This site is supported by the advertisements on it, please disable your AdBlocker so we can continue to provide you with the quality content you expect.

Correlation between lens size and focal length (aka why is 50mm lens the smallest)

Discussion in 'General Photography Discussion' started by HansA, Dec 7, 2017.

  1. HansA

    HansA Member

    Joined:
    Dec 1, 2017
    Messages:
    47
    Likes Received:
    6
    Location:
    Indonesia

    -Return to Top-

    If we take a look, a 50mm equivalent lens is usually the smallest in size pretty much universally among different manufacturers. I wonder why this is so?

    I get the notion that longer focal length lenses are usually longer and wider lenses are usually wider in size, but why is the "minimum point" 50mm equivalent and not 35mm for example? I see that the the 50mm is called "normal" and it's supposed to be the closest focal length to stereoscopic human vision but I'm not sure if that has anything to do with the lens size.

    If I could hazard a guess, it has something to do with the standard 35mm film (size)?
     
  2. HansA

    HansA Member

    Joined:
    Dec 1, 2017
    Messages:
    47
    Likes Received:
    6
    Location:
    Indonesia

    -Return to Top-

    After some browsing, I read that apparently they can't make lenses with shorter focal length shorter and shorter in size due to the angle of the light rays hitting the sensor (they have to be as perpendicular as possible especially in digital cameras with their sensor). Thus the many elements added for correction which affects the lens's physical size.

    On a related note, perhaps 50mm equivalent lenses should theoretically have the best IQ due to supposedly having the least amount of elements for correction and the smallest in size (larger lens elements introduce more micro defects) ?
     
    Last edited: Dec 7, 2017
  3. Narsuitus

    Narsuitus Premium Member

    Joined:
    Jan 9, 2015
    Messages:
    1,682
    Likes Received:
    755
    Location:
    USA

    -Return to Top-

    I have noticed that 40 and/or 45mm lenses, for the 35mm film size, tend to me smaller than the 50.
     
    HansA likes this.
  4. GregWard

    GregWard Premium Member

    Joined:
    Mar 21, 2014
    Messages:
    4,253
    Likes Received:
    1,589
    Location:
    London

    -Return to Top-

    For a start I'm not entirely sure it's true. Often a moderate wide angle can be smaller I think. Secondly there is a lot here to do with the nature of the lens design. If you think about it a lens could be just one piece of glass. So why do we even have multiple elements in a lens? Well that obviously has to do with correcting optical defects. Plus there are factors around maximum aperture etc. Just compare a Nikon 50mm f1.8 with the f1.4 version. To get even more silly - add in a Zeiss Otus.

    Then I think (although I confess to not being even remotely an expert) that there's a set of issues around retrofocal designs. A key dimension to recognise here is the "flange distance". This is the distance between the lens mount and the sensor (or film plane). On DSLR cameras this has to be quite long (so the flappy mirror doesn't collide with the back of the lens). On mirrorless cameras it's typically much shorter and on Fuji X it's very short. Now focal length is defined as the distance between the rear nodal point of the lens and the sensor. So what happens when the focal length is shorter than the flange distance? Well clearly the rear nodal point needs to be BEHIND the back of the lens. So that's where you need a retrofocal design. I believe that creates optical complications that require more elements (so a bigger lens) to compensate.

    Net you end up with a sweet spot where shorter focal lengths allow (all other things being equal) to be smaller BUT only up to a point. I also suspect that this "point" (sweet spot) should be strongly influenced by the flange distance.
     
  5. HansA

    HansA Member

    Joined:
    Dec 1, 2017
    Messages:
    47
    Likes Received:
    6
    Location:
    Indonesia

    -Return to Top-

    Yeah, I agree with you that lenses need to have a number of correctional elements and are related the camera body they are made for which determines the final physical size.

    But still, earlier I said that I think the smallness of the usual 50mm equivalent lenses has something to do with 35mm film size, but why the 50mm equivalent GF63mm on Fuji's medium format is also the smallest of the bunch like this picture shows?

    GF-ROADMAP.jpg

    Does this mean that 35mm film lens designs carry over to all of its equivalents? Why would this be so? There could be something wrong with my train of thought.
     
  6. JonPB

    JonPB Member

    Joined:
    Jul 29, 2015
    Messages:
    37
    Likes Received:
    23
    Location:
    Portland, Oregon, USA

    -Return to Top-

    Longer lenses tend to be larger lenses. Focal length is, in the most abstracted terms, the distance of the lens from the sensor/film. Of course, practice is much different -- retrofocus lenses are longer than their focal length, telephoto lenses are shorter, and only pinhole lenses really come close to matching their theoretical "thin lens" approximation. The point, though, is that longer focal length lenses start life with a severe handicap when it comes to designing small lenses. All things equal, a 100mm lens system will be twice as long as a 50mm.

    So why isn't a 25mm half the size of a 50mm? That's more complicated.

    As Greg mentioned, SLR cameras have a limitation in that lenses must be mounted far enough away from the film to allow the mirror to swing up and down, so shorter lenses need to be retrofocus designs, making them larger than their slightly longer, non-retrofocus counterparts. Remove the SLR mirrorbox and you'd expect to see 35mm lenses that are smaller than their 50mm counterparts, which is what you see among rangefinder designs. So, if the flange distance represents the limiting factor, then you'd expect lenses near that length to be the smallest offerings. Given that the optical center of the lens system tends toward the center of the optics, which are mounted forward of the flange, you'd expect lenses slightly longer than the flange distance to be the smallest offerings in any system. Examples can be seen when considering the flange distances of 46.5mm for Nikon F, 45.46mm for Pentax K, 44mm for Canon EF.

    Then there are rangefinder lenses, where 35mm lenses are smaller than their 50mm counterparts, but the Leica M mount has a flange distance of 27.8mm, so that falls in line with the flange-distance-limitation argument even if it runs counter to the 50mm-is-smallest argument. But then, there are the flange distances of 26.7mm for Fuji G and 17.7mm for Fuji X. So, while focal length is important in determining how large a lens is, it isn't the decisive factor.

    Wider lenses are also more demanding. They have to bend light more strongly, which is easier to do through the many elements of a larger lens than through fewer elements of a smaller lens. It isn't just about bending light, which all lenses do, but bending light through the center of the lens (axis), bending light at the edge of the lens but on the same plane as the axis and subject (off-axis), and bending light that passes off to the side of that plane (skew) so that all three focus on the same image point. The stronger the bend is, the more likely those three types of rays are to diverge. And that's before even getting into the problem of working with digital sensor stacks. Getting all that light to behave is an enormously complicated problem not only in designing the theoretical optics but also in finding appropriate kinds of glass and machining and then assembling them to work together as expected -- which is why 35mm rangefinder lenses, while smaller than their 50mm counterparts, still cost more.

    I think it comes down to a balancing act involving physical limitations (e.g., flange distance), cost, and image quality that favors "normal" lenses. Specifically, I don't believe there's anything magical about 50mm, nor in approximating "human vision" (whatever that means; I could go on quite the digression about that). Rather, in my understanding, there's an inflection point in lens design around where the focal length equals the image circle diameter, and this inflection point defines the "normal" focal length. That's about 29mm for 16x24, 43mm for 24x36, and 55mm for 33x44. Lenses just slightly longer than that offer a convenient combination of offering very high image quality with a practical field of view and compact size without being too expensive.

    In summary, "normal" lenses aren't always the smallest, but they offer a nice balance of conveniences for both lens designers and photographers.
     
    HansA and GregWard like this.
  7. GregWard

    GregWard Premium Member

    Joined:
    Mar 21, 2014
    Messages:
    4,253
    Likes Received:
    1,589
    Location:
    London

    -Return to Top-

    Then there's also an issue around (as it were) "un-bending" when the flange distance gets too short and the sensor size is (relative to the flange distance) too big. Trying to get light rays out to the furthest corner of the sensor with a very short flange distance is, so I believe, yet another challenge to be faced. Sometimes this is done through precise positioning of the micro-lenses on the sensor I believe.
     
  8. HansA

    HansA Member

    Joined:
    Dec 1, 2017
    Messages:
    47
    Likes Received:
    6
    Location:
    Indonesia

    -Return to Top-

    It's interesting that you introduced the various flange distances into the equation, Jon, which was something that didn't cross my mind.
     
  9. GregWard

    GregWard Premium Member

    Joined:
    Mar 21, 2014
    Messages:
    4,253
    Likes Received:
    1,589
    Location:
    London

    -Return to Top-

    This...

    Please login or register to view links

    ... can be helpful.
     
    HansA likes this.
  10. Finder

    Finder Premium Member

    Joined:
    Aug 31, 2017
    Messages:
    164
    Likes Received:
    103
    Location:
    Florida, USA

    -Return to Top-

    An normal lens, one with a focal length equal to the diagonal of the format, means that the image image viewed from the standard viewing distance, a distance equal to the diagonal of the print, will produce an apparent perspective similar to a human viewing the original scene. It has nothing to do with the angular view of the human visual system, which is greater than a normal lens. In 35mm, technically, 43mm is normal.
     
    HansA likes this.
  11. HansA

    HansA Member

    Joined:
    Dec 1, 2017
    Messages:
    47
    Likes Received:
    6
    Location:
    Indonesia

    -Return to Top-

    Dang, the article says that the flange focal distance error has to be in the neighborhood of 0.01 mm. It's a wonder that this can be achieved considering that the lens isn't fixed to the body on an interchangeable lens camera system. It's for film cameras with their manual lenses though and I wonder if this is not critical for digital cameras due to their AF system.
     
  12. FMW

    FMW Premium Member

    Joined:
    May 11, 2017
    Messages:
    205
    Likes Received:
    87

    -Return to Top-

    Most wide angle lenses are retrofocus designs. Basically they are reversed telephotos. The reason is that they need to be far enough away from the film plane to make room for the swinging mirror. Normally designed wide angles would project too far into the camera body.

    The curious thing to me is that most of the Fuji X wide angles are also retrofocus designs. They would be better if they were not. I assume that there are enough retrofocus designs around the industry that it saves time and money to design wide angles that way.
     
  13. GregWard

    GregWard Premium Member

    Joined:
    Mar 21, 2014
    Messages:
    4,253
    Likes Received:
    1,589
    Location:
    London

    -Return to Top-

    Might be a detour to the main thread - but I think a fixed prime lens camera will often substantially out-perform an interchangeable (especially interchangeable zoom) lens. I guess that ("the article says that the flange focal distance error has to be in the neighborhood of 0.01 mm") could be part of the reason?
     
  14. Finder

    Finder Premium Member

    Joined:
    Aug 31, 2017
    Messages:
    164
    Likes Received:
    103
    Location:
    Florida, USA

    -Return to Top-

    Apparently it is. Otherwise your lenses would not work so well.
     
  15. GregWard

    GregWard Premium Member

    Joined:
    Mar 21, 2014
    Messages:
    4,253
    Likes Received:
    1,589
    Location:
    London

    -Return to Top-

    On a mirrorless camera the Autofocusing is all done on the sensor. This applies whether the AF is contrast, or phase, based. So the sensor would compensate for any small distance variations as long as the lens mount was still parallel to the sensor. Of course it could be different if the lens mount was twisted in some way.

    In a DSLR the phase detect AF takes place via a special module in the bottom of the mirror box. If this is not PRECISELY positioned vs the sensor (i.e. exactly the same distance from the lens mount) this can, and will, result in small errors in AF. If these were too high then the camera should fail quality control and be re-aligned.

    However, there's a similar effect in the exact positioning of the lens mount on the lens. Again this should be within an agreed tolerance level. On occasion you can be unlucky and the two tolerances can be individually acceptable but unacceptable as a compound. So this is why higher level DSLR's allow micro-tuning of the AF system. The sensor based AF on the mirrorless means this shouldn't be necessary on those cameras.
     
  16. JonPB

    JonPB Member

    Joined:
    Jul 29, 2015
    Messages:
    37
    Likes Received:
    23
    Location:
    Portland, Oregon, USA

    -Return to Top-

    Two other factors in fixed-lens precision is that the rear elements can be much closer to and much larger than those in an interchangeable lens system. An interchangeable lens version of (e.g.) the X100 or RX1 lens would be far too vulnerable to scratches, and adding the lens and camera housing between those optics and the camera body would make the camera itself somewhat larger. I think this mostly explains the difference between the X100 and XF23/2.

    Another element that I'm less sure about is that fixed-lens cameras can have sensor stacks that are much thinner -- as there's less dust, cleaning, and other opportunities for damage and thus customer complaints -- which allow steeper ray angles to achieve the same optical quality, which allows a smaller lens of the same specification and quality. I should look for tear-downs of those cameras to see if they have the same filter stack as their interchangeable lens sistren. [Edit: tried. Didn't find anything to convince me either way, but was quite amused by the photo in iFixIt's Step 36 at Please login or register to view links ]

    But, yes, I understand the Sony E mount has a rather generous tolerance for sensor-lens spacing because the autofocus mechanism can compensate for that. I have no idea how Zeiss accounted for that with its Loxia lenses, and I have no idea how tight the tolerances are for the X and G systems -- or, for that matter, in any fixed-lens camera.

    I agree completely that "normal" has nothing to do with the angular view of the human visual system, yet that myth persists, which bugs me, so I try to avoid that topic. :)
     

Share This Page