Klarinet Archive - Posting 000217.txt from 2001/12

From: "Keith" <100012.1302@-----.com>
Subj: [kl] RE: klarinet Digest 9 Dec 2001 09:15:00 -0000 Issue 3502
Date: Sun, 9 Dec 2001 16:10:15 -0500

Walter,
<snip>
Let me say this, then we can debate it. I have made quite a few barrels out
of cocobolo, delrin, and grenadilla.

Multiple people, including several professional clarinetists, several
professional musicians who are not clarinetists, TWO very notable University
teachers, and one doctoral student, have all noticed and exclaimed over the
tonal qualities of the cocobolo barrels. Several have offered to buy one
right off my clarinet!

Now, I use exactly the same tooling (including the same reamer), as close as
possible to the same measurements, and exactly the same methods in each, yet
the cocobolo barrels are the ones that get the attention.

Why?

My deduction is that the material makes the difference.

What deduction would you make?

Walter
www.clarinetxpress.com

------------------------------
<endsnip>

I would conclude that the barrels are indeed different in some way. But I
wouldn't know without further measurements and tests whether the difference
is in the innate properties of the material (density, elasticity, damping,
anisotropy), or whether the necessarily different response of the materials
to your tools (ie surface roughness and topography) has caused the
acoustical difference.

I'm not denying the validity of your observations for an instant. These are
the kind of thing that set a scientist running to formulate and test
hypotheses (there are five in the sentence above).

Incidentally, in response to to other comments, it is not a necessary
requirement of the scientific method that there be a single-variable
experiment. While this is the easiest to deal with, there is a whole
statistical methodology that deals with multivariable experiements, how
these are designed and interpreted in order to extract the effects of
individual variable (see Cox, title somthing like "The Analysis of
Scientific Exeperiments"). It is much more efficient than single-variable
experiments when there are a lot of variables.

Keith

---------------------------------------------------------------------