Zoom mm and f range

[dje]

Jeremy all the reponses so far have been insightful and helpful IMO. However I think what needs to be stressed is that lens design is a complex business and there are invariably trade-offs required between cost, physical size and performance. The lens deisgners do not limit the max aperture of a zoom lens at the longer focal length because they want to but because they have to in order to achieve acceptable distortion figures (geometric distortion and chromatic aberation) - within the budget contraints of that particular product.

Dave

[Jeremy87]

They're not limiting it at a long local length, because in that case it's the narrow FOV that limits it.
They are limiting it at a short focal length, because the wide FOV would naturally give a higher f-number if it wasn't limited.

I just don't get it. Let's simplify the thing into just one lens with variable focus (magic) so it can move along the optical axis and zoom. Its total FOV is constant, but the farther from the sensor it is, less light (from a smaller FOV) will hit it. The lens can't vary its size, so a wider FOV (lens closer to sensor) will always have more light hitting the sensor and be more bright.
Then there's the aperture stop that artificially limits it further (makes the lens effectively smaller), but when wide open we are limited by the lens itself.

Some people have pointed out that it's easier when it's constant, so my only conclusion so far is what I suspected from the beginning - it's by design to protect the user from making "stupid" things (no offence to anyone, it's just how many devices out there are made), or to put it more nicely, to make life easier where you don't have to adjust things manually.

Also what Dave said above, I guess a wider FOV is more sensitive for imperfections near the lens edges, but a bad lens is still not a feature I'd pay more money for.

I admit that I may not know a lot of things, 2 hours before I made this thread I knew pretty much nothing about cameras except the bare essentials on how to operate my mom's point-and-shoot. I just happened to get caught in a Wikipedia stream which I found interesting =).

[Manfred M]

They're not limiting it at a long local length, because in that case it's the narrow FOV that limits it.
They are limiting it at a short focal length, because the wide FOV would naturally give a higher f-number if it wasn't limited.

I don’t think that the people who have answered have quite understood your question. I think the question you are asking should be why does a lens stops down when it is wide open at the short end of the focal length. You feel that having the lens wide open would get you better performance, i.e. a faster lens.

It is important to remember that the lens designer has to make many compromises and trade-offs to get the optical performance that is required from a specific lens. This means dealing with aberrations, distortion, flare, etc. Because a zoom lens has fixed lens elements, the maximum speed it can deliver is indeed determined by the size of the lens openings, without the iris stopping down.

In your first posting you mentioned that you were looking at an f/3.5 – 5.6 18-200mm lens. I know the Nikon version of that lens reasonably well. Its list price is $840 Cdn where I live and is made for crop frame cameras. It is a mid-range lens, from a cost standpoint. It is a wide range zoom going from wide angle at 18mm to a mid-range zoom on the long end. If you look at the specs, you have an 11.1 zoom on it.

One generalization one can make about optics is that the wider the angle of view, the more the light has to bend, which introduces distortion. The way to correct the distortion is to add more lens elements; which adds cost.. The other way to reduce distortion, as every photographer knows, is to stop down the lens. I rather suspect this is what you are seeing when you zoom out; the lens designer has traded off a bit of lens speed at the wide end, to get a bit less distortion..

I think you are confusing FoV with DoF. The FoV is highly variable with any zoom lens and is totally related to the focal length you are using and the crop factor of your camera.

Some people have pointed out that it's easier when it's constant, so my only conclusion so far is what I suspected from the beginning - it's by design to protect the user from making "stupid" things (no offence to anyone, it's just how many devices out there are made), or to put it more nicely, to make life easier where you don't have to adjust things manually.

No; it is to ensure that the lens itself performs well and that image quality is good enough.

Also what Dave said above, I guess a wider FOV is more sensitive for imperfections near the lens edges, but a bad lens is still not a feature I'd pay more money for.

I admit that I may not know a lot of things, 2 hours before I made this thread I knew pretty much nothing about cameras except the bare essentials on how to operate my mom's point-and-shoot. I just happened to get caught in a Wikipedia stream which I found interesting =).

Again, I can’t agree with your statement. You are paying for a mid-range lens. If you need the speed and quality, you would have to get a better lens. Every lens manufacturer designs to a price point. Quality (optical and mechanical), performance, etc are all balanced off to build a lens for a price that people are willing to pay.

[Jeremy87]

In your first posting you mentioned that you were looking at an f/3.5 – 5.6 18-200mm lens. I know the Nikon version of that lens reasonably well. Its list price is $840 Cdn where I live and is made for crop frame cameras. It is a mid-range lens, from a cost standpoint. It is a wide range zoom going from wide angle at 18mm to a mid-range zoom on the long end. If you look at the specs, you have an 11.1 zoom on it.

That was just an example lens I saw mentioned on a website. I personally don't even own a camera. Like I said, I just happened to get caught in a "Wikipedia stream" you know, one interesting page links to another...).
Plus I already have some theoretical knowledge from studying physics, I just wanted to map that knowledge into camera vocabulary and gain some more practical knowledge for whenever I can afford a non-cellphone camera in the future. And whenever I use a friend's camera, I like to know what's happening and why when I take the photo.
This is all for the sake of knowledge, not because it's useful to me right now =).

I think you are confusing FoV with DoF. The FoV is highly variable with any zoom lens and is totally related to the focal length you are using and the crop factor of your camera.

No, I was ignoring DOF as I already know how that works.

One generalization one can make about optics is that the wider the angle of view, the more the light has to bend, which introduces distortion. The way to correct the distortion is to add more lens elements; which adds cost.. The other way to reduce distortion, as every photographer knows, is to stop down the lens. I rather suspect this is what you are seeing when you zoom out; the lens designer has traded off a bit of lens speed at the wide end, to get a bit less distortion..

Again, I can’t agree with your statement. You are paying for a mid-range lens. If you need the speed and quality, you would have to get a better lens. Every lens manufacturer designs to a price point. Quality (optical and mechanical), performance, etc are all balanced off to build a lens for a price that people are willing to pay.

So the reason really is just lens imperfections?
The "pay more" comment was directed to an earlier answerer who would pay extra for a lens with a constant f-number. So I said (or meant) that if I pay more, I expect a better quality lens that can tolerate a higher f-number at wider angles. Not necessarily wide open (if it indeed leads to aberrations, unless it's really expensive), but not constant.

[Manfred M]

That was just an example lens I saw mentioned on a website. I personally don't even own a camera. Like I said, I just happened to get caught in a "Wikipedia stream" you know, one interesting page links to another...).
Plus I already have some theoretical knowledge from studying physics, I just wanted to map that knowledge into camera vocabulary and gain some more practical knowledge for whenever I can afford a non-cellphone camera in the future. And whenever I use a friend's camera, I like to know what's happening and why when I take the photo.
This is all for the sake of knowledge, not because it's useful to me right now =).

Theoretical knowledge is good, but it makes a lot more sense when you get real-life shooting under your belt. I've studied a lot of physics...

No, I was ignoring DOF as I already know how that works. .

In that case you don’t quite get field of view. A camera lens, depending on its focal length sees a cone from the image sensor. Field of view is sometimes known as angle of view. My 8mm fisheye lens has a field of view of 180 degrees, i.e. it “sees” pretty well everything in front of the lens. The other extreme would be my 150-500mm lens at 500mm has a field of view of just 5 degrees, so it has a very narrow FoV.

So the reason really is just lens imperfections?
The "pay more" comment was directed to an earlier answerer who would pay extra for a lens with a constant f-number. So I said (or meant) that if I pay more, I expect a better quality lens that can tolerate a higher f-number at wider angles. Not necessarily wide open (if it indeed leads to aberrations, unless it's really expensive), but not constant.

Any lens has imperfections, so that is a bit of oversimplification. I prefer to leave it as a design decision made by the lens design team.
As for a constant f/ lens, I have three of them. The reason for shooting with fast, good lenses is to take advantage of the very shallow depth of field to throw the background out of focus and get a more pleasing composition. This is especially important for portraiture, where a good fast zoom gets wonderful shallow depth of field results. Ask any professional portrait photographer which lens they use, it will likely be the f/2.8 70-200mm. Ask a wedding photographer the same question, and the most likely response is the f/2.8 24-70mm lens.

[Jeremy87]

Theoretical knowledge is good, but it makes a lot more sense when you get real-life shooting under your belt. I've studied a lot of physics...

Theoretical is all I can afford at the moment, so it will have to do.

In that case you don’t quite get field of view. A camera lens, depending on its focal length sees a cone from the image sensor. Field of view is sometimes known as angle of view. My 8mm fisheye lens has a field of view of 180 degrees, i.e. it “sees” pretty well everything in front of the lens. The other extreme would be my 150-500mm lens at 500mm has a field of view of just 5 degrees, so it has a very narrow FoV.

I don't know what I've said that suggests I don't understand FOV...
A longer (effective) distance to the sensor gives a narrower FOV (plus you may want to adjust the focus accordingly).

Any lens has imperfections, so that is a bit of oversimplification. I prefer to leave it as a design decision made by the lens design team.
As for a constant f/ lens, I have three of them. The reason for shooting with fast, good lenses is to take advantage of the very shallow depth of field to throw the background out of focus and get a more pleasing composition. This is especially important for portraiture, where a good fast zoom gets wonderful shallow depth of field results. Ask any professional portrait photographer which lens they use, it will likely be the f/2.8 70-200mm. Ask a wedding photographer the same question, and the most likely response is the f/2.8 24-70mm lens.

So the portrait photographer wants a f/2.8 70-200mm. This probably means the front lens is 200mm/2.8 = 70mm wide.
But at the widest FOV, you're only using 70mm/2.8 = 25mm of the lens width. Wouldn't it be nice to use a little bit more of it for that shallow DOF? With a perfect lens, f/1.0 is possible, and while it most likely isn't perfect, being limited to f/2.8 seems a bit wasteful.

[Manfred M]

I don't know what I've said that suggests I don't understand FOV...
A longer (effective) distance to the sensor gives a narrower FOV (plus you may want to adjust the focus accordingly).

Its because FoV has nothing to do with the distance to the sensor. You are confusing simple lens design that you see in introductory physics with the way a modern lens is designed. The distance from the lens element to the imager has nothing to do with focal length. FoV is 100% dependent on the focal length of the lens.

So the portrait photographer wants a f/2.8 70-200mm. This probably means the front lens is 200mm/2.8 = 70mm wide.

Correct,( in theory). In practice, a slightly larger element is used. If my filter size is any guide. the front element of my 70-200mm lens is actually 77mm in diameter. There has to be enough material to actually mount the lens element in the lens body, so some of the glass is not used for optical purposes.

But at the widest FOV, you're only using 70mm/2.8 = 25mm of the lens width. Wouldn't it be nice to use a little bit more of it for that shallow DOF? With a perfect lens, f/1.0 is possible, and while it most likely isn't perfect, being limited to f/2.8 seems a bit wasteful.

Actually, in theory I believe the limit is around f/0.5 for a standard lens (one can do better if you start introducing mirrors into the design). In practice, the optical (quality) and cost trade-offs are such that very few commercially available lenses are anywhere near f/1, and none of these are zoom lenses.

The only current mainstream (if you call Leica mainstream) lens is the Leica Noctilux-M 50 mm f/0.95. Cosina manufactures a few models, primarily aimed at the MFT market, and there are some smaller players making speciality lenses. Optical performance (resolution) wide open is not as good as lenses that are a stop or two slower.

It’s only wasteful, as you put it, if the output were of sufficient quality to make it useful, and unfortunately, it is not, otherwise some lens manufacturer would be doing so.

[Jeremy87]

Its because FoV has nothing to do with the distance to the sensor. You are confusing simple lens design that you see in introductory physics with the way a modern lens is designed. The distance from the lens element to the imager has nothing to do with focal length. FoV is 100% dependent on the focal length of the lens.

Well, obviously the physical distance isn't 300mm on a 300mm lens, it would be huge. That's why I used the term "effective distance", because the physical distance can be shortened by adding the right set of lenses inbetween.

Actually, in theory I believe the limit is around f/0.5 for a standard lens (one can do better if you start introducing mirrors into the design). In practice, the optical (quality) and cost trade-offs are such that very few commercially available lenses are anywhere near f/1, and none of these are zoom lenses.

The only current mainstream (if you call Leica mainstream) lens is the Leica Noctilux-M 50 mm f/0.95. Cosina manufactures a few models, primarily aimed at the MFT market, and there are some smaller players making speciality lenses. Optical performance (resolution) wide open is not as good as lenses that are a stop or two slower.

It’s only wasteful, as you put it, if the output were of sufficient quality to make it useful, and unfortunately, it is not, otherwise some lens manufacturer would be doing so.

So, and I find I'm repeating myself, it all comes down to the quality of the lens not being able to handle large FOVs with too low f-numbers.

[herbert]

I might try a summary:

The focal length is defined by the angle of view, i.e. over what angle can you see.

The amount of light that can be focused by the lens is a key defining point. This is expressed as the effective entrance pupil. It is not the size of the smallest physical aperture (hole) inside the lens but instead is expressed using the extent of the focused image back-projected to when it first enters the lens, i.e. on the front element of the lens. This is done for convention and allows us to define the f-number.

The f-number is the ratio of the focal length to the effective entrance pupil of the lens.

Therefore the effective entrance pupil is upper bound limited by the physical size of the front of the lens, but it may be smaller because not all the light that hits the front element is focused into an image. E.g. Some glass may be used just to make the edge of the lens and hold it in position, some light may not make it past the physical aperture inside the lens.

As you move the lens elements the amount of light hitting the front element that can be focused changes, i.e. the entrance pupil changes.

In some lenses the entrance pupil changes disproportionately to the focal length (variable f-number). This is the easy way to do it.

In other lenses the entrance pupil changes proportionately to the focal length (fixed f-number). This is the harder way to do it.

---

In all that I think the key point is that when the focal length changes, you have a different lens. Therefore the focusing is different and the light that can be focused is different. So the entrance pupil is different. So the f-number is different.

Going back to the original suggestion, why not make the physical hole much bigger? Imagine that you had all the glass elements in place but removed the physical aperture. Would the f-number change? If the extra light can be focused. Yes. If it cannot be focused then you have no image and so there is no point.

We know that when a lens is stopped down the quality improves. So we can infer that wider apertures perform worse. Thus there must be a part of lens design where a subjective measure on the quality of the image is performed with different apertures. A physical aperture is then chosen that prevents the image from becoming too poor.

The bottom line:

With a zoom lens the physical aperture must be chosen so that the minimum image quality is achieved across the zoom range. However the physical aperture that is chosen may not be the best aperture for each focal length. If you want that, then you will have to buy a fixed focal length lens.


Alex

[Manfred M]

Well, obviously the physical distance isn't 300mm on a 300mm lens, it would be huge. That's why I used the term "effective distance", because the physical distance can be shortened by adding the right set of lenses inbetween. .

Correct – photography, like most other fields has its own jargon, and you are not using the same terminology as photographers would use, so your meaning is not always clear to me when trying to respond to your comments.

So, and I find I'm repeating myself, it all comes down to the quality of the lens not being able to handle large FOVs with too low f-numbers.

I would say it a bit differently. What you call “quality” is not really an accurate description of the issue. Quality is a good way of describing how well something is designed and built. The basis of comparison could be the theoretical limits

Quality has nothing to do with the laws of physics, which are the real constraints here. Light travels in a straight line. Different wavelengths travel at different speeds as they travel through the optical component (think of light travelling through a prism and showing all of the colours of the rainbow). This phenomenon is known as dispersion. Light reflects and refracts as it passes through the interface where the materials have different indices of refraction. Thin film interference (a.k.a. anti-reflective coatings) can only do so much.

The more the light bends (i.e. a larger diameter lens element = small f-number), the greater the level of dispersion, and the more difficult it is to correct for this effect.

That is not to say that quality plays no part in the discussion, but ultimately, the physics of optics are the main constraint.