Re: TD 648 -- Multiple keyboards, new designs (Arthur W. Green)

Hello, there, and in Tuning Digest 648, Arthur W. Green
<goshawk@crosswinds.net> offered some very interesting remarks as part
of a dialogue on the option of multiple keyboards for microtunable
synthesizers. Here I'll try to respond on some points, including a Web
page reference to a keyboard design intended to let solo keyboardists
approximate the richness of a large ensemble performance.

[On the problem of more lavish instrumentations which may exceed, for
example, the 16-note limit of the TX802]

> Fortunately, MIDI has allowed the seamless integration of more
> synthesizer modules to accomodate this. Assuming you can keep the
> channel assignments separate, nothing stopping you from running one
> TX802 (or any MIDI synthesizer) per keyboard to keep the polyphony
> limits "isolated". If one part were to exceed the limit, it would
> not affect the other playing parts. A more expensive solution, but
> in my opinion, the most desireable.

For people who want the effect of something like a large organ where
one can literally "pull out the stops," or even say to approximate in
live performance a Venetian ensemble of around 1600 with maybe 12
instruments, this option sounds like a very handy solution.

However, for what I'm usually involved in, no more than four parts
with one or at most two "instruments" on a part -- the 16-note limit
is less of a problem. For example, a typical registration for me might
be one manual as something like a usual organ (e.g. cello + flute),
and the other as a regal (a reed organ which sounds like a crumhorn,
very pleasantly "buzzy").

If I assign 3 sound generators to the regal manual, then this
"instrument" can support up to 6 notes at the same time, leaving
another 3 generators (6 notes) for one of the voices of the "organ" on
the other manual and 2 generators (4 notes) for the other voice.

If I play three-part music (typical of the 13th and 14th centuries),
or four-part music with two parts on each manual, there's likely no
problem at all. If I play four parts on a single manual, then six
notes should still be enough pretty much to allow for any momentary
overlapping of notes -- the voice with only two generators (4 notes)
might occasionally drop out at points of overlap, but this may not be
too noticeable -- at least for me.

If each manual is mapped to an "instrument" with a single "voice" --
e.g. one flavor of "harpsichord" for the lower manual, another for the
upper -- then each instrument has four generators and eight notes,
more than enough for playing four parts at once on either manual.

[On adapting a tuning with more than 12 notes per octave to two
keyboards by mapping different subsets of the tuning to each keyboard
in repeating 12-note octave patterns]

> Unfortunately, this assumes your synthesizers allow full keyboard
> assignment (keyboard-scaling) as opposed to merely octave
> assignment. A lot of what would have been really wicked microtonal
> machines ended up being impeded by the octave-scaling limitation
> (e.g. Kurzweil K150), but with your idea, this obviously wouldn't
> really pose a big problem as you said.

Typically I use this solution for Pythagorean and meantone tunings of
between around 13 and 16 notes where only some of the accidentals
differ -- an equivalent of the "split-key" instruments of the
15th-18th centuries (e.g. G#/Ab, Eb/D#).

A more adventurous application is to have a 24-note Pythagorean or
meantone keyboard with the two keyboards a Pythagorean comma or
meantone diesis apart. The former kind of tuning I call "Xeno-Gothic,"
because it includes a 12-17 note Pythagorean tuning as described in
14th-15th century sources, plus extra notes for things like cadences
from near-7-limit intervals to the stable 3-limit ones. John Chalmers
helped to give me the idea for this when he mentioned to me that a
Pythagorean tuning eventually approximates 7-limit as well as 5-limit
ratios.

The 24-note meantone tuning, specifically 1/4-comma, is basically a
subset of Nicola Vicentino's 31-note archicembalo -- and fortunately,
a very usable subset. As I'll discuss below, Vicentino's own
archicembalo or "superharpsichord" of 1555 provides a partial
precedent for this kind of layout: the first and fourth "ranks" of his
keyboard have the same usual seven diatonic notes, but a meantone
diesis of about 1/5-tone apart.

Incidentally, the early 17th-century keyboards of Scipione Stella and
Fabio Colonna described in the latter's treatise of 1618 have a
feature mentioned in this discussion: some "redundant" keys repeating
the same note in more than one rank for more flexible fingering.
Colonna's design -- for which Stella also claimed credit -- has five
ranks of seven keys, with the ranks tuned in steps of 1/5-tone. Thus
while there are 31 steps per octave on this circulating keyboard
(likely 1/4-comma meantone), there are 35 keys per octave, with some
redundancy.

Of course, there's also 17-tone equal temperament (17-tet), a tuning
which John Chalmers also called to my attention.

> This is precisely the reason beyond the obvious "part" advantages
> that I advocate Jacob Duringer's "Monolith". I request that you
> would take some of your time by checking out
> http://www.electronic-mall.com/heavenbound/
> as I think (and hope) it will tweak your interest as it did
> mine. The email address appears to be dead. I have his current
> e-mail address somewhere if you are interested (just can't locate it
> at the moment).

To this Web page I have a few "first blush" responses.

First of all, while certainly appreciating the special art of live
performance, I would like strongly to affirm the legitimacy and value
of multitrack recording and overdubbing.

Building up a texture in melodic layers, a process analogous to some
medieval practices of layered composition, is an artform which permits
attention to each part. A solo keyboardist playing everything at once,
no matter on how sophisticated a keyboard with however many
simultaneous timbres or "instruments," is a different matter.

Maybe this is in part a question of "keyboardist's pride" -- a bit
like the ancient story of someone who discovered the power of riding a
horse, and was intoxicated by "horse pride." There is indeed a power
in being able to create the effect of a polyphonic ensemble with the
movements of one's own fingers on a keyboard -- whether the musical
engine is a medieval pipe organ or an electronic synthesizer.

However, the subtle cooperation of a group of singers and players, one
to a part, has its own charm and perfection. So does the patient
process of overdubbing (or more recently also sequencing), where again
it is possible to give each melodic line its own focal attention.

With this reservation, I would say that the "x/y" concept of keyboards
with one axis for the usual pitch variations on a single manual, and
another for timbre, is a very interesting one. In fact, the Y axis
might represent differences of either timbre or pitch -- or in some
cases, both at once.

Really, we can go back for example to 16th-century keyboard design to
see this "x and y" concept in practice. A two-manual organ with
12-note manuals tuned in unison -- e.g. a usual meantone -- has timbre
as the y-dimension.

However, Vicentino's archicembalo and also his arciorgano or
"superorgan" he advertised in 1561 use the y-dimension as one of
pitch. In a full 38-note version, what we have is basically two
keyboards with 19 notes per octave each. The usual diatonic ranks --
the white keys on the two manuals -- have the same usual arrangement,
but are tuned a diesis or 1/5-tone (what I call a "fifthtone") apart.

Interestingly, with a two-manual setup, sometimes I use the
y-dimension for pitch and timbre simultaneously. For example, to play
an early 15th-century piece, I might have one manual in a usual
"organlike" registration and a Pythagorean tuning of Eb-G#, which
provides regular and active Pythagorean thirds and sixths. The other
manual in Gb-B, in a contrasting regal registration, provides smooth
Pythagorean diminished fourths, or near-pure schisma thirds and
sixths, for written note-spellings involving sharps (e.g. E-Ab for
written E-G#).

For a German piece from the Buxheimer Organ Book, I find shifting my
hands between the manuals from phrase to phrase so that sometimes the
upper manual plays the upper part(s), and sometimes the lower part(s),
adds a pleasant variety of colors, along with the nuances of
intonation.

In an archicembalo-like tuning of more than about 16 or 17 notes per
octave, however, I tend to use a single registration for both manuals
-- because otherwise the jumping back and forth can get a bit
distracting.

Please let me just add that I see a place both for more or less
"conventional" keyboards which emulate traditional organ/harpsichord
technique (or the technique of the more recently introduced piano),
and for specialized keyboards with various "honeycomb" arrangements.
The latter keyboards may be a technological invitation to new music.

Anyway, I'm really enjoying this dialogue, which gets into the
interaction between synthesizer technology, alternate tunings, and
questions of basic musical values in solo and ensemble performances
alike.

Most appreciatively,

Margo Schulter
mschulter@value.net

> Hello, there, and in Tuning Digest 648, Arthur W. Green
> <goshawk@crosswinds.net> offered some very interesting remarks as part
> of a dialogue on the option of multiple keyboards for microtunable
> synthesizers. Here I'll try to respond on some points, including a Web
> page reference to a keyboard design intended to let solo keyboardists
> approximate the richness of a large ensemble performance.
>
> [On the problem of more lavish instrumentations which may exceed, for
> example, the 16-note limit of the TX802]
>
>> Fortunately, MIDI has allowed the seamless integration of more
>> synthesizer modules to accomodate this. Assuming you can keep the
>> channel assignments separate, nothing stopping you from running one
>> TX802 (or any MIDI synthesizer) per keyboard to keep the polyphony
>> limits "isolated". If one part were to exceed the limit, it would
>> not affect the other playing parts. A more expensive solution, but
>> in my opinion, the most desireable.
>
> For people who want the effect of something like a large organ where
> one can literally "pull out the stops," or even say to approximate in
> live performance a Venetian ensemble of around 1600 with maybe 12
> instruments, this option sounds like a very handy solution.
>
> However, for what I'm usually involved in, no more than four parts
> with one or at most two "instruments" on a part -- the 16-note limit
> is less of a problem. For example, a typical registration for me might
> be one manual as something like a usual organ (e.g. cello + flute),
> and the other as a regal (a reed organ which sounds like a crumhorn,
> very pleasantly "buzzy").
>
> If I assign 3 sound generators to the regal manual, then this
> "instrument" can support up to 6 notes at the same time, leaving
> another 3 generators (6 notes) for one of the voices of the "organ" on
> the other manual and 2 generators (4 notes) for the other voice.
>
> If I play three-part music (typical of the 13th and 14th centuries),
> or four-part music with two parts on each manual, there's likely no
> problem at all. If I play four parts on a single manual, then six
> notes should still be enough pretty much to allow for any momentary
> overlapping of notes -- the voice with only two generators (4 notes)
> might occasionally drop out at points of overlap, but this may not be
> too noticeable -- at least for me.
>
Yes, you're absolutely right. I had completely forgotten that the TX802 (as
can many of Yamaha's multi-timbral synthesizers of that vintage) can set
limitations on its pool of voices (and hence, provide part isolation from
voice stealing). So, the effect would be identical as you point out. Sorry
about that!

> If each manual is mapped to an "instrument" with a single "voice" --
> e.g. one flavor of "harpsichord" for the lower manual, another for the
> upper -- then each instrument has four generators and eight notes,
> more than enough for playing four parts at once on either manual.
Sure, unless you abuse the sustain pedal and utilize "note chorusing" (and
other polyphony-chewers) like I do. :D

>
> [On adapting a tuning with more than 12 notes per octave to two
> keyboards by mapping different subsets of the tuning to each keyboard
> in repeating 12-note octave patterns]
>
>> Unfortunately, this assumes your synthesizers allow full keyboard
>> assignment (keyboard-scaling) as opposed to merely octave
>> assignment. A lot of what would have been really wicked microtonal
>> machines ended up being impeded by the octave-scaling limitation
>> (e.g. Kurzweil K150), but with your idea, this obviously wouldn't
>> really pose a big problem as you said.
>
> Typically I use this solution for Pythagorean and meantone tunings of
> between around 13 and 16 notes where only some of the accidentals
> differ -- an equivalent of the "split-key" instruments of the
> 15th-18th centuries (e.g. G#/Ab, Eb/D#).
>
> A more adventurous application is to have a 24-note Pythagorean or
> meantone keyboard with the two keyboards a Pythagorean comma or
> meantone diesis apart. The former kind of tuning I call "Xeno-Gothic,"
> because it includes a 12-17 note Pythagorean tuning as described in
> 14th-15th century sources, plus extra notes for things like cadences
> from near-7-limit intervals to the stable 3-limit ones. John Chalmers
> helped to give me the idea for this when he mentioned to me that a
> Pythagorean tuning eventually approximates 7-limit as well as 5-limit
> ratios.
Wow, that's slick! I am going to have to take a look at that.

>
> The 24-note meantone tuning, specifically 1/4-comma, is basically a
> subset of Nicola Vicentino's 31-note archicembalo -- and fortunately,
> a very usable subset. As I'll discuss below, Vicentino's own
> archicembalo or "superharpsichord" of 1555 provides a partial
> precedent for this kind of layout: the first and fourth "ranks" of his
> keyboard have the same usual seven diatonic notes, but a meantone
> diesis of about 1/5-tone apart.
>
> Incidentally, the early 17th-century keyboards of Scipione Stella and
> Fabio Colonna described in the latter's treatise of 1618 have a
> feature mentioned in this discussion: some "redundant" keys repeating
> the same note in more than one rank for more flexible fingering.
> Colonna's design -- for which Stella also claimed credit -- has five
> ranks of seven keys, with the ranks tuned in steps of 1/5-tone. Thus
> while there are 31 steps per octave on this circulating keyboard
> (likely 1/4-comma meantone), there are 35 keys per octave, with some
> redundancy.
Interesting, perhaps I should take a look at that to see perhaps how
redundant keys might be applied to some of the instruments mentioned in our
discussion.

It would be nice perhaps to be able to figure out some "reliable" way to
implement it in respect to each tuning by rank in respect to an adjacent
ranks' tunings, rather then merely taking potshots in the manual mapping of
the tuning of each rank in the synthesizer or the controller instrument.
Perhaps, then it would seem so much less arbitrary in nature.

>
> Of course, there's also 17-tone equal temperament (17-tet), a tuning
> which John Chalmers also called to my attention.
>
>> This is precisely the reason beyond the obvious "part" advantages
>> that I advocate Jacob Duringer's "Monolith". I request that you
>> would take some of your time by checking out
>> http://www.electronic-mall.com/heavenbound/
>> as I think (and hope) it will tweak your interest as it did
>> mine. The email address appears to be dead. I have his current
>> e-mail address somewhere if you are interested (just can't locate it
>> at the moment).
>
> To this Web page I have a few "first blush" responses.
>
> First of all, while certainly appreciating the special art of live
> performance, I would like strongly to affirm the legitimacy and value
> of multitrack recording and overdubbing.
>
> Building up a texture in melodic layers, a process analogous to some
> medieval practices of layered composition, is an artform which permits
> attention to each part. A solo keyboardist playing everything at once,
> no matter on how sophisticated a keyboard with however many
> simultaneous timbres or "instruments," is a different matter.
>
While, I don't appreciate the public's lack of distinction between "man and
machine" in regard to the synthesizer, and most certainly an instrument with
any kind of special enhancements to enable more complex live performance by
a given individual is somewhat intrigueing, I fully agree with you. There is
no question in my mind that despite the public's apprehension to studio
techniques (including the use of multitrack recording and overdubbing), the
studio has fostered more innovation via experimentation than I think was
previously possible under a "live performance only" scenario.

> Maybe this is in part a question of "keyboardist's pride" -- a bit
> like the ancient story of someone who discovered the power of riding a
> horse, and was intoxicated by "horse pride." There is indeed a power
> in being able to create the effect of a polyphonic ensemble with the
> movements of one's own fingers on a keyboard -- whether the musical
> engine is a medieval pipe organ or an electronic synthesizer.
>
I will agree as well that the concept of controlling such an array of
instruments can be quite intoxicating at times, and perhaps "horse pride"
nails the concept right on the nose. =)

Overall though, I am rather tired of finding a very conservative and biased
public in regard to electronic music. It seems the public often finds the
synthesizer an icon of the "dehumanization" of music as an art and a
significant reason for the downward spiral of music in the modern day. I
think it may be possibly obvious by the recent large decline of synthesizers
as a mainstay of any kind of mainstream work.

While, I will try not step into this realm of unattractive cynicism for now,
it is clear to me that whether or not Mr. Duringer's invention enables more
sophisticated solo performance may be negligible at this time. The public
will probably never view this invention as anything more than an oddity, as
it is an unwaivering belief at the moment by the public that the synthesizer
does not require the same kind of skill, prowess and stellar musicianship
that a "real instrument played by a real musician does". Until this changes,
I am not sure what can be said, since it seems to me the "problem" with
electronic instruments isn't the instruments themselves.

Hence, why I think the Monolith's other "advantages" should be given a good
look. Since, even if I am completely wrong regarding the public's attitude
(which is certainly very possible), perhaps with the most excellent
combination of today's electronic instruments can we see some sort of
starting block for keyboard evolution in regard to "replacement" schemes
utilizing traditional techniques.

> However, the subtle cooperation of a group of singers and players, one
> to a part, has its own charm and perfection. So does the patient
> process of overdubbing (or more recently also sequencing), where again
> it is possible to give each melodic line its own focal attention.
>
> With this reservation, I would say that the "x/y" concept of keyboards
> with one axis for the usual pitch variations on a single manual, and
> another for timbre, is a very interesting one. In fact, the Y axis
> might represent differences of either timbre or pitch -- or in some
> cases, both at once.
With the flexibility of MIDI and modern synthesizers, and of course the
assignment of X/Y 'matrices' on the instrument itself, it seems to me that
any sort of scheme can be left up the "performer". I rather like this idea,
despite its lack of any sort available standard that I can recall.

Obviously, with this sort of flexibility "on call", I don't think you
necessarily need to box yourself into any particular "application" or
"technique" to find this combination viable, in my opinion.

>
> Really, we can go back for example to 16th-century keyboard design to
> see this "x and y" concept in practice. A two-manual organ with
> 12-note manuals tuned in unison -- e.g. a usual meantone -- has timbre
> as the y-dimension.
>
> However, Vicentino's archicembalo and also his arciorgano or
> "superorgan" he advertised in 1561 use the y-dimension as one of
> pitch. In a full 38-note version, what we have is basically two
> keyboards with 19 notes per octave each. The usual diatonic ranks --
> the white keys on the two manuals -- have the same usual arrangement,
> but are tuned a diesis or 1/5-tone (what I call a "fifthtone") apart.
>
I will definitely agree that it is certainly not the first time this concept
has been seen in practice. But, I do think with the level of technology we
have today with electronic instruments, perhaps it might be wise to try this
concept yet again, as I think it might be more promising this time around.

> Interestingly, with a two-manual setup, sometimes I use the
> y-dimension for pitch and timbre simultaneously. For example, to play
> an early 15th-century piece, I might have one manual in a usual
> "organlike" registration and a Pythagorean tuning of Eb-G#, which
> provides regular and active Pythagorean thirds and sixths. The other
> manual in Gb-B, in a contrasting regal registration, provides smooth
> Pythagorean diminished fourths, or near-pure schisma thirds and
> sixths, for written note-spellings involving sharps (e.g. E-Ab for
> written E-G#).
>
> For a German piece from the Buxheimer Organ Book, I find shifting my
> hands between the manuals from phrase to phrase so that sometimes the
> upper manual plays the upper part(s), and sometimes the lower part(s),
> adds a pleasant variety of colors, along with the nuances of
> intonation.
Yeah, I could imagine. You have obviously done significantly more
experimentation and made more use of this idea than I have, so I am very
pleased to see that it has worked out well for you. :D

>
> In an archicembalo-like tuning of more than about 16 or 17 notes per
> octave, however, I tend to use a single registration for both manuals
> -- because otherwise the jumping back and forth can get a bit
> distracting.
I figured that to be a problem as well. Somewhat of a hassle at times, but
nonetheless mostly effective as it would seem from what you tell me. Cool!

>
> Please let me just add that I see a place both for more or less
> "conventional" keyboards which emulate traditional organ/harpsichord
> technique (or the technique of the more recently introduced piano),
> and for specialized keyboards with various "honeycomb" arrangements.
> The latter keyboards may be a technological invitation to new music.
I do see this as well, and while I hope you will agree with me, as it is my
earnest belief that the standard "keyboard" schemes available perhaps are in
need of some kind of definite evolution if we are to see such an evolution
of any kind in the music we create and perform.

I have received a few replies to our previous posts regarding Starr Labs'
"MicroZone". As you said, my first inclination was to see that it probably
invited considerable experimentation as well. But as it seems we both agree,
it does not make a absolutely logical transformation from the traditional
keyboard schemes available and their respective techniques (what, are these
all so useless now?). So, with the obvious problem of manufacture of the
Monolith aside, I see as much of a place for instruments like the Monolith
as I do for the MicroZone.

>
> Anyway, I'm really enjoying this dialogue, which gets into the
> interaction between synthesizer technology, alternate tunings, and
> questions of basic musical values in solo and ensemble performances
> alike.
As am I! Much appreciated. =)

>
> Most appreciatively,
>
> Margo Schulter
> mschulter@value.net
>
>
>
-- Art