a capella tuning

Thanks to David J. Finnamore and those others who have done empirical
testing of a capella tuning. While I am anxious to find out more through
such testing, my own experience with choirs leads me to believe that,
unless you are talking about the Hilliard Ensemble or Anonymous 4 or some
other professional specialists in early choral chamber music, it is for
practical purposes meaningless to speak of tuning systems in choral music
with the same precision that we do with keyboards.

First of all, singers in choirs virtually never agree on a particular
pitch. To a limited extent, you *want* this "spread" of pitches around a
hypothetical center pitch to give the richness of the aptly-named "chorus
effect." In my experience, this spread can range 15 cents or more, i.e. the
distance between the 5/4 and equal tempered third. Second, there is
vibrato, which, with Western singers untrained in early music, frequently
has a deviation of a semitone.

OK, so, ideally, the spread and all the vibratos should still center around
a theoretical "ideal" pitch which spectral software should identify (given
a low enough sampling rate and a window size that is large enough to be
precise but small enough not to mask pitch drift). Still, intonation
effects, particularly beats, are greatly masked by these phenomena. (How
many times have I heard the vibrato get wider at intonationally sensitive
points...)

Third, the tuning of an interval in an a capella context frequently depends
on the melodic as well as the harmonic context. Melodic motion up,
especially in the case of the leading tone, will tend to sharpen a pitch,
for example. Dissonant or ambiguous chords will allow for a greater
deviation from JI, and perhaps that is even desirable. I'm sure that the
timbre of vowels will also affect intonation. Then there are pitch scoops,
drift over time, instability as breath goes, and so on.

So, until I see more empirical evidence to the contrary, I will continue to
believe that to expect the kind of consistency and accuracy necessary to
see commatic shifts in most choirs is quite unrealistic. I seem to recall
that Seashore's old study with college choirs found this.

One point about the Renaissance, though. There is evidence that so-called
"a capella" music of the period was in fact frequently accompanied by
instruments ad libitum. Of course this practice varied in time and
geographic region, especially once the Reformation took hold. If the choir
was accompanied by an organ, well, then they would of course tend to that
tuning. If the choir was accompanied by wind instruments, then those
instruments, too, had a certain amount of intonational inconsistency.

I remember back in my crumhorn-playing days that those beasts had easily a
range of a minor third either way on a given fingering, just by varying the
air pressure. A good deal of our rehearsals was given to intonation, but to
expect that we could accurately and consistently represent a distinction
between 12TET and JI would be asking a lot.

So I hope to see more spectral studies!

Bill

^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
^ Bill Alves email: alves@hmc.edu ^
^ Harvey Mudd College URL: http://www2.hmc.edu/~alves/ ^
^ 301 E. Twelfth St. (909)607-4170 (office) ^
^ Claremont CA 91711 USA (909)607-7600 (fax) ^
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

>From: alves@orion.ac.hmc.edu (Bill Alves)
>
>Thanks to David J. Finnamore and those others who have done empirical
>testing of a capella tuning. While I am anxious to find out more through
>such testing, my own experience with choirs leads me to believe that,
>unless you are talking about the Hilliard Ensemble or Anonymous 4 or some
>other professional specialists in early choral chamber music, it is for
>practical purposes meaningless to speak of tuning systems in choral music
>with the same precision that we do with keyboards.

Its probably better not to go in trying to see which "already conceived"
tuning system is being used, but investigate without any a priori
assumptions what kind of intervals do people sing and play when they are not
constrained by instruments with disctrete tuning? I'm sure this depends on
lots of things like who they are, what culture, thier level of training, the
kind of music etc. I'd really be surprised if there was nothing to learn
from an investigation like this.

>
>First of all, singers in choirs virtually never agree on a particular
>pitch. To a limited extent, you *want* this "spread" of pitches around a
>hypothetical center pitch to give the richness of the aptly-named "chorus
>effect." In my experience, this spread can range 15 cents or more, i.e. the
>distance between the 5/4 and equal tempered third. Second, there is
>vibrato, which, with Western singers untrained in early music, frequently
>has a deviation of a semitone.

The spread is certainly there but the energy is usally centered in a
narrower portion of the band, mostly in the middle. A group of people
singing together, unless hopeless, will no doubt be synergistically
influencing each other to create a "gestalt" tuning for the part. This
includes the "chorus effect" but a chorus effect centered on the intended
pitch.

>
>OK, so, ideally, the spread and all the vibratos should still center around
>a theoretical "ideal" pitch which spectral software should identify (given
>a low enough sampling rate and a window size that is large enough to be
>precise but small enough not to mask pitch drift). Still, intonation
>effects, particularly beats, are greatly masked by these phenomena. (How
>many times have I heard the vibrato get wider at intonationally sensitive
>points...)

I suppose it might be hard to pin down the frequency of Renata Squatto or
some other singer with a vibrato big enough to drive a truck through. Not
all singers sing like this. I was just listening to a recording of emma
calve from1908 where she sings a little french ditty with just about the
purest steady tone ive ever heard (too bad shes singing with a piano).
>
>Third, the tuning of an interval in an a capella context frequently depends
>on the melodic as well as the harmonic context. Melodic motion up,
>especially in the case of the leading tone, will tend to sharpen a pitch,
>for example. Dissonant or ambiguous chords will allow for a greater
>deviation from JI, and perhaps that is even desirable. I'm sure that the
>timbre of vowels will also affect intonation. Then there are pitch scoops,
>drift over time, instability as breath goes, and so on.

This is >exactly< the phenomenology that needs to be examined.

>
>So, until I see more empirical evidence to the contrary, I will continue to
>believe that to expect the kind of consistency and accuracy necessary to
>see commatic shifts in most choirs is quite unrealistic. I seem to recall
>that Seashore's old study with college choirs found this.

Theres only one way to know for sure.

>So I hope to see more spectral studies!

Id be willing to put up spectral studies as part of my website or at least
link to anyones studies on their own site. Maybe it could be called
"Proceedings of the First Aristoxenus Memorial Conference for the
Spectrographic Investigation of Musical Praxis" :-)

dante

At 14:34 6/9/99 -0800, you wrote:
>From: alves@orion.ac.hmc.edu (Bill Alves)
>
>Thanks to David J. Finnamore and those others who have done empirical
>testing of a capella tuning. While I am anxious to find out more through
>such testing, my own experience with choirs leads me to believe that,
>unless you are talking about the Hilliard Ensemble or Anonymous 4 or some
>other professional specialists in early choral chamber music, it is for
>practical purposes meaningless to speak of tuning systems in choral music
>with the same precision that we do with keyboards.
>
>First of all, singers in choirs virtually never agree on a particular
>pitch. To a limited extent, you *want* this "spread" of pitches around a
>hypothetical center pitch to give the richness of the aptly-named "chorus
>effect." In my experience, this spread can range 15 cents or more, i.e. the
>distance between the 5/4 and equal tempered third. Second, there is
>vibrato, which, with Western singers untrained in early music, frequently
>has a deviation of a semitone.
>
>OK, so, ideally, the spread and all the vibratos should still center around
>a theoretical "ideal" pitch which spectral software should identify (given
>a low enough sampling rate and a window size that is large enough to be
>precise but small enough not to mask pitch drift). Still, intonation
>effects, particularly beats, are greatly masked by these phenomena. (How
>many times have I heard the vibrato get wider at intonationally sensitive
>points...)
>
>Third, the tuning of an interval in an a capella context frequently depends
>on the melodic as well as the harmonic context. Melodic motion up,
>especially in the case of the leading tone, will tend to sharpen a pitch,
>for example. Dissonant or ambiguous chords will allow for a greater
>deviation from JI, and perhaps that is even desirable. I'm sure that the
>timbre of vowels will also affect intonation. Then there are pitch scoops,
>drift over time, instability as breath goes, and so on.
>
>So, until I see more empirical evidence to the contrary, I will continue to
>believe that to expect the kind of consistency and accuracy necessary to
>see commatic shifts in most choirs is quite unrealistic. I seem to recall
>that Seashore's old study with college choirs found this.
>
>One point about the Renaissance, though. There is evidence that so-called
>"a capella" music of the period was in fact frequently accompanied by
>instruments ad libitum. Of course this practice varied in time and
>geographic region, especially once the Reformation took hold. If the choir
>was accompanied by an organ, well, then they would of course tend to that
>tuning. If the choir was accompanied by wind instruments, then those
>instruments, too, had a certain amount of intonational inconsistency.
>
>I remember back in my crumhorn-playing days that those beasts had easily a
>range of a minor third either way on a given fingering, just by varying the
>air pressure. A good deal of our rehearsals was given to intonation, but to
>expect that we could accurately and consistently represent a distinction
>between 12TET and JI would be asking a lot.
>
>So I hope to see more spectral studies!
>
>Bill
>
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>^ Bill Alves email: alves@hmc.edu ^
>^ Harvey Mudd College URL: http://www2.hmc.edu/~alves/ ^
>^ 301 E. Twelfth St. (909)607-4170 (office) ^
>^ Claremont CA 91711 USA (909)607-7600 (fax) ^
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Professor Alves,

What you have described in your post sounds very much like what might has
been going on in traditional Turkish music, both instrumental and vocal.
Pitches forming the underlying 'scale' for a given 'maqam' seem to be fuzzy
clusters centered around a 'medoid', a term used in cluster analysis, and
a particular pitch that needs to be played in that cluster at a particular
instant during the course of an improvisation depends on melodic context.
I like to characterize this feature as 'music having memory', going back
several sounds in the sequence. Clearly keyboard or fretted instruments
are totally out of question in Turkish music. Carl Seashore's work early
in the century was one of principal sources of inspiration for me in the
beginning.

The work I am doing currently could be considered a pedestrian style
'spectral study' of instrumental Turkish music based on improvisations.
Hopefully some results will be published in the near future. There may
be a common universal thread tying many musics on the issue of tuning.
I am craving for more discussion on this list on 'fuzzy' dynamic scales
as opposed to fixed, deterministic scales.

Sincerely
Dr. Can Akkoc
Alabama School of Mathematics and Science
1255 Dauphin Street
Mobile, AL 36604
USA

Phone: (334) 441-2126
Fax: (334) 441-3290
Web: http://199.20.31.100/GIFT/