"Screwy E-M5 IBIS": Theory and evidence

Based on several long threads on this forum, at least some users experience problems with the IBIS system on the E-M5 at certain shutters speeds (around 1/100 s).

http://forums.dpreview.com/forums/read.asp?forum=1041&message=41516235

http://forums.dpreview.com/forums/readflat.asp?forum=1041&message=41589267

http://forums.dpreview.com/forums/read.asp?forum=1041&message=41640571

http://forums.dpreview.com/forums/read.asp?forum=1041&message=41666879

When I received my E-M5 about a month ago, I didn't even get around to testing for the problem (as I had planned) before it hit me in the face. I was shooting in the dusk with my Minolta MD 85/2 at a speed of 1/100, just one notch below the 1/EFL standard rule, and noticed that every shot came out slightly blurred in spite of IBIS being turned on and however hard I tried to keep the camera steady.

Since then, I have spent considerable time (I think I might be approaching a thousand exposures) trying to pinpoint the exact nature of the problem and under what circumstances it does and does not manifest itself. I have done this mainly because I want to know how to work around the problem in the best way possible. When actually shooting, I am used to knowing a) what is a reasonably safe shutter speed (with this or that lens and at this or that FL if a zoom) from a camera-shake point of view and b) what is the slowest speed that is even worth trying in the hope that I get lucky. But I also did it because I think it might help us all if we can pinpoint exactly what is going on here. So here is a brief summary of my findings this far in the hope that it might be of some help to others.

First, it seems reasonable to assume that the shutter is the ultimate cause of the problem. How could we otherwise explain the fact that the problem is worst at a shutter speed of about 1/100 and becomes less noticeable at lower as well as higher speeds? Since, regrettably, the E-M5 does not have a purely electronic shutter except in video mode, however, we cannot conclusively test for the possibility that this basic premise is wrong.

At any rate, given the premise that the shutter is the root cause, I can think of at least three different ways in which it might cause blurry images:

1. The shutter causes vibrations that fools IBIS to behave in an inappropriate manner.

2. The shutter causes a shock that is too strong for the electromagnets that hold the E-M5 sensor in place, thus causing unwanted uncontrolled sensor movement.

3. The shutter causes camera shake of a kind that IBIS is unable to counteract.

When I first heard about the problem, I found the first of the three hypothesis most plausible. However, in my own testing, I haven't found a shred of evidence to support it.

First, if this hypothesis were right, we should expect the problem to manifest itself with the camera rigidly fixed. Since my trusty old Velbon broke down during vacation, I don't have a real tripod to experiment with at the moment. But I tried placing the camera on a stable table and shoot with 2s self-release. Everything looked just fine. I also tried placing the camera on our tiled (on top of concrete) kitchen floor so as to eliminate any kind of vibration dampening. Again everything looked just fine.

Second, in none of the test configurations I have tried so far (Minolta MD 85/2, Panasonic 100-300 at 100 and 300, Olympus 40-150 at 40 and 150, Panasonic 14-45 at 45, and Olympus 45/1.8) have I found any systematic evidence that the problems are worse with IBIS on than with IBIS off, as should be expected if the first hypothesis is correct.

If this conclusion is correct, it is good news as well as bad news. It is good news in the sense that keeping IBIS on at all times won't make the situation any worse. It is bad news in the sense that there is less hope that Olympus will be able to fix the problem via a firmware fix (e.g., a fix filtering out the troublesome vibrations from the IBIS control circuit).

The tests I conducted with the camera resting on a tiled floor with concrete underneath also, to my mind, rules out hypothesis two. This test scenario should maximize the kind of completely unelastic shutter shock that might cause the sensor to move uncontrollably.

This leaves hypothesis three as the most credible at the moment. This hypothesis is compatible with the evidence already mentioned above. It is also consistent with another experiment I performed with the E-M5 plus 100-300 placed on an Ultrapod II mini-tripod (which is not heavy and sturdy enough for this combo) and shooting with 12s self-release. In this case, I saw problems in a certain shutter range (from about 1/20 to about 1/200), with IBIS on as well as off, but no problems below or above that range. If indeed hypothesis three is right, we should perhaps rename the problem from "screwy IBIS" to "screwy shutter".

(Continued in next reply)