Why variable aperture?

[Joe Pineapples II]

The new 16-35 mm SL zoom is huge, heavy, and expensive, and if it is like similar recent offerings from Leica, it will be an under-corrected optical design that relies on software distortion correction, yet it has a variable maximum aperture like a cheap kit lens! It seems a bit underwhelming.

J.

 

[Jared Willson]

The new 16-35 mm SL zoom is huge, heavy, and expensive, and if it is like similar recent offerings from Leica, it will be an under-corrected optical design that relies on software distortion correction, yet it has a variable maximum aperture like a cheap kit lens! It seems a bit underwhelming.

J.

Surely that's not a surprise? The 24-90 was variable aperture and physically very large. The 90-280 was variable aperture and physically very large. Why would the 16-35 be any different?

As far as the use of software for correcting distortion, lateral CA, and vignetting... Within limits, I actually like the concept. For a given size and weight, it allows the final image to be better than it otherwise would be. The same approach is used by:

- Leica

- Nikon

- Canon

- Fujifilm

- Sony

Basically, anyone who designs lenses and bodies to work in concert has adopted this approach. It leaves the optical engineer the ability to give higher priority to addressing other factors such as longitudinal chromatic aberration, spherical aberration, spherochromatism, astigmatism, and coma. It's really all about the final image, and it's not like you were going to use an SL lens on a non-Leica camera, so there really isn't much of a down side. I suppose one could claim they should have addressed the issues in the lens design rather than in software, but something else would have to give.

As far as the variable aperture goes, to get a zoom to have a fixed minimum focal ratio, you basically need to design the lens to have the maximum aperture at the telephoto end of the range, then mask it for shorter focal lengths. That requires more glass and more expense than what Leica is doing. The lenses are already heavy and expensive enough, in my view. I wouldn't want to carry around a 90-280 f/2.8 constant aperture lens. Heck, I wouldn't even want to carry the current 90-280.

So, am I underwhelmed? Nah. The 24-90 is unquestionably the best zoom I have ever owned. It's a huge improvement on the Nikon 24-70 f/2.8, for example.

 

[Joe Pineapples II]

Surely that's not a surprise? The 24-90 was variable aperture and physically very large. The 90-280 was variable aperture and physically very large. Why would the 16-35 be any different?

As far as the use of software for correcting distortion, lateral CA, and vignetting... Within limits, I actually like the concept. For a given size and weight, it allows the final image to be better than it otherwise would be. The same approach is used by:

- Leica

- Nikon

- Canon

- Fujifilm

- Sony

Basically, anyone who designs lenses and bodies to work in concert has adopted this approach. It leaves the optical engineer the ability to give higher priority to addressing other factors such as longitudinal chromatic aberration, spherical aberration, spherochromatism, astigmatism, and coma. It's really all about the final image, and it's not like you were going to use an SL lens on a non-Leica camera, so there really isn't much of a down side. I suppose one could claim they should have addressed the issues in the lens design rather than in software, but something else would have to give.

As far as the variable aperture goes, to get a zoom to have a fixed minimum focal ratio, you basically need to design the lens to have the maximum aperture at the telephoto end of the range, then mask it for shorter focal lengths. That requires more glass and more expense than what Leica is doing. The lenses are already heavy and expensive enough, in my view. I wouldn't want to carry around a 90-280 f/2.8 constant aperture lens. Heck, I wouldn't even want to carry the current 90-280.

So, am I underwhelmed? Nah. The 24-90 is unquestionably the best zoom I have ever owned. It's a huge improvement on the Nikon 24-70 f/2.8, for example.

Lumping all those manufacturers into the same basket is going a bit far IMO. I don't know about all the offerings, but Canon's two standard 24-70 mm "L" zooms are reasonably conventional, with medium distortion at the wide end that can be corrected in PP, but perfectly useable without correction. The Leica SL 24-90 zoom has software correction "baked-in" to the DNGs, and without it, the distortion is unacceptably high. So the design approach is a bit different.

I'm not claiming that is a bad approach to rely on software corrections in designing a lens+camera optical system, but I'm questioning where is the payoff? It seems to result in lenses that are bigger, heavier, slower, and more expensive. I know the answer will be that they are simply "better" in some obvious and yet unmeasurable Leica way, but the smaller, lighter, faster, cheaper Canon 24-70 mm f/2.8L II is pretty darned good, and I'm getting old and cynical.

J.

 

[bosjohn21]

The new 16-35 mm SL zoom is huge, heavy, and expensive, and if it is like similar recent offerings from Leica, it will be an under-corrected optical design that relies on software distortion correction, yet it has a variable maximum aperture like a cheap kit lens! It seems a bit underwhelming.

J.

the short answer to why variable apertures are in the trade offs. If it was a constant f3.5 the front element would be about twice the size to get enough light for that rating. if you think its big now a constant will be much bigger/

there are also some distortion issues that are more easily and better addressed using a variable aperture.

as for cheap kit I have never used one so I have no opinion

 

[Joe Pineapples II]

the short answer to why variable apertures are in the trade offs. If it was a constant f3.5 the front element would be about twice the size to get enough light for that rating. if you think its big now a constant will be much bigger/

there are also some distortion issues that are more easily and better addressed using a variable aperture.

But Canon's "standard" zoom is smaller, lighter, faster, and it has a constant aperture and less optical distortion, compared to the equivalent SL lens, so I don't buy your story!

as for cheap kit I have never used one so I have no opinion

They are plastic and mass-produced, but optically they represent astonishingly good performance / $.

J.

 

[bosjohn21]

the rule of thumb is the faster the lens the bigger the front element and this is true of zooms. the shorter the zoom that is wider sorry the bigger the front element will need to be to collect the light. I have no idea what canon does. I think they could make an f4 constant aperture zoom in a reasonable size

 

[Khun_K]

The new 16-35 mm SL zoom is huge, heavy, and expensive, and if it is like similar recent offerings from Leica, it will be an under-corrected optical design that relies on software distortion correction, yet it has a variable maximum aperture like a cheap kit lens! It seems a bit underwhelming.

J.

Surely that's not a surprise? The 24-90 was variable aperture and physically very large. The 90-280 was variable aperture and physically very large. Why would the 16-35 be any different?

As far as the use of software for correcting distortion, lateral CA, and vignetting... Within limits, I actually like the concept. For a given size and weight, it allows the final image to be better than it otherwise would be. The same approach is used by:

- Leica

- Nikon

- Canon

- Fujifilm

- Sony

Basically, anyone who designs lenses and bodies to work in concert has adopted this approach. It leaves the optical engineer the ability to give higher priority to addressing other factors such as longitudinal chromatic aberration, spherical aberration, spherochromatism, astigmatism, and coma. It's really all about the final image, and it's not like you were going to use an SL lens on a non-Leica camera, so there really isn't much of a down side. I suppose one could claim they should have addressed the issues in the lens design rather than in software, but something else would have to give.

As far as the variable aperture goes, to get a zoom to have a fixed minimum focal ratio, you basically need to design the lens to have the maximum aperture at the telephoto end of the range, then mask it for shorter focal lengths. That requires more glass and more expense than what Leica is doing. The lenses are already heavy and expensive enough, in my view. I wouldn't want to carry around a 90-280 f/2.8 constant aperture lens. Heck, I wouldn't even want to carry the current 90-280.

So, am I underwhelmed? Nah. The 24-90 is unquestionably the best zoom I have ever owned. It's a huge improvement on the Nikon 24-70 f/2.8, for example.

If the objective of a lens is reasonable size and optimized optical performance and arrive at a variable aperture then it is what it is. Fixed aperture zooms have more value in the past when camera automation is more limited and autofocus system requires brighter aperture for less efficient phase detection chips to work but nowadays it is less important, IMHO.

The 24-90 is a perfect example, although still big (and by comparison makes the 90-280 reasonable in size) - but it is part of the design concept that goes along with the whole SL system and I do not particular miss it is not a fixed aperture zoom.

 

[JT26]

I agree to be honest Joe. I find it very underwhelming.

An Leica SL, the three zooms and 50 lux that they offer, will be quite the lump to carry around. They always design super long lenses, and I dont know why...

 

[Joe Pineapples II]

If the objective of a lens is reasonable size and optimized optical performance and arrive at a variable aperture then it is what it is. Fixed aperture zooms have more value in the past when camera automation is more limited and autofocus system requires brighter aperture for less efficient phase detection chips to work but nowadays it is less important, IMHO.

The 24-90 is a perfect example, although still big (and by comparison makes the 90-280 reasonable in size) - but it is part of the design concept that goes along with the whole SL system and I do not particular miss it is not a fixed aperture zoom.

I just don't buy these kinds of rationalisations. Older AF systems didn't require faster lenses to work - that simply isn't true. And once metering systems measured light TTL, they compensated for variable aperture automatically. A Pentax Spotmatic meter would work fine with a variable aperture zoom.

The SL is a mirrorless system and it has baked-in software lens correction - two technologies that should make the lens designer's job easier; but the lenses are heavier, larger, slower, (and much more expensive - well, it is Leica) than their DSLR equivalents. Shouldn't there be a benefit along at least one of the dimensions of the measurable performance envelope?

J.

 

[Jared Willson]

Surely that's not a surprise? The 24-90 was variable aperture and physically very large. The 90-280 was variable aperture and physically very large. Why would the 16-35 be any different?

As far as the use of software for correcting distortion, lateral CA, and vignetting... Within limits, I actually like the concept. For a given size and weight, it allows the final image to be better than it otherwise would be. The same approach is used by:

- Leica

- Nikon

- Canon

- Fujifilm

- Sony

Basically, anyone who designs lenses and bodies to work in concert has adopted this approach. It leaves the optical engineer the ability to give higher priority to addressing other factors such as longitudinal chromatic aberration, spherical aberration, spherochromatism, astigmatism, and coma. It's really all about the final image, and it's not like you were going to use an SL lens on a non-Leica camera, so there really isn't much of a down side. I suppose one could claim they should have addressed the issues in the lens design rather than in software, but something else would have to give.

As far as the variable aperture goes, to get a zoom to have a fixed minimum focal ratio, you basically need to design the lens to have the maximum aperture at the telephoto end of the range, then mask it for shorter focal lengths. That requires more glass and more expense than what Leica is doing. The lenses are already heavy and expensive enough, in my view. I wouldn't want to carry around a 90-280 f/2.8 constant aperture lens. Heck, I wouldn't even want to carry the current 90-280.

So, am I underwhelmed? Nah. The 24-90 is unquestionably the best zoom I have ever owned. It's a huge improvement on the Nikon 24-70 f/2.8, for example.

Lumping all those manufacturers into the same basket is going a bit far IMO. I don't know about all the offerings, but Canon's two standard 24-70 mm "L" zooms are reasonably conventional, with medium distortion at the wide end that can be corrected in PP, but perfectly useable without correction. The Leica SL 24-90 zoom has software correction "baked-in" to the DNGs, and without it, the distortion is unacceptably high. So the design approach is a bit different.

I'm not claiming that is a bad approach to rely on software corrections in designing a lens+camera optical system, but I'm questioning where is the payoff? It seems to result in lenses that are bigger, heavier, slower, and more expensive. I know the answer will be that they are simply "better" in some obvious and yet unmeasurable Leica way, but the smaller, lighter, faster, cheaper Canon 24-70 mm f/2.8L II is pretty darned good, and I'm getting old and cynical.

J.

I can’t tell you exactly what factors the optical designers were balancing when they came up with the Leica SL zooms or which factors received the highest priority. However, I will say that the Leica 24-90 is about the same weight and length as the Nikon 24-70, but you get an extra 20mm of reach with the Leica and significantly better optical performance wide open. That is in exchange for a variable aperture and, of course, Leica pricing.

Can one make different compromises and get a result that might suite any given photographer better? Sure. But I am generally happy with the approach to the SL zooms. If I want smaller and less bulk, I grab my CL instead. For me, the SL and it’s lenses are more about image quality rivaling a good prime coupled with fast AF performance and a robust design. That’s exactly what I would want in a workhorse camera and it’s lenses.

 

[Erik Kaffehr]

The new 16-35 mm SL zoom is huge, heavy, and expensive, and if it is like similar recent offerings from Leica, it will be an under-corrected optical design that relies on software distortion correction, yet it has a variable maximum aperture like a cheap kit lens! It seems a bit underwhelming.

J.

...

As far as the use of software for correcting distortion, lateral CA, and vignetting... Within limits, I actually like the concept. For a given size and weight, it allows the final image to be better than it otherwise would be. The same approach is used by:

- Leica

- Nikon

- Canon

- Fujifilm

- Sony

You can add Hasselblad to that club.

Digital corrections make a lot of sense, mostly regarding distortion and lateral color. But those corrections loose resolution.

What I don't understand with Leica is rather that they keep 24MP on 24x36 mm and 37.5 MP on 45x30. Doing software correction throws away resolution and introduces artifacts, so it is better to start with a high resolution.

The other side of the coin is that high resolution, or rather a high contrast at the pixel level causes artifacts, fake detail. Most vendors combine low resolution sensors with an antialiasing filter, even if Nikon and Sony have dropped that filter at 36 MP and above.

So, sharp lenses, 24MP and no aliasing filter should not play well.

Best regards

Erik