[Winder]

Maybe so, but there's much more to effective SR than gross movement distance.

The difference might simply be faster processors, better algorithms, and more sensitive MEMS sensors than were available in the A900 era.

I, for one, would lose interest in a Pentax FF that required lens-based stabilization. IBIS is one of the primary reasons I bought into the system in the first place.

I agree. My preference is for IBIS, but my preference does not change that it may or may not be possible.

It is true that there "might" be a way to do it, but right now, given the statement by Pentax, and looking at how much movement there is in the only existing FF IBIS on the market I am inclined be believe that Pentax views this as one of the real challenges they spoke about.

Pentax developed an OIS system and they have shown it with the 90mm HD 645 lens. There is no way Pentax is going to sell enough 645D lenses to recover the development cost of that system in the near future. If Pentax develops a FF K-mount I think it will use the newly developed OIS.

Pentax has not released ANY mid-long fast glass in recent years. Part of that I believe is due to not having an AF system that is accurate enough to handle a DA* 135mm F/1.8, but it may also be because they were developing OIS and they did not want release pro-grade glass without this feature. If the AF in the K-5II is as good as Pentax claims, then they now have 3 of the key technologies to start making competitive professional grade glass again. The 3rd being improved coatings. Now if they will release the key 4th technology being SDM-II.

HD D-FA* 135mm F/1.8 OIS SDM-II Limited
HD D-FA* 85mm F/1.4 OIS SDM-II Limited
HD D-FA* 50mm F/1.2 OIS SDM-II Limited

Those all sound pretty good to me.

---

Last edited by Winder; 09-27-2012 at 04:44 PM.

[TomTextura]

HD D-FA* 135mm F/1.8 OIS SDM-II Limited HD D-FA* 85mm F/1.4 OIS SDM-II Limited HD D-FA* 50mm F/1.2 OIS SDM-II Limited

Drool. I'll take one of each please. It would be so awesome if Pentax came out with these. Can't say I wouldn't rather have IBIS but OIS will be fine if it is what it takes to bring a Pentax FF to market. I know SDM gets knocked a lot, which I kind of get and kind of don't since I have had good experience with both my DA* f/2.8 zooms. I would, however, really like the focusing to be faster. My two zooms seem to lag pretty far behind any of my screwdrive lenses when it comes to AF speed. The thing I like most about SDM over screwdrive is it's silence, which is great for going somewhat unnoticed when shooting, a feature that wedding photographers and such surely appreciate. The shutter on the K-5 is so quite that coupled with an SDM lens no one can know if a picture was taken or not unless they are looking right into the lens and manage to see the shutter close for that split second.

[KHatfull]

Is there a single photographer using any system that would ever consider any APS-C body at $2k?!

Yes, I would. I shoot primarily macro and wildlife. APS-C is a free focal length and magnification upgrade for me, there's never too much. Give me a body with additional DR and even better high (and low) ISO performance and sure, it would be worth it. Give me a less noisy sensor with smaller photosites to bump up my resolving power, good to go.

FF works better for some things, not as well for others. To suggest that in all cases FF is superior and therefore intrinsically worth more than APS-C monetarily is incorrect IMHO.

One day it may come to pass that FF is the "standard" and APS-C is the "speciality". In fact I believe that to be the case.

[TomTextura]

To suggest that in all cases FF is superior and therefore intrinsically worth more than APS-C monetarily is incorrect IMHO.

Didn't mean to suggest that. My thinking was that an APS-C at that price would need to offer something additional, like the lightening fast FPS or incredible AF I mentioned, to merit the high price tag since, well I could be wrong but, it seems that any sensor that can be made for APS-C can be upscaled for FF and if that is the case, then the focal length multiplier benefit of APS-C is negated. I would actually really like to see continued advancements to top end APS-C models and have them receive all, or most, of the goodies that FF pro models get but at a reduced price. Personally, I mostly like APS-C for the lower price not for the smaller sensor and all the image characteristics that come with that. That's because my shooting style often leans toward the wide side in both terms of focal length and aperture. I do appreciate lighter and less bulky camera and lens combos though. There's pros and cons to every tool. That said, I do think that price should generally match performance.

---

Last edited by TomTextura; 09-28-2012 at 07:09 PM.

[Winder]

What would it take for an APS-C camera to sell for $2,000?
1. Tank like construction.
2. Canon 1DX AF.
3. VF from the Sony A900.
4. 10 fps. with the D4's buffer.
5. 36MP APS-C sensor (no AA) with the performance of the D600.
6. 1/250 Flash Sync.

If Pentax had the glass that would sell to professional sports and wildlife shooters all day long for $2,000.

[Fogel70]

I agree. My preference is for IBIS, but my preference does not change that it may or may not be possible.

It is true that there "might" be a way to do it, but right now, given the statement by Pentax, and looking at how much movement there is in the only existing FF IBIS on the market I am inclined be believe that Pentax views this as one of the real challenges they spoke about.

Pentax developed an OIS system and they have shown it with the 90mm HD 645 lens. There is no way Pentax is going to sell enough 645D lenses to recover the development cost of that system in the near future. If Pentax develops a FF K-mount I think it will use the newly developed OIS.

Pentax has not released ANY mid-long fast glass in recent years. Part of that I believe is due to not having an AF system that is accurate enough to handle a DA* 135mm F/1.8, but it may also be because they were developing OIS and they did not want release pro-grade glass without this feature. If the AF in the K-5II is as good as Pentax claims, then they now have 3 of the key technologies to start making competitive professional grade glass again. The 3rd being improved coatings. Now if they will release the key 4th technology being SDM-II.

HD D-FA* 135mm F/1.8 OIS SDM-II Limited
HD D-FA* 85mm F/1.4 OIS SDM-II Limited
HD D-FA* 50mm F/1.2 OIS SDM-II Limited

Those all sound pretty good to me.

On 135/1.8 it might be possible to fit OS, but I doubt that there is enough space in the optics of 50/1.2 & 85/1.4 to fit those extra leses needed for OS. At least if these leses are designed for DSLR.

If redesigning the optics to fit OS they will probably end up much slower. And I really doubt that 50/2.8 & 85/2.8 with OS is preferable over 50/1.2 & 85/1.4 without OS.

[vonBaloney]

How about a whole new approach that doesn't need the extra space? Here is a method that records the sensor movement and removes the motion blur afterward:

Image Deblurring using Inertial Measurement Sensors

[Fogel70]

How about a whole new approach that doesn't need the extra space? Here is a method that records the sensor movement and removes the motion blur afterward:

Image Deblurring using Inertial Measurement Sensors

I doubt "software" stabilization can replace "hardware" solutions, but this type of stabilization might have a use in low end cameras such as mobile phones.
I don't think it's possible to recover detail lost in the blur, so for high end cameras it will not be good enough.

[vonBaloney]

I doubt "software" stabilization can replace "hardware" solutions, but this type of stabilization might have a use in low end cameras such as mobile phones.
I don't think it's possible to recover detail lost in the blur, so for high end cameras it will not be good enough.

It is hardware and software -- it doesn't just try to remove the blur based on the blur in the image -- it would have a sensor that is recording the vibrations during the shot and then reverse-blur it "knowing" exactly the direction and magnitude of the blur. The proof-of-concept stuff looks pretty good. Better than nothing anyway.

Another advancement, which I expect to see sometime (but could be a decade or two away) is for sensors/processors to be able to incrementally capture at super-fast rates -- so for instance a 1/1000 sec shutter speed might be made up of data that could actually be individually separated into slices that are only 1/100000 sec. So then we could have captures that are more like negative film -- you could never blow out the highlights (unless you wanted) because that excess data would just be thrown away while all the information from the shadows (which takes longer to build up) would be preserved. A huge dynamic range would be possible, but also all those 1/100000 slices could be post-aligned to eliminate vibrations. So like the current system but the sensor would be in-effect adjusted after the shot rather than during it (or both). And by that point still and video will have totally merged as you'll be able to pull high quality stills from any video stream.

[starbase218]

While reading all of the interviews over Photokina, one item stuck out. One of the Pentax execs said that the DA lenses were not FF just larger to accommodate the movement the sensor's IBIS. This means that if Pentax does come out with a 24 Megapixel FF camera, it can not have IBIS as the sensor would go out of the image circle of even the best FA lens out there.
This is the conundrum for Pentax and why we are not seeing any hint of FF coming from their lips. They went the IBIS route with APC because it could be done with a slight increase in lens size while all the other FF and APC manufacturer's went with in lens stabilization. IBIS can not be done on FF so what is Pentax to do. Release a non-stabilized body to go with all the non-stabilized lenses. This has to be the crux of the problem and why of all the manufactures using Sony sensors, Pentax is the lone camera company with no FF.

Just because you thought of it, it has to be the reason?

Also, the Pentax exec in question is John Carlson, and the thing he suggested was that none of the Pentax DA lenses cover 36x24mm, which is not true (as I think has been proven). It may still be that these lenses don't cover FF + SR, but he did not say that. And the thing he did say was not true, as far as I know.

Anyway, I don't see the problem with this. The thing is, Pentax never advertised DA or DA* lenses as being APS-C compatible. If, a year from now, you buy a Pentax FF body, walk into a store and say "I want a lens for this body", the sales person won't be able to help you. That's the problem. (not counting the few existing FA/DFA lenses, which John Carlson never spoke about at all)

Besides, I never drive to an IBIS hotel in an Armoured Personnel Carrier.

[RobA_Oz]

One thing that would help the discussion on the OP's original assertion are the actual sizes of the pre-digital lens image circles. The question of how much advantage SR gives comes down to how many stops of shake reduction can be attained with a given system, and how much sensor movement that requires.

If you look at Falk Lumo's paper on the subject (two years old, now, and largely base on the system in the K-7) you'll see that we are really talking about small movements with IBIS. The movements discussed in that paper (of the order of 0.01mm) are less than the size reduction of the new D600 sensor. Even if the sensor movements were up to 0.5mm in any direction, how much of a problem would that be? The tolerance on image circle size would have to be extremely tight for significant corner loss to occur, and even if it was zero, lopping 1mm off the 24mm frame height represents around a 4% linear reduction in image size. Not so long ago, we used to be content with a 95% viewfinder. You could even use software to crop the image in-camera, if you don't like clipped corners. Even 92% is a lot bigger than 43% (the APS-C area crop) and the real figure is possibly greater than that, depending on what the lens image circle tolerance range actually is, and how much SR you want to use.

Of course, things have moved on in the couple of years since Falk wrote that paper, so any further data would be welcome. From what we have, though, I doubt IBIS is really the roadblock that's been suggested. Existing lenses (even including film legacy lenses) may be a limitation, but they shouldn't bring development of a 35mm sensor camera body to a complete halt.

[Laurentiu Cristofor]

SR is nice to have but not a deal breaker. It is more important to come with an appealing body and a good set of lenses. If they are trying to figure all the issues now, they'll never issue a product.

[Winder]

If you look at Falk Lumo's paper on the subject (two years old, now, and largely base on the system in the K-7) you'll see that we are really talking about small movements with IBIS. The movements discussed in that paper (of the order of 0.01mm) are less than the size reduction of the new D600 sensor. Even if the sensor movements were up to 0.5mm in any direction, how much of a problem would that be? The tolerance on image circle size would have to be extremely tight for significant corner loss to occur, and even if it was zero, lopping 1mm off the 24mm frame height represents around a 4% linear reduction in image size. Not so long ago, we used to be content with a 95% viewfinder. You could even use software to crop the image in-camera, if you don't like clipped corners. Even 92% is a lot bigger than 43% (the APS-C area crop) and the real figure is possibly greater than that, depending on what the lens image circle tolerance range actually is, and how much SR you want to use.

Look at how much the Sony A900 SR moves the sensor. Does this look like 0.5mm?

Youtu.be

Youtu.be

[RobA_Oz]

Look at how much the Sony A900 SR moves the sensor. Does this look like 0.5mm?
Sony Alpha A900: Super SteadyShot on FF - YouTube

Some actual measurements would be good to know, not to mention what part of the cycle is involved with the image circle, because there has to be movement associated with acceleration and deceleration. Thanks for digging that one out. I'd only looked for information on the Pentax system, and Falk is one of the few people here concerned with objective measurement, rather than vague generalities.

[RobA_Oz]

And a bit more technical reading on the K-7 SR system: Study of the Effectiveness of Shake Reduction in the Pentax K7

Like Falk's article, this one only deals with sensor movements of the order of 5-10 pixels. Again, some hard information on Pentax's lens image circle sizes would be helpful.