Hello - and thank you for your elaboration upon the cembal d'amore.
 
A fine point, but one not commonly understood, is that _all_
clavichords must be quiet and inefficient of necessity, because
the length of the string is defined by the same tangent which
excities. it. The point of contact between string and tangent is
defined as a node, or the _end_ of the string; the other end is of
course at the bridge pin. The node is the most inefficient place
possible to put energy into the string; try plucking a harpsichord
string exactly at its end.
 
It occurred to me to say that the clavichord works at all only
because the tangent has a finite width, not a perfectly sharp
top, and so the node is confused. But I think that must be wrong;
it is probably not energy _transverse_ to the string, as in piano
and harpsichord, which excites the string, but _longitudinal_
energy which stretches the string a little, which then rebounds.
 
Do the physicists among us have an intuition on this? In any case,
it's the energy input at the node, and not (as some innocents have
supposed) all that cloth around the strings that makes them quiet.
 
                                        David Calhoun, Seattle