PDA

View Full Version : Canon EF-S lenses


Doug Kerr
June 8th, 2011, 11:21 AM
Most here are familiar with the Canon EF-S lens series. But sometimes the entire story may not be grasped clearly, or correctly. I thought I would review the topic here.

***********

The Canon EF-S lens series is a subgroup of the EF series. EF-S lenses are only intended for use with, will only work on, and will only physically mount on Canon EOS digital cameras having the smallest of the three current sensor sizes, the one often identified as "APS-C" (or "1.6x").

The prominent differences in their specifications compared to EF-series lenses proper are:

The minimum diameter of their image circle (the boundary of the image they form on the focal plane) is smaller.

Their minimum back focus distance (the distance from the rearmost point on the lens to the focal plane, for focus at infinity) is shorter. (In fact, the "S" in EF-S is said to be evocative of "short back focus".)

We can easily realize the significance of the former, but what about the minimum back focus distance?

The smaller the back focus distance, the more the lens protrudes into the camera. Lenses with a short focal length, if designed in a "classical" way, will have a small back focus distance, and thus a significant protuberance.

In an SLR camera, we cannot have the lens protrude an arbitrary distance into the camera, as beyond a certain point it will interfere with the reflex mirror as it swings up (or maybe even with in in its normal position).

Thus, in the design of short focal length lenses, special design techniques (such as the so-called "retrofocus" design) must be employed to increase the back focus distance (and thus reduce the protuberance of the lens into the camera to an acceptable amount). There are penalties in so doing - a necessary compromise.

For a smaller format camera, the reflex mirror is typically smaller, and thus a greater protuberance of the lens can be accommodated; that is, a shorter back focus distance is allowed.

Thus, in a lens intended only for such use (such as an EF-S lens), the degree to which special techniques of design to increase the back focus distance are applied can be reduced, and their bad side effects minimized.

We might hope that both of these unique, "relaxed" specifications lead to lenses that, for given optical parameters and a given performance bogey, will be less expensive, lighter, and/or smaller than if they were suited for general application to the range of EOS cameras.

Is the mount different

Basically, no.

However, EF-S lenses are designed with a rear protuberance (partly comprising a rubber collar) that will interfere with a feature on the mirror box of cameras with the two larger format sizes ("APS-H"/"1.3x" and "full frame") so the lens cannot be mounted there. (The use of the rubber part makes the attempted but failed penetration benign.)

Can we cheat?

Suppose that for certain EF-S lenses, we could excise the rubber collar (and maybe also machine off some other material) so that the lens would now physically mount on a larger format camera. What then?

Well, in general:

There would be the risk, if not the certainty, that the reflex mirror would strike the rear protuberance of the lens proper.

The reduced image circle would mean that there would be substantial vignetting of the images.

Equivalent focal length

Under the equivalent focal length concept, we express the field of view of a certain lens, with a certain focal length, when used on a camera with a certain format size, by stating the focal length of lens that, on a full-frame 35-mm camera, would give that same field of view. This value is called (to give its full name) the "full-frame 35-mm equivalent focal length" of this lens when used on a camera with "this" format size.

It is not a focal length of the lens, under any situation. The focal length of a lens is a basic optical parameter. It does not depend on what type of camera (if any) the lens is mounted to. The terms in which it is expressed are not in any way dependent on any presumption as to what kind of camera it might be used on.

Ef-S lenses, like all EF-family lenses, are marked with their (nominal) focal length (what some people would call the "real focal length", but of course this is the only focal length a lens has).

Some people believe that, since EF-S lenses are only intended for, and suitable for, operation on cameras with a very small range of sensor sizes, smaller than full-frame 35-mm size, they should be marked with (or designated by) their "full-frame 35-mm equivalent focal length". Thus, these people say, we could most directly recognize the field of view the lens would give under the familiar metric of full-frame 35-mm equivalent focal length.

But doing so would cause at least these two problems:

Many properties of a lens in use (such as depth of field) depend on the focal length, not any "equivalent focal length". If our lenses were marked with an equivalent focal length, predicated on a certain sensor size, we would have to convert that to the (real) focal length before we could use it in depth of field calculations and the like.

Suppose we had an EF-S lens with a focal length of 50 mm, and (under the doctrine suggested by these commenters) it was marked "80 mm". We would get familiar with its behavior with respect to field of view, and when it would be an appropriate choice for a certain photographic task. Now we buy an EF lens with a focal length of 80 mm (and marked "80 mm"). It would have a much different field of view that the first lens, even though they bore the same marking. It would be as if some of the wrench sockets in our toolbox were marked with the size across the flats, and others with the size across the points.

Best regards,

Doug

fahim mohammed
June 9th, 2011, 01:11 AM
Most here are familiar with the Canon EF-S lens series. But sometimes the entire story may not be grasped clearly, or correctly. I thought I would review the topic here.

***********
....snip....

But doing so would cause at least these two problems:

Many properties of a lens in use (such as depth of field) depend on the focal length, not any "equivalent focal length". If our lenses were marked with an equivalent focal length, predicated on a certain sensor size, we would have to convert that to the (real) focal length before we could use it in depth of field calculations and the like.

Suppose we had an EF-S lens with a focal length of 50 mm, and (under the doctrine suggested by these commenters) it was marked "80 mm". We would get familiar with its behavior with respect to field of view, and when it would be an appropriate choice for a certain photographic task. Now we buy an EF lens with a focal length of 80 mm (and marked "80 mm"). It would have a much different field of view that the first lens, even though they bore the same marking. It would be as if some of the wrench sockets in our toolbox were marked with the size across the flats, and others with the size across the points.

Best regards,

Doug

Hi Doug,

I thought depth of field also depended on the lens to subject distance ( besides the aperture and FL ).
( e.g to realize the same dof of a full-frame 35mm camera with a 50mm lens at f=1.4 and same distance on a smaller sensor camera would need something very expensive and big!).

Additionally, it is the perspective that would really matter if we were to use the 35mm equivalent fov values.

One can of course crop a ' full-frame ' photo to approximate the equivalent fl on a small sized sensor, but
it would ( and this to me is important ) the distance relationships between objects would change.

The EF and EF-S nomenclature used by Canon refers to attributes/physical properties/limitations/use
of various lenses on their different DSLRs and SLRs.

Similar differences are also applicable to other brands, but with different nomenclature/s.

Best.

Doug Kerr
June 9th, 2011, 09:08 AM
Hi, Mohammed,

I thought depth of field also depended on the lens to subject distance ( besides the aperture and FL ).
Oh, indeed, it does, and also on the criterion we adopt as to how much blurring from imperfect focus is to be considered "acceptable" (the circle of confusion diameter limit).

Additionally, it is the perspective that would really matter if we were to use the 35mm equivalent fov values.

One can of course crop a ' full-frame ' photo to approximate the equivalent fl on a small sized sensor, but it would ( and this to me is important ) the distance relationships between objects would change.

Not if the shot were taken from the same location (which I assume is why we might want to crop it).

Note that the the perspective effects of a shot - one way to describe those distance relationships - depend only on the location of the camera, not at all on the focal length.

The reason we tend to think that it is affected by focal length is that there may be a relationship between the focal length we use and where we take the shot from.

For example, let's suppose we initially consider a portrait setup where we shoot from a certain distance with a 35 mm lens. But we find that the perspective effects are not "handsome".

Instead, we decide to use an 85 mm lens. Now we must shoot from a greater distance to get about the same "framing".

The perspective effect is now quite different, but this is a result of the greater shooting distance rather than the change in focal length.

If we instead had shot from the new distance but with the 35 mm lens, and then cropped out the desired framing of the subject from that image, we would have that exact same new perspective effect.

Now, we can consider the situation in which we take two shots of a scene, from the same location (and of course with the camera pointing in the same direction) with these two rigs:

"Full-frame" format, 80 mm lens (80 mm ff35 equivalent).

"1.6x" format, 50 mm lens (80 mm ff35 equivalent).

Comparing the two images, we find that:

The "amount of scene material" included in the two is identical, owing to the equal field of view (or, if we prefer, owing to the equal ff35 equivalent focal length) along with the same camera location.

The relative sizes of objects in the image - a manifestation of the "perspective" - is identical, owing to the same camera location (not due to anything involving focal length or ff35 equivalent focal length).

I know you know all that. I just thought I would take the opening to "dissert"!

The EF and EF-S nomenclature used by Canon refers to attributes/physical properties/limitations/use of various lenses on their different DSLRs and SLRs.

Similar differences are also applicable to other brands, but with different nomenclature/s

Quite so.

Best regards,

Doug

fahim mohammed
June 9th, 2011, 09:19 AM
Hi, Mohammed,

.....

Note that the the perspective effects of a shot - one way to describe those distance relationships - depend only on the location of the camera, not at all on the focal length.
.....

Best regards,

Doug

Unfortunately not quite, Doug.

Place a camera at a fixed distance. Put a wa lens. take a shot.
Now put a tele, same cam position.

The perspective shall change. To give a sense of increased space and distance between subjects
a wa is usually utilised.

To compress the spatial relationships between subjects, I tend to use a tele.

One of the reasons a 50mm lens on a 35mm full frame camera is so called is that it tends to maintain the perspective relationships between subjects almost similar to the normal human vision.

Regards.

Doug Kerr
June 9th, 2011, 09:47 AM
Hi, Fahim,

Unfortunately not quite, Doug.

Place a camera at a fixed distance. Put a wa lens. take a shot.
Now put a tele, same cam position.

The perspective shall change.
Not so.

If we look at the two images, and find out that in one, a certain distant person has exactly half the height as a certain near person, then exactly the same will be true on the other image.

Said another way, if we take a crop from the WA shot that embraces the same material as all of the tele shot, and print them both to the same size, and lay one over the other, we will find that the positions of every object in one print are identical to their positions in the other print.

(I have a nice series of frames that illustrates this, but I can't put my hands on it at the moment.)

Now perhaps you have a different concept of what "perspective" means than I. I'd be glad to hear of that.

To compress the spatial relationships between subjects, I tend to use a tele.
Yes, but that is because (to give an example) to compress the apparent distance relationships between a series of five trees along the side of a road, we must shoot from a substantial distance from all five. Then, to make the framing effective, we will want a longer focal length lens.

If instead, we used a wide angle lens, we would probably shoot from closer to the trees. It is that, rather than the difference in focal length, that would result in this shot having the sense of "increased space and distance between subjects" of which you speak in connection with a WA shot.

One of the reasons a 50mm lens on a 35mm full frame camera is so called is that it tends to maintain the perspective relationships between subjects almost similar to the normal human vision.

We hear that a lot, but in fact when we look into what it really means, we come up short.

Does it mean that the field of view of the human eye is nearly that of a full-frame 35-mm camera with a 50 mm lens? Not even close.

Does it mean that the "perspective" we note in direct view of a scene from a certain place is the same as the perspective in a shot taken from that same place with a full-frame 35-mm camera with a 50 mm lens? Yes. And it will be the same for a shot taken from that same place with a 35 mm lens.

So, to those who hold that "the perspective afforded by the use of a 50-mm lens on a full-frame 35 mm camera is similar to that afforded by the human eye", I ask, "Tell me more. What do you mean by that?"

Best regards,

Doug

fahim mohammed
June 9th, 2011, 09:58 AM
Doug, if the size of a subject remains the same ( which can only be done by moving the camera placement) forward or backward..then what you say is correct. Else, no.

The dof which was referred to previously is dependent not only on the subject to camera distance, but the
FL and the aperture used. It also depends on the size of the sensor as I mentioned previously.

As to your last sentence.." So, to those who hold that "the perspective afforded by the use of a 50-mm lens on a full-frame 35 mm camera is similar to that afforded by the human eye", I ask, "Tell me more. What do you mean by that?";

The so-called ' standard lens ' is the focal length that offers the closest perspective to ' normal' vision when viewed on a 10 x 8 inch print held at arm's length. 50mm is the normal ' standard ' lens for 35mm full frame cameras. That is how they are named since before my time. Whether that nomenclature is correct or not scientifically is for the most part irrelevant; the test:

Ask anyone what a 35mm full standard lens is..industry, shop or users. If they say anything other than
a 50mm; I shall buy you a coffee.

Doug Kerr
June 9th, 2011, 11:25 AM
Hi, Fahim,

The dof which was referred to previously is dependent not only on the subject to camera distance, but the
FL and the aperture used. It also depends on the size of the sensor as I mentioned previously.
It depends on the size of the sensor only if we choose a COCDL that is a fixed fraction of the sensor dimensions. But we often do, and so our normal conclusion is as you say.

The so-called ' standard lens ' is the focal length that offers the closest perspective to ' normal' vision when viewed on a 10 x 8 inch print held at arm's length.

Well, if we take an image shot with a 50 mm lens, and make a print on 10" x 8" paper (with the image occupying 10" x 6.67" - the image is 3:2 in aspect ratio), and view that print from a distance of 13.9", then the size of the objects seen on the print will be the same as the size of the objects seen directly by the eye from the place where the camera was.

If we consider viewing "at arms length" to be at 20" from the eye (perhaps more reasonable), then this rationale would suggest that the "standard lens" should be considered to have a focal length of 72 mm.

There is no real issue of perspective here. The perspective seen by the user (manifested as the ratio of the visual sizes of two objects, one more distant than the other, or by the alignment in the image of features on two such objects) would be unchanged as we changed the viewing distance, the print size, or the focal length of the lens that was used, so long as the comparison was to direct human viewing of the scene from the same location as the camera.

At one time is was very common on full-frame 35-mm cameras to use a 50 mm lens on non-interchangeable lens cameras (and as the "kit" lens for interchangeable lens cameras). It is likely that this was based on the fact that this seemed to be the most useful, overall, to the user for a wide range of photographic "tasks". This most likely led to the description of a 50-mm lens as "normal". (It certainly was "normal" - you could go into any camera store at the time and see that in the showcase!)

(Later, 45 mm became common on serious non-interchangeable lens cameras.)

The business of the supposed relationship to the "perspective" of the human eye was probably contrived as a way to rationalize this practical situation. It actually has no technical meaning - it can't be either true or not.

The relationship you describe based on the apparent size of objects in a print of a certain size viewed at a certain distance is meaningful, and gives the well-known answer ("50 mm") if we assume a 10" x 6.67" image (perhaps on 10" x 8" paper) and a viewing distance of about 13.9".

Best regards,

Doug

fahim mohammed
June 9th, 2011, 12:02 PM
Hi Doug,

In that case you could form a committee to have the name ' standard ' lens not to mean what is currently meant. I am sure there is a misrepresentation law in trade and commerce. For a 35mm full frame one
could petition for a 72mm FL. Or whatever the scientific calculations deem it to be.

I shall stand behind you in this ( a little far! ) to lend an international dimension to the petition.

I am sure the ' standard ' would then be changed. When everyone realizes their gross mistake.

Take care and best regards.

Doug Kerr
June 9th, 2011, 12:09 PM
Hi, Fahim,

To quote, by literal translation, from La Marseillaise:

"The bloody standard we raise!"

Thanks for your inputs.

Best regards,

Doug

Doug Kerr
June 9th, 2011, 12:32 PM
Here are some images that demonstrate that the "perspective" of a shot is influenced not by the focal length of the lens used but only by the point from which the shot was taken (which is why that is called the point of perspective, actually).

First let me give the definition I use for perspective:

Perspective - the phenomenon in a drawing or photo in which objects having the same actual size will have different sizes on the image depending on their distance.

A corollary is that points in the three-dimensional scene that lie on a line through a particular point in three-dimensional space (the point of perspective) will have identical positions in the image.

From that I will give the definition I use of "two shots having the 'same perspective' ". It is when:

the relative sizes of all the objects (that appear in both images) on the image is the same for the two images.

and

the relative alignment of all the objects on the image is the same for the two images.

Those who have some other concept of what "same perspective" means will find a different result that I. I'll be glad to hear of alternate concepts of "same perspective".

In photography we often speak of differences in perspective in the sense of the appearance in the image of strings of objects at different distances in a scene. My demonstration will use just such a scene = a string of mailbox columns on my street.

All shots were taken with same camera (an EOS 40D), from the same point and with the same pointing direction (a tripod was used), and with the same lens (an EF 24-105 mm f/4L USM), at three different focal lengths.

The shots with the two smaller focal lengths were then cropped to present essentially the same scene material for ease of visual comparison.

Here we go.

http://dougkerr.net/images/technical/Perspective_105-00R.jpg
Douglas A. Kerr: Mailboxes on Charlie Way 105-00
Focal length 105 mm; full frame presented.

http://dougkerr.net/images/technical/Perspective_047-01R.jpg
Douglas A. Kerr: Mailboxes on Charlie Way 047-01
Focal length 47 mm; cropped to match scope of 105 mm image.

http://dougkerr.net/images/technical/Perspective_024-01R.jpg
Douglas A. Kerr: Mailboxes on Charlie Way 024-01
Focal length 24 mm; cropped to match scope of 105 mm image.

You will note that the relative heights* of the columns at different distancesare the same for all three images, as is their relative positions on the images.

All three images exhibit "the same perspective", as we should expect. They were all taken from the same point; that is, they have the same point of perspective.

*Their absolute heights on the image are the same, too, but only since I chose to present the images at different magnifications by way of cropping. Had I not done that, what I said just above would have been equally true, and the three images would have still exhibited identical perspective.


Quod erat demonstrandum.

Best regards,

Doug

fahim mohammed
June 9th, 2011, 12:54 PM
To talk about three dimensional space on a printed photograph has no meaning. The printed photo is
only two dimensions. Period.

My definition of perspective is how objects appear to me through the viewfinder. To me they take on different shapes and sizes when I stand in the same place and put on lenses of different focal lengths.

For me that is sufficient to enable me to take a photograph with the focal length that gives me the
' perspective ' I want from a given place. For me that is indeed Q.E.D.

Regards.

Doug Kerr
June 9th, 2011, 01:10 PM
Hi, Fahim,

My definition of perspective is how objects appear to me through the viewfinder. To me they take on different shapes and sizes when I stand in the same place and put on lenses of different focal lengths.
For me they don't have different shapes. They certainly get bigger and smaller. And if that to you is a manifestation of "different perspective" . . .

Moving from this view in the finder:

http://dougkerr.net/images/technical/Perspective_105-00R600.jpg

to this one:

http://dougkerr.net/images/technical/Perspective_047-00R600.jpg

I certainly see things "smaller", but hardly of different shapes, either individually or as groups.

Best regards,

Doug

Doug Kerr
June 9th, 2011, 01:22 PM
Hi, Fahim,]

For me that is sufficient to enable me to take a photograph with the focal length that gives me the ' perspective ' I want from a given place.
Nothing wrong with that.

For me it is sufficient to be able to use a lens with the focal length that gives me the framing I want once I have decided to shoot from a particular place for possibly many reasons, one being to get the perspective I want.

Now, here's a challenge.

Suppose I show you some images with no details of how the shots were taken.

Do these two have the same perspective?

http://dougkerr.net/images/technical/Perspective_105-00R600.jpghttp://dougkerr.net/images/technical/Perspective_047-00R600.jpg

I suspect you will say no. OK.

Do these two?

http://dougkerr.net/images/technical/Perspective_105-00R600.jpghttp://dougkerr.net/images/technical/Perspective_047-01R600.jpg

If not, how are they different?

If so, isn't it amazing that we can change the perspective of an image by just cropping it and presenting it at a greater magnification.

Maybe we should be able to do that with our politicians.

Best regards,

Doug

fahim mohammed
June 9th, 2011, 02:08 PM
"A corollary is that points in the three-dimensional scene that lie on a line through a particular point in three-dimensional space (the point of perspective) will have identical positions in the image."

My point still holds. To talk about theoretical three dimensional concepts has no relevance to a
printed photograph on a sheet of paper that only has two dimensions.

The ' perception ' of depth or otherwise is just that; a perception. There is no ' real depth ' either on a printed photograph or in the photographs that you have put up as illustration.

Coverging parallels and verticals are a perception. That is the camera/lens combination perspective from a given point taken from a given angle with a certain focal length and tilted at a certain angle ( or not ).

Take a wide angle lens. Say 15mm on 35mm full frame. take a photo very near a person's face. There is distortion. I cannot for all practical purposes create a similar rendering with a telephoto lens whatever I do. It is the nature of the beast. The focal length. Now One can reduce or crop till ones heart's content; but no way can that perspective be reproduced in a form that is presentable
or printed say 8 x 10 in for a 35mm full frame with sufficient resolution.
An exercise in numbers maybe, but really of no practical utility.

I have already agreed that changing a position would alter perspective, the way you define it.

That a telephoto compresses and a wide angle gives me the ' perception ' of increased distances between objects is what my ' perspectives ' are. Not bigger or smaller. 1:1 or 1:2, for that I might go into macro/micro + close-up lens territory. I rarely venture into those.

Similarly I still maintain that the dof is dependent on the subject to camera/lens distance, The focal length, and the aperture being used. There is no way ( economically and practically ), except as an exercise in numbers, that a small sensor ( non-full frame 35mm ) camera can have a lens attached that would give me the same depth of field for a given camera position as say a 75mm f/1.4 attached to a full frame 35mm camera. That is the perspective that I talk about. Because I do use dof and the perspective I get with full frame ( 35mm ) on film or on digital.
Of course, bar the Leica no one has a 75mm f/1.4 that is common.

Try to get a lens that would give the same dof as a Canon 85/1.2 on a full frame using a cropped sensor camera.

Regards.

Here is an illustration of how the perspective changes with the focal length, while keeping an object the same size..
http://www.photozone.de/focal-length-and-perspective#

This is what I mean. Maybe I should have put this up before.
Best.

Doug Kerr
June 9th, 2011, 03:14 PM
Hi, Fahim,



Here is an illustration of how the perspective changes with the focal length, while keeping an object the same size.
If we change the focal length, but keep the object the same size in the image, we must have changed our distance from the object.

It is that change in distance which changes the perspective. It of course flows, in this scenario, from our change in focal length (having first decided that we wish to maintain the object size on the image).

If we want to say, "If we keep the object size in the image constant, then a change in focal length will lead to a change in perspective", that is certainly true.

Best regards,

Doug

fahim mohammed
June 9th, 2011, 03:18 PM
I said that here..

" Doug, if the size of a subject remains the same ( which can only be done by moving the camera placement) forward or backward..then what you say is correct. Else, no. "

Regards.

Doug Kerr
June 9th, 2011, 04:49 PM
Hi, Fahim,

I said that here..

" Doug, if the size of a subject remains the same ( which can only be done by moving the camera placement) forward or backward..then what you say is correct. Else, no. "
Indeed, but I wasn't able to be certain of the context, or be sure what I said that you would agree with under that proviso!

Best regards,

Doug

fahim mohammed
June 9th, 2011, 04:58 PM
Hello Doug,

In that case sir, I shall attend to our chief bartender!!

Regards.