M9, last of the line?

[rosuna]

Advertisement (gone after registration)

Changing the flange to sensor distance has absolutely nothing to do with the lens design and the angle of incidence.

 

It is related. Go to a "strong" case. Compare a 21mm for M system and a 21mm for a reflex system. The retrofocus design of the reflex lens allows for a more perpendicular angle of incidence.

[rosuna]

Which is exactly what Leica did on the M8 and M9 - have a look and you can see the protruding mount.

 

They already saved a bit of thickness that way then. It is a good trick, isn't?

[AlanG]

Yes, but the relative lens-to-flange distance (relative to the diagonal of the format) is much larger than that of the M system for the first two cases. The sensors are quite small, so the entire mechanism is smaller than would be necessary for a 24x36mm sensor. The 5D has the advantage of a larger lens-to-flange distance, so the light rays reach the sensor more perpendicularly, and the filters, cover glass and microlenses don't bring the problems the M system has to deal with.

 

 

I think you have it completely backwards. The fact that the M has a short flange to sensor distance gives it much greater freedom for lens design than in a DSLR where the lens can't intrude into the mirror box. Why do you think some wide angle lenses require a mirror lockup on an SLR? Because they protrude past the mirror despite the larger flange to focal plane distance.

 

There is nothing forcing some Leica wide angle lenses to be designed the way they are. E.g - reaching into the camera mount somewhat. The only reason Leica chose this design was to make them as small as possible. They could just as easily have used other lens designs that had longer barrels to compensate for the thinner body or made retrofocus lenses. I can't see why you don't get this. A lens from a DSLR will work perfectly well on an M if you have the correct adapter. And two otherwise identical 50mm lenses - one for an SLR and one for an M, could exist and work the same, but the one for the M would have a slightly longer barrel to make up for the shorter flange to focal plane distance.

 

The flange to sensor size has nothing at all to do with where the optical elements are placed (within physical limits.) If a new M body was made that was 10 mm thinner, all you'd need is a 10mm spacer in front of the body for the lenses to behave exactly the same. Try to picture a view camera. It basically has a ground glass for focusing that is replaced by film or a sensor for a picture. There is no fixed flange to focal plane distance at all. This is determined by wherever you adjust the front standard to focus a particular lens. Any kind of lens design can be accommodated as long as the bellows can extend far enough or the standards can be squeezed together close enough. The mechanical way this is achieved - bellows, thicker body, lens barrel size, is irrelevant as long as you locate the optics the correct distance from the sensor plane. A thinner camera body gives you the same flexibility. An example is why Leica R lenses can be used on Canon bodies but not the other way around.

 

As for IS... it can either be made to fit or it can't. It all comes down to how much Leica wants to redesign the camera body, their engineering skills, and if they have access to the technology. The only company that currently has this in a full frame camera is Sony, so I don't know if Leica can license this and I doubt they can develop it on their own. A dusk shaker looks pretty easy in comparison... there is definitely a cover glass in front of the sensor that could be shaken... not stirred. (They probably have to license this technology.)

[jaapv]

Alan is perfectly right: Erwin Puts :

The second[...] argument for the change to this newer [retrofocus] type of design is their greater optical potential. As soon as the designer has familiarized himself with the inherent characteristics of retro-focus designs, (s)he has more opportunities for correction

[jaapv]

I must confess that about two years ago I would have been firmly in the thinner-body camp. However, nowadays, when I take a film M in my hands I find it simply lacks some thickness. Clearly that design does not hit the ergonomic optimum. The LTM bodies are fine as they are, but the M3 to M7, in retrospect, are simply too thin.

[stunsworth]

I must confess that about two years ago I would have been firmly in the thinner-body camp. However, nowadays, when I take a film M in my hands I find it simply lacks some thickness

 

Personal preference I guess. I'd prefer something the size of my M6, but I'm realistic enough not to expect that to happen.

[jaapv]

Advertisement (gone after registration)

It certainly is personal preference, I have size 10 for gloves But it is a matter of habituation as well.

[rosuna]

Of course Alan, you can put an extension tube and use lenses designed for reflex cameras on a M body. You can even design lenses that way for the M system. You can use 4/3 lenses on a micro 4/3 body too (it is a similar case). But that is nonsense. You would have to pay a very high price for avoiding the angle of incidence problem: very large lenses mounted on extension tubes. So, in theory, you are not forced to use lenses designed specifically for the M system, but the M system only has sense with lenses designed for it. I can't see why you don't get this. The M system imposes -for optimum use of the advantages of it- a particular set of constrains for its lenses.

 

The lens-to-flange distance of the M system is 28mm, so any lens under 28mm should bring a solution similar to a retrofocus design. For instance, the 21mm lenses, pre-ASPH (from 1980) and ASPH. But even in those cases the angle of incidence will be more acute than that of a 21mm retrofocus design for a reflex camera, which has to be far more radical... due to the much longer lens-to-flange distance.

 

Now we return to the starting point. Leica may design lenses for use with extension tubes (or including themselves some kind of extension tube) on the M mount, of course, but that would make the camera uncomfortable to use. It would be nonsense from a practical point of view. The system's parameters impose constraints.

 

The angle of incidence of the light rays on the filters and microlenses bring many problems to the M8 and M9. I think the "fringe" problems of many M lenses on digital bodies may be due to this to a large extent, and not to chromatic aberrations of the lens. You cannot move the sensor for image stabilization and change the angle of incidence without worsening those problems on one side of the frame. You cannot do it without problems even with R lenses with R-to-M adapters because the M sensor has special microlenses designed for M lenses. Those microlenses incorporate an estimation of the angle of incidence depending on the distance from the axis. I guess that estimation is some kind of average, because the angle of incidence has to change if we compare wide angles and tele lenses for the M system (even considering wide angles may be retrofocus designs and teles are normal lenses extended). But R lenses are completely different. Only an engineer with real information might make this clear, but my feeling is that there are problems here.

 

The dust removal may be of two types: movement of the whole sensor; or movement of a filter in front of the sensor. Considering the second type, you have to separate the filter and the sensor, at least for a fraction of a millimeter. That introduces a new jump glass-air, and reduces even more the distance between the filter and the lens. I don't say it is impossible to do. I don't know. But it seems to be problematic.

[ho_co]

... You cannot move the sensor for image stabilization and change the angle of incidence without worsening those problems on one side of the frame ... because the M sensor has special microlenses....

Excellent point! Brilliant! It's obvious once pointed out, but I hadn't seen it! Thanks, Rubén!

[ho_co]

Alan is perfectly right: Erwin Puts :

 

Jaap, I respectfully disagree. You're citing from page 102 of the Leica Lens Compendium, web version. It's one of Erwin's famously subtle arguments. If you read the rest of the paragraph, he's arguing that:

 

1) The M5 was the cause of the move away from symmetrical designs. (Alan has read back today's standards into the past on this one.)

 

2) Once understood, retrofocal designs offer a new degree of freedom for the designer.

 

3) Nonetheless, other parameters of the rangefinder's design mitigate against this supposedly increased freedom.

 

To me, he's saying, 'yes, symmetrical designs are advantageous, but when you're forced to use a retrofocal design, you can do things that can't be done with symmetrical designs.'

[inte]

Herr Karbe has already gone on record saying that those lenses outresolve the current sensor AND the next generation.

 

I would LOVE to see this. Any documents online with some sort of quote this effect, or the data Karbe was referring to?

[AlanG]

I don't know what Rosuna or some of you are arguing about. I did not say that Leica should change any lens designs. I only said that the thinner the body, the more options lens designers have as you can always extend a lens barrel, but it may be harder to recess a lens.

 

For the last time Rosuna, try to get this... the flange to focal plane distance is only a factor if it is too great to support a given lens design. Simply having a thin body does not imply you need to use any particular lens design. The fact that some Leica lenses are not very retrofocus and thus have a higher angle of incidence simply has to do with a decision by lens designers and has nothing at all to do with the flange to focal plane distance. Some lenses recess into the camera and some do not.

 

The main reason Leica may reconsider using retrofocus lens designs is to avoid the current problems some w/a lenses have on the M9. Most likely the resulting lenses will be larger and heavier but will not require a greater flange to focal plane length. They simply will extend the lens elements out further by making a slightly longer barrel. Anyone who looks at images from the new Canon 17 and 24 TSE lenses will see what high image quality can be achieved with a retrofocus design and how large those lenses turned out to be.

 

The hope is that there may be some kind of compromise or radically different design that will work for smaller lenses on the M9 or at least on future M digital cameras. And making the flange to focal plane smaller and the lens mount larger will avoid any possible physical restrictions to achieving those designs. For all we know a successful smaller w/a design could be telecentric and also have a very large rear element only a 1/2 inch from the sensor. Why restrict this possibility?

 

Maybe this will help as it shows the relationship between focal length, object distance and image distance. The flange to sensor distance in the M is purely arbitrary in order to make the body a given thickness. It could be larger or smaller and the same lenses would work on it as long as the actual lens elements are located the same distance from the sensor.

 

#1 Site To Learn About Light

[hoppyman]

I would LOVE to see this. Any documents online with some sort of quote this effect, or the data Karbe was referring to?

Inte that quote is from the interview with Peter Karbe which is published on the Luminous Landsape website. Obviously there is NOONE better qualified to speak on the subject You can see the MTF diagrams for yourself of course in the PDFs for each lens on the Leica Camera site.

Interview with Leica's Peter Karbe on Vimeo

There are other interviews there too and the one with Stefan Daniel is particularly instructive, as I mentioned, earlier in this thread. Nothing like hearing from the people that set policy or run the divisions to put forum speculation and opinion in perspective.

[dpattinson]

Was just chatting with a dealer, and he's under the impression that the impending EVIL 'R' solution will also mount M lenses.

 

Also mentioned 'variant M9' bodies being in the pipeline, but was pretty sure this didn't mean variant sensor.

[jaapv]

Jaap, I respectfully disagree. You're citing from page 102 of the Leica Lens Compendium, web version. It's one of Erwin's famously subtle arguments. If you read the rest of the paragraph, he's arguing that:

 

1) The M5 was the cause of the move away from symmetrical designs. (Alan has read back today's standards into the past on this one.)

 

2) Once understood, retrofocal designs offer a new degree of freedom for the designer.

 

3) Nonetheless, other parameters of the rangefinder's design mitigate against this supposedly increased freedom.

 

To me, he's saying, 'yes, symmetrical designs are advantageous, but when you're forced to use a retrofocal design, you can do things that can't be done with symmetrical designs.'

Maybe I misunderstood Alan. i thought he was saying retrofocus design gives the designer more options, so the rear protrusion is not the main argument for the choice. Point three I do not understand, as it is just one of the choices a designer makes. Lenses designed after the M5-CL era have retrofocus characteristics as well.

[jaapv]

Of course Alan, you can put an extension tube and use lenses designed for reflex cameras on a M body. You can even design lenses that way for the M system. You can use 4/3 lenses on a micro 4/3 body too (it is a similar case). But that is nonsense. You would have to pay a very high price for avoiding the angle of incidence problem: very large lenses mounted on extension tubes. So, in theory, you are not forced to use lenses designed specifically for the M system, but the M system only has sense with lenses designed for it. I can't see why you don't get this. The M system imposes -for optimum use of the advantages of it- a particular set of constrains for its lenses.

 

The lens-to-flange distance of the M system is 28mm, so any lens under 28mm should bring a solution similar to a retrofocus design. For instance, the 21mm lenses, pre-ASPH (from 1980) and ASPH. But even in those cases the angle of incidence will be more acute than that of a 21mm retrofocus design for a reflex camera, which has to be far more radical... due to the much longer lens-to-flange distance.

 

Now we return to the starting point. Leica may design lenses for use with extension tubes (or including themselves some kind of extension tube) on the M mount, of course, but that would make the camera uncomfortable to use. It would be nonsense from a practical point of view. The system's parameters impose constraints.

 

The angle of incidence of the light rays on the filters and microlenses bring many problems to the M8 and M9. I think the "fringe" problems of many M lenses on digital bodies may be due to this to a large extent, and not to chromatic aberrations of the lens. You cannot move the sensor for image stabilization and change the angle of incidence without worsening those problems on one side of the frame.

You cannot do it without problems even with R lenses with R-to-M adapters because the M sensor has special microlenses designed for M lenses.

Those microlenses incorporate an estimation of the angle of incidence depending on the distance from the axis. I guess that estimation is some kind of average, because the angle of incidence has to change if we compare wide angles and tele lenses for the M system (even considering wide angles may be retrofocus designs and teles are normal lenses extended). But R lenses are completely different. Only an engineer with real information might make this clear, but my feeling is that there are problems here.

 

The dust removal may be of two types: movement of the whole sensor; or movement of a filter in front of the sensor. Considering the second type, you have to separate the filter and the sensor, at least for a fraction of a millimeter. That introduces a new jump glass-air, and reduces even more the distance between the filter and the lens. I don't say it is impossible to do. I don't know. But it seems to be problematic.

All R wideangles, up to and including the fisheye, work without any problem on the M9. No edge problems, no vignetting, The results are as good or better than the equivalent M lenses.

[Guest]

Maybe Leica becoming it's best competitor?

I changed from DSLR mainly due to IQ/size, as a logic consequence after going from 1Ds to 5D.

What if we saw the next champion from Solms (I'm saving-saving, but not for a non-chrome cam for over 5k) being a true M which will be...

SMALLER :) !

 

The RF which I find perfect-PERFECT (to the extent that I wish young Leicaners to celebrate in the mid-fityes of this century ), the range finder is the limiting size factor.

 

What if the M xy would begin it's elegant curve very close to the right of the accessory RF window and hence be 2-3cm less broad? And still house a FF chip, with the dials L&R of the LCD on the back still being in centre?

 

Once at it, it could be 2mm thinner than the M1-M7, too. With the M-mount being the most protrusive feature. Not on only users with a medium glove size would agree.

 

Serious: digi-miniaturisation will be there, in what? 2 years maybe.

And it would be more useful for M-shooters than any hybridisation of features, that we do without very well now, and which would lead to a mandatory new lens line for a clientele, who likes evolutionary improvements.

If they would venture into a half compatible lens line with a semi-lame AF, they'll have the New Coca-Cola or the Porsche 928.

Hassy did well not to start a compact (or RF), instead of perfecting the core product.

My 2 cents.

 

Cheers!

[AlanG]

Maybe I misunderstood Alan. i thought he was saying retrofocus design gives the designer more options, so the rear protrusion is not the main argument for the choice. Point three I do not understand, as it is just one of the choices a designer makes. Lenses designed after the M5-CL era have retrofocus characteristics as well.

 

"3) Nonetheless, other parameters of the rangefinder's design mitigate against this supposedly increased freedom."

 

I'm guessing he meant - "You can only make the lenses so large before they block too much of the viewfinder."

[jaapv]

Well, although I feel less than confident second-guessing Erwinesque pronouncements, that is true, and that might be easier to keep under control with a wider mount.

[Advertisement]

Hallo,

Look here: Leica M

[ho_co]

Maybe I misunderstood Alan. i thought he was saying retrofocus design gives the designer more options, so the rear protrusion is not the main argument for the choice. Point three I do not understand, as it is just one of the choices a designer makes....

 

Again, I was just saying that my interpretation of Erwin's text differs from yours.

 

The paragraph from which you quoted relates to the first retrofocal 21. I think the retrofocal Midland 28 mentioned on p 104 preceded the initial retrofocus 21.

 

Puts says that it was the M5 that brought about the first retrofocus designs. In the paragraph you mentioned, he then goes on to speak of the "second, more disguised" argument. That's why I disagree that "the rear protrusion is not the main argument."

 

... Lenses designed after the M5-CL era have retrofocus characteristics as well.

Correct.

 

 

... I'm guessing he meant - "You can only make the lenses so large before they block too much of the viewfinder."

I think that's the most likely point he was making, though of course that explanation is a bit restrictive.

 

 

 

EDIT: I notice that some of my edits are not sticking. I had a couple corrections that dropped out with later edits.