Hammers fit to Strings

[Michael J. Wathen]

Here is some more scoop on the sound samples.  One of the samples, I
believe it is the fit.wav has a little jingle sound in it after about one
second.  This sound is coming from the id tags on my dog's collar as he
shakes his head.  I am assuming that means that he doesn't like that sound
and would prefer the unfit hammer.

The tuning is the same for both notes.  I took the sample for the fit
first.  Then I took out the action and filed a little off one side.  I
checked it to make sure it was no longer fit.  I then recorded it.  So, I
believe that the two samples were taken about five minutes apart.

A first hearing the sounds are indistinguishable to my ears.  If I listen a
little longer I can tell them apart.  My preference goes against the grain,
I believe, of what people are aiming for.  The unfit is what I prefer
(here).  It audibly appears to have a longer singing tone.  The fit has no
dirt in it and is a little uninteresting.

There are several things to look for in the graphs.  The best way to make a
comparison is to actually print them.  You would need some kind of
clipboard program to capture what is on the screen.  

The next question is very important.  It is actually heeded advice I
received from Harold Conklin a while back.  After you think you have
isolated in your minds ear the sounds it is then that you go to the
analyzer.  Here is the question: do the graphs agree with what you hear?
The tendency is to through away our judgment in favor of blind trust of
measuring equipment.  Piano sound is extremely complex. It is so difficult
to measure in a way that allows one to distinctly discriminate between to
pianos.  There are so many factors that put tinted glasses to the vision.
Are you actually measuring the equipment.  Is there some sort of bias to
the microphone and sampling device?  Am I only sampling one spot in space?
What is the microphone was moved about six inches to another spot in space?

To look for in Spectrum.

Notice the continuity of the spectrum lines.  The unfit has discontinuities
amounst several of its partials.  This most like accounts for the wangyness
that is typically identified with unfit strings.  However, also notice that
over time the partials of the unfit note end up stronger.  From this we
might conclude that the partials periodically  disappearing indicates that
energy is not be transferred out of the string as fast hence a longer
sustain.  Also, notice in the spectrum the non harmonic noise content by
looking for spots or clusters on the graph that are not part of the partial
lines themselves.