Camera- Vs. Lens-Based Stabilization
A friend is researching her first DSLR. Can anyone point me to information regarding the merits of camera-based vs. lens-based image stabilization here on dgrin?
I've found some resources through Google, but I'm interested in what some of our experts think.
The cost factor is obvious, but I can't help but feel there must be much more to it than that.
Any resources much appreciated.
p.s. I did a search and came up blank.
I've found some resources through Google, but I'm interested in what some of our experts think.
The cost factor is obvious, but I can't help but feel there must be much more to it than that.
Any resources much appreciated.
p.s. I did a search and came up blank.
0
Comments
The reason is that with longer focal length lenses the image jitter is exaggerated, which is pretty easy to see even in the viewfinder, and the imager based systems have to move much more than currently available to control the jitter in very long focal lengths. Optical systems have placement of an optical corrector close to the nodal point of the lens and simply require more angular motion with longer focal lengths, which means larger actuators but that's easier to include in larger, telephoto lenses.
A major advantage of optical IS is that the effect is visible in the viewfinder, which allows you to judge the ideal moment of action and stabilization for capture.
The major advantage of imager based systems is that it works (potentially) on any lens that works with the camera.
Moderator of the Cameras and Accessories forums
NEW Smugmug Site
http://www.jonathanswinton.com
http://www.swintoncounseling.com
Your research on in-body IS should focus on the brands that feature it: Olympus, Pentax (Samsung), and Sony (Minolta prior to Sony).
I have enjoyed the in-body IS that my Pentax K10D provides. I keep it turned on by default nowadays. After about 9000 shots, my experience tells me that it works quite well on both wide-angle and longer focal lengths. I use it regularly with a new-but-heavier SDM 50-135mm lens and an even heavier, classic 200mm f2.8 telephoto. In general I get a stop or two benefit, but more importantly I get increased psychological confidence that I can capture better shots in more challenging environments.
Also important, almost miraculous, is that in-body IS works with ancient lenses, some of which are available relatively cheaply. Right now I’m shooting with a 40+ year old screw-mount Mamiya/Sekor macro lenses that has color rendition and bokeh new lenses cannot touch.
M
http://www.popphoto.com/cameras/4615/image-stabilization-special-stop-the-shake.html
In body uses motors on the sensor to move the sensor to compensate for the shake. It's pro's are that it works with all the lenses for the small (if any) increased cost for the system.
The cons are that you cannot see the stabilization in the viewfinder, so the image in the viewfinder is still blurry. It's also not as effective as the lens based system. I think they are working on a panning mode for the body system, but many do not have that type of stabilization (I may be wrong on this though).
The lens based systems is found in Nikon and Canon bodies. It's pros are that it's designed for a specific lens and tends to reduce the shake more than the body system. You can also see the stabilized image in the viewfinder which is nice when you try to compose the image. Most if not all allow you to stabilize the lens as you pann and follow moving objects.
The cons are that it's much more expensive. For Canon lenses, it's anywhere about $70 for the IS in the kit lens to about $500 on a prograde lens. They tend to be heavier than non stabilized lenses.
I like how the lens system is set up, but for a persona who has to take costs into consideration, much prefer the body system.
Having said all this, I would not make my camera purchase decision on which stabilized system it employs. I'd look at image quality, lenses available, ease of finding accessories, and how the camera feels in your hand/ergonomics.
Also given the choice, I'd rather have a very fast lens to image stabilization as well.
M
The Olympus E-510 and E-3 have a panning mode.
http://www.jonathanswinton.com
http://www.swintoncounseling.com
That goes both ways though, if you have bought a 70-200 2.8 IS, it's not like you need to spend R&D-money/implementation-money on IS again for quite some time? You can just change the body.
Canon EOS 30D, Canon 50mm f/1.4, Sigma 70-200 f/2.8, Sigma 18-50 f/2.8, Tokina 12-24 f/4. Sigma 1.4 TC, Feisol 3401 Tripod + Feisol ballhead, Metz 58 AF-1 C, ebay triggers.
Not really; in-body IS is a generalist system so it's designed for each new camera and then produced in manufacturing runs that far exceed the numbers of each lens design. But as in-lens IS is a specialist system it needs to be optimized for every lens to get the best possible results and then is only produced in limited numbers, making it more R&D intensive and more expensive per unit. As others have already said, the tradeoff for the higher expense is that the results for lens-based IS, especially for long telephotos, is better. Today.
On the other hand, yes, you're right in that when you buy a 70-200 with IS, it probably won't be re-designed for a very long time. Meanwhile, we will both have upgraded our camera bodies a couple of times, with improvements to the stabilization technology (in mine) with each upgrade.
I was just looking at this PDF: http://www.usa.canon.com/app/pdf/lens/Lens_Extender_chart_new.pdf . Two things jump out at me: one, there are a LOT of lenses that would benefit from in-body stabilization. The other is that the only stabilized lens that's been discontinued and replaced is the 75-300 f4-5.6; the rest are as good as they ever were.
there are. at pop photo few months ago. it doesnt matter. as long as youre taking good pictures.
- Ansel Adams.
Not only would I agree with this statement, but would like to note that one things I often see missing in discussions on Image Shake reduction, be it in body or in lens, is the fact that nothing beats understanding and using good photographic techinque. With good techique the need for IS on short glass is virtually eliminated. On longer glass the need is greatly reduced.
Fast lenses with IS would be ideal in my view as you can hence freeze motion with speed, isolate subjects better with speed, and get brighter viewfinder with a fast lens. IS would also help when the lighting is dim. But this is can be very expensive in the lens based system. Pretty much free with in body system (which I personally would prefer, though I use a Canon).
I too see a lot of misunderstanding with IS. IS basically lets you handhold a few shutter speeds slower and reduce the chances of camera shake.
This is not a guarentee against camera shake though. As your shutter speeds get lower, there is a greater chance of camera shake, IS helps to reduce the percentage of shots with camera shake at a relatively lower shutter speed. IS however does not give a better image. Many equate IS with IQ. You put a crappy lens with IS, you get soft image b/c of optics. If you don't compose your image, you get a shake free image with bad composition or lighting. If your subject is moving and you use IS with low shutter speed, the image is blurry from subject movement.
IS in panning mode can smoothen the background so it can work in blurring the background of panned shots while keeping the panned subject more stable and hence sharp, assuming it has a panning mode and it's used. However, better panning technique will probably win over a non skilled shaky shooter using IS.
I don't think IS is a fad, it's uniformly available by all including Canon and Nikon (in their point and shoot), but suspect that Canon/Nikon will eventually introduce it for a lot less in their lens for dslr bodies (like the new kit lenses with IS/VR) or eventually introduce it in the body like Oly/Panasonic/Pentax etc if competition dictates it, as Olympus did with live view and sensor cleaning.
I don't think it's the biggest thing to hit digital phtography as noise reduction and improved color rendition, but do see it as a significant and useful feature and hope that Canon introduces it in their bodies.
However, I would place image quality (resolution/color rendition/ noise control/dynamic range), performance/robustness, ergonomics/ease of use, and the whole camera system before Lens vs body IS.
Having said all that, most casusal users usually get a kit lens and maybe a consumer grade telephoto lens and that's it. If this is the case, any of the current dslr will be more than fine. Canon and Nikon have already introduced IS/VR versions of their kit 18-55mm and 55-250mm zooms for consumers so the lens vs body IS system is more moot, other than the cost issue. For those folks, I think going by price, feel, and features would be fine. Honestly, I really don't know of any current dslr for entry level that is a stinker. I think all are good and the shooter would be the limiting factor.
Tee Why:
You seem to be responding to an earlier post that I deleted. Sorry about that. I appreciate your thoughtful and well-crafted replies.
I see what you're saying about technique vs. IS. I'm not sure I agree, but I'm still grappling with it. I don't understand things as quickly and clearly as I did when I was younger!
I DO think IS improves IQ. This debate seems to be one of semantics. Clearly IS doesn't affect composition, focus, lighting, or soft glass. I would think this would go without saying, and no reasonably knowledgeable consumer asserts that IS does anything but help to correct for motion blur. But by eliminating motion blur, it improves IQ in that regard. Image sharpness, as it relates to camera shake, is an aspect of image quality. Again, there's this issue of conflating what IS does and doesn't do (I can see that you're quite clear on this).
These "feature" debates often become polemic. It's clear that IS is not the be-all end-all. Obviously it improves some images under some conditions. A skilled shooter with an IS lens will produce better images than a non-skilled user with the same lens. Will an IS lens eliminate the need for skill and technique? It narrows the gap--as did autofocus, auto film loading, TTL flash metering, and any number of technical feature you want to name--going back as many decades as you'd care to count. But a skilled shooter is always going to produce better images. Again, I think this goes without saying. I guess there are a lot of uninformed consumers running around with lots of exaggerated expectations, but that's not who's contributing to this thread, or for the most part, this forum.
My original question was what the big difference between in-camera and in-lens IS. I can see now that the in-lens IS is probably superior, but that in-camera offers some advantages, primarily financial. It may be that Canon and Nikon will react to Olympus and Pentax, but if I were betting, I'd say not. I don't even think it's safe to assume that because Olympus got to market first with live-view that Canon followed as a reaction. I worked for both Nikon and Minolta for a number of years. How, when, and why these companies do what they do is a complicated matter not easily understood even by people who work for them. The smaller companies are often first--look at Samsung's APS-C sensor compact camera. Canon and Nikon will have one. Are they reacting to Samsung? Of course not. They're developing one according to their standards. And when it's introduced, it will likely be superior and more expensive.
Anyway, thanks again for the reply, and for the all the replies in this thread. I never coming here asking a question without leaving with a thorough education.
NEW Smugmug Site
Minor correction needed here.
Motion blur is not (directly) affected by the use of IS. IS is used to counter the effects of camera shake. Short exposure duration (high shutter speed) is the most effective counter for motion blur. Since camera shake is also reduced with short exposure durations, it's usually more important to make use of higher ISOs to engage the faster shutter speeds.
The "pan" mode of some IS systems does allow smoother pans which may allow smoother motion blur when desired, but the systems are actually turning off one axis of IS correction, so smooth holding technique remains important in "pan" mode.
Moderator of the Cameras and Accessories forums
Myself, I prefer better shooting technique. It's universal, and works with any brand gear in any situation with any lens.
http://www.chrislaudermilkphoto.com/
And camera shake leads to motion blur. What's the distinction? I still say this is semantics.
I give up.
NEW Smugmug Site
If I understood Ziggy correctly, "motion blur" refers to the blur that is caused by the subject moving + slow shutter speed for that subject. "Camera shake" is the blur that is caused by the photog.
Comments and constructive critique always welcome!
Elaine Heasley Photography
Tommy boy - it's not semantics at all. We just need to be a little more specific.
If the motion blur is caused by the camera movement, you're correct.
What I believe we're talking about here is motion blur caused by the subject moving. IS can't do anything about that. So if IS lets me take great hand held shots at 1/15th of a second, I will probably see more motion blur in my shots of people. People can move pretty far in 1/15th of a second. Especially active kids.
http://georgesphotos.net
I've always refered to the blurring caused by camera shake as "motion blur." I can see where that would be confusing and possible a misnomer. When I worked in the camera business, that's how my contemporaries refered to it. I guess "camera shake" is a more accurate term.
It's abundantly clear to me--and always has been--that IS doesn't have anything to do with subject movement. I don't even understand where that idea comes from, though apparently some consumers do suffer this misunderstanding.
I appreciate everyone's efforts to clarify. I sometimes find these protracted discussions frustrating (my problem, not yours) when so many voices are trying to be heard, often seemingly at cross purposes.
Thanks again.
NEW Smugmug Site
My viewpoint is slightly different. Most of the answers you've received are addressing the technical aspects. They may or may not be right, but I don't think it really makes any difference.
Here's how it breaks down for me.
Regardless of the facts, this is a big marketing issue for Canon and Nikon. So big, in fact, that both have introduced a kit lens with built in IS.
The in-body IS camp has done such a good job of selling the feature that they've indirectly sold a bunch of IS lenses for Canon and Nikon. Before the marketing of in-body IS, there wasn't much awareness or need for the feature except in longer focal lengths. Now, it's a must have feature for everyone.
The killer for the magazines and web sites is nobody really has a good way of measuring the effectiveness of IS. So, we get the "IS may improve xxx lens by 2-4 stops, your milage may vary" statements in comparisons and reviews. I'm not being critical here. It's the best they can do until they figure it out.
Here's my big question -
Why does this need to be an either/or feature? Why not put it in every body and also offer it as an option in long focal length lenses? If it's in the body, and I have a lens that does a better job of IS, I could just turn off the body-based IS and turn on the lens based IS. Best of both worlds.
Well the answer to that is pretty simple from the camera makers viewpoint.
If the camera maker puts it in the body, then very few photographers will pay the price differential to get IS in the lens to get the tiny differential in the few cases where lens based would be slightly better.
By offering both I've just given customers a way to directly compare the effectiveness of the two features and probably eroded the possible market for my very expensive and profitable IS lenses.
By not offering body-based IS I have the opportunity to upsell all of those casual users the IS version of the kit lens. Think of all of the non-hobbiest slr users you know. Most of these people never buy a second lens for their camera. The kit lens is all they will ever have. You've just opened this whole group of people to a profitable step-up at time of sale.
The in-body IS guys are selling the feature as a differentiator at the point of sale to gain a little market share. They have this feature the other guys don't have that will save you money in the future, or improve the lenses you already have.
This looks like a winner for everyone -
The in-body IS camera makers have a new feature to sell.
The lens-only IS makers have new products to sell.
The consumer wins in either case at the small expense of making the decision.
For the grand majority of folks that will only have one or two lenses, it just doesn't make any difference.
For the high end, cost-is-no-object people, they can rest assured that they have the best, most expensive solution.
There is the middle group that will have just a few lenses. Every lens will be a significant cost/performance decision. These folks are probably better off with in-camera IS.
Whew.... I'll step down from my soap box now.
http://georgesphotos.net
I agree that many sources describe camera shake as a type of motion blur.
I was taught that camera shake is defined as that type of blur or smear induced at and by the camera and lens. Everything is blurred or smeared to pretty much the same degree in camera shake, both the subject and the background.
Motion blur is a blur or smear caused by an object in motion and the camera held still or the result of tracking the subject through panning causing the subject to appear sharp while blurring or smearing the background. The distiction is that the effect is usually intentional and indicates a sense of relative motion or speed by the subject.
While it's possible to have both types of blurring in the same image, I think it valuable to distinguish between the two for the sake of discussion and prevention or induction of the effect.
PhotoShop has a "Motion Blur" effect which I believe is really a misnomer and disservice and may lend to the confusion. I believe it should be called a "tracking blur" as that would be more descriptive. (Adobe never asked my opinion however.
I'm open to alternatives as long as we can all agree on terminology that works to accurately describe the situation.
Moderator of the Cameras and Accessories forums
Thanks. I think "we" (you and I) know what we are discussing. I struggle with terms to keep things clear (as mud sometimes).
Moderator of the Cameras and Accessories forums
Canon and Nikon are a little unique in that they both produced IS lenses for the film market before the digital market matured. They both have a vested interest in the technology of lens based IS and typically they have used the technology for longer focal lengths first.
Minolta was the first dSLR manufacturer to recognize that an in-camera IS system could be used as a marketing advantage. It didn't take long for other manufacturers to realize the marketing advantage in camera based IS and how they had to apply some sort of technology or lose market share to those who had the technology.
Both Nikon and Canon have introduced a new type of IS control mechanism for their low-end and short focal length zooms which should offset some of the entry level market that was in jeopardy. I do expect to see more mid-level (prosumer if you will) lenses with IS as well.
I would not be surprised to see a teleconverter with built-in IS produced by someone in a short time.
Moderator of the Cameras and Accessories forums
When you are hand holding a camera pointed at a subject, the sensor is moving but (to first order) the lens image is sitting still. This is because the part of the camera shake which causes blur in the image (under most circumstances) is rotation about the optic center of the lens. Both kinds of IS use a gyroscope to mesure rotation of the camera and compensate for it. In body IS moves the sensor to track the image; in lens moves a specially designed optical element which causes the lens image to track the motion of the sensor. Here is the key thing to understand: the limit on IS performance (whether in body or in lens) is the distance over which the IS system can track the motion of the sensor.
In the case of in body IS, there are piezo elements which shift the sensor to track the image. The film plane, mirror box and imaging circle of the lens must all be made larger to accomodate the motion of the motion of the sensor. In the case of in lens IS, there is no impact on body design because the optic element adjusts the position of the image before it reaches the mirror box.
From a pure performance point of view, the primary advantage of in-lens IS is at longer focal lengths. At the focal length gets longer, the IS optical element moves farther from the film plane which means small adjustments of that element can adjust for quite large motions of the sensor. To get the the phenomenal IS performance of the Canon and Nikon supertelephotos with in body IS would require an oversized camera body to allow for very large sensor movements and the price paid for such a design (beyond a very expensive and bulky body) would be vignetting as the sensor moved outside the imaging circle of the lens.
Another thing worth noting is that in-body IS gets more difficult as the sensor gets larger both because the required sensor movements get larger and also because larger sensor are bigger. Today (to my knowlege) the cameras with in-body IS are all in the 4/3s system with its relatively small sensor. In body IS will be more difficult and more expensive to implement on the larger APS-C and 1.5 crop sensors; Canon and Nikon may yet release crop bodies with in camera IS, but expect both the cost and weight differential to be larger than it is for the 4/3s system. As for in body IS for full frame 35mm, my guess is that we will never see it; the mirror box for a full frame body is already jam packed and making room for sensor movement is likely not practical.
Finally, as for technique. Your goal, when using IS, is to limit the maximum rotation of the camera to keep it within the range if the IS system. When an IS system is rated for 3 stops, that is based (I presmume) on some kind of random walk model of camera motion. However, if you can limit the maxium rotation of your camera (by, say bracing against something fixed) you can get very long exposure times out of an IS system. Technique does not become irrelevant when you start using IS; rather (at least in my experience) the potential benefits from good technique actually become more dramatic.
M
Yep...my first DSLR was the Konica Minolta 5D with in-body IS.
Comments and constructive critique always welcome!
Elaine Heasley Photography