Rumor - Three new Leica M lenses tomorrow?

[pgk]

Advertisement (gone after registration)

Let me ask you a question: if lenses designed for film don't work well with a cover glass, why should lenses designed for a cover glass work well on film? Don't believe everything Karbe says. He works for Leica after all and Leica wants you to believe that their new M lenses work equally as well on film and digital bodies, and you're free to believe whom ever you want

 

Well, most films probably have lower tolerances than higher MPixel sensors so will accept discrepancies with less visual impact. Any problems due to redesign for a cover glass (of which I am far from convinced because such a design parameter would affect all future sensor decisions and who knows what will happen in sensor design in the future?) will probably be more than compensated for by the flatter field and other corrections made due to advances in lens design. So I don't really see why newer lenses shouldn't work as effectively on film as older ones myself.

[jaapv]

It is quite possible that the improvements in lens design "forced" by sensors will result in improvements on film as well. There is no reason at all to think either-or.

It was the same way with the advent of retrofocus wideangle lenses. Initially needed to clear the mirror of SLR cameras, it turned out  that it gave more degrees of freedom to improve lens design in general.

[edwardkaraa]

Jaap and pgk, the digital optimization focuses the image periphery on a different plane, and introduces negative astigmatism to counter for field curvature and astigmatism caused by the cover glass. On film you would get both field curvature and astigmatism. The argument that film is more forgiving doesn't hold if you scan at high resolutions. Maybe some film users don't mind, as there are also digital users who don't care how their lenses perform in the borders/corners. Anyway I'm done justifying the obvious. There are too many articles on the net about this. It's not like I'm coming up with some revolutionary crazy ideas

[Peter Branch]

Well, most films probably have lower tolerances than higher MPixel sensors so will accept discrepancies with less visual impact. Any problems due to redesign for a cover glass (of which I am far from convinced because such a design parameter would affect all future sensor decisions and who knows what will happen in sensor design in the future?) will probably be more than compensated for by the flatter field and other corrections made due to advances in lens design. So I dan't really see why newer lenses shouldn't work as effectively on film as older ones myself.

 

As far as I am aware, with the exception of some very exotic specialised film types, all films have a "Supercoat" which is analogous to a cover glass. Colour film structures, though obviously completely different, are in many ways as complex as those of a digital sensors,

 

There are many reasons why a lens design optimised for film can benefit from being re-optimised for a digital sensor.  Some aberrations that are relatively unimportant for film are more important in the digital world and vice versa.

 

That said I very much doubt, outside a laboratory and using micro-film, that differences between the two versions of a lens can be detected on film but they probably can be detected with a 24Mp sensor.  

[jaapv]

Jaap and pgk, the digital optimization focuses the image periphery on a different plane, and introduces negative astigmatism to counter for field curvature and astigmatism caused by the cover glass. On film you would get both field curvature and astigmatism. The argument that film is more forgiving doesn't hold if you scan at high resolutions. Maybe some film users don't mind, as there are also digital users who don't care how their lenses perform in the borders/corners. Anyway I'm done justifying the obvious. There are too many articles on the net about this. It's not like I'm coming up with some revolutionary crazy ideas

Film IS more forgiving as this is all about acceptance angles, which is not applicable to film. If you minimize the angle of incidence on a sensor to reduce the  degradation outside the centre (remember, this is about a planparallel surface) it will have no effect whatever on film.

 

Your hypothesis would only hold true if one would introduce deliberate astigmatism towards the edges to counteract the   aberrations introduced by the cover glass. It would be silly to do so for interchangeable lenses, as they must be able to work on any (future) sensor, irrespective of the optical properties of the cover glass.

 

Such a design could only work on a single lens-sensor module - like the Q  for instance (not that I know whether Leica pulled such tricks there).

[pgk]

The argument that film is more forgiving doesn't hold if you scan at high resolutions. Maybe some film users don't mind, as there are also digital users who don't care how their lenses perform in the borders/corners. Anyway I'm done justifying the obvious. There are too many articles on the net about this. It's not like I'm coming up with some revolutionary crazy ideas

Ummm. But optimisation for digital sensors will be a compromise too, due to varying thickness of cover glass/coatings, offset micro lens positioning, etc.. So optimising a lens by paying more attention to flat field requirements for example would IMO counter any other 'pro-digital' characteristics which are 'anti-film'. Theory is fine but film is more forgiving in practice in my experience anyway.

That said I very much doubt, outside a laboratory and using micro-film, that differences between the two versions of a lens can be detected on film but they probably can be detected with a 24Mp sensor.  

 

The most pertinent comment. Testers often prefer to use extreme resolution B&W films but these are akin to lab conditions. In the real world I too doubt any detectable differences will occur. I'm still dubious about Leica deciding on a specific digital requirement though as it would tie them to a specific sensor characteristic if they were pedantic about it. 

[wattsy]

Advertisement (gone after registration)

Your hypothesis would only hold true if one would introduce deliberate astigmatism towards the edges to counteract the   aberrations introduced by the cover glass. 

 

 

Unless I'm misunderstanding (which is perfectly possible), I think this is roughly what Edward is saying Leica have done with the new lenses. Personally, I think the theory is, at best, pure speculation extrapolated from some marketing statements and anecdotal findings that the new 28 Summilux performs better around the edges on the SL. IMO Edward's theory both over-eggs the significance of the changes Leica have made with the newly released lenses and credits the Leica engineers with more sophistication than they probably deserve (not unlike Tim Ashley claiming that the more refined M240 rangefinder is deliberately set-up to mitigate the field curvature of the 35 Summilux FLE).

[edwardkaraa]

It is in fact documented that Zeiss has indeed introduced astigmatism and field curvature on the loxia line from the original ZM designs to counteract the effect of the thick sensor glass of the A7 series. I have no idea what Leica has done. It could be all marketing bs. But if what they claim is true, and knowing the lens size remains comparable, I am guessing this is what they have done as well, but my guess could be wrong. The summilux 28, as mentioned by wattsy, is an example of a recent M lens that may support my guess.

[wlaidlaw]

The 18mm SEM, a lens introduced well into the digital era, also performs slightly better at the edges on the SL, than it does on the M240. This is another argument to say that Leica may be building in deliberate astigmatism for thicker cover glass. OTOH, did Leica know the sensor specification of the SL at the time both the 18SEM was designed - I doubt it, so you might have expected it to have been optimised for the M9 or M240. All this points to other factors being at work. 

 

Wilson

[jaapv]

I think you are perfectly correct in your understanding, Ian  - but there is no indication whatever that Leica took this route. Certainly no statement from Leica, we will have to wait for a next interview with Peter Karbe for that, I suspect.

 

I would be surprised if Leica had introduced counter-aberrations for five reasons:

 

1. They set high store on (retro) compatablity.

2. They still build film M cameras, and have repeatedly stated that their lenses perform as well on film as they do on sensors

3. It would jeopardize forward compatablity with future sensor designs which might not even use a cover glass, or of a different thickness, or of a different refractive index. Leica lens designs are meant for decades rather than years

4. Different sensors within their own brand ( i.e. M, T and Q) would need different corrections.

5. If it came out the resulting Internet storm would give reputation damage, far more than the silly hoohah about the completely legitimate decision to use digital corrections on the Q.

[jaapv]

The 18mm SEM, a lens introduced well into the digital era, also performs slightly better at the edges on the SL, than it does on the M240. This is another argument to say that Leica may be building in deliberate astigmatism for thicker cover glass. OTOH, did Leica know the sensor specification of the SL at the time both the 18SEM was designed - I doubt it, so you might have expected it to have been optimised for the M9 or M240. All this points to other factors being at work. 

 

Wilson

Agree. We might simply be seeing different sensor designs here.

[edwardkaraa]

Making lenses perform on film as well as on sensors equally should be taken with a grain of salt. By keeping the cover glass as thin as possible, they can reduce the performance difference to negligible amounts, the sweet spot being around 1mm thick, which is the minimum thickness they can get away with without compromising the physical resistance and IR filtration.

[jaapv]

The lenses perform as well on both media- whether the film or the sensor performs better or worse is another matter.

[edwardkaraa]

The lenses perform as well on both media- whether the film or the sensor performs better or worse is another matter.

Finally we can agree

[flyalf]

I think you are perfectly correct in your understanding, Ian  - but there is no indication whatever that Leica took this route. Certainly no statement from Leica, we will have to wait for a next interview with Peter Karbe for that, I suspect.

 

I would be surprised if Leica had done this for five reasons:

 

1. They set high store on (retro) compatablity.

2. They still build film M cameras, and have repeatedly stated that their lenses perform as well on film as they do on sensors

3. It would jeopardize forward compatablity with future sensor designs which might not even use a cover glass, or of a different thickness, or of a different refractive index. Leica lens designs are meant for decades rather than years

4. Different sensors within their own brand ( i.e. M, T and Q) would need different corrections.

5. If it came out the resulting Internet storm would give reputation damage, far more than the silly hoohah about the completely legitimate decision to use digital corrections on the Q.

Thank you,

 

I wish this could be better communicated and highlighted by Leica. For me these are convincing arguments.

 

P.S. Should it be "I would not be surprised if Leica had done this for five reasons:"?

[jaapv]

For clarity I'll edit

[Herr Barnack]

That could be characterised as "rubbing salt into the wound" 

I suppose it could be - but it was not intended as such.  My apologies go out to jvansmit if he perceived that way.

[CheshireCat]

Don't believe everything Karbe says.

 

 

Blasphemy !

 

It is important to note that there are many different ways to "optimize for digital".

For example, "more telecentric" lenses will improve performance on digital without any side effect on film.

 

The SEM 21 is a lens "optimized for digital". Leica used a "retrofocus-like" design (their words) to reduce problems on digital sensors. The lens lives up to the expectations and is a big step ahead from other film-era 21mm lenses, even on film.

 

Note that the SEM 21 is a new design, while these new 28mm lenses are old (film) designs revised for digital, and as Edward says, their performance on film may be different.

[gpwhite]

Blasphemy !

 

It is important to note that there are many different ways to "optimize for digital".

For example, "more telecentric" lenses will improve performance on digital without any side effect on film.

 

The SEM 21 is a lens "optimized for digital". Leica used a "retrofocus-like" design (their words) to reduce problems on digital sensors. The lens lives up to the expectations and is a big step ahead from other film-era 21mm lenses, even on film.

 

Note that the SEM 21 is a new design, while these new 28mm lenses are old (film) designs revised for digital, and as Edward says, their performance on film may be different.

 

I hope my question does not prompt excommunication from Wetzlar   :

 

My 21 Elmarit ASPH six-bit (i.e., produced during Leica's digital era) is excellent on the SL, with corners fully crisp by f/4.8. I have not shot an SEM 21 on my SL (do not own that lens). Can you comment from your experience with both lenses on an SL that explains why you find the SEM "is a big step ahead from other film-era 21mm lenses"?

 

I am not challenging your evaluation, rather I am very curious. The pre-ASPH 21mm Elmarit was poor on anything, IMHO. But the 21mm ASPH six-bit works beautifully on the SL, and this lens has a) quite a steep ray angle and none of the secret sauce ascribed by some posts here regarding the Summilux 28mm.

 

This thread is great reading, but I am just speculating about what is conjecture? Maybe 50/50 (+1 28 SX; -1 21 EL-ASPH), or do we have other evidence to push off from "semi-ambivalence"?

[Advertisement]

Hallo,

Look here: Leica M

[Leicaiste]

I

 

My 21 Elmarit ASPH six-bit (i.e., produced during Leica's digital era) 

 

 

 

 

The ELMARIT-M 21 mm f/2.8 ASPH. was designed before 1997, more than 9 years before the release of the M8.