Sorry to say that careful testing has revealed a fatal flaw in my (new) sample of the Sony A900: At normal shutter speeds in normal light, pictures are routinely less sharp with SteadyShot turned on than off.
This is similar to an effect Carl Weese uncovered with an early Pentax (K20? Memory does not serve*), and it was not true of the earlier A900 I used, so it is not an inherent flaw of the camera—it's a sample variation.
I will be attempting to call B&H Photo's Customer Service today to see about returning this unit for exchange. If I have missed that window, then I'm thrown to the vagaries of Sony's warranty repair.
Given the cost of the camera, I'm feeling a little freaked out about this.
Mike
*I had the identical Pentax camera body at that time and could not duplicate Carl's results—my camera was fine. Again, I believe that was a case of sample variation.
UPDATE: B&H's Customer Service was great—they were very helpful and have issued a mailing label for the return, and will be exchanging the camera for another one. Further posts about the A900 will be slightly delayed, however.
In the meantime, a Panasonic G3 is en route to TOP World Headquarters, so we might have some postings about that camera coming up.
Send this post to a friend
Please help support TOP by patronizing our sponsors B&H Photo and Amazon
Note: Links in this post may be to our affiliates; sales through affiliate links may benefit this site. More...
Original contents copyright 2011 by Michael C. Johnston and/or the bylined author. All Rights Reserved.
Featured Comment by Brad Nichol: "Mike, sorry to hear you got a dud. I would like to add to perhaps clear things up for those who seem to have some doubts about the sensor stabilization on the Sony...yes, it absolutely works as intended.
"I have carried out several exhaustive tests on the A900, with all sorts of lenses, shutter speeds, with stabilization on and off and the camera both on and off the tripod. Despite Sony's warnings, and accepted wisdom, it even works well when the camera is tripod mounted; it does not seem to degrade the image at all.
"Once you have used the Sony with stabilization turned on you will find it very hard to justify not having it on a camera. It is very very rare to get blurry shots due to movement.
"So to all those who doubt Mike's technical skill, if there is in issue it is real and nothing to do with settings, lenses, method, and it is definitely not typical of how the A900 performs."
Mike replies: Indeed. I did all those tests the first time I used the A900 back in 2008, because I needed to be sure of what I was saying. I can state categorically that, working properly, the A900's SteadyShot doesn't degrade resolution. Really! After I ran my own tests back then, I was sufficiently impressed that I simply left SteadyShot on all the time.
I also did the tests by the book with my original Konica-Minolta 7D, the camera that originally sold me on this feature. Again, it acquitted itself with flying colors—the sample I owned, anyway. The 7D has since gone considerably wonky, such that I hardly ever use it any more, but its Anti-Shake still works fine.
Rots of Ruck with Sony repair service in the US.
Posted by: Bill Mitchell | Wednesday, 21 September 2011 at 11:32 AM
Sorry to hear about that Mike, that's not normal as you point out. Hope you get it exchanged soon.
Posted by: David Anderson | Wednesday, 21 September 2011 at 11:58 AM
Well, you could try a different lens and see if the SteadyShot on that one works any better.
Oh, yeah. Sony puts it in the body, not the lens. Sorry.
Posted by: Speed | Wednesday, 21 September 2011 at 11:59 AM
Many of us (we?) Nikon shooters have long since discovered that using VR (Nikon's version of stabilization that is on-lens) can inhibit perfect sharpness even when speeds are adequate.
Maybe you were just lucky with the prior A900 and shot at a speed/lens/SteadyShot combo that did not affect sharpness??
I've pretty much accepted that there will never quite be any substitute for a tripod - even though I hate using the things.
Posted by: Andrea B. | Wednesday, 21 September 2011 at 12:00 PM
Did you pay attention to your shutter speeds? Last year Thom Hogan explained Nikon's VR and advised against using it above 1/500s because it can reduce the sharpness sometimes.
Posted by: Christian Beck | Wednesday, 21 September 2011 at 12:17 PM
If it's anything like VR (half shutter press activated), it needs a bit of time to spool up - so if the shutter gets pressed too soon then exposure starts while the stabilisation is getting into position.
Posted by: Kwasi | Wednesday, 21 September 2011 at 12:23 PM
Seems I read not to turn on VR with a tripod or bean bag, etc. as the system will attempt to correct vibration that is non-existent and in fact create slight blur. Don't know those details of your test.
Posted by: David Zivic | Wednesday, 21 September 2011 at 01:07 PM
C'mon, give me a little credit here, guys. I'm satisfied I tested the camera competently and that the problem I'm describing actually is a problem.
Mike
Posted by: Mike Johnston | Wednesday, 21 September 2011 at 01:08 PM
Are you using a tripod along with SteadyShot? I have heard some talk about getting fuzzy images using this method.
Posted by: Ed Steinerts | Wednesday, 21 September 2011 at 01:13 PM
I for one completely trust your diagnostic skills.
And plus, I'm looking forward to reading about your experience with the G3. I have a G2 and love everything about it except the quality of the color and oh yeah, the highlight clipping. Minus 2/3 stop whenever the sun shines.
Posted by: Paris | Wednesday, 21 September 2011 at 02:09 PM
I have to admit that while I've made great use of KM/Sony's IS over the last 6 years, it bugs me that my camera has a sensor that moves. My KM 7D exhibits a little discoloration of high ISO images after I've been shooting with IS for a while (presumably sensor heating) and I know I occasionally see odd image artifacts that don't look like camera shake or lens softness or OOF, but I suspect are due to a sensor reacting to the fact that I just picked it up, and not realizing I put it to my eye and am now holding it steady.
So I prefer to turn off IS when i don't need it. That's why I don't like that Sony is now hiding SSS on/off in the menu on newer models, instead of on an easily accessible switch.
Ultimately, it's one of those "simplicity" things. I don't like the potential problems associated with the complexity of the system, nor the occasional user errors I cause; the luddite in me would prefer a solidly locked down sensor that doesn't budge ! But six years of sharp images, from tack sharp in good light to much-better-than-without-SSS at slow shutter speeds in low light, it works.
Posted by: Dennis | Wednesday, 21 September 2011 at 02:15 PM
Plausible explanation: electronics are not identifying the lens focal length correctly. Using older lenses, even with M42 ones, the DR in my K5 works great, but if I change a lens and forget to dial in the new focal length I get blur. Even if its switching a 50 for a 35, you get minimal blur but the difference is clear with a shot were the correct focal length is used.
Posted by: Max | Wednesday, 21 September 2011 at 02:34 PM
Nikon has had at least three versions of VR in various lenses. The official advice about using VR on a tripod varies by version. So, be careful of broad advice on this topic off the Internet! You need advice specific to a particular lens.
Mike, best of luck getting a working camera! The complexity inherent in stuff like IS/VR/SteadyShot is mind-boggling, when I think about it I'm surprised it ever works at all.
Posted by: David Dyer-Bennet | Wednesday, 21 September 2011 at 03:14 PM
This is a rather complex engineering problem. The ever-vigilant Falk Lumo has documented quirks in the Pentax K7's IS algorithm (which seem now to be corrected).
Posted by: Semilog | Wednesday, 21 September 2011 at 03:30 PM
Frankly I can't believe that people think that a moving sensor (regardless of design), is going to improve image sharpness. Sounds more like marketing hype than mechanical engineering.
Manufacturers such as Leica, Hasselbled, and Phase One, to name just a few, go to great lengths to very accurately align a sensor for optimum performance.
You may notice that none of the above mentioned manufacturers have any form of image stabilisation based on a moving sensor.
The laws of physics just don't change to suit marketing departments.
Posted by: Seascape | Wednesday, 21 September 2011 at 03:36 PM
I have the same issue with a Pentax K-x & an Olympus E-P1, both with in-body IS. Roundabout 1/100sec the image stabilization system can produce less sharp pictures. I only turn IS on when i need, otherwise it's turned OFF.
Posted by: XebastYan | Wednesday, 21 September 2011 at 04:31 PM
Not directly germane to your post, but it suddenly occurs to me that I never saw anything you wrote on your experiences using the Pentax K5. Did I miss your comments, and if so, could you provide a link?
Posted by: Ted | Wednesday, 21 September 2011 at 07:32 PM
Mike, anyone who's done tests will tell you the same thing: stabilization will soften your images. Period. If you don't need it, turn it off. If you do need it, turn it on.
Posted by: the other James | Wednesday, 21 September 2011 at 08:39 PM
"stabilization will soften your images. Period."
Other James,
A common enough misconception, but not so.
Mike
Posted by: Mike Johnston | Wednesday, 21 September 2011 at 08:46 PM
I found that if I pressed the shutter release quickly without pausing to focus on my Pentax K10 , it seemed like the camera would move the sensor just as I was making the exposure. I hated pausing with the release half pressed so I sold the camera.
On the other hand I routinely shoot at 1/20 with a Nikkor 105mm lens on my Sony NEX and usually it's pretty sharp.
Posted by: hugh crawford | Wednesday, 21 September 2011 at 10:04 PM
I once had an Olympus E-P1 that exhibited the exact same phenomenon. Unfortunately turning IS on and off meant a foray into its menus every time.
Posted by: Carsten Bockermann | Thursday, 22 September 2011 at 12:20 AM
Dear Seascape,
Your understanding of the laws of physics is insufficient. This is an engineering problem. Image stabilization is a well-understood and well-established technique. It does not work under all circumstances, all the time. That is also well-known. But it definitely works well, a lot of the time.
~~~~~
Dear other James,
I've done tests. I quite disagree with this anonymous "anyone" you refer to.
pax / Ctein
Posted by: ctein | Thursday, 22 September 2011 at 01:01 AM
Yep, if you have a steady hand IS works exactly like a tripod roughly over 1/10th second (as it is intended), from what I've got with Canon point and shoots and Pentax slr's. The fact that it sounds impossible doesn't make it less effective.
If it's on and mated with a suitable lens it's a camera issue.
Posted by: Max | Thursday, 22 September 2011 at 06:59 AM
Mike, I have an Olympus E5 and the only way to get perfect sharpness, specially with the SHG digital zuiko lenses is: using a tripod, focusing manually, rising the mirror and turning the sensor stabilizer off. If I leave the stabilizer on, with the same setting, I don't get perfect sharpness. The same happens with my previous Oly E3 and with the EP2. I realized that the IS is good only when yo are shutting between 1/8 to 1/60 of a second to get just decent sharpness.
Regards.
Posted by: Marcelo Guarini | Thursday, 22 September 2011 at 08:19 AM
Dear Ctein,
My point about moving sensors was not about the fact that under certain circumstances they may improve an image capture.
My point was simply that a sensor that incorporates movement, would be extremely difficult to optimally align in a moderately priced consumer product, not to mention the reliability of such a design.
Regards
DK
Posted by: Seascape | Thursday, 22 September 2011 at 11:46 AM
This is just one of those topics where people have vastly different experiences, even with the same camera. I have an E5 and shoot landscape with it and achieve perfect sharpness using IS in most conditions. I turn it off for long exposure tripod work, and off for flying birds, running dogs, etc. I wish that DPR would more thoroughly test in-camera IS in their reviews of higher end cameras, since people always are curious.
Posted by: John Krumm | Thursday, 22 September 2011 at 01:08 PM
I've tested the IS somewhat rigorously myself on Canon, I've fiddled around with it on Nikon, and I've seen a number of tests (of varying quality) online that all show image stabilization having an effect on sharpness when it's not needed--particularly when you're on a tripod. It's not always a big difference, but it is there.
Personally, I leave IS on all the time if I'm handholding, even when I'm using flash to freeze motion--if only to keep the viewfinder image stead.
I am not just a random poster repeating lore I've heard about IS being a bogeyman. I'm one of those "anonymous anyones" that Ctein mentioned...and there seem to quite a few others posting in the comments here as well.
Posted by: the other James | Thursday, 22 September 2011 at 09:13 PM
Dear Seascape,
I understood your point. I do not agree, nor do my tests.
pax / Ctein
Posted by: ctein | Thursday, 22 September 2011 at 09:32 PM
Congratulation for the G3 which is heading for the TOP headquarters. Don't forget to order a spare battery in time. I have been using the G3 for two months now, but I am still waiting for a spare battery to be delivered. There is a shortage, at least in Europe. The same applies to lenses. Very annoying.
Posted by: CMS | Friday, 23 September 2011 at 05:06 AM
The way how my A900 (bought new in Spring 2009) renders hand-held shots taken at 1/250 s in a hurry with a Minolta AF Apo HS 400 mm G lens + TC 1.4× Apo II (= 560 mm) sharp to the single pixel level keeps impressing me.
Posted by: 01af | Friday, 23 September 2011 at 04:54 PM