Do image stabilzers weaken the IQ of a lens?

[sebastian_klamt]

<p>Hello everybody,<br>

in another tag I asked about the digital slr I consider to buy. Now I'm quite sure that it'll be the Eos 60D camera.<br>

But talking about the lens I aspire ( the Tamron 2,8/17-50mm), someone wrote that the version with image stabilizer (VC) has a weaker IQ than the version without.<br>

So can you generalize that image stabilizers cost quality?<br>

Kindest "exposural" greetings from Germany,<br>

Seb</p>

[jaydesi]

<p>I don't know that you can make the generalization, especially considering there are many lenses sold that only come in IS or non-IS format, you don't have a choice.</p>

<p>A better way to think of it is whether the IS will allow you to get an in-focus shot in low light where the non-IS wouldn't due to camera shake. In that case, the non-IS could be far superior optically, but the resulting image will be quite poor compared to the IS version of the lens.</p>

<p>Odds are good that in real shooting, you wouldn't likely notice the IQ difference between the lenses, and IS is very handy. </p>

[david_henderson]

<p>By no means all IS lenses are available in a non-IS form, so its often difficult to disentangle the effect of IS vs other variable parameters. However the Canon 70-200 f4 is. I've heard of people choosing the non IS version because its cheaper, or because they don't value IS, or to part fund a f2.8 version rather than the f4. I haven't seen anything to the effect that that selection has been made to improve image quality.</p>

<p>Of course even here the effect of IS is tangled up with decisions to use a tripod or not. If for example you decide that because of IS you'll take a shot hand-held with a telephoto at say 1/30, then that image may well not be as sharp as you would have achieved with a tripod- so you could say that the availability of IS has reduced image quality even though its not the lens's "fault". You might add a further complexity if for example this shot were to have been made at f4 with narrow dof in order to shoot handheld, whereas if on a tripod you'd have chosen f8 gaining better sharpness and more dof.</p>

<p>Equally someone intent on hand-holding could make a series of images in the range (say) 1/15 to 1/60 and point out that they're notably better than he'd expect to achieve without IS. </p>

[paddler4]

<p>I don't think you can generalize. it is safer to look at reviews for each lens. For example, David mentioned the 70-200 f/4 IS. If I recall correctly, the reviews generally say that it is sharper than the non-IS.</p>

[scott_ferris]

<p>Short answer. From an ultimate IQ point of view, yes. From a real world image point of view, no.</p>

<p>Longer answer. More elements mean more problems for lens designers, more reflections, more light transmission loss etc etc. It is no secret that the non IS versions of a lens generally test slightly better.</p>

<p>But that isn't the whole story, few of us shoot on a test bed with manual focus, live view and a test chart. Say, for example the non IS lens tested at 100%, the IS version might be 90% of that, in real world use handholding will take away 25% from the non IS lens but only 10% from the IS one, that makes the IS lens "better" in real world use. The next question is how noticeable is the initial 10% difference? Say you put both on sturdy tripods and took a very detailed landscape, obviously the IS would be switched off due to tripod use, now take your two images and print them at 20"x30", how many people could see a difference? Very few.</p>

<p>The conclusion is, whilst theoretically IS does weaken the ultimate IQ of a lens, in real world shooting the benefits it gives far outweigh the theoretical disadvantages. If you were thinking of getting a 70-200 f2.8 to live on a tripod in a studio and always shoot with studio light the non IS is a better buy and might, perceivably get you noticeably sharper images, for almost every other use, the IS version will get you better images. Just check out<a href="http://www.the-digital-picture.com/Reviews/ISO-12233-Sample-Crops.aspx?Lens=242&Camera=453&Sample=0&FLI=3&API=0&LensComp=103&CameraComp=453&SampleComp=0&FLIComp=3&APIComp=0"> this link,</a> it shows that the answer is not clearcut, lens design is a series of compromises. IS good!</p>

[Marcus Ian]

<p>The short answer is that yes, it is harder to build a lens w/ IS that produces a high level of IQ, than an equivelant lens w/o the technology. In many circumstances however, the technology allows a 'net' gain in IQ, as it allows much improved IQ in situations where handheld slow shutter speeds are required. So you can't really say that 'IS costs IQ' because it's rare in the field to get optimal levels of sharpness out of <em>any</em> equipment unless you have the skill and time and conditions to allow it.</p>

<p>For IS, the design challenges are numerous, and (at least w/ Canon) the designs have gone through numerous 'generations' each time improving the responsiveness and capability of the equipped lenses. They have also been utilizing the technology for darn near 20yrs. Tamron and Sigma are still on their 'first generation', which has performed considerably better than Canon's 'first gen' though they've also had 20yrs of tech improvement (plus Canon's experience) to build on.<br>

However, basic physics can help explain why it's difficult to build such a system (and make it work!) in the first place. As well as why the 'faster' a lens is, the harder it is to apply improved IQ. The faster a lens is, the larger it's smallest internal lens must be (as well, adding focal length also increases the minimum lens diameter). For example, the 70-200/4 has a minimum aperture of 50mm, a 70-200/2.8 has a minimum aperture of ~71mm. Depending on the details (thickness, curvature, etc.), such a change in diameter can <em>easily</em> equate to double the mass from the f4 to the f2.8 moving element. Therefore the IS system must be not only capable of moving the moving lens as dictated by the controller, but be able to so rapidly, and precisely. The larger that single element is, the harder that mission is to accomplish. <br>

Add to that the decreased DOF in a faster lens, and you've compounded the problems presented.</p>

<p>An excellent example of that is the Sigma 70-200/2.8 HSM. They have two versions, one w/ OS (their version of IS), and a predeccesor without. The OS equipped unit is noticeably softer WO than the non-OS version - in good light, and during tests. Of course if the light is dim, and your shutter speed (even at f2.8) declines to the 1/50 range at the tele end, you are unlikely to see any benefit at all from the 'sharper' lens.<br>

The Tamron 17-50/2.8 VC vs. non-VC is a similar story. The VC equipped lens is (in ideal shooting conditions) softer than the non-VC version. But if you are shooting at the limit and w/ available light, you are unlikely to see that difference. Your hands will suck the sharpness out of any lens if shot slowly enough.</p>

<p>Frankly though, I think it's unlikely that you'll see any lens's upper limit in sharpness and IQ anytime soon. It takes time (and practice) to optimize the skills required for that, so if I were you, I'd go for the VC unit, despite the slight penalty in 'ultimate' IQ.</p>

[sebastian_klamt]

<p>Hi,<br>

it's amazing how fast you get qualified answers to your questions, thanks a lot!<br>

Of course i don't think that IS is to improve the IQ of a lens!<br>

But used to classic FD lenses and cameras, I could cope with going without an IS! So in my opinion IQ beats IS! For I'm not a papparazo, I can use a tripod if nescessary. And for example concerning, the Tamron 17-50mm, there are two versions. So I aim to get the best IQ! And I'd like to know what you can say about the quality of Tamron lenses in general! </p>

[mark_pierlot]

<p>Sebastian, a further consideration is that IS is more useful for longer focal lengths than for shorter ones, and 17-50mm is not particularly long.</p>

[arie_vandervelden1]

<p>Go get the 17-50/2.8 non-VC. It's a superb lens. Super sharp and great value.</p>

[Marcus Ian]

<p>Sebastian - If you value ultimate IQ, and can live w/o the VC, the non-VC version is the one to get. The 17-50/2.8 is a marvelous lens for crop sensor units. It's incredibly sharp and produces imagery on par w/ Canon's finest equivelant (the EF-S 17-55/2.8 USM IS). <br>

The only hiccup you may find is that, since you are coming from FD, MF is not as refined as on the USM (w/ FTM) lenses. </p>

<p>As far as the overall quality of Tamron glass, despite a dubious past, their modern products have proven to be pretty darn good. Sure, most aren't <em>quite</em> up to their 'L' equivalent, but considering they are ~1/2 the price, that's not really too much to ask. They've just started putting their "USD" (their version of USM) lenses out, so what has long been the only valid complaint about the brand in general (that their AF kind of sucks) may soon be a thing of the past. I guess we'll see, but I'm waiting to see how their announced 24-70/2.8 USD VC performs, and if it performs decently, I plan on 'upgrading' my 24-70/L.</p>

[mike_earussi1]

<p>IS, or any other camera companies version, can result in a lowering of the IQ primarily due to the increase in flare the extra IS element(s) add. Just make sure you check put the various test sites regarding flare effects before you buy. My favorite one for testing flare is this one:<br>

<a href="http://www.optyczne.pl/Testy_aparat%C3%B3w_Testy_obiektyw%C3%B3w_Testy_lornetek_Inne_testy.html">http://www.optyczne.pl/Testy_aparat%C3%B3w_Testy_obiektyw%C3%B3w_Testy_lornetek_Inne_testy.html</a></p>

<p>sample page: <a href="http://www.optyczne.pl/106.9-Test_obiektywu-Canon_EF_70-200_mm_f_2.8L_IS_USM_Odblaski.html">http://www.optyczne.pl/106.9-Test_obiektywu-Canon_EF_70-200_mm_f_2.8L_IS_USM_Odblaski.html</a></p>

 

[Robin Smith]

<p>Generally I agree with with what is said above, but it is worth saying the Canon IS 70-200mm zooms in general have better performance on the test bench than the non- IS versions, so this cannot be said categorically. Reason would suggest that an IS lens has something else to go wrong, which could also be a deciding factor.</p>

[zml]

<p><em>> Frankly though, I think it's unlikely that you'll see any lens's upper limit in sharpness and IQ anytime soon.</em><br>

<em></em> <br>

<em></em>You have obviously never shot with EF 400/2.8 L IS (either generation...), 300/2.8 L IS (ditto), etc., or with their Nikon counterparts.</p>

[zml]

<p><em>> a further consideration is that IS is more useful for longer focal lengths than for shorter ones, and 17-50mm is not particularly long.</em></p>

<p>Right...that must the reasoning behind the new Canon EF 24/2.8 IS optics.</p>

[steve_feldman4]

<p>Sebastian, Your concern with image quality should expand to considering only cameras which allow you to do "AF Microadjustment" for each lens you own. Just about every lens will front or back focus because of manufacturing tolerances. If your camera does not allow you to compensate, your photos will never be as sharp as the lens is capable of delivering. Unfortunately, the the EOS 60D does not have the AF Microadjustment feature. The 50D and 7D do, just as 2 examples.</p>

[craig_meddaugh]

<p>Another example of people over-thinking things. Sure, all else being equal (and it never is) adding more elements makes a designers life much more difficult. However, we can't simply say that more elements makes for a worse lens as if this were true, we'd still all be shooting Cooke triplets. More elements typically make for a better lens so long as internal flare and transmission losses are well controlled and the design is well thought out.</p>

<p>That's the theory. For the rest of us, the real world will suffice and for all of the non-IS versus IS lenses I can think of, the non-IS provides a small advantage in terms of <strong>bench</strong> IQ. Even given this difference, none of the lenses are different enough that anyone will notice any real difference in prints smaller than a couple of feet on a side. So even the small advantage in bench performance is irrelevant for 99.9% of us. </p>

<p>As for field performance, unless you marry a tripod and have impeccable technique, sharpness losses due to motion blur will cancel out the exceedingly small differences in optical performance. This makes the choice pretty clear. </p>

 

[Sanford]

<p>IS may make you attempt to take photos you would ordinarily pass on. </p>

[g dan mitchell]

<p>There are various theories that predict that IS will result in lower resolution... but, meanwhile back in the real world, some of the IS lenses are among the very best available in terms of resolution. For example, compare the IS and non-IS f/4 70-200mm L zooms, etc.</p>

<p>Dan</p>

[Stock-Photos]

<p>Can we assume that when a lens has IS, more elements are added? Or, can they use an element that had to be there anyway?</p>

[Marcus Ian]

<p>In fairness, seven <em><strong>years</strong></em> elapsed between the introduction of the 70-200/4 L and the 70-200/4 L IS. I sure would <em>hope</em> they'd be able to improve upon the design in that kind of time frame.</p>

 

[g dan mitchell]

<p>Marcus, the fact remains that the 70-200mm f/4 IS lens is a tremendously good optic - among the very best zooms made. And if that isn't enough, by all reports the 70-200mm f/2.8 IS L II is even better. </p>

<p>Again, while people can think up theoretical reasons why an IS lens <em>should</em> or <em>could</em> be less sharp, it seems rather silly in the face of so much evidence that some of the very best lenses you can obtain have IS.</p>

<p>Dan</p>

[plangereis]

<p>I had a choice to buy either the Canon 70-200 f4 with or without IS. It was an easy decision. I feel you have a little more flexibility with IS, but then the lens I was considering was a much longer focal length. I guess it depends how you are going to use the lens (hand-held vs on a tripod). Good luck. Bye the way, the 70-200 f4 IS lens is amazingly sharp, and I have been able to hand hold it at fairly slow shutter speeds, and still get very sharp images (down to about 1/30 sec at 200mm). try doing that without IS! Good luck!</p>

[rishij]

<blockquote>

<p><em>> a further consideration is that IS is more useful for longer focal lengths than for shorter ones, and 17-50mm is not particularly long.</em><br>

Right...that must the reasoning behind the new Canon EF 24/2.8 IS optics.</p>

</blockquote>

<p>The only purpose of those new introduced wide-angle primes with IS, in my opinion, is for video. Where IS is of utmost convenience.</p>

[sarah_fox]

<p>70-200/4 L IS -- 20 elements in 15 groups<br>

70-200/4L non-IS -- 16 elements in 13 groups</p>

<p>Whether these additional elements and groups amounts to poorer image quality is anyone's guess. After all, if that were the case, then this rather crappy lens might outperform the above two lenses...</p>

<p>EF 75-300 III -- 13 elements in 9 groups</p>

<p>I've seen all sorts of discussions of this lens vs. that lens, and nobody ever discusses the number of elements. Rather, lenses are just weighed on the merits of how they perform in whole. The ONLY time anyone goes into a hysterical rant about extra elements and reflecting surfaces is when a filter is discussed.</p>

<p>In practice, I suspect the difference between the two lenses lies simply in the relative merit of the entire optical design. If you want a counter-example, the IS version of Canon's EF-S 18-55 is optically superior to their non-IS version.</p>

<p> </p>

[david_henderson]

<p>Number of elements used to be a discussion point, but it always seemed that the higher quality more expensive lenses always had more elements and the numbers were often quoted in marketing materials to support IQ expectations. Now it seems that some people at least think that more elements means less image quality. Maybe things have changed, or equally likely, maybe its just a totally theoretical straw being grasped to support a contention that is IMO unlikely to be meaningfully true. </p>