That depends very much on the material, for solid structures, or the mechanism by which the damping is created.
It's probably simplest to talk about the bandwidth of resonances for homogenous steel structures.
Over the range 3-2000 Hz the internal damping doesn't seem to change much, that is the %age bandwidth of the resonances stays roughly the same.
Cast iron has the reputation of offering more damping than steel, I can't remember if the high frequencies are more heavily damped, I think that's the way it works.
Rubber's damping varies. In all sorts of ways. On Monday I'll look up some sample graphs.
Mechanisms can easily produce frequency dependent damping - we explicitly use things called hydrobushes that are very heavily damped at some frequencies, yet are virtually undamped at others. A so-called friction bush is a type of hydrobush that uses fluid in shear, it behaves much more like electricpete's classical viscous damping equation (as a consequence it has very high transmissibility at high frequencies).
Another way of adding damping at some frequencies not others are harmonic dampers or tuned absorbers. The wiki has an article on them.
Cheers
Greg Locock
SIG

lease see FAQ731-376 for tips on how to make the best use of Eng-Tips.