Leica M Lenses -- what's up next?

[michaelwj]

Advertisement (gone after registration)

Are you willing to trade the 2,0 for a 4,0 aperture in this case?

 

To quote an unknown photographer ( but very good violin player ) : " Ich slep mich noch lieber am Tode " roughly translated: I'll rather die, than go for lightweight lenses...

 

If you want big and heavy, get an SL, or any DSLR.

 

The M has always had small lenses with moderate apertures in order to allow light to the RF windows for starters.

[Dirk Mandeville]

Horses for courses. If the goal is a small, lightweight bi focal length lens, f/4 makes sense. Not sure I would have need for such a lens, but no doubt plenty of people would love them. I wonder if they would be concerned that such a lens might cannibalize sales of the other lenses in those focal lengths. May not be worth it to them financially to put the resources into developing such a lens.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[pico]

Are you willing to trade the 2,0 for a 4,0 aperture in this case?

 

Certainly, gentlemen to not go out in the dark.

.

[gcspeed]

I have no idea what the plans are but a new collapsable 50 summicron with current optics and minimum focus of 0.7m would make me pull out my wallet.  The extra portability without having to give up speed as with the 2.4 summarit or 2.8 elmar would be great.  

[james.liam]

Are you willing to trade the 2,0 for a 4,0 aperture in this case?

 

To quote an unknown photographer ( but very good violin player ) : " Ich slep mich noch lieber am Tode " roughly translated: I'll rather die, than go for lightweight lenses...

 

 

Got a 21 SEM and it's incomparable. To travel with a pair of light bi-Elmars and a fast 35 or 50 when it's called for? Count me in for the trade (up). 

[Jared]

How many lenses do they have room for before the six bit code runs out!

[Mark Pope]

Advertisement (gone after registration)

By my reckoning the M can handle up to 255 lens codes. The 6 bit code on its own provides 31 discrete values. Factor in the frame lines you could get up to 255.

But then it's early and I haven't had any coffee yet :-)

Edited March 1, 2017 by Mark Pope

[pop]

By my reckoning the M can handle up to 255 lens codes. The 6 bit code on its own provides 31 discrete values. Factor in the frame lines you could get up to 255.

But then it's early and I haven't had any coffee yet :-)

How so? Three times 2^6 = 3 * 64 = 192.

[Mark Pope]

There are three frame line which values will use 2 bits - for example: 28/90 00, 35/135 01 and 50/75 11. Prepend these to the 6 on the lens gives 8 bits, which gives 256 values - 0 to 255.

It's purely speculation on my part, but when you think about how the MATE works on a digital M, I can't think how else it could work.

Edited March 1, 2017 by Mark Pope

[wattsy]

It's a bit late now but wouldn't it have been better – more future proof and possibly allowing for other features – for Leica to have introduced some kind of ROM contact for the M mount? Unless I misunderstand how these things work (which is quite possible) presumably a ROM chip could be embedded in an M lens without interfering with the ability to use that lens on a film camera. The 6-bit system is ingenious in a way but it is also restricted (it can only communicate one piece of data to the camera, no matter how sophisticated the lens might become) and seems a little Heath Robinson to me. For the M system, Leica has seemed reluctant to fully embrace electronic opportunities; preferring, instead, hybrid mechanical solutions such as the half-baked, and now abandoned, M7 DX reader.

[Exodies]

There are three frame line which values will use 2 bits - for example: 28/90 00, 35/135 01 and 50/75 11. Prepend these to the 6 on the lens gives 8 bits, which gives 256 values - 0 to 255.

It's purely speculation on my part, but when you think about how the MATE works on a digital M, I can't think how else it could work.

2bits but only 3 different values are detected.

[Stein K S]

Since I just bought a silver chrome on brass 50 Lux ASPH I'm sure Leica are about to introduce a silver chrome on aluminium lightweight version.

If so, you where just in time then... Congrats!

Edited March 1, 2017 by Stein K S

[adan]

Mark - the limiting factor is not how many bits you have in which to store data - but how much data you have to store in the bits.

 

There are only 3 possible frameline positions - 28/90, 35/135, 50/75. Therefore you can't use all of the potential of your additional 2 bits (4 encodings) - it is only 75% efficient. You can encode frameline positions 1, 2, and 3, as 01, 10, and 11 (direct conversion from decimal to binary) - but there is no frameline setting corresponding to 00 (no frameline selected).

 

It's like a camera sensor that can distinguish 16384 tones, or 14 bits. Sure, to hold that, you need 2 bytes (16 bits - 64000+ possibilities) of storage. But that won't magically improve the sensor's capabilities. You just end up with "buckets" that are never used.

 

@ wattsy - study up on the troubled introductions of the R4, and the R8/9 motor, (not to mention the rather kludgy M4-2 motor and M5 meter system) and you'll understand why Leica is always more comfortable if they can machine brass, grind glass, and brush paint - and avoid pushing electrons around.

 

Heck, they had to farm out the electronics of the digital R back and M8 to Imacon and Jenoptik.

 

As to "future-proof" - which do you think will still be functional in 2050? A system using black and white paint and simple IR LEDs/sensors, or a system based on a 2005-era chip? My understanding is that Leica is already have difficulty sourcing "1985" electronics for M6 meters.

Edited March 1, 2017 by adan

[Paulus]

If you want big and heavy, get an SL, or any DSLR.

 

The M has always had small lenses with moderate apertures in order to allow light to the RF windows for starters.

 

We are comparing Bi-Elmar with two summicrons or summiluxes. Another comparison is IMO out of context. It's the difference between a bottle of water or not. Not so much..

[wattsy]

@ wattsy - study up on the troubled introductions of the R4, and the R8/9 motor, (not to mention the rather kludgy M4-2 motor and M5 meter system) and you'll understand why Leica is always more comfortable if they can machine brass, grind glass, and brush paint - and avoid pushing electrons around.

 

Heck, they had to farm out the electronics of the digital R back and M8 to Imacon and Jenoptik.

 

As to "future-proof" - which do you think will still be functional in 2050? A system using black and white paint and simple IR LEDs/sensors, or a system based on a 2005-era chip? My understanding is that Leica is already have difficulty sourcing "1985" electronics for M6 meters.

 

 

I know a fair bit of Leica history, thanks – I don't need to "study up" on the introduction of the R4. The problems you mention (many of which date back to the 1970s and will have involved management and engineers long gone – some in a final sense) doesn't seem to include the ROM chips on the R lenses. Have there been any problems with this system? Presumably these ROM chips (some of which must be nearly 30 years old) are still functional and support is unaffected by the unavailability of 1985 electronics for M6 meters?

 

What is more useful going forward? A "system using black and white paint and simple IR LEDs/sensors" to communicate a single piece of information or one "based on a 2005-era chip" (presumably easily and cheaply swapped out if it became desirable) that opens up the possibility of communicating other information about future lenses that may enable, for example, digital corrections that go beyond the vignetting and colour cast correction (based on semi-guesswork) rooted in the M8-9 era?

[JonPB]

One potential area of improvement is in minimizing breathing. This would be useful as it would make framing more precise regardless of focus distance. It would also leverage expertise perhaps developed by Sonderoptic with cine lenses.

 

If the APO 50 has taught us anything, it is that there are many of us who are willing to pay a large premium for subtle improvements in technical rendering. Making faster lenses with Elmar or Super-Elmar quality would not be unwelcome. 

 

Also, if the new 28 Summaron has taught us anything, it is that some people will be a large premium for modern coatings and the reassurance that comes with new manufacture. With such variability in the used market, that's not a small selling point. Leica has a huge back-catalog of interesting and enjoyable lenses. Maybe make one available per year, available by special order for one year only?

 

I'm not opposed to live-view-only zooms and long lenses for the M mount -- a range of collapsible, modest speed, long lenses would be intriguing -- but I suspect marketing them would be prohibitively confusing to many M customers. Perhaps simply branding them as Visoflex would make it clear that they're designed for EVF use? Heck, they might even be fully backwards compatible then.

 

As for the 6-bit design, 192 lenses is quite a few; going by the wiki's M lenses x ordinal page, the 28 Summilux is the 189th lens, including at least M39, M, R, and Visoflex designs. At that pace, we should be good for another 80 years or so, give or take a few decades. I doubt I'll live long enough to see them run out of codes. Still, ROMs are incredibly durable, and I don't see any practical difference between designing an optical system to read 6-bit codes off a lens versus an electrical circuit that reads a relatively simple ROM chip. I suspect any electrical engineer worth their degree could, given basic specifications, input the correct voltage and make sense of the output. This is simpler than designing a system to work with the I/O system on a Raspberry Pi, which is to say that any motivated 5th grade class could figure it out. So future readability doesn't concern me, so long as Leica publishes specs (which would be a novelty for them, I fear). But, I suspect a chip would simply take more space inside the lens, and small size is the Barnack/M forte. So I guess I'm pro-6-bit. That said, I don't care either way: I prefer to use lens profiles in post than Leica's built in, best guess, all purpose correction system, though I wish it were an option to only use lens detection/manual selection for EXIF only. I'll be happy using any M lens that is compatible with the M3.

 

Speaking of film cameras -- I think Leica should develop a film scanner that does justice to modern technology. But that's another topic. :-)

 

I'm also superbly happy with Kölsch-era lenses, so I don't have a horse in the lens-design race. My comments are therefore somewhat moot. Still, you've read this far, so--

 

Cheers,

Jon

[TRIago]

Please bring the 50mm DR Summicron back. The exact same way it was inside and outside. But this time with the near range working for digital. Danke.

 

 

Sent from my iPhone using Tapatalk

[wattsy]

Still, ROMs are incredibly durable, and I don't see any practical difference between designing an optical system to read 6-bit codes off a lens versus an electrical circuit that reads a relatively simple ROM chip. I suspect any electrical engineer worth their degree could, given basic specifications, input the correct voltage and make sense of the output. This is simpler than designing a system to work with the I/O system on a Raspberry Pi, which is to say that any motivated 5th grade class could figure it out. So future readability doesn't concern me, so long as Leica publishes specs (which would be a novelty for them, I fear). But, I suspect a chip would simply take more space inside the lens,

 

 

Yes, I think size may have been a good reason why Leica decided in favour of the 6-bit system back in 2005/6. The lens recognition system had to be backwards compatible with almost all lenses that have ever been in the M system and I guess there may not be enough room in a typical densely packed M lens for a ROM chip. My assumption has been that a tiny chip could be embedded in the flange itself (within a machined pit) but maybe the flange is too thin to allow that?

Edited March 2, 2017 by wattsy

[Mark Pope]

Mark - the limiting factor is not how many bits you have in which to store data - but how much data you have to store in the bits.

 

There are only 3 possible frameline positions - 28/90, 35/135, 50/75. Therefore you can't use all of the potential of your additional 2 bits (4 encodings) - it is only 75% efficient. You can encode frameline positions 1, 2, and 3, as 01, 10, and 11 (direct conversion from decimal to binary) - but there is no frameline setting corresponding to 00 (no frameline selected).

 

It's like a camera sensor that can distinguish 16384 tones, or 14 bits. Sure, to hold that, you need 2 bytes (16 bits - 64000+ possibilities) of storage. But that won't magically improve the sensor's capabilities. You just end up with "buckets" that are never used.

 

@ wattsy - study up on the troubled introductions of the R4, and the R8/9 motor, (not to mention the rather kludgy M4-2 motor and M5 meter system) and you'll understand why Leica is always more comfortable if they can machine brass, grind glass, and brush paint - and avoid pushing electrons around.

 

Heck, they had to farm out the electronics of the digital R back and M8 to Imacon and Jenoptik.

 

As to "future-proof" - which do you think will still be functional in 2050? A system using black and white paint and simple IR LEDs/sensors, or a system based on a 2005-era chip? My understanding is that Leica is already have difficulty sourcing "1985" electronics for M6 meters.

Fair point. So Pop is right. 192 discrete values ( note to self to let caffeine- blood level stabilise in future :-) )

[Advertisement]

Hallo,

Look here: Leica M

[michaelwj]

We are comparing Bi-Elmar with two summicrons or summiluxes. Another comparison is IMO out of context. It's the difference between a bottle of water or not. Not so much..

Sorry, my comparison was not with a proposed bi-elmar, but with a bi-summicron, which would be huge provided it spanned a useful focal range like 28-50 or even 35-50, rather than a pretty meaningless 75-90 for instance which would not be much larger than the 90.