The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

S2 Chromatic Aberration - how big an issue is it?

JanRSmit

New member
Of course. Or a really poor sensor design, or a really poor lens design. At the end of the day, it could be any of these, but for Leica's sake, we hope it's just a conversion software issue -- and IMO that is the likeliest issue. (Phase files can have excessive fringing too if you convert them in LR...)
I use a D700 with leica r lenses, and do get purple fringing and ca in certain cases. Also before buying the D700 i had a D3 with the nikon 24-70 2.8 zoom to test, and in a particular occasion made a picture of tree backlit by early morning sun, shot wide open at 28mm. It shows heavy purple finging and ca.
Minute apart i made a picture with my M6 and elmarit 28mm 2.8, wide open, and kodak portreta negative film. Also purple fringing and some ca, but less.
With my D700 and leica glass, i get in extreme backlit cases ( or strong sun reflections) purple fringing, like sun reflecting via small waves on water surface. I use lightroom for my raw conversion and development.
I took these issues as a given based on extreme light conditions. With this thread i am getting more interested, and plan to look at those raws (and scan's) again and see if other developers like C1 and NX2 do a better job on this particular issue.
Note that in the nikon D3/24-70 images, i find CA issues present much stronger that in the case of leica glass.

P.S.: in one case, still under investigation, it looks like it is the LCD monitor i use (EIZO), with dark and strongly saturated blue/purple color, it appears the LCD monitor is causing what looks like CA.

To summarise, despite the D700 filters in front of its sensor there are issues, purple fringing does also occur when using film. So if a particular body(electronics,firmware)/lens/raw developer does not show these issues, it would be very interesting to find out why!
 

robmac

Well-known member
As per Jack's comments...

While God knows CEOs have egos and not all are as mature as people would like to believe, think in terms of logical cold-hearted business reasons: a) why a deal might make sense, b) why it could unravel and c) what would happen and how both players would react if it was unwound at the 11th hr and the 'unwinding' turned into a (very public) dog's breakfast.
 

narikin

New member
Let's clarify this a bit further:

1) CA is a LENS aberration caused by an inability to bring all three primary colors (or secondary colors) of the visible light spectrum into precisely the same focus point on the imaging medium. Moreover, the effect is usually spread laterally, and software is exceptionally good at being able to re-align the three separate visible primary color channels and bring them into one common point -- and why it is thus "relatively easy to correct for."

2) Purple fringing is primarily a lens aberration, however it is accentuated by a digital SENSOR's excessive UV and IR sensitivity. It is [usually axial and] caused by the lens' inability to bring the short UV and long IR bands outside the visible spectrum to the same focus point as above. The distinction is that these bands are outside normal visible spectrum and it's virtually impossible to correct for them in lens design as the spread of wavelengths is far wider than the total visible spectrum to begin with. Thus, internal camera IR/UV cut filtration is probably the best approach to attenuate these; external filtration probably the second best approach. IOW, if a camera exhibits this trait, the lens is probably not the place to place the blame, rather the design of the sensor's filtration is the more relevant culprit.

3) Sensor bloom is, or rather can be, another form of fringing and can be perceived as green, red, cyan or magenta (purple) depending on the sensor design. It is caused primarily by a pixel getting over-saturated and bleeding off to adjacent pixels. The resultant aberration can vary from spots to ghosting to streaks to edge fringing. Moreover, it's primarily a CCD issue as most CMOS sensors have anti-blooming gates by design. Though many CCD's do use anti-blooming gates in their design, the attenuation isn't as efficient as with CMOS, thus we tend to see it more commonly in cameras using CCDs. Fortunately, good software will attenuate this anomaly too, the key being its ability to isolate it before the desaturation process as a global desaturation is not a viable method.

Edit: The above explanations are simplified in the interest of keeping them easy to understand. Here is a good source for those that want to investigate optical issues in more depth: http://toothwalker.org/optics/chromatic.html
thanks Jack,

yes and as OP on this, I should state that the CA we are talking about primarily here, besides 'purple fringing' is 'Longitudinal Chromatic Aberration', which can be found in many fast lenses, but Leica S2 70mm is a particularly egregious example when wide open, and no software correction is supplied.

see towards the bottom of this review for examples on an otherwise very good Nikon lens: http://www.dpreview.com/lensreviews/nikon_50_1p4g_n15/page4.asp

mind you this is at f1.4 rather than f2.5, and costs $410 not $4500...
 

narikin

New member
As a 40 megapixel MFD camera - The S2 doesnt do anything better than Hasselblad - and can't do a lot of things Hasselblad does - really it is that simple for me.
respectfully disagree - because of its form factor (mid size dSLR, totally weatherproof, fast frame rate) it can go places and take images that nobody with a bazooka camera - be it Hassy or Mamiya or Phase or whomever - can take. It is quieter, smaller, less obtrusive and weather proof.

suitable taped up it looks like an ordinary SLR, not a super-pro one, so there's a lot you can do with this that you cant with a 'Blad or Phase.

ymmv of course. I own Phase, but would buy an S2 also if they could give me good results at wider apertures. sadly they can't - and thats the point of this thread.
 

georgl

New member
There are no <f2.8 lenses with better correction than the S-lenses (besides fast teles like the Apo-Summicron 180) I'm aware of. So when you're not happy with these results, you'll have to stop down, period.
David has some nice 100% off-axis crops @f2.5:
http://www.drivehq.com/file/df.aspx/shareID6403983/fileID542438486/542438486.jpg
http://www.drivehq.com/file/df.aspx/shareID6403983/fileID424688221/424688221.jpg

Normally, even 100% crops from 35mm-12MP-cams + >>1k$ lenses aren't this sharp and lack aberrations to this extend.

Mainly Zeiss introduced some major breakthroughs in medium format-lens-design 20-50 years ago - besides Leica and Rodenstock/Schneider- for technical cameras (I agree, these lenses are magnificent and better than the S-lenses @f2.5, but slow and not usable for SLRs) no one cared about investing any money or know-how to introduce significant improvements since then.
Hasselblad ordered new designs but instead of improving old Zeiss-designs, they ordered mostly similar performing designs from Fuji and added software correction as the "one and only"-solution, then they started to introduce lenses which are not even usable on bodies introduced a few years ago to a new system - just to sell more new bodies! Now they introduce a back which cannot be used meaningful (no, I don't buy a 60MP-back instead of the 50MP to use a crop) with these lenses just introduced and still call it a modular system! That's what I was trying to say. Of course, a 3k$ Fuji-lens will perform reasonable under most circumstances anyway.

Maybe Leica needs some serious talk with Adobe for adapting LR3 - but choosing the difficult and expensive way to create new lenses (besides the new form-factor of the body itself) shouldn't be ignored. It's not some magical system which outperforms larger (and more expensive) 60MP-designs or makes every aberration magically disappear, but it offers outstanding performance in a quite ergonomic, small form-factor - not to make classic MFDBs superfluous but to add something to the market that wasn't there before.
 

ptomsu

Workshop Member
Well said Mark.

There were other reasons as well to the breakup of Leica and Phase One besides the conflict of interest on software registration. The claims from Phase One agents and dealers at Photokina (hours after the announcement of a strategic alliance) of deep involvement in developing the S2 camera system and Leica making lenses for Mamiya mount certainly didn't help build any goodwill. I was standing in Leica booth when Leica managers found out what was being said and stormed over to Phase One. The rest of the show, reps spent their time in interviews emphasizing that the S2 was designed and built 100% by Leica, without help from any partners. Talks with Adobe started about a month or two afterwards towards the end of 2008.

As you said, the past is the past. Let's look to the future.

David
David, all,

I could not agree more - lets look at the future and see what will happen.

Nevertheless for the time being, I find my needs 100% covered by Hasselblad, the H system, Phocus and the genial HTS 1.5 which I am using with my 28 HCD for landscape panoramas. This in combination with PS is just delivers the result right out of the box with minimal corrections in post processing. Really happy about that!

Would love to see Leica to come close to this level with their S System and whatever SW combination for post processing.
 

ptomsu

Workshop Member
Re the Phase/Leica debacle:

The reason as I heard it was simpler, and made far more sense from a historical POV: The nit that killed the deal was that Leica would have had to release to Phase all of their internal, proprietary image processing secrets for their raw protocol, and at the 11th hour, Leica management decided against this. If you look historically at the relationship between Leica and Panasonic on the smaller sensor cameras, we know this same mentality existed at Leica; the Leica versions of the small Panny cameras always seemed to produce slightly better images -- when pressed, Leica did admit the cameras were mechanically identical and it was their own internal secret sauce that delivered the superior images.

Now Leica is stuck with an open output format a-la DNG, but one where Adobe has their own internal secret-sauce components to muddy the waters... So now Leica has to implement their secret sauces BEFORE the data gets DNG'd, but in a fashion where the benefits will be present in the DNG that Adobe sees openly --- a non-trivial task I fear. I suspect given enough time that Leica will get it sorted, but how long will it take? And will the final solution be as elegant as it might have been had they opened up to Phase for CaptureOne? Personally, I doubt it...
Jack,

fully agree! And as you mention, C1 would have been the logical choice. Well today they could even knock on Hasselblad doors for Phocus, which (as you all know I am a Hasselblad shooter) in my opinion has the even superior processing SW WRT IQ. Just my view of course. And yes - I know that this is all daydreaming.

Wish Leica all the best for fixing their issues with the S System.
 

narikin

New member
There are no <f2.8 lenses with better correction than the S-lenses (besides fast teles like the Apo-Summicron 180) I'm aware of. So when you're not happy with these results, you'll have to stop down, period.
(a) what about the Zeiss Planar 80mm AF introduced for Sinar M system in 2008?
http://www.zeiss.com/c12567a8003b58b9/Contents-Frame/858dbbbbd2fb78a6c125711800592377

(b) "you'll have to stop down, period" is not an answer. Good lens design couple with software solutions make my Phase 80mm-D lens very usable at wide apertures with much less CA that the Leica S system shows. So - you don't "have to stop down", but rather, you have to avoid a Leica S camera, for big aperture shooting, sadly.
 

fotografz

Well-known member
Don't get me wrong. I am not whitewashing the issue with the S2 and fringing/CA. I am trying hard to put it in perspective and clarify that in real-world use (versus testing) it only shows up in a very low percentage of photographs. It takes several factors to get the CA to become evident and most photographers simply don't get those factors to line up.

Take the example photo of the glass embedded in the wall posted earlier. I downloaded the DNG (thank you Guy for making it available) and tried to do a simple fix in LR3. I couldn't correct the image and started to get bummed out. Then I realized the image was shot at f2.8 in bright sunlight (1/3000 sec.). I would have never shot that scene at f2.8 if I had that much sunlight. I would have been stopped down to around f5.6 or f8 to get additional depth of field and still had plenty of shutter speed for a sharp hand-held photo. I would have stopped down to give myself some margin to compensate for body sway and not holding the camera perfectly parallel to the wall.

Take Marc's example of the plant with the door light behind it posted previously. I can't say for sure what f stop he used, but I bet the lens was wide open (f2.5) or very close to it. That scene is in the Don Cesar hotel in St. Pete Beach, Fl which is a beautiful hotel. I expect if Marc were shooting that scene for the hotel as a client he would have shot it much differently (e.g., portable lighting, smaller f stop for depth of field, etc.) and CA would not be present.

Testing cameras and lenses provides important information. However, it is important to keep this information in perspective as to how one expects to use the camera and lenses.
You are assuming that the tests were not done with a specific purpose related to real world conditions. In reality that was exactly what I was doing with the S2. My intent was to see if this camera could eliminate the need for a 35mm DSLR and a MFD kit for weddings ... get it down to one SLR kit and my Leica M9s.

I shoot wedding in venues like that all of the time, which made the Don Cesar test scenario very relevant for me. If I were shooting for the hotel as a place of business, it would be a whole different scenario ... off-camera lighting/softboxes, camera on-tripod, optimal f-stops, assistants lugging stuff, etc ... (and now most likely a H4D with a HTS for PC control and stitched Panoramic images).

But at weddings in a beautiful venue like that, people are the subject ... and they are moving. It's usually hectic and unpredictable. Often no time for setting up lights ... use of diffused on-camera fill coupled with dragging the shutter and wide open apertures is the rule, not the exception (when shooting inside). This is even more critical when you consider that the S2 isn't all that great at higher ISO's.

Personally, my hope is that the S2 will get there, and improve to the point that I'll take another look at it for what I shoot ... in the real world.

-Marc
 

markowich

New member
You are assuming that the tests were not done with a specific purpose related to real world conditions. In reality that was exactly what I was doing with the S2. My intent was to see if this camera could eliminate the need for a 35mm DSLR and a MFD kit for weddings ... get it down to one SLR kit and my Leica M9s.

I shoot wedding in venues like that all of the time, which made the Don Cesar test scenario very relevant for me. If I were shooting for the hotel as a place of business, it would be a whole different scenario ... off-camera lighting/softboxes, camera on-tripod, optimal f-stops, assistants lugging stuff, etc ... (and now most likely a H4D with a HTS for PC control and stitched Panoramic images).

But at weddings in a beautiful venue like that, people are the subject ... and they are moving. It's usually hectic and unpredictable. Often no time for setting up lights ... use of diffused on-camera fill coupled with dragging the shutter and wide open apertures is the rule, not the exception (when shooting inside). This is even more critical when you consider that the S2 isn't all that great at higher ISO's.

Personally, my hope is that the S2 will get there, and improve to the point that I'll take another look at it for what I shoot ... in the real world.

-Marc
marc,
after some more hours of comparism i am now convinced that -at equal f stops- the H system with Phocus processing has little to no advantage over the S system, except of course resolution in the case of the 50mpx Hassy system. CA/fringing is more apparent in S shots because you are tempted to shoot wide open and the lenses are generally faster than the Hasselblad counterpart. this except the HC 100mm f 2.2, which shows a lot of CA wide open, even properly processed in PHOCUS (see my posting above in the same thread). to be more specific, the S 35mm is a spectacular lens, the 180mm as well. wonderful resolution up to the corners, great bokeh. the S 70mm lens is in my view less spectacular and shows the most color artifacts of all S lenses.
the question is: what would the S system be like with proper software? i think nothing less than spectacular. i hope leica will go the extra mile and do proper software for their flagship system.
peter
 

gogopix

Subscriber
marc,
...the question is: what would the S system be like with proper software? ..peter
Dear Peter,

from what you say, better than the other systems.. However, it may just be one of the things we minimize and live with. At the same time that we all agree that only 1 in 20-50 shots is a real keeper, we worry that the killer shot will be ruined.
Comments here suggest setup, lighting and composition may be more important than anything.

Interesting comment about the 70mm; I wondered why it couldn't 'blow away' my Contax 80mm 2.0 lens (that can get a little soft WO). but it seemed pretty good for the 2000+ shot I took in Ireland.

Enjoy; I look forward to your experiences with the rest of the line.

best
Victor
PS Here are 100% from S2 (LR defaults) , S-K 35mm P65+ (C1), and S2 (output from ACDSEE!) strong back sunlight.
Yes SK is half the fl but also 2x from Phase, thought there might be some CA/fringe
The last one shows that the default in LR (1st one) CAN take out the routine fringing/CA (obviously seen in the poor raw conversion of ACDSEE)
 
Last edited:

markowich

New member
Dear Peter,

from what you say, better than the other systems.. However, it may just be one of the things we minimize and live with. At the same time that we all agree that only 1 in 20-50 shots is a real keeper, we worry that the killer shot will be ruined.
Comments here suggest setup, lighting and composition may be more important than anything.

Interesting comment about the 70mm; I wondered why it couldn't 'blow away' my Contax 80mm 2.0 lens (that can get a little soft WO). but it seemed pretty good for the 2000+ shot I took in Ireland.

Enjoy; I look forward to your experiences with the rest of the line.

best
Victor
PS Here are 100% from S2 (LR defaults) , S-K 35mm P65+ (C1), and S2 (output from ACDSEE!) strong back sunlight.
Yes SK is half the fl but also 2x from Phase, thought there might be some CA/fringe
The last one shows that the default in LR (1st one) CAN take out the routine fringing/CA (obviously seen in the poor raw conversion of ACDSEE)
victor, very interesting comparism. it shows the difference that a reasonable RAW converter (LR) can make over a crappy one (ACDSEE). i figure that an excellent one over a reasonable one could make the same level of difference.
peter
 

fotografz

Well-known member
marc,
after some more hours of comparism i am now convinced that -at equal f stops- the H system with Phocus processing has little to no advantage over the S system, except of course resolution in the case of the 50mpx Hassy system. CA/fringing is more apparent in S shots because you are tempted to shoot wide open and the lenses are generally faster than the Hasselblad counterpart. this except the HC 100mm f 2.2, which shows a lot of CA wide open, even properly processed in PHOCUS (see my posting above in the same thread). to be more specific, the S 35mm is a spectacular lens, the 180mm as well. wonderful resolution up to the corners, great bokeh. the S 70mm lens is in my view less spectacular and shows the most color artifacts of all S lenses.
the question is: what would the S system be like with proper software? i think nothing less than spectacular. i hope leica will go the extra mile and do proper software for their flagship system.
peter
Peter, personally I'd have to agree that the H4D/40 has no significant image advantage over the S2. I did find the S180 to be excellent and a serious reason to consider the S2 by itself. If the 35 follows suit as you say, then all the better. Again, I will subjectively say that to date, I've not seen one single S2 image that makes me jealous from a IQ POV ... or any other criteria for that matter. Still waiting and watching.

In my case, CA or fringing to some degree would not have been a deal breaker. As others have said, it rears it's ugly head only on a smallish % of images anyway. I do admit that it was a surprising amount with the 70mm .... more than anything I've ever encountered with the H/C100/2.2. (BTW, that S2 70mm hotel lobby shot I posted was not shot wide open ... it was at f/3.4. Try the HC100/2.2 at 3.5 and see if your results are still similar).

I also agree that one would be tempted to shoot wide open when it's there. Yet, the 180/3.5 is no different for my applications than the HC210/4 in terms of relative speed to focal length.

The relatively minor issue in a supposed area of optical superiority that cost a significant premium, coupled with a lack of versatility, questionable higher ISO performance, and the need to crop the S2 files to fit most common usage verses the capture ratio of the other 40 meg backs ... and add Hasselblad's timely improvement of their True Focus AF system .... and the advantages of a $40K swap out, or even addition to existing kit, become less apparent.

In short, to double my cash outlay, the S2 had to blow me away, and it didn't.

Not everyone can afford multiple MFD systems ... or, in my case to pull it off, in addition to dumping all 35mm DSLR gear, I would have to off my entire M9 system ... which just isn't going to happen.

As the song goes ... If I were a rich man ... ;) ... which up until a few years ago I was, and thanks to the economic melt-down no longer am ... so we adjust to the real world and shifting priorities. A Hassey Multishot was a workhorse, money making priority over a passionate luxury item that added nothing to the output and siphoned off a considerable amount of incoming capital.

Timing is everything, sometimes it works for you :) sometimes it doesn't. :(

-Marc
 

markowich

New member
Peter, personally I'd have to agree that the H4D/40 has no significant image advantage over the S2. I did find the S180 to be excellent and a serious reason to consider the S2 by itself. If the 35 follows suit as you say, then all the better. Again, I will subjectively say that to date, I've not seen one single S2 image that makes me jealous from a IQ POV ... or any other criteria for that matter. Still waiting and watching.

In my case, CA or fringing to some degree would not have been a deal breaker. As others have said, it rears it's ugly head only on a smallish % of images anyway. I do admit that it was a surprising amount with the 70mm .... more than anything I've ever encountered with the H/C100/2.2. (BTW, that S2 70mm hotel lobby shot I posted was not shot wide open ... it was at f/3.4. Try the HC100/2.2 at 3.5 and see if your results are still similar).

I also agree that one would be tempted to shoot wide open when it's there. Yet, the 180/3.5 is no different for my applications than the HC210/4 in terms of relative speed to focal length.

The relatively minor issue in a supposed area of optical superiority that cost a significant premium, coupled with a lack of versatility, questionable higher ISO performance, and the need to crop the S2 files to fit most common usage verses the capture ratio of the other 40 meg backs ... and add Hasselblad's timely improvement of their True Focus AF system .... and the advantages of a $40K swap out, or even addition to existing kit, become less apparent.

In short, to double my cash outlay, the S2 had to blow me away, and it didn't.

Not everyone can afford multiple MFD systems ... or, in my case to pull it off, in addition to dumping all 35mm DSLR gear, I would have to off my entire M9 system ... which just isn't going to happen.

As the song goes ... If I were a rich man ... ;) ... which up until a few years ago I was, and thanks to the economic melt-down no longer am ... so we adjust to the real world and shifting priorities. A Hassey Multishot was a workhorse, money making priority over a passionate luxury item that added nothing to the output and siphoned off a considerable amount of incoming capital.

Timing is everything, sometimes it works for you :) sometimes it doesn't. :(

-Marc
marc,
i totally agree. i would have never sold the M equipment for the S2. it was just a luxury acqusition (some women do buy hermes bags and suitcases, no?---))), which i happened to be able to afford at the moment.
the hassy is ugly but it does work. i once tried the phase camera but didn't like the feel of it at all.
anyway, if i needed to make money out of the camera, it would be hasselblad all the way-until the S system is complete, then one has to reassess.
peter
 

Jack

Sr. Administrator
Staff member
I'll say it again: *IF* Leica had simply made the S2 Mamiya-mount compatible, I would be all over it as a second body right now, even with the processing foibles. And I would probably own the S70 as my main prime instead of the 80 LS. I am confident that had they done that, Phase would have built a good processing module on their own for the camera -- and even if not, I am confident I could get a passable set of processing parameters and capture profiles suitable for my own needs.

But then I wake up and realize they never asked me :ROTFL:
 

gogopix

Subscriber
Dear Jack

I am sending you a lathe and a block of aluminum...:D

Unless there is something nasty in the electronic interface (as with Contax :angry: ) shouldn't be an issue. Registration distance seems small enough. Hasselblad V already done...

Have at it!

all the best

Victor
 
Top