Re: GG:has the electronic future arrived?

["Bradley Lehman" <bpl@umich.edu>]

Sivan wrote:
>This article in the New York Times was recently pointed out to me:
>
>http://www.nytimes.com/library/national/science/022200sci-music-composers.h
tml
>
>Someone created a program that allows a user to interpret music
>electronically; it seems to be something that Gould prophesied!  For
>more accurate and detailed information connect to the above link.

Interesting!  The concepts outlined in the article make some sense in terms
of shaping an overall general sound.  That's already a large step forward:
not playing all the notes equally.

Now turning from theory to the actual examples provided:

Mendelssohn's Octet Opus 20 Scherzo.  This one comes off the best of the
three, being the most mechanical piece (relative to the other two) to begin
with.  Not bad, though I miss a large-scale shaping from phrase to phrase, a
sense of natural breathing.  It also doesn't seem to take into account the
physical act of bowing in two directions: the fast repeated notes here
especially sound strange.

Mozart's Piano Concerto K453, 3rd Movement.  Also not too bad, though again
the large paragraphs don't breathe.  I miss playfulness and whimsy.  But the
biggest oversight is a fundamental element of musical grammar.  The program
clobbers the resolution note of every appoggiatura (and this movement has
hundreds of them at the ends of phrases) rather than resolving quietly.
This error is as musically serious and obtrusive as using the word "the" at
the end of a sentence, or putTING acCENTS on the wrong syllaBLE.  You'd
think that given all the trouble they've gone to in this program, especially
with regard to dynamics, they'd teach it how to recognize this.

Beethoven's String Quartet Opus 132, 3rd Movement.  Disaster.  No sense of
organic forward motion or architectural shape, even when Beethoven's tempo
changes are programmed in and the dynamics are reasonable.  The music
doesn't seem to mean anything.  There are some botched appoggiaturas in this
piece, too, and those consistent trills sound silly....

And what's that continuous and consistently heavy vibrato about?  I think
it's about 300% too much, and real players don't use it so predictably.  Up
through the early 20th century (yes, 20th), vibrato on stringed instruments
was considered an ornament (highlighting selected notes within a phrase),
not something to be used as a constant color to the sound.  It was also more
for soloists than for chamber or orchestral players: a way for a soloist to
stand out from the texture.  In chamber music especially, if everybody uses
heavy vibrato simultaneously it merely clogs up the sound.  Why have the
computer simulate something that is intrusive, ugly, and anachronistic?
Vibrato here is merely substituting for the computer's lack of ability to
shape phrases meaningfully without it.  (Well, I guess this one comes down
more to taste and to study *outside the score* than the other elements.)

In the illustrations Clymer notes that "Composers have different ways of
using vibrato.  Each note has a different vibrato partially because of the
pitch of the next note."  Uh huh, based on what?  Durations, pitches,
general speeds, and to some extent dynamics may all be derived from a score
and lead to a general performance style, which is the point of the article.
But where does he get the assertion about different composers writing
different amounts of vibrato into their music?  Sure, there exist some 17th
century pieces for flute (by Hotteterre) and for viola da gamba (by Marais)
that specify exactly which notes should receive which of several types of
vibrato.  The notation in bars 95-100 of the fifth Brandenburg, first
movement, is arguably for flute/violin vibratos rather than trills.  Some of
CPE Bach's clavichord music has vibrato notations.  In every case these are
ornaments.  None of this leads convincingly to an extrapolation of how a
computer should add a *continuous* vibrato filter component to every long
note in Mozart, Beethoven, and Mendelssohn.

Anyway...looks like a promising start, but I remain dubious...especially if
the program would ever attempt French baroque music without learning the
grammar.  It's bad enough for human performers to attempt French-influenced
pieces such as Bach's French and English suites without being very familiar
with French style.   (Oops, don't go there on the GG list...Our Hero might
be implicated......)  At least the human performers make up for it with
imagination, creativity, and musical instinct, which computers don't have.
(Whew, close catch there.)

GG himself of course raved about "Switched-On Bach," but there was a human
musician behind every note and every decision on that album.

Bradley Lehman, http://www-personal.umich.edu/~bpl
Dayton VA