Impact of the image circle size on lens design

[Rory]

I've noticed that the longer telephoto Olympus lenses are not any smaller than their full frame equivalents even though the projected image circle for MFT is half the diameter of full frame.  For example:

Lens                                                                    Max Aperture   Weight (lbs)   Length (in)       Front element (in)   Format Width (mm)

Olympus 300mm f/2.8 ED                                              2.8             7.24               11.2                    5                        18
Nikon AF-S NIKKOR 300mm f/2.8G ED VR II                    2.8             6.39               10.5                    5                        36
Canon EF 300mm f/2.8L IS II USM                                  2.8             5.17                 9.8                    5                        36

Olympus ED 300mm f/4 IS PRO                                       4               3.25                 8.9                 3.64                      18
Canon EF 300mm f/4L IS USM                                         4               2.62                 8.7                 3.5                        36
Nikon AF-S NIKKOR 300mm f/4D IF-ED                            4               3.17                 8.8                 3.5                        36

What I would like to know is how the required image circle (sensor size) impacts the overall lens size.  For example, if Olympus built a 500mm f/4 lens, without using diffractive optics, how small could they theoretically make it.  The Canon 500/4 has a  5.75" front element, is 15" long and weighs 7 lbs.

I fully understand that the apparent focal length will be different on the MFT cameras.  I'm just interested in the impact of the image circle on lens design.

[TonyVentourisPhotography]

I'de love to hear some insight into this as well. 

[pegelli]

(Front) lens diameter is set by the aperture required, not by the size of the image circle.
Length is set by the optical design but mostly not a big variable either (with a telephoto) based on the need to be well corrected.

[Rory]

The part that confuses me is if the front element is the same size and one image circle is much smaller why isn't it brighter?

[pegelli]

Are you sure the image circle is much smaller? The camera they are used on has a different size sensor but are the image circles significantly different?
And even if the image circle is smaller it's caused by some light being absorbed by the black internals in case there is a tighter restriction in there.
Since the aperture and focal length are the same the center of the image cannot become brighter when you compare the lenses.

[Jack Hogan]

The part that confuses me is if the front element is the same size and one image circle is much smaller why isn't it brighter?

Hi Rory,

It depends on what you mean by 'brighter' and the assumptions underlying the setup.  For instance in an 'equivalent' situation with the same sized front element the number of photons per capture is intuitively about the same for all formats:



However as you can see effective f-number (f/D) is half for the micro FT camera versus Full Frame with a constant D - therefore mFT gets a higher Exposure aotbe (that is it gets more photons per unit area).  More here.

Jack

[Jack Hogan]

For example, if Olympus built a 500mm f/4 lens, without using diffractive optics, how small could they theoretically make it.  The Canon 500/4 has a  5.75" front element, is 15" long and weighs 7 lbs.

To provide equivalent FF performance to a 500mm f/4 mFT lens all Canon would need is a f/8.  So interestingly telephoto lenses performing equivalently are not necessarily smaller/lighter in smaller formats.

Jack

[pegelli]

Hi Rory,

It depends on what you mean by 'brighter' and the assumptions underlying the setup.  For instance in an 'equivalent' situation with the same sized front element the number of photons per capture is intuitively about the same for all formats:



However as you can see effective f-number (f/D) is half for the micro FT camera versus Full Frame with a constant D - therefore mFT gets a higher Exposure aotbe (that is it gets more photons per unit area).  More here.

Jack

Interesting picture and concept, but not fully applicable to the question of the OP.

In his question he is keeping front lens size (aperture number) and focal length constant.
In your picture the focal length gets reduced as the sensor format shrinks, which indeed makes the image on a smaller sensor brighter (more photons per unit area).

But if you keep the focal length and front lens size constant the photons per unit area remains constant irrespective of sensor size
 

[Jack Hogan]

Interesting picture and concept, but not fully applicable to the question of the OP.

In his question he is keeping front lens size (aperture number) and focal length constant.
In your picture the focal length gets reduced as the sensor format shrinks, which indeed makes the image on a smaller sensor brighter (more photons per unit area).

But if you keep the focal length and front lens size constant the photons per unit area remains constant irrespective of sensor size

Sure, then exposure is the same in both formats but the number of photons captured vary (FF capturing more).  On the other hand imho one is then figuratively comparing apples to oranges, which instead the setup in the picture above avoids.

Jack

[pegelli]

Sure, then exposure is the same in both formats but the number of photons captured vary (FF capturing more).  On the other hand imho one is then figuratively comparing apples to oranges, which instead the setup in the picture above avoids.

Jack

I agree, your picture explains very well why at "equivalent" focal length a MFT lens can be so much smaller at the same aperture vs. the same "equivalent" focal length of a FF lens.
And I also agree that's a much more apples-to-apples comparison because they result in the same picture (except for sensor differences and difference in DOF).

[TonyVentourisPhotography]

Rory, if I understand correctly...after seeing this conversation...you are really asking about image circle due to how we often depict it.

For example A picture of an image circle from a full frame lens that covers a 36mm sensor completely.  A m43 sensor would fit nicely inside of that and use only the center portion and relieve itself from the edge quality per say.  Therefore could the manufacture build a lens better or smaller, or differently due to that factor? 

That is interesting to think about it that way because that seems quite applicable.  I am an architectural photographer by trade and "image circle size" is something that actually determines what lenses I buy or use.  For example on a tech cam I prefer using lenses that allow me to shift 15 - 20mm in any direction.  That is opposed to say another lens that can only shift 10mm.  The lens I prefer might cast a 90mm image circle as opposed to one that casts a 60mm image circle.  Same reason APS-C lenses that cast a 24mm image circle can't be used on a 36mm sensor. (or they can but there has to be cropping)

This is interesting because normally you would think the lens just casts a bigger circle the further away it gets.  However, lenses do a have a hard stop in their exit where you will see the lens body instead of light.  If you look at the rear of a Hasselblad lens and the rear of an Olympus prime... there is a serious size difference. 

So is this what you are asking, how this plays into lens design?

[Rory]

Thanks for explaining what I was thinking Tony. That is precisely what I want to know.

[Jack Hogan]

For example A picture of an image circle from a full frame lens that covers a 36mm sensor completely.  A m43 sensor would fit nicely inside of that and use only the center portion and relieve itself from the edge quality per say.  Therefore could the manufacture build a lens better or smaller, or differently due to that factor? 

Well, they are indeed designed specifically for m43, so for a telephoto lens the short answer is 'yes' for differently and 'not meaningfully' for lighter and smaller. 

Of course in a non equivalent situation everything is possible - but then the images produced by the two formats would not be directly comparable.  For instance in the quoted example the field of view would be much smaller in the smaller format so one image could literally show apples and the other apples & oranges.

Buyers and camera manufacturers alike make some assumptions based on intended usage when choosing/designing a lens.  So it behooves the photographer comparing images produced by different formats to see how they do when capturing the same image: the two systems need to be set up in the same spot, looking at the same scene, with the same field of view, the same shutter speed, DOF, etc, apples to apples.  That's what's called an equivalent setup.

It turns out that, if we stay away from the extreme limits of the working envelope of either format, a telephoto lens designed specifically for a smaller format is often not (much) smaller and lighter in that case.  The explanation starts with the drawing above and gets quite involved.

Jack

PS A bit of a simplification, but an easy way to see why is to compare two of the lenses at the top of the OP:
Lens                                                                    Max Aperture   Weight (lbs)   Length (in)       Front element (in)
Olympus 300mm f/2.8 ED                                             2.8             7.24               11.2                    5   
Canon EF 300mm f/2.8L IS II USM                                  2.8             5.17                 9.8                    5 

If you add a 2X III Extender to the Canon it turns it into a 600mm f/5.6 which is equivalently comparable to the Oly 300mm f/2.8.  The combined lens-extender specs for the Canon would then be

Canon 300mm f/2.8L + 2X III Extender                           5.6             6.0                 11.9                    5

That's with an extender.  If the lens were built with those specs to start with it would be smaller and lighter than that.

[Mousecop]

A better option here is to compare 3rd party lenses to native ones, as most of the 3rd party are actually APS lenses adapted for the M43 mount.

Rokinon 12mm f/2.0, manual focus, 255g, 2.85 x 2.28", 67mm filter
Olympus 12mm f/2.0, autofocus, 130g, 2.20 x 1.69", 46mm filter

Sigma 19mm f/2.8 DN, autofocus, 160g, 2.39 x 1.80", 46mm filter
Olympus 17mm f/2.8, autofocus, 71g, 2.24 x 0.87", 46mm filter

Sometimes it makes a big difference, sometimes it doesn't. It depends a great deal on how the lens is designed, how much software correction is used, maximum aperture, and so forth.

[Jack Hogan]

Sometimes it makes a big difference, sometimes it doesn't. It depends a great deal on how the lens is designed, how much software correction is used, maximum aperture, and so forth.

Yes, comparing equivalent apples to apples in this case means using lenses designed each for their own format and to similar IQ standards.  One such lens designed for a larger format and mounted on a smaller one will typically perform worse than a lens of similar cost designed specifically for the smaller format.
Jack

[Rory]

(Front) lens diameter is set by the aperture required, not by the size of the image circle.
Length is set by the optical design but mostly not a big variable either (with a telephoto) based on the need to be well corrected.

So, what is the variable then, given the requirement for different image circle sizes - the distance from the rear lens element to the sensor plane?

[pegelli]

So, what is the variable then, given the requirement for different image circle sizes - the distance from the rear lens element to the sensor plane?

Diameter of the lens between the aperture and the sensor or exit pupil size. The size of the rear element can be (but does not necessarily has to be) an indication

[Rory]

Diameter of the lens between the aperture and the sensor or exit pupil size. The size of the rear element can be (but does not necessarily has to be) an indication

Thanks.

[qwz]

Please compare a 300mm F4 4/3 lens with 600mm F4 135 format lens 'cause they has the same Field of View and Aperture.
Also, closing focus/maximum magnification matters - and for particular Olympus Zuiko 300mm F4 Pro it almost macro - so lens will be big, but not even close to 600mm full-frame.

[Rory]

Yes I understand the whole equivalence thing thanks.