Felix Delbruck writes:

>But Prof. Rosen is right in the sense that timbre and volume are not
>independent

It is possible (I don't know if it's the case though) that nonlinearity in
the response of a struck string (as in the piano's mechanism) may cause
significant changes in timbre with less significant changes in volume,
especially at the extremes of dynamic range.  I.e., the perceived volume
may "saturate" (if only because of the logarithmic perception of loudness
as a function of amplitude) while the harmonic content changes enough to be
clearly distingushed.  I also suspect this would be far more likely to be
a property of the specific instrument rather than any specific pianist's
technique.

len.