[interested_observer]

I made the "lens-based system is better for longer focal lengths" comment, essentially based on the moment arm effect. For a unit of movement with the sensor, there is a much larger movement in longer focal lengths when compared to the shorter focal lengths. That said, since I like wider angle lenses, I still see and perceive value in using SR or stabilization in the shorter lengths. Logically you need to move the sensor more to compensate for a movement change in a short focal length - however, with the shorter focal lengths, you should not be seeing a large set of movement change in the scene, in that it is closer. So it all works out. When I was designing a star tracking system for a telescope, and Pentax's floating sensor is similar but not as complex, the lever arm or moment arm effect really comes in to play - especially when you have a minimum focusing distance of the Moon.

Lowell - I was typing before my morning coffee - which is always dangerous. I intended to just say stabilization systems - not lens-based. My fault for not reading the reply closely enough before hitting [send]. I agree with you, and that is why I bought into Pentax in the first place. By the way - excellent image! Tomorrow morning it will be open mouth, insert coffee cup.....

---

Last edited by interested_observer; 12-14-2011 at 07:58 PM.

[sjwaldron]

In-camera SR has a large number of benefits over lens based besides just stabilization. The K-5 is really solid with SR, but can do things like shift the sensor in steps for use in framing on a tripod as well as things like astro tracking with the GPS accessory. Not to mention in video we get SR with any lens, which means old manual focus lenses that excel at being used for video are even better. Both methods probably have their positives, but I think in-camera has the leg up for the sheer level of things that can be done with it.

[VisualDarkness]

I still don't get why the lens based on offers "better technology" for stabilizing longer lenses. Isn't it just an assumption people do without facts? I mean, the technologies should be able to counter the same amount of motion (maybe even faster with the low mass of a sensor) and it's just up to the camera and lens makers to utilize it properly.

I could see a difference in that the cost of in-body must be somewhat cut to keep the camera price down, but lens-based offers a possibility for the companies to put great technology in premium lenses where money isn't as big of a limiting factor.

[twitch]

If ever Canon or Nikon released a DSLR with IBIS then it would become the greatest thing since sliced bread. Until then it's either inferior (if they have in lens stabilisation for that focal length) or unnecessary (if they don't have a lens of that focal length with stabilisation). They can argue any of these things with great conviction, whatever argument suits their cause at that time, and "prove" they are right.

[VisualDarkness]

I don't understand people do this, as it renders the survey useless:
"To give lens based IS a fair chance, the questions assumed that it was slightly more efficient then body based IS, though that may not always be true."
Image Stabilization Survey Results - Body or Lens based IS?

[Raffwal]

The independent test I mentioned was in the German Foto-Magazin in Feb 2010, I believe. I don't have the magazine, but read it at the time. One can question their methodology, but the conclusion was more or less that the lens-based stabilization was somewhat more efficient than IBIS. And also that the IBIS by Sony was a bit more efficient than that of Pentax.

[Rondec]

Falconeye has made the comment that sensor based image stabilization has the potential to be better than lens based stabilization. I haven't seen much improvement between the K7 and K5, although each generation of camera before that seemed to introduce some benefit.

I think the benefits are simply (1) smaller lenses (2) stabilized primes. Very few primes in Canon/Nikon's portfolio are stabilized. The ones that are are quite pricey. (3)Decreased cost. Well, or possibly just going without stabilization on a number of lenses. Just looking at Sigma's 70-200 f2.8 -- with OS at B and H it is 1399, without OS it is 799. (4)One less thing to break in your lens. I know several shooters who have had an IS/VR lens go haywire and require servicing. On the other hand, reports of in body SR failures seem to be few and far between. This is particularly so since you update your body periodically, while you are likely to keep your lenses for longer. (5)Stabilization gains whenever you purchase a new body.

I think Canon/Nikon went with in lens stabilization because it was the only option when you were shooting film. They continue to use it, because (a) they have promoted it as better than in body stabilization for long enough that their shooters actually believe it and (b) they get a premium price for every IS/VR lens sold.

I think Olympus and Sony actually do in body stabilization better than Pentax and I hope that over time, Pentax works to improve it, because it certainly has a lot of potential to make shaky hands steadier.

[Lowell Goudge]

the question is, I think, not which is better, as I don't think a user can tell, but how much do you expect with each new body.

Certainly if I can go out and shoot with a 500mm lens hand held, at less than 1/10 the "rule of thumb" for shooting speeds hand held, and do it repeatidly, then there is not a great need for improvement. I could not do this with my K10D, With my K7 although I have pulled similar shots, it took some effort, the shot I posted here with my K5 I was not even thinking about it. In fact, it was not until I looked back at the image that I was aware of the actual shutter speed.

While my shooting technique and perhaps years of experience shooting hand held without the benifit of shake reduction certainly help, anyone can go out and get great shots using shake reduction to their benefit.

To make the system "better" in my perspective would require more data than can be derrived from the simple input of focal length. This is where an in lens has the advantage, because it can do more to fine tune the system to the mass of the lens, physical size, nodal point etc..... Unfortunately, the use would be to compensate to the maximum of people with bad technique, like trying to shoot with a 500mm lens and use live view AKA point and shoot style. I doubt it would do a whole lot for me, My shots are now starting to be limited by subject movement.

What I would like, in terms of SR is a means to customize the function. Allow more or less rotational gain, or more or less vertical and lateral gain. just to allow us to make it suitable for our individual shooting styles.

[Stone G.]

Canon IS and Nikon VR "are arguably better than Pentax in-body SR for the longer focal lengths (> 300mm)"? That just may be so and at least I cannot prove otherwise.

What I can and do say is that being a user of long focal length lenses since 1982 (350mm, 300mm and 200mm Tamron Adaptalls plus 1.5X and even 2X teleconverters) it is unquestionable that Pentax SR works very well with these (old and manual) long lenses also. For me the effect is significant. And I still use these lenses and TCs a lot.

So, in-body SR is not only the thing I have, it is also the only thing I really want!

And then there is the "Astrotracer Feature" of the O-GPS1. Where would that be without Pentax SR? (Sigh, wish I could upgrade to K-5 right here and now.....)