Dave Etchells at Imaging-Resources has posted a very interesting news item about the site's technical IS (image stabilization) testing, announcing both the I-R's first published IS test result (for the Canon 70–200mm ƒ/4 IS USM lens) and a white paper on its extensive testing methodologies (said to be for "ultra techies." I admit I haven't read all of it yet, being no sort of techie, much less an ultra one). Dave promises more IS tests results in weeks and months to come, calling his site's testing "the first...series of
comparative, quantitative IS tests performed using an open methodology
available anywhere on the 'net."
Here's a brief excerpt: "Having worked with roughly 25
different IS systems over the last two years of development, we've been
a little surprised at just how wide a range of performance we've seen,
and also at just how much difference the photographer makes
in IS system effectiveness: Relatively steady and shaky shooters will
often have rather different experiences with IS systems. For this
reason, we measure performance with two very different shooters, and
also indicate how they did relative to the standard '1 over the focal
length' rule of thumb for a 'safe' shooting speed. This will let our
readers determine where their experience with a given IS system might
fall."
We'll be looking to I-R and SLRgear.com for more.
(Thanks to Eolake)
Shutterbug have some comparisons as well. They claim that the Tamron 18-270 blows the socks off of the Canon 24-105. They also like the Pentax K20.
Posted by: Clayton Lofgren | Sunday, 05 April 2009 at 06:02 PM
I look forward to this. It might end the battle of whether lens- or body-based IS is better.
Nah, that won't happen. But I still look forward to the testing. If nothing else, it will make manufacturers think twice before claiming outrageous f-stops of improvement. It might also keep my Canonite friend from telling me that he gets sharp images at 300mm shooting at 1/10th s thanks to his lens-based IS.
Posted by: Miserere | Sunday, 05 April 2009 at 08:50 PM
@Miserere: It probably won't be enough to end the battle but the white paper does mention that lens-based stabilisation works a bit better than body-based stabilisation does.
Posted by: Thomas | Monday, 06 April 2009 at 08:43 AM
Miserere,
...Assuming it exists. Most of the lenses I'm interested in aren't available with IS.
Mike
Posted by: Mike Johnston | Monday, 06 April 2009 at 09:06 AM
I like Dave Etchells and have always been very impressed with his review site, which both goes beyond the numbers and has the statistical data to back up any conclusions.
But I wonder how scientific the IS testing really is if it relies on two *human* testers who will age, hurt their arm or hand, have good days and bad (e.g., coffee, lack of sleep), and so on.
Would it be that hard to invent a device that holds a camera and can wobble/shake at various specified speeds, frequencies, and degrees? (No "My first Velbon was like that" jokes, please.) Not that *I* could put together such a device, but it seems like it would be relatively easy for the mechanical-assembly kind of person.
Posted by: Robert Noble | Monday, 06 April 2009 at 09:32 AM
I've been having a hard time figuring out if my expectations / hopes are reasonable, and been generally finding the IS in my Nikkor 70-200/2.8 unimpressive. I think I like the IS in my Sigma 120-400 better. For that matter, I think I like the IS in my Panasonic DMC-LX3 better, too (but the focal lengths and holding situation are so different I don't know if that means anything). So other people trying to do systematic testing is all to the good.
On the one hand, human testers look likely to introduce significant variables. On the other hand, I'm not at all sure we can characterize human shake accurately enough to make a valid mechanical tester, either. A well-enough controlled protocol with rest times and such specified, and some retesting, will help figure things out.
Posted by: David Dyer-Bennet | Monday, 06 April 2009 at 10:02 AM
For the record, I have yet to encounter any sort of IS that is as effective (for me) as that in my original Konica-Minolta 7D.
Mike
Posted by: Mike Johnston | Monday, 06 April 2009 at 10:05 AM
"Would it be that hard to invent a device that holds a camera and can wobble/shake at various specified speeds, frequencies, and degrees? (No "My first Velbon was like that" jokes, please.) Not that *I* could put together such a device, but it seems like it would be relatively easy for the mechanical-assembly kind of person."
The Trick is the "various specified speeds and frequencies." Any simple system with springs or rubberbands will only vibrate at a few frequencies You could probably have three such systems: A low frequency, mid frequency and high frequency. There are random vibration testing systems, but they tend to be big things with motors and feedback systems designed to test military systems, so a bit of overkill for testing a single SLR.
Posted by: KeithB | Monday, 06 April 2009 at 11:09 AM
I’ve owned a 70–200mm ƒ/4 IS for a few months now. It's the only zoom I own and it's a very good lens. I use a 35mm F/1.4 for 90% of my shooting. Yeah it's a fast lens but as I often find myself at 1/15 sec at f/1.4 ISO 3200; I know that IS in the body could only be a good thing for me.
Posted by: sean | Monday, 06 April 2009 at 01:18 PM
The comments I've seen so far only confirm my continuing bias against relying on IS for sharp images in low light. There doesn't seem to be any way to get consistent and repeatable results with it. Mileage varies from brand to brand, lens to lens and photographer to photographer. IS may be better than shooting hand-held without IS, but when you need guaranteed sharpness there's still no substitute for a good (sorry for cussing, Mike) tripod.
Posted by: Gordon Lewis | Monday, 06 April 2009 at 01:53 PM
I've always thought the Olympus E-3 to be the best of the body based IS, after having owned pretty much every other one out there. (7D, A100, A700, K10D, and K20D) But then there's a friend of mine who says the K20D is the absolute most effective for her. Me, I'd put the Pentax solutions at the bottom of my list for effectiveness...
It is pretty neat just how much of an effect the person has on the effectiveness of the system.
Posted by: vincent | Monday, 06 April 2009 at 03:39 PM
Mike wrote:
Miserere,
...Assuming it exists. Most of the lenses I'm interested in aren't available with IS.
That's an important point, Mike. I like primes, so I'm pleased I shoot Pentax. But no matter how good/bad the IS, it's like sex: when it's good, it's great—but if it's bad, it's still better than nothing.
Posted by: Miserere | Monday, 06 April 2009 at 06:45 PM