When OIS Lenses are Put onto Sensor-shift IS Bodies..

Here is a very interesting latest experiment by the Japanese Impress DC Watch:-

http://dc.watch.impress.co.jp/docs/review/special/20090908_313599.html
(Article in Japanese, Google Translation in English Here)

The following are tested and compared:-

1. Sigma 50-200mm F4-5.6 DC OS HSM on K-7, and;
2. Panasonic Lumix G Vario 45-200mm F4-5.6 Mega OIS on Olympus E-P1.

And then both lenses are zoomed to the tele side, i.e., 200mm and a target at a distance of 5 metres was shot. 30 shots was taken for each test for each of the four combinations of enabling the lens-based and/or body-based IS for each combo. The amount of blurred shots against successful ones were then counted.

As the K-7 has an multiplying factor of 1.5X whereas that of the Oly is 2X. As such, the effective focal lengths, in traditional 135/film sense, are just 300mm and 400mm respectively for the K-7 and the E-P1 under this test. To make the test "fair" enough and meaningful, the tester then set the shutter speeds at 2 stops slower for both cameras, i.e., at 1/80s and 1/100s respectively for the two cameras.

One thing I think of a minor flaw of the test is that the focal length should be the right thing to adjust between the cameras, instead of adjusting the shutter speed. It is just because with different effective magnifications of the shot target, actually it is indeed somehow unfair to compare for the obtained results, for whether the shot is blurry or not. The design of this test favours more to the K-7 actually, as the magnification of the effective focal of 300mm is just lower than that of 400mm for the same target placed at the same distance of 5m away.

But anyway, let's look at the results. For the sake of easy reading and I think not many of my readers here do read Japanese (non the Google translation), I copy and re-type a summary table below for the obtained results:-

Condition (Lens / Body IS On or Off)	K-7 Results (Number of Successful / Failed Shots)	Successful Rate (%)	E-P1 Results (Number of Successful / Failed Shots)	Successful Rate (%)
On / Off	17 / 13	56.67	9 / 21	30
Off / On	11 / 19	36.67	20 / 10	66.67
On / On	16 / 14	53.33	5 / 25	16.67
Off / Off	5 / 25	16.67	17 / 13	56.67

I've also calculated and added the successful rate in percentage in my above table to give you a better idea on the actual performance.

Now the results are really very interesting..

1. Without any IS, the K-7 gives a lot of blur shots but the E-P1 is not that bad actually even with everything turned off. Well, one possible reason for this is because the E-P1 is mirrorless.

2. With body IS on but lens turned off, the performance of the K-7 is not much satisfactory and only marginal improvement is achieved. With solely body IS, the E-P1 has been able to improve the successful rate only by a very marginal factor, also.

3. With solely the lens IS, now the K-7 (but actually the Sigma!) performs the best, the successful rate is the same as what the user can get with the E-P1 with body IS. In contrast, with the Pana lens to do the OIS on the Oly, the result is really poor. The successful rate drops far below than what everything is turned off! Is that the Pana lens is incompatible with the Oly actually? Or is it that Pana lens is just a poor performer for its IS ability?

4. With both body and lens IS turned on (which is non-sense actually), the K-7 combo does still work. The successful rate is only just second to using the lens IS alone with a small difference and it is still far better than what the K-7 body IS/SR alone can do and of course better than when there is no IS of any at all. In fact, this result can be explained and it does tally with the previous findings on the effectiveness of the (Sigma's) lens against (Pentax's) body IS. The K-7's body IS is found to be ineffective in this test whilst the Sigma is shown to be working better. When both systems counter-react, the Sigma is still able to "overcome" and thus as a whole it does still work! On the other hand, the Oly/Pana combo gives the worst result when both IS systems are put into work, which is indeed not a big surprise as chaos has been created and blur is introduced.

So, after all these. The conclusions are clear. With the K-7, it is better to use the Sigma in-lens OIS technology than using the Pentax's SR! (Well..) With the Oly, it will do no big harm to forget about any IS completely as turning on the body IS only improves the successful rate by a small fraction, whereas using the lens IS could give worse results!