I´m heartened to see that others have experienced the same strange effects that I have too: several different tests with cameras on tripod and mirror locked up or not locked up have not always delivered the expected results of the images WITH the mirror locked up being always better than the others.
As a matter of principle, when mlu is possible I´d use it, but it is no guarantee of best results. In fact, I have found that if I actually put some pressure down on the camera with one hand and trip the shutter with the other, a fairly good result will follow.
But with all of this, I believe that Bernard touched on a vital point: the intrinsic vibration-style of the tripod itself. And that seems to vary somewhat according to the material upon which its feet are based too: very hard floor tiles, for example, appear to set in motion a ´brittle´ sort of mechanics whereby the tripod feels very stiff but very movement-prone. Odd; has anybody else found this too?
Perhaps the worst case I´ve come across has been with the Pentax 67 ll, where with mirror locked up, on a heavy Gitzo, the shutter has bounced so much that the edges of the frame have been hopeless. Didn´t keep the camera long... a further big loss of money.
Perhaps we are nearing a situation where regardless of optics, sensor or film, mechanical limitations will bring us up against a brick wall which we will only pass by changing how we do things to capture images. Somewhere, I read of the idea of using no shutter at all, but of having exposure made via the trick of switching the sensor on and off for the time required to achieve exposure. Neat, if it can be done; fewer moving parts to go ass over tit. It might come from Canon, seeing how they don´t favour easy MLU ;-)
Just re-read JeffKohn´s last post: is this the same shutter you refer to as an electronic one? I have always assumed the exposing function of an electronic shutter to be exactly the same as a mechanical one: curtains.