What's the significance of lens construction?

[Kajuah]

I am trying to learn more about what I see and what things mean when reading up on choosing a certain lens for my camera. I look up a lens like this one;

http://www.canon.ca/digitalphotograp...?id=961&cid=12

And I can understand and interpret all of the specification information aside from lens construction. I don't know and would like to know (if someone would be so kind to educate me) what this is, what to look for, what significance it has when buying a lens, and just generally anything else that you know about the topic :P sorry if that's a bit general, I'm a bit of a moron when it comes to mathematical values (ironically I choose the art which is centered around manipulating those mathematical values to get pictures! :idea: )so when I see something like " 17 elements in 12 groups " I go

[SmartWombat]

More elements, more groups, more chance for reflections internally.
Though with a Canon lens it ought to be anti-reflection coated internally.

EF-S means that it can't be used with the professional bodies, only the 20D and the Rebel bodies.

What's more of interest to me is the aperture.
1:/4-5.6 means as you zoom from 17mm to 85mm the aperture changes from f/4 to f/5.6
Not much better at wide angle and for me too small an aperture for low light shooting.

It's got full time manual focus with the ring-type ultrasonic motor, even in AF mode.
The Image Stabilisation is good for static subjects in low light, but won't help you for moving victims.

[Kajuah]

Ahh I see now, thanks for that..I don't understand (unless exception of patents) why all camera makers have their own little name for the same definition (do you know what I mean?)

Canon - nikon for example:

EF-S - AF-S
IS-VR

Same thing, different silly abreviation.

[Frog]

I thought more groups and/or elements made it better?

[Franglais]

I thought more groups and/or elements made it better?

You could have a lens with just one element - but it wouldn't be very good. Each element in a lens "bends" the light but also introduces some errors (couloured fringes, distortion). The lens designer uses several elements with opposite errors so that finally the errors cancel one another out and you get a near-perfect image. The elements may be cemented together in one group or in several groups with a space between each group. On most prime lenses 4-6 elements is enough to get an excellent result.

In a zoom lens there is at least one group of elements that moves in order to change the focal length. On a VR/IS lens there is yet another group that moves to do the image stabilisation trick. The result is a lens with up to 20 elements.

Unfortunately each time light enters or leaves a lens element there is some dispersion resulting in loss of contrast, so there is a limit on how many elements you can have in a lens and still get a usable image. Thanks to progress in lens coatings and glass technology, you can now get very complex zoom lenses that give results that are almost as good as simple prime lenses (the zoom lens still has more distortion).

Example: my 18-200VR gives me better results than my 5 element 28mm f2.8 prime

[Kajuah]

So just so I understand...a balanced element to group ratio = good?

[deckcadet]

Ahh I see now, thanks for that..I don't understand (unless exception of patents) why all camera makers have their own little name for the same definition (do you know what I mean?)

Canon - nikon for example:

EF-S - AF-S
IS-VR

Same thing, different silly abreviation.

No. EF-S and AF-S are completely different. EF-S would be similar to DX, however Nikon's DX lenses can be used on all Nikon DSLRs to date... even the full frame ones, in DX crop mode.

AF-S is the equivalent of USM on the Canon side.

[Kajuah]

ah i see now...switching between systems is like learning the languages of the two ! ironically canon and nikon are both asian ...

[Franglais]

So just so I understand...a balanced element to group ratio = good?

I am not an expert in optics but I don't think having a balanced element to group ratio is important. What counts is the results.

For years I have been collecting lens reviews from my favourite photography magazine (Chasseurs d'Images - sorry it's in French), and testing my own gear. My general conclusion is:

- expensive=good
- cheap=can be good but some manufacturers are better than others (Nikon are particularly good)

[Wild Wassa]

Why are there so many elements in a lens and why are they in groups? ... and following on from Franglais's first reply.

In a nut shell, ... but I'm happy to delve deeply into the theory if you like.

Compound lenses are necessary because single elements can not be highly corrected of aberrations.

Groups of elements consist of positive elements (convex lenses) and negetive elements (concave) which neutralize many of each other's aberrations, combining to produce an almost perfect positive lens.

These aberrations are only corrected for a limited range of object to image ratios. With the lens you are looking at, the Canon EFS 17-85mm f1:4-5.6, as you zoom the lens you change the subject to image ratio. The aperture not only changes the amount of light reaching the image plane, it also has a function called, being an effective aperture that helps to correct aberrations or unfortunately induce aberrations.

When the lens is at 17mm a minimum aperture of f4 will adequately reduce aberrations but when the lens is being zoomed out to 85mm if remaining at f4, aberrations will occur. By reducing the maximum aperture towards the higher focal lengths to f5.6 those aberrations are then reduced.

You might like to see this lens test programme for the EFS 17-85mm f1:4-5.6 showing the Nyquist Frequency which gives the maximum theoretical resolution of the lens at the different zoomed focal lengths in relation to the different apertures over the entire field of coverage. In this test you can set varying focal lengths and f-numbers and scan the field of coverage, to find the optimum settings for the lens. The test will give a relationship to sharpness over coverage in line pairs and shows changes in chromatic aberration and brightness fall off.

Set to 35mm at f5.6 the lens is close to optimum, from the centre of the image out to the edges of the field of coverage. At 17mm at the edges of coverage ... well check the test and find the most effective aperture.

The test is not forgiving on this lens, nor should it be ... I have the EFS 17-85mm f1:4-5.6 and I'm delighted with it. I think it is a good lens. The test showed me the weeknesses to avoid.

http://www.dpreview.com/lensreviews/..._c16/page3.asp

You will need Adobe's Flash Player installed, downloads can be done from the test page if you don't have Flashplayer. Also allow some time for the test page to load ... it was almost a minute and a half.

Warren.

[Kajuah]

Thanks warren, that explanation was perfect for me, I totally understand now

[Wild Wassa]

Kajuah, you are welcome Mate.

Warren.

[Loupey]

All else being equal, the number of elements in so many groups is simply a general measure of how well "corrected" a specific lens has been designed.

But because no two lenses are ever equal, this specification is not really useful.


As for zooms, there are so many zoom ranges and aperture sizes provided by so many manufacturers that the #elements isn't a valid comparison.

As for primes, it may be helpful but only when comparing similar lens designs of the same focal lengths (i.e. a 50mm f/1.2 verses a 50mm f/1.4 verses a 50mm f/1.8). But definitely not when comparing a 50mm f/1.8 verses a 28mm f/1.8.


One thing it does do is it generally adds weight to a lens. Although my 300mm f/4 is pretty darn light for having 15 elements (I had to check the #elements).