Image Stabilisers: Optical or Mechanical?

Canon, Nikon and Panasonic all stabilise their lenses. On Pentax, Sony and Olympus cameras it's the bodies that are stabilised. Samsung has used both systems, but in two incompatible camera ranges. In other words, none of these cameras can be used to compare optical and mechanical stabilisers because the results could be corrupted by any number of variables, from the grip and handling to the weight of the lens. Now things have changed because Sigma recently decided to use the same optical stabilisation system across its latest lenses, body be damned. So we decided to jump on the occasion to analyse the effects of optical and mechanical stabilisation on a single camera with a single lens; the only thing that changed was which stabiliser we turned on.

No Stabiliser

If you want to measure how effective an image stabiliser is, there's really only one way to do it: turn it on, take a series of photos at progressively slower shutter speeds until the shots come out blurry, and then compare the results from each series to those obtained using no stabiliser.


For our experiment we used a Pentax K-5, which has a mechanical sensor-shift stabilisation system—unlike Minolta/Sony's system, Pentax's also corrects rotations—and a Sigma 50-500mm OS HSM lens, which Sigma launched with a stabiliser in 2010. We took sets of ten photos and divided the resulting images into three categories: sharp (good detail even at 100%), passable (adequate for 20 x 30 cm prints) and blurry.


There's a rule of thumb that says that when taking photos freehand with no stabiliser, one should use a shutter speed that is greater than the inverse of the equivalent focal length. That means that with 500mm (750mm equivalent) you need to use a shutter speed of 1/750 s to be guaranteed a sharp image.

We took our photos with feet planted and arms held stiff, and got a good success rate up until 1/250 s. Up until that point all of the 20 x 30 cm pictures came out acceptable and eight out of ten were fine even after serious cropping and could be made into good extra-large prints. But it doesn't take long for the quality to drop. Already at 1/250 s, nothing but good karma will give you a sharp image on the first try.

Mechanical Stabiliser

Using only the mechanical stabiliser from the K-5 body, here's what we got:


This is a clear improvement over the no-stabiliser results, but it isn't earth-shattering either. Basically, it limits your chances of getting entirely blurry throwaway shots, seeing as they only appear at 1/60 s, but anyone who crops a lot may be disappointed to see that at 1/250 s and 1/125 s only half of the pictures turned out "sharp".

For 20 x 30 cm prints, with mechanical stabilisation you essentially gain two shutter speeds.

Optical Stabiliser

Using only the optical stabiliser from the Sigma 50-500mm lens, we got this:


The difference here is much greater than with the mechanical stabilisation. We got an over 70% success rate all the way up until 1/60 s. And for people who make regular-sized prints without doing any major cropping, they can even go up to 1/60 s without any worries. That said, starting at 1/30 s the number of passable photos isn't all that better than with the mechanical stabiliser; there's just less of a transition between the good shots and the bad shots.

Both Stabilisers Simultaneously

Camera brands advise against using more than one stabiliser at a time, arguing that multiple systems could interfere with each other. We decided to put it to the test and turned both stabilisers on.


The success rate becomes critical at just 1/250 s. Judging by these results, two systems is indeed less effective than one. In fact, it's hardly better than using no stabiliser at all.

Conclusion

Of course, this wasn't an exhaustive test. For a definitive answer one would need to try several different combinations of lenses and bodies with differing focal lengths and users. That said, it's worth noting that our results were similar to those we obtained a couple years ago when the press was comparing Canon and Nikon's optical stabilisers to Minolta/Sony and Pentax's mechanical stabilisers.

Based on our measurements, optical stabilisation wins. It's roughly one shutter speed more effective than mechanical. Both systems are a considerable improvement over using no stabiliser, especially when it comes to reducing your number of throwaway shots. But at the end of the day, image stabilisers are most useful for separating the wheat from the chaff, not ensuring picture-perfect pictures on every shot.