Re: Re[2]: GG: influence of harpsicord

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

On Wed, 20 Aug 1997, Mark Williamson wrote:

>      How nice that this has prompted so many thoughtful comments. 
>      Here is a follow-up question:  have any of you ever heard of 
>      an instrument called a "fretted harpsichord?"  The reason I 
>      ask is that harpsichords tend to go out of tune very quickly 
>      (because the pins and frame will not hold tension as well as 
>      a modern piano).  But I seem to recall reading about a 
>      "fretted harpsichord" with fewer strings and an internal 
>      fretboard.  And the interesting thing about that is that 
>      with a fretted instrument, you just tune the fewer strings 
>      to a single pitch and all the other notes are tuned by the 
>      mathematics and geometrics of the fretboard.  Have you ever 
>      heard of such an instrument?

Intriguing idea, but no, because it's impossible (I think).  As you point
out below, a mechanism would *first* have to stop the string at the proper
point, then the jack and plectrum would pluck the string...an engineering
nightmare in any century including our own.  I can think of at least the
following problems: 

  - How would such an instrument play fast passages?  The mechanism would
have to release the stop *after* the dampers have stopped the string, else
you'd have pitch fluctuation; if you needed that note or another note on
the same string right away, quick repetition would be almost impossible. 
Also, the precision of all this stopping mechanism (independent of the
speed problem) would make the instrument more problem-prone than it would
be worth. 

  - How would the stopping mechanism not make a tone, the same volume as
on a clavichord?  It would have to be a hard enough substance to get a
cleanly ringing string, but soft enough not to activate string vibrations. 
The stoppers would have to retract far enough that they don't buzz against
the string when not engaged.  With all that to worry about, why not just
build a good fretted clavichord instead, which is difficult enough *and*
more musically rewarding?  (I mean, a good honest clavichord instead of a
hybrid stopping harpsichord; a good honest harpsichord would also be more
musically rewarding than a hybrid.)

  - How would you achieve tonal evenness, since a pluck at a different
point on the effective sounding length of a string gives a different tone
color?  A builder arranges the bridges and pluck point very carefully to
get tonal control; move the pluck point 1mm either direction and the tone
changes significantly.  To combat this, each key would have to have its
own jack, with these two or three jacks all lined up at precise places to
pluck the correct point (percentagewise) along the string; for two notes a
half step apart, they'd be so close together that they would hit each
other.  But then how would you get only one of the two or three plectra to
touch the string?  They'd have to move back and forth, another engineering
nightmare. 

  - Such an instrument would not allow any flexibility of tuning, either
to a different temperament or to a different pitch level.  One would have
to mess with all the frets to change anything.  Furthermore, the frets
couldn't possibly be straight across the instrument; there would be
hundreds of fret snippets which would take care of only 1-3 strings each. 
(I've seen "well-tempered" guitars which do this.)

In short, it would be a completely impractical harpsichord.  Too much to
worry about, and the required effort/money/maintenance would be much
better invested in getting a simpler design set up perfectly.  There are
plenty of variables there to be controlled already. 

Now, clavichords, on the other hand, are very commonly "fretted."  The
strike point becomes a bridge, as the end of the sounding length, with the
tangent staying in contact with the string.  (No problem with tonal
unevenness, because the contact point is always an end.) One can do a bit
of adjustment to different temperaments by bending a tangent slightly to
the right or left, so it strikes a different point.  Mine ("double
fretted") has several notes sharing strings: C/C#, E/Eb, F/F#, G/G#, and
B/Bb.  From tenor F to the bottom, there is no sharing.  "Triple fretted"
clavichords have some instances where three notes share one string. 

>      Jeff Dod (musician/chemist on this mailing list) recently 
>      sent me some very interesting tapes that mention this 
>      concept:  the lecturer talks about the various temperaments 
>      in use in the 18th century not being what we call "equal" 
>      temperament, but he notes that many lutes and other fretted 
>      instruments were tuned to what we would call equal 
>      temperament, this being so easy to do and being strictly a 
>      function of mathematics and not hearing (except to tune the 
>      six lute strings to a single pitch which is quite easy).  
>      With respect to the harpsichord, I am less interested in the 
>      temperament on which the fretboard would be based than I am 
>      in the possibility of quickly retuning an instrument that 
>      may need to be tuned as often as once a day.

What's the problem?  With practice, an experienced harpsichordist can
retune the entire instrument from scratch (three sets of strings) in 15-20
minutes.  If it's just a matter of touching up only a few notes, that
takes only maybe five minutes tops to go through the whole instrument
fixing minor problems.  It's a common thing to do at intermission.  It's
also possible *between pieces* to change a few notes...for instance, if a
certain piece has a bunch of Ab's but no G#'s, it doesn't take long to
move those notes up a bit to work better as Ab's.  I've done a piece in
performance where I needed only *one* specific D# and I didn't want to
change all the Eb's to D#'s, so I quickly tweaked just that one, then back
at the end of the piece. 

There are also some harpsichords and organs which have split accidental
keys: the front half plays the sharp, the back half plays the flat.  These
keyboards let one play better in tune either way, as well as having other
expressive options of playing the wrong one on purpose!  (To get a high
leading tone, or for other reasons.)

And there are "short octave" harpsichords and organs: in the bass, because
the sharps and flats are rarely needed, those levers play other notes
instead.  In June I played a concert on an organ in Germany (built
mid-1730's) where on the manuals it went CDEF (no black keys between
them); in the pedal it was even weirder, with the note that looked like D
playing C, and the note that looked like D# playing D.  The C and C#
pedals didn't exist (but the C was playable on the "D").  On my Italian
virginal, the bass naturals go all the way down to C, but the "Eb" plays
B, the "C#" plays A, and the "B" plays G.  A fair amount of 16th-17th
century harpsichord music is written with such a layout, which allows the
player to get great effects such as playing parallel tenths, or ending a
piece with G-G-D or A-A-E all played with the left hand. 

Back to lutes and viols: they generally have movable frets so the player
*can* adjust tuning to the key of a given piece, better than equal
temperament would afford.  One can slant the fret across the fingerboard,
or put it straight, or divide its strands to go to separate places.  So,
simply tuning all six strings to a single pitch (or octaves thereof)
wouldn't gain us much; the player is still responsible to play in tune
during the piece, and an experienced player wouldn't want to be limited to
equal temperament.  Especially so when playing with wind instruments that
are physically built to be in something other than ET. 

>      If you have any useful tidbits, pass them on.  Perhaps my 
>      memory is failing me, but it would be mechanically possible 
>      to build such an instrument using 18th century technology:  
>      there would only be one quill which would be activated by 
>      the depression of any key, and each key would be connected 
>      to depress a particular string at a particular place on the 
>      fretboard (and that string would be depressed momentarily 
>      before activation of the quill).  You would not be able to 
>      play an unlimited number of notes at one time, however . . .

As I noted above, there are the tone color, speed, and reliability
problems, and the quietness of action, and the limitation of notes that
can be played simultaneously.  With all that, why bother?  (OK, the techie
in me appreciates the novelty, but the musician in me would rather have a
*real* harpsichord than one of these.)

For more on keyboard temperament, see
http://www-personal.umich.edu/~bpl/temper.html, which includes a
spreadsheet to calculate the effects of different temperaments. 

Sorry about the low GG content of this posting, folks.

Bradley Lehman ~ Harrisonburg VA, USA ~ 38.44N+78.87W
bpl@umich.edu ~ http://www-personal.umich.edu/~bpl/