Klarinet Archive - Posting 001355.txt from 1999/05

From: James Pyne <jpyne@-----.edu>
Subj: Re: [kl] Even Partials (was Bradley's Question)
Date: Sat, 29 May 1999 12:59:15 -0400

This is in response to Dee Hays post - Thu, 27 May 1999 20:08:49 -0500.

Dee,

A goal of research is, of course, to support or disprove previously held
assumptions. This is why scientific "fact" changes so much over time. There
is the possibility that your reply could rely on dated information. I would
very much like to know the exact source and date of the article that you
refer to. It may be that, in the research arena, I am operating on the
basis of newer information.

Basically you seem to question the following three things, and I will
answer them in this order.

*** A. The existence of powerful equal partials in clarinet spectra
*** B. My understanding, methodology and accuracy concerning measurement
processes
*** C. How partials effect the listeners perception in terms of source
identification
----------

*** A. The existence of powerful equal partials in clarinet spectra.

To support my position (that there is abundant even-partial content in
clarinet tonal spectra) I'll list three recent research presentations that
I have done. They included slides of computer generated visualizations
showing powerful even-partial content in clarinet tones of all registers,
chalumeau, clarion, and altissimo. We have produced tonal spectra (showing
the even-partial content) here at OSU on at least three separate analysis
systems and with a variety of hi-quality microphones. I believe our
measurement techniques are viable and respected by the scientific community.

These are the three presentations:

1. Acoustical Society of America (133rd Meeting) Penn State, June 1997
Lecture: "The evolution of single- and double-reed musical instruments: A
study of spectral characteristics", with David Butler. Abstract published
in The Journal of the Acoustical Society of America, Vol. 101, No. 5, Pt.
2, May 1997

(My co-researcher in this study is David Butler, "The Musicians Guide to
Perception and Cognition", Schirmer Books 1992. Obviously he accepts the
accuracy of the spectral characteristics (clarinet and oboe tones) that
form the basis of this study.

2. Acoustical Society of America (131st Meeting) Indianapolis, April 1996
Chair, Session 2aMU, Musical Acoustics: " Reed Instruments: Research and
Performance"
Lecture: "Techniques for defining clarinet reed quality via computerized
light transmission analysis" Abstract published in The Journal of the
Acoustical Society of America, Vol. 99, No. 4, Pt. 2, April 1996

3. Acoustical Society of America (126th Meeting) Denver, Oct. 1993
Lecture: "The interaction of pitch, harmonics, and the equal loudness
curves in the timbre analysis of musical tones." Presented in Session 4pPP,
Psychological, Physiological and Musical Acoustics. Abstract published in
The Journal of the Acoustical Society of America, Vol. 94, No. 3, Pt. 2,
September 1993

The audience for these presentations consisted largely of physicists and
other researchers who specialize in some area of musical acoustics. A
question and answer period takes place at the end of each presentation and
is typically used to argue questionable points. The plausibility of the
existence of even partials has not been questioned. Nor has the accuracy of
their relative intensity measurements been an issue.
-------------

*** B. My understanding, methodology and accuracy concerning measurement
processes

In response to your comment on the analysis of overblown tones:

>Secondly the 4th line D is NOT the 1st partial. All the clarion notes are
>the 3rd partial of the fundamental chalumeau note (for clarion D, the
>fundamental is the low G). The chalumeau is counted as the first partial.
>So you need to play the chalumeau and measure the partials from there.

I have analyzed many spectra on Low G that contain even as well as odd
partials and am well aware of their relationships.

However in the case of 4th line D... it could be considered a calculated
3rd partial based on tube length. But to play 4th line D, low G must be
successfully suppressed by the performer or a multiphonic would be heard.
This means Low G no longer serves as the driving frequency (fundamental).
The term fundamental is defined as the frequency of the generator driving
the system. In this case it is the reed vibrating at 523Hz and producing
partials based on that fundamental. Additionally this degree of suppression
means that patterns of air (containing information about partial content)
at the frequency of Low G no longer move to our ears or to a microphone. So
in "real life" 4th line D functions as a new fundamental (the driving
frequency), developing its own patterns of air movement. To do this the
clarinet and performer use the calculated 3rd partial range of the tube as
a "window of opportunity", to produce and stabilize the resulting written
4th line D.

If this were not true how would one account for Altissimo A. It is played
with the tube length (chalumeau fingering C) that should produce Bb. Yet
once fingered produces, in the spectrum, appropriate partials for that
overblown "fundamental" (altissimo A). In fact, if we could produce tones
only at the exact frequencies of those overtones present in the chalumeau
mode we could never play "in tune".

So overblown tones in the clarion and altissimo may be analysed, not
necessarily calculated, as spectra with a new functional
fundamental/partial 1. Sound must reach our ears or a microphone via moving
air to be analysed. In this research we look into, through the careful
analysis of tonal spectra, the nature of the complex movement of air-borne
sound, and how that might actually interact with the mechanism of hearing.
--------------

*** C. How partials effect the listeners perception in terms of source
identification

>For starters if the even partials were
>that strong, the clarinet would sound more like an oboe as that is a very
>significant element in what makes an oboe sound like an oboe.

Recognition is a very complicated process. Certainly it is not exclusively
dependent on strengths of relative harmonics, that is to say the relative
intensity of resonance peaks (partials) having integer
relationships...1,2,3,4... The formulas for music sound synthesis would be
much easier if the process were that simple. To my knowledge there is not
complete understanding of how we accomplish the recognition task relative
to musical instruments and say, the way we recognize the voices of many
different people

---Jim Pyne

>I must take issue with your statements as they are completely at odds with
>the article that I remember reading. For starters if the even partials were
>that strong, the clarinet would sound more like an oboe as that is a very
>significant element in what makes an oboe sound like an oboe.
>
>Also if the even partials were this strong, the earliest spectrum analyzers
>(used in acoustic research on oboes and flutes) would have picked them up
>and they did NOT. These early attempts incorrectly "confirmed" the
>hypothesis that the clarinet sound did not contain even partials while oboes
>and flutes did.
>
>Secondly the 4th line D is NOT the 1st partial. All the clarion notes are
>the 3rd partial of the fundamental chalumeau note (for clarion D, the
>fundamental is the low G). The chalumeau is counted as the first partial.
>So you need to play the chalumeau and measure the partials from there.
>
>Dee Hays
>Canton, SD
>

James Pyne, professor
Clarinet Studio/Research Group
School of Music
The Ohio State University
1866 College Road
Columbus, Ohio 43210
pyne.1@-----.edu
Tel: 614 292 8969
Fax: 614 292 1102
http://www.arts.ohio-state.edu/Music/Clarfest

-------------------------------------------------------------------------
Unsubscribe from Klarinet, e-mail: klarinet-unsubscribe@-----.org
Subscribe to the Digest: klarinet-digest-subscribe@-----.org
Additional commands: klarinet-help@-----.org
Other problems: klarinet-owner@-----.org

   
     Copyright © Woodwind.Org, Inc. All Rights Reserved    Privacy Policy    Contact charette@woodwind.org