Klarinet Archive - Posting 000756.txt from 2004/07

From: ormondtoby@-----.net (Ormondtoby Montoya)
Subj: Re: [kl] Re: Measuring sound character
Date: Mon, 19 Jul 2004 20:56:29 -0400

Bill Hausmann wrote:

> The FIRST thing we need to do is establish
> whether or not a measurable difference even
> EXISTS.

The on-line graphs that I cited earlier:

http:hughes38.som.ohio-state.edu

are useful because they demonstrate the *complexity* of information that
is included in the graph of a tone.

It seems to me that the question is not whether a difference exists
between two tones --- since no two tones are identical. Rather, the
question is whether a researcher can identify which of the hundreds
(thousands? millions? billions?) of differing data points in a pair of
graphs correspond to something that a human being can identify as being
unique --- regardless of whether other human beings can recognize the
same uniqueness and regardless of whether human beings can agree on a
verbal description of the unique feature,

During my admittedly brief off-list discussions with Dr. Pyne, he felt
(at that time) that the answer to this question [my paraphrase
thereof] is: "No, there's simply too much data in a graph. Should we
look at individual peaks or valleys in the graph, or at relationships
between certain peaks or valleys, or at summations of entire bands that
include a peak or valley, or at relationships between summations of
various bands, or at smoothness of the slopes that lead toward certain
peaks or valleys, or at relationships between the smoothnesses of slopes
that lead to certain peaks or valleys......etc etc etc?"

---------------------------------------------------------------------
Klarinet is a service of Woodwind.Org, Inc. http://www.woodwind.org