I've always thought of them like this:
Static rollover - as helopat says, reaching or exceeding the angle at which the helicopter will topple over rather than drop back onto the skids, ie when a line straight down from the CG goes outside the skid base.
Dynamic rollover - imagine you're taking off, one skid catches on an obstruction and you start to roll.
Say for the sake of argument that full left cyclic can tilt your thrust vector 15degrees to the left of the mast. If the fuselage reaches 15 degrees right tilt with a skid caught, full left cyclic will only point the rotor thrust straight up, so it's not going to help you.
Any more fuselage tilt will turn any rotor thrust into a pro-rolling force, and the only option open to you is to dump the collective to get rid of the force, although it may well be too late by then.
Apparently dynamic rollover can happen very quickly because the rotor thrust is actually dragging you around the pivot point.
Just read the initial post again - I see you were asking for experts...definitely not one, but hope that helped anyway.