There are 50mm f/1.4 lenses that fit inside a coffee cup, and 50mm f/1.4 lenses that take two hands to lift. So why are some lenses so large?
It’d be tempting to think this is caused by the need for speed, and yes, the light transmission of a lens is determined to an extent by its size, but that’s probably not the fundamental issue.
An f-stop is the focal length divided by the size of the “entrance pupil,” which strictly speaking means the size of the aperture as viewed through any lens elements that are in front of the aperture.
That’s why many lenses have what looks like a magnifying glass on the front, which makes the iris look bigger. It’s pretty intuitive that if the iris looks bigger, more light will fit down the hole and the lens will be faster.
In the end, what this means is that a 50mm lens only needs to be at least, and in practice a bit more than, 25mm across to achieve f/2, or 36mm to achieve f/1.4. So why are they often bigger?
Reason one – they’re not
If you look down a 50mm cinema lens, you’ll often see a lot of stepped, conical metal with a fairly small bit of glass at the bottom. Why is the shell so much bigger than the glass?
Convenience. There’s a great desire for all the lenses in a set to have their focus and iris gears in the same position, so the diameter of the lens tends to be the minimum diameter required by any of the lenses in a set, which will tend to be determined by the shortest focal length.
The length of the lens, likewise, tends to be the minimum length required by any of the lenses in the set, which will tend to be the longest focal length. So, they’re generally as wide as the widest one, and as long as the longest one, which means that they’re all somewhat bigger than they individually need to be. That’s not big glass, that’s just a big housing.
Reason 2 – Stay away from the edge
Sometimes there are reasons for the glass elements to get larger. On bigger sensors, we’re using more of the image projected by a lens, and the edges of the picture are created by the edges of the glass. That’s where lenses are often least ideal.
Make a lens bigger, and we take more of the image from what might be considered the middle of the elements. At the same time, cameras have become significantly sharper than 35mm ever was, so problems are more visible. Whether we need 8K or not, it’s also true that a larger lens element means that any flaw is simply smaller in comparison to the image overall.
It’s easy to look at stills lenses such as the popular Nikkor AI-s, and wonder that a 50mm lens can be that small and achieve a sub-f/2 minimum aperture while still covering the full 8-perf stills frame.
The huge caveat is how well it does that. These lenses rarely meet massively sharp full-frame cameras like Alexa LF or Venice; they’re far more likely to go on DSLRs that are heavily compressed, or on super-35 devices like Ursa Mini or FS7 which only see the central area. Most people don’t routinely perform side-by-side comparisons of old stills lenses alongside modern full-frame glass.
Reason 3 – Retrofocus
Ever notice that lenses often get physically shorter as focal length slides down toward 50mm, and then get longer again? The shorter the focal length of the lens, the closer it needs to be to the film or sensor to form an image, and that’s a pain.
On lenses made for mounts designed for film cameras – which includes EF and PL – there’s a lot of space behind the mount to accommodate a viewfinder mirror.
This means the lens sometimes can’t get close enough to the sensor for things to work out easily. The solution is a retrofocal design, which we won’t go into here other than to say that it requires more lens elements and thus more space.
A particularly good example of this is the ENG broadcast lens. They’re often incredibly wide – single-digit millimetres – and they clearly can’t get that near the sensors as the chips are mounted on an optical splitter block. There are other requirements for broadcast lenses which make things even more complicated, but this is part of the reason ENG zooms are quite long, and some of this applies to primes too.
This is partly why modern mounts, for mirrorless cameras, place the lens closer to the sensor. Micro four-thirds and Sony E are well known shallow mounts, but the reason Canon developed RF was not to make everyone buy new lenses, or at least not entirely to make them buy new lenses; it was because a shallower mount makes for much easier lens design. RF lenses of the same performance as EF lenses can be cheaper and smaller. Similar considerations might have influenced the design of Arri’s LPL.
There are other reasons lenses can be large. Some are fairly obvious – stills lenses may have inbuilt electronics, image stabilisation, and other optional extras which tend to bulk things up, but some of it is determined by optical physics and usability.
And yes, at least some lenses are big because big lenses cost more.