Explain EF-S to me, Also explain crop factors

[timmyquest]

Ok, sorry about the poor drawing (or at this skill levle mabye it's called a drawling). I'm almost certain this isnt the case but i'm basing this on the fact that an EF-S lens will not project on an entire 35mm frame. So if this indeed isnt the case, explain to me what exactly it does.


http://www.questphotos.com/efs.jpg


Now to my crop question. Does the 1.6/1.5/1.3 crop factors of todays DSLR's provide the same effect that cropping a photo does on image quality? That is, a loss in image quality?

[cmM]

Now to my crop question. Does the 1.6/1.5/1.3 crop factors of todays DSLR's provide the same effect that cropping a photo does on image quality? That is, a loss in image quality?
A full frame photo is bigger to begin with. If you crop that to the image size produced by a 1.6x sensor, you'll get the same image... I think (not sure). So I don't think it affects quality.

[Jon]

Lenses are designed to cover a certain field of view both in front and in back. The EF-S lenses are designed so they can only project to cover a 22x16 sensor. This is a combination of vignetting and not bothering to correct for distortion beyond those limits. So, the EF-S will vignette if you use it beyond the DR/10D sensor, and even if you were to remove the lens elements from the lens mount, while maintaining their correct relationship, you'd see horrendous aberrations outside the image circle.

Is the effect the same as cropping a photo? Not really. There are a lot of variables that may come into play, but in simplest terms, if you crop a film image, you're blowing up the grain and the (limits of) resolution beyond what you'd get in a full-frame print of the same dimensions. With a 1.3 or 1.6 x crop factor, you're not blowing up the "grain" (sensor cell size) or resolution limits; you just haven't captured as much of the image as would have been possible with a 1:1 crop. You're using all the image quality that was captured, at the best level you can for that image. If you had film and digital sensors with the exact same resolution, and you used the exact same lens to take the pictures, and you printed the exact same part of each picture to the exact same size, the quality would be comparable (allowing for differences inherent in digital and film). But if you took (with a zoom, say) the exact same picture (field of view) and printed only a portion from each corresponding to the 1.6x sensor's area, the digital would show better quality (Dreb full frame, 35 mm. applying a 1.6x crop and using central 22x16 mm of the neg.).

[timmyquest]

Lenses are designed to cover a certain field of view both in front and in back. The EF-S lenses are designed so they can only project to cover a 22x16 sensor. This is a combination of vignetting and not bothering to correct for distortion beyond those limits. So, the EF-S will vignette if you use it beyond the DR/10D sensor, and even if you were to remove the lens elements from the lens mount, while maintaining their correct relationship, you'd see horrendous aberrations outside the image circle.


Right, so "normal" lenses, that were originally designed to cover a 35mm frame are now being used on "film" that is smaller then 35mm, so part of the image is not seen, thus giving the crop. So if the EF-S is designed for the smaller CMOS sensors that Canon is currently using...is the 1.6 crop factor still applied? If so, why? How?

[Jon]

The "crop factor" is really shorthand to let us compare the coverage of the, currently widely-known, 35 mm. format to that of the digital cameras. We know what a 28 mm. lens "looks like" in 35 mm; so to get that effect, we need a lens of FL 28/1.6, or 17+ mm. Conversely, our 28 mm. lens, on our 1.6x crop body will "look" like a 28 * 1.6 = 45 mm. lens on a 35 mm. camera.

[Tom W]

Right, so "normal" lenses, that were originally designed to cover a 35mm frame are now being used on "film" that is smaller then 35mm, so part of the image is not seen, thus giving the crop. So if the EF-S is designed for the smaller CMOS sensors that Canon is currently using...is the 1.6 crop factor still applied? If so, why? How?

The 1.6X "crop factor" still applies. What the factor says is that the lens you are using offers the same field of view that a certain lens would give on a 35-mm film-sized sensor. You're not changing the actual focal length of the lens.

Lets take the 17-85 for example. It is a 17-85 lens - the focal length as measured from the "null" point (near the aperture) of the lens to the focus point (the sensor). The lens, when used on a 1.6X-sized sensor will provide the same field of view that a lens 1.6X as long would provide on 35 mm film. In this case, the 17-85 lens on the 1.6 sensor gives the same field of view (and thus image coverage) that a 27-136 mm lens would give you.

This is independent of the image circle diameter.

[timmyquest]

I know this is the case as 50mm on my 18-55 is the same as my 50 f/1.8 but what i dont understand exactly then is what the advantage to EF-S is other then a cheaper way to build lenses. I dont know what it physically does, nor do i understand how it would not project fully on 35mm film (i know this is the case by putting the lens "on" my rebel 2000) but it still gets cropped on the drebel/"d20".

Why cant they make it NOT get cropped, wouldnt that make more sense?

[Tom W]

I know this is the case as 50mm on my 18-55 is the same as my 50 f/1.8 but what i dont understand exactly then is what the advantage to EF-S is other then a cheaper way to build lenses. I dont know what it physically does, nor do i understand how it would not project fully on 35mm film (i know this is the case by putting the lens "on" my rebel 2000) but it still gets cropped on the drebel/"d20".

Why cant they make it NOT get cropped, wouldnt that make more sense?

The primary advantage of the EF-S are cost and size. Canon can make a much wider lens for a lot less money and in a smaller package with the EF-S mount, because they don't have to make it as big. It is cheaper partly because of its diameter - it doesn't have to be as big around because it doesn't have to cover as large an area.

I think what you're doing is confusing an image cropped by too-small of an image circle (which an EF-S lens would do on larger sensors) with the "crop-factor" which is really just a multiplier used to show approximate image coverage relative to a 35-mm film base.

[timmyquest]

So there is no relation to the focal length to the filed of view until you consider the size of the frame it's being projected on?

the only reason 25mm is 25mm on 35mm film is because of the 35mm film?

A 50mm Med. Format image is differnt then a 35mm(film) 50mm image?

[Jon]

It's easier to design a lens to "cover" a smaller area than a larger one. For instance, you can get a 75-300 mm. f/4-f/.6 IS zoom from Canon to cover 35 mm for about half of what a Rodenstock f/5.8 lens (without shutter) for a 5x7 view camera costs. Or a T-Mount 300 f/5.6 for under $100. (chosen for max. aperture - I know what the 300 f/2.8 costs!) Or a 100 mm macro for a Pentax 67 is $1200, while the Canon 100 mm. macro is about $470. The larger the area you're trying to keep good image quality over, the more difficult the lens design becomes. You see this within 35 mm. lenses if you look at the tests - edge results are invariably worse than the center results in terms of both aberrations and distortion. So, by designing for the sensor area of a DR, Canon's optical engineers were able to reduce the area over which they needed to worry about image quality. It's the same kind of trade-off they have to make in designing lenses for speed vs. economy. Compare the costs of the 50 f/1.8 and 1.4. Not all that's due to cheaper construction. Some is due to design, and to more complicated glass.

[Jon]

So there is no relation to the focal length to the filed of view until you consider the size of the frame it's being projected on?

the only reason 25mm is 25mm on 35mm film is because of the 35mm film?

A 50mm Med. Format image is differnt then a 35mm(film) 50mm image?

Bingo. In medium format (6x6), 80 mm is "normal"; compared to them, 35 mm. has a 1.6x crop factor. If we ever all start using the same sensor size in digital, we won't need to think about "crop factors" either. It'll just be "normal", "wide", or "tele".

[Tom W]

So there is no relation to the focal length to the filed of view until you consider the size of the frame it's being projected on?

Yes! Focal length is an actual measurement. Image sensor size is a measurement. Field of view is determined by the ratio of the two.

the only reason 25mm is 25mm on 35mm film is because of the 35mm film?

A 25 mm lens is 25 mm regardless of where you put it (even there!!) :). But the field of view that it presents is related to the other half of the equation, the sensor/film size.

A 50mm Med. Format image is differnt then a 35mm(film) 50mm image?

Yes! 50 mm on medium format is a little bit of a wide angle, while it is considered "normal" on 35 mm film. It is a little telephoto on a 1.6X sensor, and would be a supertelephoto on my S-400.

[timmyquest]

Ok. I always knew where the focal length came from i guess i was just too set on what it actually meant...er something.

[Tom W]

Ok. I always knew where the focal length came from i guess i was just too set on what it actually meant...er something.

That's because 99% of the world shot 35 mm film for a long time. So, in an effort to not confuse everybody, the camera industry (or someone) decided that it would be a good idea to relate everything digital to the 35 mm format. Which, of course, tends to confuse everything instead.

They could have just explained the relationship in the first place, and people would understand.

[timmyquest]

exactly, i've used 35mm cameras my entire life.

What i really find ammusing is P&S'ers who ask me "How many X is that lens?"

Talk about a stigma of poorly educated consumers.

I was at the bears training camp a few weeks ago with one such person who happened to be a friend of mine, i pointed to a big 300mm f/2.8 and said "hey, that lens is $4,000 and it's only a 1x lens!"

[Tom W]

exactly, i've used 35mm cameras my entire life.

What i really find ammusing is P&S'ers who ask me "How many X is that lens?"

Talk about a stigma of poorly educated consumers.

I was at the bears training camp a few weeks ago with one such person who happened to be a friend of mine, i pointed to a big 300mm f/2.8 and said "hey, that lens is $4,000 and it's only a 1x lens!"

Just tell him that in "binocular" talk, its a 6X lens. Or tell him it has "digital zoom". :)

[hmhm]

A 50mm lens doesn't know what type of camera it is connected to. It projects an image onto a plane behind it, that image will be a recreation of the scene in front of the lens. It will roughly look like a circle, as the image fades to darkness as you get further away from the middle of the image.

Now, photographers have a tendency to mount lenses onto cameras, and cameras are dark boxes with either film or digital sensors inside. The lens doesn't know which, or how big it is, it can't look behind it to check, it always look ahead, all it knows is how to act like a 50mm lens, and that doesn't depend on what's behind it.

If the photographer happens to put a really tiny digital sensor behind the lens, then it will only capture a small rectangle out of the middle of that projected image made by the 50mm lens. The net effect is a very small angle of view. We call a lens that produces a very small angle of view a "telephoto" lens.

If the photographer puts a large piece of film behind that 50mm lens, then it will capture a very large rectangle out of that projected image, with the net effect being a very wide angle of view. We call a lens that produces a very wide angle of view a "wide-angle" lens. Note that our same 50mm lens might be called a "telephoto" when mounted on one camera, and a "wide angle" when mounted on another. It doesn't mind this at all, it understands, though sometimes the photographers get confused.

The photographer has to be careful, though, because different designs for 50mm lens will result in projected images that start to "vignette" (fade to darkness at the edges) at different distances from the middle. Some lenses will make "big circles" and some "small". The images they project are always the same, they just differ in how much of that image is there before it fades to darkness. If the photographer tries to use a piece of film or sensor that's bigger than the circle the lens makes, then his capture will have dark corners, or even show only a small image in the middle.

So a 50mm lens is a 50mm lens, and the angle of view that is captured depends on the size of the film or sensor used to do the capturing, and you need to ensure that the 50mm lens you use produces an image circle big enough to fully cover the film/sensor you plan to use to make that capture.

A 50mm EF-S lens would produce an identical image as a 50mm EF lens, except that it _might_ produce an image circle that's smaller than the EF lens. Or it might not, but let's just say that the EF-S lens is "allowed" to make a smaller circle, since we're assured that it will only be used with smaller sensor cameras. And if this assumption allows the designer of the EF-S lens to produce it more cheaply, well, then that's the whole idea.
-harry

[timmyquest]

You'd of saved yourself a lot of time if ya read the thread.

But thanks for the post anyways ;-)

[robertwgross]

Another way of looking at it is this:

Suppose you had two equivalent 50mm lenses, but one had a hunk of glass in it that was one inch in diameter, and another one had a hunk of glass in it that was two inches in diameter. If the first one projected its usable image onto a small sensor correctly, it would work fine on just those small sensors. If you tried to make the first one project onto a full size sensor (or film), then you would get "crap" around the edges, or worse.

If the second lens projected its usable image onto a full frame sensor or film, then that works fine. If you let it project onto a small sensor, then it still works fine, but there is some image that drops off outside the small sensor, and that is effectively cropped off. Think of it as light that is wasted. No harm done.

Both of these lenses work fine in their intended application. The one difference is that the first one with the small glass is going to be cheaper to build and lighter to carry. The second one is more expensive to build and heavier to carry, but it is more universal in that it can be used on full frame or small sensor.

I carry a film camera and a digital camera around and use my lenses interchangeably between the two. I don't think I would like a "digital lens" that I could not use effectively on my film camera.

---Bob Gross---

[kre84u]

Great thread. An additional question. I understand I can't use an EF-S lens with a full-frame Canon body. Is this absolutely true, due to depth of lens incursion into body or lens mount differences, or is it just a consideration that the different configuration would cause vignetting?

[tonylong]

EF-S lenses and "crop" cameras are manufactured with an EF-S mount and "full frame" and 1.3 crop cameras are not, to the EF-S lens will only mount on a "crop" camera without being modified. "Crop" cameras do, though, have the hardware capability to mount an EF lens.

There are a couple considerations with EF-S lenses -- the inner element is often made to sit closer to the opening of the camera and can be hit by the larger mirror of the "full frame" body and also the smaller optical construction of the EF-S lens can cause vignetting. There are some modifications that can be made, but only if you know what you are doing. And, there's no getting around the vignetting issue -- in fact, third party lenses that are "made for digital" can fit on a larger camera but show the vignetting and you have to factor that into your shooting.

[czeglin]

Great thread. An additional question. I understand I can't use an EF-S lens with a full-frame Canon body. Is this absolutely true, due to depth of lens incursion into body or lens mount differences, or is it just a consideration that the different configuration would cause vignetting?
The S in EF-S stands for "short back-focus" I believe. Parts of the lens protrude into the camera body closer to the sensor plane. On cameras with smaller sensors, the mirrors are smaller as well, and this isn't a problem. But if you managed to put the EF-S lens onto a full-frame camera the part of the lens that protrudes inward might get hit by the mirror as you take a photo. There are people who have modified EF-S 10-22mm lenses to fit on FF cameras, but they risk mirror damage if they use them at certain focal lengths where the lens intrudes.

[kre84u]

Thanks to both of you. You answered the main question I had. One additional, if I might.

If the EF-S lens, as opposed to a similar focal length EF lens, intrudes more deeply into the camera body, does this in any way offset the "crop" factor? Or is the measurement taken, not from the back of the lens, but from the aperture plane? I'm not trying to be difficult. I guess I'm looking for what advantage comes to Canon (not me) from this difference in construction.

[Jon]

No. It won't offset the "crop" factor. The measurement (by which I assume you mean the focal length) of a lens is not influenced by the depth of the camera body, or the distance between the sensor and the rear of the lens. A designer will start out by defining the focal length, desired maximum aperture and the back-focus (distance from rear of lens to sensor) and set about building the lens design from there. The advantage to Canon in developing EF-S lenses (and Sigma in their DC lenses, Tamron in the Di II, etc.), as explained above, is that it's easier and cheaper to both design and to manufacture a lens that covers only a smaller area (of the APS-C sensor of a T1i, for instance) rather than the 2.5x larger (areally) sensor of a 5D 2, 1Ds3 or 35 mm film camera.

[SkipD]

Thanks to both of you. You answered the main question I had. One additional, if I might.

If the EF-S lens, as opposed to a similar focal length EF lens, intrudes more deeply into the camera body, does this in any way offset the "crop" factor? Or is the measurement taken, not from the back of the lens, but from the aperture plane? I'm not trying to be difficult. I guess I'm looking for what advantage comes to Canon (not me) from this difference in construction.The APS-C format cameras such as Canon’s Digital Rebel series and their 20D through 50D series plus the new 7D have a smaller sensor than a 35mm film frame. If you limit the lens selection to those lenses designed to fill a 35mm frame (such as Canon’s EF series lenses), you will find that there are no ultra-wide-angle lenses for the APS-C camera.

The shortest zoom lens focal length in the EF lens family is 16mm. There are two primes that are a bit shorter, and one of those is a "fisheye" lens. NONE of these lenses are what the average photographer would call "affordable".

To design an ultra-super-wide-angle lens such as a 10mm (non-fisheye) lens for a 35mm film camera is a VERY expensive proposition, which is why there are none.

By making some changes to the design criteria - reducing the "film" area to be covered by the lens, and allowing the lens to project deeper into the mirror box (moving the rear element of the lens closer to the "film"), it becomes much more economically possible to design lenses for the task. Thus, the EF-S family of lenses was born, the “S” standing for Short back focus.

The EF-S lens mount is purposely designed to be different from the standard EF lens mount so that you cannot mount the EF-S lenses on cameras that were not specifically designed for them. If you modified the mount of an EF-S lens to be able to put it on a 35mm film camera, there would be a high probability that the mirror would crash into the rear element of the lens at certain focal lengths (the mirrors in the APS-C cameras are significantly smaller than those in 35mm cameras and “full-frame” DSLR’s). Also, the 35mm film frame would have a dark circle around the edges and the image would be inside the circle (known as vignetting).

Now that you have some understanding of what the EF-S lenses are all about, you need to understand that focal length is focal length is focal length. An EF 50mm prime lens designed for a 35mm camera, when used on an APS-C camera, will provide you with EXACTLY the same image size as an EF-S 18-55 lens set to 50mm. If you could cobble together a mount for a Hasselblad 50mm lens and use it on a 20D, you would again have the same size image as a result.

NO CHARACTERISTIC of any lens changes when you mount it on different format cameras. Focal length (or focal length range for zooms) never changes. Aperture range never changes. The only thing that would change is the apparent field of view, and that change is not a function of the lens but it is a function of the size of the sensor or film that will record the image. There are some changes to the depth of field, but that is also a function of the size of the sensor or film that records the image.

EF-S lenses, by the way, will only fit on the Digital Rebel series cameras (300D, 350D, 400D, 450D, and 1000D), the 20D, 30D, 40D, 50D, and the 7D as of this writing.