I think I've sorted this out and I'm embarrassed it took so long. First, I should not have been calling the sound I heard hiss. More accurate would be hum or buzz or something between the two, a drone maybe? That was carelessly worded.
I moved the mic off the Olympus E-M5.2 cold shoe onto a separate holder and the "drone" sound went away, in the sense that the voice quality produced was as good as we get from our old Canon camcorder, which resolves my wife's issue. I didn't understand why this should work until I came across an interweb posting that stated that even when IBIS is turned off, the motors still run to keep the sensor in place. I had never had to think about this before and simply assumed (mistake) that when IBIS is turned off, the sensor is "parked" to a mechanically fixed location. It seems not to be.
Btw, my E-M5 mark 3 behaved exactly the same way.
I set the camera's internal mic input gain to -6 db and upped the output of the mic. I used an H1 in pass-through for this test and if I set its output to 70 and above (or 60 if closer to the person speaking) then the audio was fine for our purpose. I'm certain it doesn't meet pro Hollywood standards, but no one is paying us the big bucks either.
I wondered why I never had this issue with the Canon camcorder, as all I did with it was to turn off its stabilization. But its lens is stabilized, not sensor, and it is a tiny lens so maybe the residual noise from that motor is low enough that it cannot be heard. Also, I am using a dedicated Canon hot shoe mic with that camcorder and Canon may be compensating for the stabilization noise, if any, some other way.
A friend wondered if this issue might be why some lower end vlogging cameras (e.g., Panasonic G100) are not sensor stabilized.
