The 200mm f2 has a 100mm rear element, its front elements are much larger than that.
Why I dont like fixed aperture zoom lenses.(39 posts) (19 voices)
What an interesting post, thanks heartyfisher! After reading everyone's comments (yes I know that's cheating :-), it seems to me the reasons for buying a fixed aperture zoom is 1) Better optical performance at the wide end, and 2) your exposure not changing as you zoom. As has been said, as someone who usually shoots in high stress, fast moving situations, having that one less thing to think about is highly attractive to me, so I suppose what I'm saying is, given the choice between a 70-200 2.8, or a 70-200 1.8/2-2.8, I'd actually prefer the slower constant.
@SquamishPhoto : "The 200mm f2 has a 100mm rear element, its front elements are much larger than that."
"1) Better optical performance at the wide end"
So you are saying that for variable aperture zooms the Wide end has poorer IQ than the longer parts of the zoom range? I don't think I see that in my variable aperture lenses VS the Fixed aperture lenses.
"2) your exposure not changing as you zoom."
I have heard this many times before, I just dont get this at all -- If I set my aperture priority to F5.6 however much I zoom or focus its still F5.6.
I think your question is coming since some point and clicks have a very wide aperture (like 1.8 or 2.8) at their wide end and a narrow aperture (maybe 10 or 22) at their long end? Is that correct?
I don't know much about lens design but I own a Tokina that is 28-70 2.6-2.8. While that lens is reasonably sharp (as sharp as my 35mm/1.8 at on DX) it is not by a long shot my sharpest lens and has chromatic aberation issues. I suspect if they could have designed a lense with a wider aperture they certainly would have so I expect there are technical challenges if you look at the design of a high quality lens.
I don't understand the people posting that having a constant max aperture is a virtue, but if the optical quality is unuseable I wouldn't want the lower aperture # to be available.
The 200mm f2 has a 100mm rear element, its front elements are much larger than that.
The largest rear element possible with the F-mount is 40.8mm, part of which is already obscured by the contact pins. (This means that a lens with an opening of f/1.0 is not possible in the F-mount, because there is not enough room for the marginal ray angle of such an opening). As for the front element: the 200mm f/2 really does require a minimum 100mm diameter lens [f=200mm; => f/2 = 100mm). With an overall diameter of the lens of 124mm, the actual front element is indeed just a touch over 100mm in diameter.
For god's sake, will someone, or rather one more person than just gasman, PLEASE understand his simple question! :-)
Everyone is referring to how unuseful that would be and how it would make the lens extremely expensive etc. pp. Did I miss a post? I guess I did, but from the posts I've read, only gasman got the point that heartyfisher was making, that to an optical layman who is not working in lens design, it appears strange to have a lens that is capable of i.e. 24-70 f/1.8-2.8 (or something like that) AS IS (!!!) be artificially downgraded to 24-70 f/2.8. The keyword is: AS IS, because for a lens to be f/2.8 at 24, the opening has to be closed down when you zoom in from 70mm f/2.8. If the opening was just the same, f/2.8 at 70mm would become f/0.96 at 24mm. Right?
All the rationalizing of how much more practical it is having fixed aperture doesn't count, as heartyfisher already said.
All the arguments that it would make the lens huge are wrong (I'd guess), as they're not picking up on the core of the question (a fixed aperture zoom is already larger at the wide end and seems to be artificially limited (closed) when going wide.
If someone reads this who thinks "What??? I answered exactly to that!", please point me to the post.
I've always been asking myself the same thing, actually.
When you design a lens there are some optimazations that have to be done for the widest aperture. If the lens you create would be a 24-70 f/2.6-2.8 it would not be much of a difference. In that case there would not be any needs to artificially limit the lens to be a fixed f/2.8.
On the other hand if your lens would be a 24-70 f/2.0-2.8 the entire design would have to be optimized for the 24 mm f/2.0. Chances are rather high that this would increase the cost of the lens.
The other option would be to design a fixed f/2.8 and not limit it at the wide end making it the 24-70 f/2.0-2.8 - just as the OP suggested. In that case the optical quality might not be very good at 24 mm f/2.0. This means bad reviews - some might even say it is not an f/2.0-2.8 lens but a fixed f/2.8 that is more expensive because it calls itself f/2.0-2.8.
So in the end it might be more of a marketing decision than a real technical issue.
I will attempt to explain why constant aperture is more complex. First of all, for those really interested, Pierre Toscani's website ( http://www.pierretoscani.com/annexeGB.html ) has a number of pages that have some of the best animations of lens principles ever seen on the web.
But let me summarize to a point. Almost all zoom lenses are actually made up of three distinct groups of lens elements: at the rear, we have a regular lens, corrected for all the good stuff, like spherical aberration, coma, image plane planarity, colour aberrations and everything else that comes to mind. That lens usually contains the diaphragm, so that it can be stopped down. As we know from Nikon's lens catalogue, a half-decent prime lens contains at least six elements, consisting of two Cook-triplets, sometimes broken up into additional elements to obtain additional benefits. In front of this sits the "variator". Pierre Toscani explains this better than anyone, and you can play with mouse-over animations that illustrate this principle http://www.pierretoscani.com/echo_telezooms_english.html . The variator can only work, if it consists of two lens groups, which move in relation to each other. The front one is a "positive" group, i.e. it concentrates rays, while the second group is a negative one, i.e. it makes rays diverge. When the light exits the variator, parallel rays are parallel again, but the variator has managed to make those ray bundles converge or diverge, to create larger or smaller images on a sliding scale. These images are now projected onto the sensor by the regular lens at the back. If we only make the variator big enough in diameter (there are limits when it comes to wide-angle constructions, but I leave this aside), the whole zoom lens will be of constant aperture, namely the aperture f-value of that proper lens. The elements of the variator are ideally Cook triplets into themselves, so that we arrive at a construction of at least 12 elements.
It gets more complicated now: Such a zoom as described above still has a number of limitations. For VR, add another Cook triplet. Also, every one of these groups moves, if focusing and zooming is carried out. We have become a demanding lot: we want the front element to be stationary (add another three elements or so), we want internal focusing (add some more) and we want wide-angle zooms, which complicate the construction even more, because we cannot use the space which is occupied by the mirror box.
Of course, the lens designers look for short cuts, to achieve zoom constructions for consumer lenses that do not have 18 or more elements, plus exceedingly complicated gears to move the groups, and to retain focus when we zoom through the range. One of the possible short cuts: let the second part of the variator be a part of the actual lens. The rays that come out of the variator are now no longer parallel; rather they look like a ray bundle somewhere inside a prime lens. If the rays are so compromised, the aperture now becomes variable, because we cannot make the inside elements big enough that they catch all of a diverging bundle at the extremes, thus limiting the f-number as the bundle opens up to emulate a tele lens.
The most dumbed-down zoom lens construction I have ever seen is the old 18-55 DX (not to be confused with the fairly decent 18-55VR that is the only lens of this range now), which uses every short-cut in the book to get down to just 7 elements. The actual lens is a Tessar-type (no triplets); out front is a large negative lens to achieve retrofocus, and two more elements make up the variator. More precisely, the negative front element does double duty as retrofocus diverger and being one half of the front variator. The rays after the variator are not parallel.
Since that abominable lens, the pendulum has swung back towards more quality. Pierre Toscani has a good rendition of the 70-300mm VR (which closes from 4.5 to 5.6 as you zoom through the range), but has otherwise fairly decent qualities. Of course, you pay for additional luxuries, and the present 70-200VRII is one of the most elegant constructions anywhere. It does cost 4 times as much for a reason...
Summing up: Constant aperture zooms have a complex construction that separates variator from lens. They cannot be opened up on the short end, because they are "lens-limited". Variable-aperture zooms are less complex, they are "variator-limited".
Did you not read my long and windy exposé on the first page? If that did not count as understanding Hearty's simple question then I don't know what would :)
As for the answers, I think it's a case of you not liking the answers more than no answers being given.
Could it be done? Probably yeah. Could it be done without making the lens bigger, more expensive? Again, probably yeah. Would it be a good idea? Most likely not. Why? Quality would suffer considerably at the wide (short) end when the lens is wide open... which would lead to bad MTF charts... which (like correlli said) would lead to bad reviews... which would lead to bad.... you get the idea.
"If someone reads this who thinks "What??? I answered exactly to that!", please point me to the post."
"Hearty, I'm not a lens design expert and don't know much about lenses beside a basic rudimentary knowledge but if I had to take a guess as to why no one makes such lenses, I'd say it's because the wide end would suffer considerably if they did. I think the problem here is that we're assuming that the aperture size at the tele end is the only limitation and ignoring any possible undesirable effects at the wide end. Current constant aperture lenses almost always will have poorer performance with aperture wide open at the wide end than at the tele end and specially at the corners. Making the aperture bigger at the wide end while keeping the size of the lens the same while at the same time maintaing the same quality is probably very hard from a lens design perspective."
@Meinrad: Great summary. Let's hope this thread is dead now because if variable apertures were better/easier/cheaper to make than fixed, every company's pro lenses would be made like that.
Corelli, Meinrad and Gabandi: Thanks very much for the answers, especially Meinards extensive explanation and the links!
@spraynpray: Oh come on :-) Seriously, the argument "If it was easy, it'd be more common" as an answer to a "Why is it like xyz" question is almost like answering "Because we've always done it like that". Did you know all this stuff Meinrad explained? I didn't, and I bet many others didn't either.
I think we should "get" a 70-200 2.8 VR II now. :-)
Thank you for this absolutely wonderful explanation. Even an old person can understand it. It also confirms why I tend toward prime lenses, as they just do not have to have as much inside, thus potentially can produce better results. My 24mm f/1.4 Nikkor G is one of the best I can think of.
I wonder if anyone has thought about why Carl Zeiss and Ernst Leitz make almost no lenses which are variable focal length. Oops, it is now called Leica Camera.....
Anyway, after a look at what lenses are offered by these two companies, certainly at the very top of the quality pyramid, one might think a zoom is a compromise to begin with. And, if I am not mistaken, wide lenses require some rather interesting corrections to avoid a wide range of undesired distortions. Also, while one might think the smaller aperture diameter required would allow a faster lens at the short focal lengths, I think when we get to f/2.8, going faster is very very expensive. This is especially true in a zoom lens. Nikon does not make a zoom faster than f/2.8, but if they did, like a 70-200mm f/2.0, yikes, it would be a $7,000 lens. Not a lot of demand for that I would guess.
If you go look at the cinematic lenses Zeiss and others offer (just do a category browse at B&H) and you will start to see the actual cost of such lenses. We're talking about $100,000 glass for high end here boys and girls.
Gabandi did a good job giving the #s (unfortunately for those who only work in imperial distances, the difference MM is completely lost on us). If you go back to past lenses (20-50 years ago) you will find F2-2.8 lenses. They didn't last for one reason or another but none of those reasons were because people didn't want it. Price, optical quality, etc is what made them go away.
I have a Tokina 28-70 f2.6-2.8 ATX PRO (an Angenieux design) that has a great "feel" but the 2.6 at 28mm it stinks - it really makes a difference that you can see. CA are terrible, sharpness drops like a rock from 2.8 to 2.6 and a whole host of other technical issues that are just unbelievably multiplied. I just leave it at 2.8. Now if I just shoot B&W with it, it has one of the greatest feels to it out of any lens I have used.
You must log in to post.