago as a motorsport photographer (Using Manual Focus!!  Ugh!!) I would expect a Nikon / Canon 400mm/500mm+ does not have very much DOF at the infinity scale.  Perhaps it is interesting to know from our readers if they feel the 35mm systems fail at infinity focussing?

To close, I am sorry to hear if you felt that it was hard to get in touch with someone at Hasselblad.  However, I have been a pretty active poster here for a couple of years and am only a PM away.

Best Regards,




David

[rainer_v]

Edit: Rainer, I've seen after writing my post that you are using the Sinar M. Coincidence? Anyway, I'd like you to comment about your feeling on this design.
Do you feel a better usability than with other MF cameras ? the M being let's say a modular philosophy and closer to a dslr design.
That would be interesting to have your thoughts. Thanks.

i have a contax setup and a sinar M  setup, both with all lenses. the contax is still my favourite body because its so handy and nice small lenses, but i never use af with it- anyway with the sinar either.
the sinar M is more modern and with the small battery it can be way better  used as with the large battery on your photo. the 28mm rodenstock is stunning, but also the other lenses are great,- maybe the best lenses i have used. the 40mm shows a tick of CA, but all are sharp even wide open, which arent the contax which want to be stopped down for maximum sharpness.
the display on the newer emotion 75 backs is nice as well . not as big as leaf and hassy but very good readeable even in sunlight. the applying of the white files in eXposure works now as i have advised it to sinar two years ago. very practical with great result and convenient workflow.

to be honest i rarely use non shiftable systems so both cameras have a quiet life and are rarely used. its a  pity, but for my work ( except for aereals and long tele lenses ) i only use shift lenses, now 50% the canon and 50% the artec,-  which is a great camera,-( no wonder that i say that right?  ), beside the fact the it was made together by sinar and my person i think its really an intelligent solution for architecture photography which provides a great workflow on site,- i still prefer the work with the artec over the mostly lcd based work with the canon and i prefer it for 1000% for any viewer based solution or any solution which requires to unmount the ground glass.

in practical conditions i have shot now several projects side by side with canons and the artec and choosed the files afterwards only selecting these which i liked more ( color rendering , moiree, resolution , general appearance and so on. ) i end up with mostly 50% for each system. thats a funny result beacuse it sais if i would use only 1 system i could take the one or the other ending up with a similar end result. have to say that i stitch often with the canon / pentax lenses and i dont do it often with the artec, teher is no need in terms of resolution and the shift ways are shorter.

thinking sometimes in buying a P65 or an A10, i am no resistant at all against this hi-tech addiction as you can easily see if you check my equipment and what i did in the past to get usuable images ( i mean also that not everyone was working in develloping soft and hardware,  motivated nearly only from his own perfectionism ). but i am not convinced .... i have a serial of big exhibitions in museums in front of me in several parts of the world, the prints i make will not be sooo huge as it is so modern right now, i will print 100 x 145 cm and some images 100 x 220cm,- and these are looking great with my emotion backs and also with ( right- left or up- down stitched ) canon images. as soon i will believe that my images will gain in any way from higher resolution or some ( at least claimed ) higher dr i will run and get such tool.
for my architecture work, which hardly could be more demanding in terms of my clients and objects i dont see any need for more resolution, but i see each time again the need for highest concentration, a lot of time to spend for the right light, a lot of tools and efforts  to get the right perspectives and a convenient and pleasing workflow which does not hinder me creating my images.
long answer - of topic - but  i hope not a too big problem to fade away a bit.

[rainer_v]

Lens cast correction has been available 3 years plus in Hasselblad software.  Can we get back on topic?

no we cant, but soon.
if i am not completely wrong phocus was announced end of may 2008 for V1 so we get two years at least.
how many years before have been sold cameras and sensors without any possibility to correct these casted files?

and i would be pleased if you answer this my question from my post, referring to the newest generation of kodak sensors in your 39/50/60 mp backs:

is there no more color shift visible between the right and the left side of the frame if e.g. an even grey colored wall is photographed, without that you are applying a white reference ?

[fredjeang]

i have a contax setup ad a sinar M  setup, both with all lenses. the contax is still my favourite body because its so handy and nice small lenses, but i never use af with it- anyway with the sinar either.
the sinar M is more modern and with the small battery it can be way better  used as well as with the large battery on your photo. the 28mm rodenstock is stunning with it, but also the other lenses are great,- maybe the best lenses i have used. the 40mm shows a tick of CA, but all are sharp even wide open, which arent the contax which want to be stopped down for maximum sharpness.
the display on the newer emotion 75 backs is nice as well . not as big as leaf and hassy but very good readable even in sunlight. the applying of the white files in eXposrue works now as i have advised it to sinar two years ago. very practical with great result and convenient workflow.

to be honest i rarely use non shiftable systems so both cameras have a quiet life and are rarely used. its a  pity, but for my work ( except for aereals and long tele lenses ) i only use shift lenses, now 50% the canon and 50% the artec,-  which is a great camera,-( no wonder that i say that right?  ), beside the fact the it was made together by sinar and my person i think its really an intelligent solution for architecture photography which provides a great workflow on site,- i still prefer the work with the artec over the mostly lcd based work with the canon and i prefer it for 1000% for any viewer based solution or any solution which requires to unmount the ground glass.

in practical conditions i have shot now several projects side by side with canons and the artec and choosed the files afterwards only selecting these which i liked more ( color rendering , moiree, resolution , general appearance and so on. ) i end up with mostly 50% for each system. thats a funny result beacuse it sais if i would use only 1 system i could take the one or the other ending up with a similar end result. have to say that i stitch often with the canon / pentax lenses and i dont do it often with the artec, teher is no need in terms of resolution and the shift ways are shorter.

thinking sometimes in buying a P65 or an A10, i am no resistant at all against this hi-tech addiction as you can easily see if you check my equipment and what i did in the past to get usuable images ( i mean also that not everyone was working in develloping soft and hardware,  motivated nearly only from his own perfectionism ). but i am not convinced .... i have a serial of big exhibitions in museums in front of me in several parts of the world, the prints i make will not be sooo huge as it is so modern right now, i will print 100 x 145 cm and some images 100 x 220cm,- and these are looking great with my emotion backs and also with ( right- left or up- down stitched ) canon images. as soon i will believe that my images will gain in any way from higher resolution or some ( at least claimed ) higher dr i will run and get such tool.
for my architecture work, which hardly could be more demanding in terms of my clients and objects i dont see any need for more resolution, but i see each time again the need for highest concentration, a lot of time to spend for the right light, a lot of tools and efforts  to get the right perspectives and a convenient and pleasing workflow which does not hinder me creating my images.
long answer - of topic - but  i hope not a too big problem to fade away a bit.

Very interesting Rainer. Thanks for sharing your thoughts.

I also agree that with the Canon, enlargements of 1m-2m are perfectly good. I find the deth a little "washed" if I can use this expression and tend to add grain in PP to compensate, but that is the Canon's look anyway.

Also in the same thoughts as yours about the upgrading to the extremes. I'm obviously not in a position to purchase a P65, but I already asked myself that question anyway and I think I would not upgrade to more than 40thish MP even if I had the B.Gates bank account. I understand however why some are doing it.

About your statement I painted red, I could not agree more. That is indeed where MF manufacturers will have to concentrate their efforts.
This has been discussed and claimed quite a lot on this forum and I hope you guys (David, Doug) are writting user's reports to the people who take the decisions at Phase, Hassy etc...

But there is also another factor that has been hardly discussed here and I'll try to explain the best I can with my English.
Take the making-off (s). Do you imagine the crane, the helicopter, all that machinery with a little dslr? Something would look wrong. What I'm talking about is the domain of the image and its impact.
All these devices are heavy to deal with, but they transmit something to client. That is not a minor factor IMO.

[Guy Mancuso]

Just to go back to the AF of the S2 and DF bodies just to clarify a few points. First I have not shot the absolute latest version of the S2 firmware so some things may have changed but what was left out here is the DF will shoot continuous focus and the S2 will not ( again this may have changed but I don't think so). So back when Jack and I did the review of the S2 and P40+ the S2 did feel faster to lock in on the center point over the DF but it was kind of weird as it did not refocus when you released the the half press and readjusted as easily as the DF body did. I also felt the DF was actually better for a lot of work where continuous was needed . In this thread I am talking about the Sensor Plus but if you look down in the article I was shooting runway with a 300mm on a monopod and using continuous focusing on the DF and it it really did a nice job of following the model and obviously the worst movement for AF is something coming straight at you and I did not miss much especcially with a 300mm AF lens of older design ( Phase there is a hint here ). Now granted not Canon/Nikon speed but it works very well here is a link if interested. http://forum.getdpi.com/forum/showthread.php?t=13379

Now I do not know the tech. data behind the scenes of levels of focus ability as I do not have that data from the OEM's and obviously I can't speak for the Hassy system in here as I am still waiting to test the HD4/40 system. Anyway this is real world and I also found the DF was a little better in low light but it was very close and both systems where very good in low light which is a big bonus in MF. The S2 does a very nice job of focusing but I had a hard time with it in the beginning until I switched it to manual than used a rear button for AF which I really liked much more. I do talk about this early on in this review http://forum.getdpi.com/forum/showthread.php?t=12243. Now I will not speak of Marks review here since he also has shot the S2 with a later version of the S2 firmware but these systems are pretty darn close, not sure I would say they are worlds apart but yes some differences for sure and the S2 does feel quicker but the lenses are also bigger and heavier and in the working with them all things being together as a package not sure I would say one has a clear advantage over the other but I will say the are far better than the AFD III, II and AFD. Again I am sure the new Hassy focus system is on par as well. It actually may have a slight advantage because of the different AF system employed now. That is a guess on my part since I have not tried it but nice reports on it for sure. Bottom line none of them are in the 35mm league and still for the most part are center point focusing.  I know I do very well with the DF and I liked the S2 on it's end once I figured how to work it to my taste, obviously this will differ among users. But as many of us know in use most of the time we are in manual mode. I NEVER shot a advertising project with AF in all my 35 years, so I have to laugh a little because I know many shooters are the same as me stick it on manual and get to work. Sorry I just have never completely trusted any AF system on the planet and a very simple logical reason they can't think. They have no idea what is going on in your brain and what you are after. End of day though the modern MF bodies are much better than even a year ago but they can all be better as well.

[David Grover / Capture One]

Thanks for the real world info Guy.

[fredjeang]

....  Perhaps it is interesting to know from our readers if they feel the 35mm systems fail at infinity focussing?



Best Regards,

David

I would say that it depends a lot on what is on this infinity.

I have experience with the Canons (but not the last generation) and the Pentax. I came to the conclusion that human brain+practise+intuition can smokes any of these AF (oh my god what I just wrote now...  )
and I'm focussing now 1000% manual with green light confirm if there is. I'm totally with Guy Mancuso on that terrain.

They try to make robots that emulate the human, but they have failed so far in reproducing the walking properlly.

But it is fair to point that Nikon seems to do a better job in that area. I have zero experience with the D3, but from so many users I heard that the D3 AF hardly fails.

[Guy Mancuso]

Fred kind of my point was trust in the AF, not sure many folks really have it regardless of who makes it they can't think and I believe that makes many shooters nervous about AF taking control. As shooters we never want to ever lose that contol over a image. They are great to have onboard and sometimes a invaluable tool no question about it and sports almost everyone depends on it but trust is a big obstacle as well.

[Rob C]

I NEVER shot a advertising project with AF in all my 35 years, so I have to laugh a little because I know many shooters are the same as me stick it on manual and get to work. Sorry I just have never completely trusted any AF system on the planet and a very simple logical reason they can't think. They have no idea what is going on in your brain and what you are after. End of day though the modern MF bodies are much better than even a year ago but they can all be better as well.

And from an even older shooter, bingo!

Rob C

[bcooter]

To keep this on topic, about comparing these cameras . . . would some tester please, please, please, shoot something pretty with one of these things?

I've seen backyards, shoots of rainy streets from rooftops, alley's with blown out sky, shots of sunny streets from rooftops, (they all look like bad Homeland Security surveillance footage).

Jez . . . it seems every "tester"  takes a gazillion dollar camera and shoots butt ugly photos.  Come on shoot something real pretty, then explain from front to back the process, workflow and little things like did the firewire connection cut off, did the software crash, did the retoucher have to cut purple ca from around an image, did the camera jam, are there lenses available, can you shoot in low light, etc. etc. etc.

Then the forums light up with all the medium format players beating each other on the head saying there camera shows 4% better micro detail and can focus on a pimple from 9,000 ft. away.

And now to take this totally off topic . . . for Rob C.  Not everyone in 'merica drives a V-8 they just wished they did, well except everyone that lives by the beach.     On my block in Santa Monica there are 7 (yes 7) toyota Prius' (all black) and I live on a short block.  7:45 in the morning is funny cause all these cats come out and flick the remote to figure out which car is theirs.   Consequently,   If I drive 24 miles out of LA half the driveways have 5 liter V-8's in the driveway.  'merica really is two countries, the cities and the burbs and the burb folks have a lot of fun.

http://www.youtube.com/watch?v=St2FCxtlV7w

[tho_mas]

would some tester please, please, please, shoot something pretty with one of these things?

the latest trend is not to post any captures. Because monitors are not appropriate to judge about images  

[fredjeang]

..................Shhhh........................
To keep this on topic, about comparing these cameras . . . would some tester please, please, please, shoot something pretty with one of these things?

I've seen backyards, shoots of rainy streets from rooftops, alley's with blown out sky, shots of sunny streets from rooftops, (they all look like bad Homeland Security surveillance footage).
..................Shhhh.........................

And it seems a generalized reality all over the world.
I've checked 2 french websites about high-end gears and they do some testings.
The contents are really good until you reach the images. Although I saw some at least well done portraitures.

But it is the facade, the leaf, the tree, the garden, the fence, the dog and in the worst case the familly shots.

They could at least post a V8 engine.

This is the General Lee European version V8 dream  
[attachment=22882:renault_...ine_a110.jpg]

[eronald]

I am sure Mark will research the topic very carefully, and I do look  forward to reading his coming article.

Edmund

Aaron,

Your aggressive tone is uncalled for. In fact since this was first raised by the Hasselblad rep and the issue came to our attention Mark has been aggressively pursuing as much technical information on the subject as he can, directly from senior engineering people at Leica and Phase One / Mamiya. Though in the past getting to high level technical contacts at Hasselblad has proven difficult, maybe now we can do so.

The net result is that when all the data is available and sorted out Mark now intends on writing something comprehensive on the subject, which I hope will benfit all of our understanding on the issue.

What I've learned so far is that while phase detection AF systems do not have "zones" the way that contrast detection systems do, they do indeed focus their lenses in discrete steps, and this is an understood issue in medium format AF design due to such systems' more shallow DOF than 35mm or smaller systems.

So I suggest that the name calling and finger pointing be shelved, and let's all wait to learn more about this subject. Mark Dubovoy is a scientist by profession, with a Phd in particle physics, and therefore well aware of  the scientific method and how to apply it to researching and writing up a topic like this.

It appears that we all have something to learn on the subject.

Michael

[JeffKohn]

Still, what I'd love to see on any comparison is a test like the Zaguto people funded. Sure it may have some bias given Zaguto sells bits for dslrs, but overall when you fill the room of "working" professionals I doubt seriously if everyone there has any agenda other than to find out the facts.

http://www.zacuto.com/shootout

This was an interesting series, even though I have no plans to shoot HD video myself.

Did anybody notice that in the last episode they measured the DR of a 5D2 RAW as being 10.5-11 stops, using the "photographic" test that MF proponents argue for (backlit 13-stop glass step wedge).  So much for the "6-7 stops for DSLR" argument.

[Guy Mancuso]

To keep this on topic, about comparing these cameras . . . would some tester please, please, please, shoot something pretty with one of these things?

I've seen backyards, shoots of rainy streets from rooftops, alley's with blown out sky, shots of sunny streets from rooftops, (they all look like bad Homeland Security surveillance footage).

Jez . . . it seems every "tester"  takes a gazillion dollar camera and shoots butt ugly photos.  Come on shoot something real pretty, then explain from front to back the process, workflow and little things like did the firewire connection cut off, did the software crash, did the retoucher have to cut purple ca from around an image, did the camera jam, are there lenses available, can you shoot in low light, etc. etc. etc.

Then the forums light up with all the medium format players beating each other on the head saying there camera shows 4% better micro detail and can focus on a pimple from 9,000 ft. away.

And now to take this totally off topic . . . for Rob C.  Not everyone in 'merica drives a V-8 they just wished they did, well except everyone that lives by the beach.     On my block in Santa Monica there are 7 (yes 7) toyota Prius' (all black) and I live on a short block.  7:45 in the morning is funny cause all these cats come out and flick the remote to figure out which car is theirs.   Consequently,   If I drive 24 miles out of LA half the driveways have 5 liter V-8's in the driveway.  'merica really is two countries, the cities and the burbs and the burb folks have a lot of fun.

http://www.youtube.com/watch?v=St2FCxtlV7w

It's been done to a certain level http://forum.getdpi.com/forum/showthread.php?t=12243  and raws to download which are still up. But I am not going to put great images up for raws to be used anywhere and YOU should know better than that. They are test files pure and simple but putting up real work is NOT something I will do especially the raws and no one testing these rigs will either and you know the reasons. The other side of this coin and I know I can speak for the rest of reviewers on this we are not getting one red nickel for our efforts either to even do these tests than take the heat because we really don't want to put up our real work of art. Sorry bud screw that I put something up really nice i want my pocket filled up really nice as well. From a guy that works for commerce I am surprised by your plea. This is a simple case and you especially can get anyone of these systems as a demo and run these test yourself and shoot all of the beautiful images you want.

[markowich]

the latest trend is not to post any captures. Because monitors are not appropriate to judge about images  

only prints 24x30 inch tell the IQ truth, according to recent LL sources.
peter

[ErikKaffehr]

Hi,

Yes I enjoyed the Zacuto tests. I'm somewhat confused about the contradictory statements on DR. My experience is that it is very well possible to extract a lot of shadow detail from my DSLRs (Sony), as a matter of fact, I was testing HDR but found that just using controls in LR (including liberal use of "fill light") I could achieve similar results from an "underexposed" picture. With HDR there was much less noise in the shadows, however.

I hope that some of the MF authors on LuLa will eventually elaborate the issue.

Best regards
Erik

This was an interesting series, even though I have no plans to shoot HD video myself.

Did anybody notice that in the last episode they measured the DR of a 5D2 RAW as being 10.5-11 stops, using the "photographic" test that MF proponents argue for (backlit 13-stop glass step wedge).  So much for the "6-7 stops for DSLR" argument.

[ErikKaffehr]

Thanks for the link!

I have seen it before but it's time to revisit!

Best regards
Erik

It's been done to a certain level http://forum.getdpi.com/forum/showthread.php?t=12243  and raws to download which are still up. But I am not going to put great images up for raws to be used anywhere and YOU should know better than that. They are test files pure and simple but putting up real work is NOT something I will do especially the raws and no one testing these rigs will either and you know the reasons. The other side of this coin and I know I can speak for the rest of reviewers on this we are not getting one red nickel for our efforts either to even do these tests than take the heat because we really don't want to put up our real work of art. Sorry bud screw that I put something up really nice i want my pocket filled up really nice as well. From a guy that works for commerce I am surprised by your plea. This is a simple case and you especially can get anyone of these systems as a demo and run these test yourself and shoot all of the beautiful images you want.

[eronald]

It's been done to a certain level http://forum.getdpi.com/forum/showthread.php?t=12243  and raws to download which are still up. But I am not going to put great images up for raws to be used anywhere and YOU should know better than that. They are test files pure and simple but putting up real work is NOT something I will do especially the raws and no one testing these rigs will either and you know the reasons. The other side of this coin and I know I can speak for the rest of reviewers on this we are not getting one red nickel for our efforts either to even do these tests than take the heat because we really don't want to put up our real work of art. Sorry bud screw that I put something up really nice i want my pocket filled up really nice as well. From a guy that works for commerce I am surprised by your plea. This is a simple case and you especially can get anyone of these systems as a demo and run these test yourself and shoot all of the beautiful images you want.

Cmon guy, that don't explain the rarity of brick wall raws

Edmund

[tho_mas]

only prints 24x30 inch tell the IQ truth, according to recent LL sources.

exactly. obviously only at 300dpi - or 360dpi? - we can see the "real" resolution of lenses/chips. is that with "capture sharpening", "print sharpening", "creatve sharpening" or without any sharpening at all? And is it C-print or ink jet? which one on which paper?
As "the web is sRGB" (    ) only prints show the uncompressed gamut of the camera (nonsense, of course - no printer will show the entire gamut of a camera, not even a point and shoot camera. "...
Man, we all know how a 100% crop looks printed (in your preferred printing technique...). Too, who says that it is impossible to post small crops as 8bit TIF... if you really care about JPG compression (that is often not a problem at all... especially not in comparisions).
Etc. ...
Strange...