Just like the Konica Minolta 7D and 5D the Alpha DSLR-A100 inherits a CCD based 'Anti Shake' system which Sony has relabeled as 'Super SteadyShot'. In contrast to optical based stabilization systems which depend on the lens having the stabilization system built-in the Minolta/Sony system instead stabilizes the CCD itself by mounting it on a platform which moves in reaction to movement of the camera body. The most significant advantage of this approach is of course that it instantly provides image stabilization to all lenses rather than having to pay a premium for lenses with built-in systems.
Photographers are taught the 'inverse of the focal length' rule which states that the slowest shutter speed you can use to avoid blur due to shaking is 1/(focal length) for an image printed at 8x10, for digital SLR's we have to use the effective focal length (so 1/80 sec would be good for a 50 mm lens on a 1.5x crop D-SLR). Most optical based anti-shake systems claim a 2 stops over this (1/20 sec at 50 mm lens, 1.5x crop), newer ones claim 3 stops. Sony claim 3.5 stops with their Super SteadyShot system (around 1/8 sec at 50 mm lens, 1.5x crop).
The next issue became testing this, we had a go at testing the Minolta system when we reviewed the Konica Minolta Maxxum 7D and despite the unscientific methods were fairly happy to report an approximate 2 stop advantage. As we still don't have a definitive test system for image stabilization systems I had to come up with another method which would at least give us an idea of how accurate Sony's claims were.
The stabilization test
The test was pretty simple (and hence I won't claim to be absolutely scientific), twenty hand-held shots were taken of a static scene, half of those without SteadyShot, half with, the shutter speed was decreased by a stop and repeated (from 1/80 sec to 1/10 sec). The lens used was the Minolta 50 mm F1.4 (producing a 75 mm equiv. FOV), the test chart was 1.2 m away from the camera.
The resulting 80 images were then inspected and given a blur score from zero to three where zero represented a very blurred image and three a sharp image with no noticeable blur (see crop examples below). Obviously the amount of blur which is acceptable will depend on your personal taste and the final image size (for instance a '2: Soft' will still look fine as a 4x6 print or in a web gallery). Example crops from these four blur scores can be seen below.
|0: Very blurred||1: Blurred|
|2: Soft||3: Sharp|
Hand-held, no stabilization (50 mm lens, 75 mm equiv.)
As you can see from the results here we managed 9 out of 10 sharp images at 1/80 sec but this dropped to just 4 out of 10 at 1/40 sec (although another 4 were acceptable if soft). In total we had 20 usable shots (13 of which were sharp) from the 40 taken.
Hand-held, with Super SteadyShot (50 mm lens, 75 mm equiv.)
Now we enable Super SteadyShot and you can immediately see an improvement, again 9 of our 10 shots at 1/80 sec are sharp, and so are 7 out of 10 at 1/40 sec and 6 out of 10 at 1/20 sec. In total we had 28 usable shots (22 of which were sharp) from the 40 taken.
The results basically state that using Super SteadyShot gave us approximately two more stops in which to work, although it wasn't a cast iron guarantee that at shutter speeds two stops slower every single shot will be sharp, it's clear that this system can only deal with certain types and amplitudes of shake (no current stabilization system can for correct for back/forth shake and only the Pentax SR system claims correction for rotational movement).