I probably shouldn't say "tack sharp" but reasonably sharp under the circumstances.
That certainly is reasonably sharp under the circumstances. I wish I could have got such good results when I was in Italy a few years ago trying to comply with the usual no-flash rules that apply in most cathedrals. My camera was the 20D and my widest lens was the Sigma 15-30 (24-48 in 35mm terms) with no IS.
I really think it's time for an IS shoot-out between the A900 and the 1Ds3 (or the 5D2 if available).
If anyone is able or willing to do this, can I suggest the following procedure.
First take a series of shots with camera and selected lens on tripod, at various shutter speeds from 1/4 to 1/125th, with MLU enabled and MLU disabled. Use remote cord, switch off IS, and allow a couple of seconds after flipping the mirror to allow any vibrations to settle down before taking the shot.
Examine the results from both cameras in order to determine at what shutter speed the effects of mirror slap are greatest.
Then proceed to repeat all the shots at the same apertures, ISO settings and shutter speeds with cameras now off tripod, hand-held in the same position as they were when on tripod. Enable IS of course.
Ideally, it would also be useful if one could carry out such tests in a studio with variable lighting. One could then use the cameras at base ISO and the the lenses at their sharpest aperture for all shots. Without variable lighting, one will have to change aperture and/or ISO as one changes shutter speed, which will of course affect the quality of the shots to some degree. However, this should not invalidate the results as long as one compares same exposures (and same ISO and aperture) on and off the tripod.
The results of such an experiment should answer the following questions.
(1) At what shutter speed (or range of shutter speeds) is enabling MLU most beneficial for the camera and lens combination used?
(2) How much resolution is lost using IS with a hand-held shot as opposed to tripod without IS?
(3) To what extent, when the camera is hand-held, are the effects of mirror slap mitigated (if at all) at those critical shutter speeds when it's worst on the tripod?
Unfortunately, I'm unable to carry out such tests myself because I don't have an A900 (or any camera with an anti-shake sensor). Nor do I have a 1Ds2 or 5D2.