Yahoo Tuning Groups Ultimate Backup tuning About the Terpstra keyboard

on 4/3/04 5:48 AM, Alexandros Papadopoulos <Alexmoog@otenet.gr> wrote:

> Hello
> I think some people from the list are very interested about this
> generalized keyboard
> that was mentioned a while ago.
> I personally can't stand it anymore not being able to play other
> tunings besides what my piano offers , and I cannot find
> any other serious solutions.
> I could have bought the Starrlabs microzone but I don't know exactly
> what will the new keyboard offer , and
> how will be different from the microzone.
>
> I know that some people here are involved in this project , so please ,
> can we have some information , news , updates on the
> progress of the keyboard making?
>
> I am sorry if I am asking too much , but this is torturing me.
> Thanks for your time

Don't forget this keyboard which was discussed some in early march:

on 3/4/04 3:22 PM, gooseplex <cfaah@eiu.edu> wrote:
> http://www.HakenAudio.com/Continuum/

One thing that I don't know whether people realize about this keyboard is
that with suitable software filtering it could be mapped to have any desired
"key" configuration you like. The documentation doesn't make this obvious
and the feature is kind of hidden, but you don't have to use the standard
left-to-right pitch information. However to do something else with it you
have to be able to write some software unless you can find some already
written that does what you need.

However: if anyone buys one of these keyboards and wants to lend it to me
for a little while, maybe I could write the necessary software to implement
a hexagonal-matrix key arrangement (for example). The open question is how
to control the tuning and I don't know whether a single "canonical" system
of finite scope has been worked out that would give say everyone on this
list everything they want without requiring infinite complexity in control.
On the other hand it could be implemented in Max or PD and ongoing tweaking
would then be easy for the end user to do, with a little help.

What I don't know is whether the playing surface permits an overlay on which
key markings could be printed. I will post an email to the manufacturer
about that. Even without it, the existing markings combined with a small
number of front-to-back rows might be adequate for some purposes.

-Kurt

🔗Afmmjr@aol.com

In a message dated 4/3/04 8:43:10 AM Eastern Standard Time,
Alexmoog@otenet.gr writes:

> I know that some people here are involved in this project , so please ,
> can we have some information , news , updates on the
> progress of the keyboard making?
>

Things are still on schedule. The first octave is now being built, before
building all 5 physical octaves. There will be virtual octaves above and below
those 5 octaves. This keyboard will work with SCALA, designed by Manuel op de
Coul. Once the prototype has been turned over to Joel Mandelbaum, another 20
will be built for sale at a reasonable cost. When those are sold, we will
build another 20.

The physical details in the modeling of the keyboard are beyond my complete
purview, involving toolings, materials, dimensions. Once there is an okay on
the single octave by the New York group, the full prototype will be built in
Toronto by Dylan Horvath. There is a team of people making this keyboard
happen.

The design by Sieman Terpstra based on some Bosanquet pioneering makes it a
more natural feel for tunings, thus escaping the entrapment of the conventional
12-tone Halberstadt keyboard. The Terpstra keyboard will make it quite
natural to explore different tunings. I expect the ability to macro quick changes
between tunings. Manuel has seen the key arrangement and has agreed to make
his free Internet tuning software available for the Terpstra keyboard.

We all await progress on this front at its soonest.

best, Johnny Reinhard
AFMM, Director

🔗Alexandros Papadopoulos <Alexmoog@otenet.gr>

This instrument is great , but I am a little worried about the 4ms latency that it has.
I wish I could try it but I don't think anybody use it in Greece...

On Apr 3, 2004, at 10:30 PM, Kurt Bigler wrote:

> on 4/3/04 5:48 AM, Alexandros Papadopoulos <Alexmoog@otenet.gr> wrote:
>
>> Hello
>> I think some people from the list are very interested about this
>> generalized keyboard
>> that was mentioned a while ago.
>> I personally can't stand it anymore not being able to play other
>> tunings besides what my piano offers , and I cannot find
>> any other serious solutions.
>> I could have bought the Starrlabs microzone but I don't know exactly
>> what will the new keyboard offer , and
>> how will be different from the microzone.
>>
>> I know that some people here are involved in this project , so please >> ,
>> can we have some information , news , updates on the
>> progress of the keyboard making?
>>
>> I am sorry if I am asking too much , but this is torturing me.
>> Thanks for your time
>
> Don't forget this keyboard which was discussed some in early march:
>
> on 3/4/04 3:22 PM, gooseplex <cfaah@eiu.edu> wrote:
>> http://www.HakenAudio.com/Continuum/
>
> One thing that I don't know whether people realize about this keyboard > is
> that with suitable software filtering it could be mapped to have any > desired
> "key" configuration you like. The documentation doesn't make this > obvious
> and the feature is kind of hidden, but you don't have to use the > standard
> left-to-right pitch information. However to do something else with it > you
> have to be able to write some software unless you can find some already
> written that does what you need.
>
> However: if anyone buys one of these keyboards and wants to lend it > to me
> for a little while, maybe I could write the necessary software to > implement
> a hexagonal-matrix key arrangement (for example). The open question > is how
> to control the tuning and I don't know whether a single "canonical" > system
> of finite scope has been worked out that would give say everyone on > this
> list everything they want without requiring infinite complexity in > control.
> On the other hand it could be implemented in Max or PD and ongoing > tweaking
> would then be easy for the end user to do, with a little help.
>
> What I don't know is whether the playing surface permits an overlay on > which
> key markings could be printed. I will post an email to the > manufacturer
> about that. Even without it, the existing markings combined with a > small
> number of front-to-back rows might be adequate for some purposes.
>
> -Kurt
>
>
>
> You can configure your subscription by sending an empty email to one
> of these addresses (from the address at which you receive the list):
> tuning-subscribe@yahoogroups.com - join the tuning group.
> tuning-unsubscribe@yahoogroups.com - leave the group.
> tuning-nomail@yahoogroups.com - turn off mail from the group.
> tuning-digest@yahoogroups.com - set group to send daily digests.
> tuning-normal@yahoogroups.com - set group to send individual emails.
> tuning-help@yahoogroups.com - receive general help information.
>
> Yahoo! Groups Links
>
>
>
>
>
>

🔗Kurt Bigler <kkb@breathsense.com>

on 4/3/04 1:25 PM, Alexandros Papadopoulos <Alexmoog@otenet.gr> wrote:

> This instrument is great , but I am a little worried about the 4ms
> latency that it has.

Ah, I didn't notice that. However...

Don't forget that since sound travels about .34 meters per ms, 4ms is about
the time it takes for sound to travel 1.3 meters. If you can move the
speakers 1.3 meters closer to you, you have compensated for the 4ms. Of
course you are piling keyboard latency on top of midi latency, audio
latency, and in a softsynth situation processor scheduling latency,
midi/audio buffering latency, etc. but it is usually possible to get all
those additional factors down to almost 1ms (IIRC) if your computer is
powerful enough.

At the other extreme, organists are acustomed to be sitting 5 to 10 meters
(15 to 30 ms) or even quite a bit more from the pipes they are playing.
When I first had the 10-meter or so experience, my reaction was "how can
anyone play this?" but I made the adjustment easy enough. Not that I am
saying there is no *difference* as a result, but this experience was very
interesting to me considering my initial reaction. As an aside, I believe
there is some relationship between being able to listen to the music I
create from a "distance" in time and my ability to improvise. But that's
another topic.

When a badly-implemented MIDI softsynth is involved (such as one that I
wrote myself) sometimes inconsistent latencies can result, and I have found
myself to be pretty thrown off by that, but reducing *audio* buffer sizes
also limits the effect of that problem.

-Kurt

🔗Kurt Bigler <kkb@breathsense.com>

on 4/3/04 5:02 PM, Kurt Bigler <kkb@breathsense.com> wrote:

> on 4/3/04 1:25 PM, Alexandros Papadopoulos <Alexmoog@otenet.gr> wrote:
>
>> This instrument is great , but I am a little worried about the 4ms
>> latency that it has.
>
> Ah, I didn't notice that. However...
>
> [blah blah blah]

I just looked more closely at the Continuum specifications and I see that
4ms is the "scan interval". This is worse than a 4ms latency because this
means the latency will be inconsistent by as much as 4ms. Curiously the
scan interval is the same for the smaller and larger versions of this
keyboard. Perhaps they do some tricks to reduce the effect of the 4ms, or
perhaps 4ms is the worst case. I sent another message to them hoping to
clarify these issues.

FWIW, I looked at the exact (numeric) details regarding inconsistent latency
in my own softsynth. When the maximum inconsistency in latency (roughly
analogous to scan interval) was 12 ms this was often a problem for me. When
I reduced it to 6 ms a couple months ago it entirely stopped bothering me.
However, when I reduced it further to 1.5 ms I really appreciated the
improvement, but this so increased the CPU load that I put it back to 6 ms
for the time being.

-Kurt

🔗Jon Szanto <JSZANTO@ADNC.COM>

Kurt,

--- In tuning@yahoogroups.com, Kurt Bigler <kkb@b...> wrote:
> analogous to scan interval) was 12 ms this was often a problem for
> me. When > I reduced it to 6 ms a couple months ago it entirely
> stopped bothering me. However, when I reduced it further to 1.5 ms
> I really appreciated the improvement, but this so increased the CPU
> load that I put it back to 6 ms for the time being.

I would profer that tolerance of relative latency issues is both context sensitive as well as part of the performer as well. I have spent my life, more than anything else, as a percussionist: when I strike something (and strike it at the time *I* want to), there is 0.0 ms latency.

When I do music that must be rhythmically accurate, no amount of latency is acceptable. When there is, as in the inputting stage, I have to fix it later (either in a sequence or moving things digitally in a mix) to satisfy my ear.

I have always equated timing with tuning: they both need, at various times, absolute accuracy. When that fails, it weakens the music.

Cheers,
Jon

🔗Kurt Bigler <kkb@breathsense.com>

on 4/3/04 11:58 PM, Jon Szanto <JSZANTO@ADNC.COM> wrote:

> Kurt,
>
> --- In tuning@yahoogroups.com, Kurt Bigler <kkb@b...> wrote:
>> analogous to scan interval) was 12 ms this was often a problem for
>> me. When > I reduced it to 6 ms a couple months ago it entirely
>> stopped bothering me. However, when I reduced it further to 1.5 ms
>> I really appreciated the improvement, but this so increased the CPU
>> load that I put it back to 6 ms for the time being.
>
> I would profer that tolerance of relative latency issues is both context
> sensitive as well as part of the performer as well. I have spent my life, more
> than anything else, as a percussionist: when I strike something (and strike it
> at the time *I* want to), there is 0.0 ms latency.
>
> When I do music that must be rhythmically accurate, no amount of latency is
> acceptable. When there is, as in the inputting stage, I have to fix it later
> (either in a sequence or moving things digitally in a mix) to satisfy my ear.
>
> I have always equated timing with tuning: they both need, at various times,
> absolute accuracy. When that fails, it weakens the music.

I understand that this is certainly dependent on a lot of things. But I
want to give feedback about this "absolute accuracy". Why? Well there is
really no use in my saying any of this unless it makes you think of
something you hadn't thought of before, and perhaps results in your giving
an even better explanation or argument on this topic. So bear with me and
apologies to everyone if this misses.

0.0 ms latency will not cut it technically. 0 ms perhaps. 0.0 ms by
technical convention implies an accuracy to the tenth of a ms. A tenth of a
ms corresponds to about 1.3 inches of distance travelled by sound. If the
distance of your ear from the drum varies by 1.3 inches, you have altered
the latency in your own relationship between your action and what you hear
by a tenth of a ms. So my point is again that everything is relative.

At first look what really should matter most is the latency that the
audience experiences. If you are playing in an orchestra, and you get into
being absolutely accurate, or even accurate to the nearest tenth of a
millisecond, then things could get really silly. If you had only one person
in the audience, and you placed all the musicians at equal distance (to the
nearest 1.3 inches) from the audience person, then you could achieve this
kind of accuracy if all the musicians were basing their timing on visual
cues (e.g. a conductor) that they all saw. WIth more than one person in the
audience, you begin to get into problems achieving consistent latencies from
the audience's perspective. With a large orchestra and a large audience,
the problems are maximized.

So since technically accurate latencies for the audience are out of the
question, then that idea has to be thrown out. Maybe what really matters is
something else.

Another place to start is to honor a less technical (but no less essential)
aspect of this, which might have some relation to what you mean by "absolute
accuracy", which has to do the the whole web of relationships between
everything involved in a musical performance. The whole scenario, with all
its imperfections and inconsistencies in latency is being dynamically
integrated by performers that are in potentially dynamic relationship with
the audience. The synergy of such a situation can be "right" in ways that
neither measurement nor science can account for, if for no other reasons
then just because it is complex beyond any possibility of analysis. This is
the magic of the moment. This can be talked about in terms like "mutual
entrainment", and this can help to understand some aspects of what goes on,
but all understanding can pale in comparison to the richness of actual
experience. But this is not about any kind of technical accuracy. It is
full of a bunch of people integrating all kinds of latencies that simply
occur naturally in the air and in the nervous system. If you are
experienced in a particular context and something is altered, it may throw
you. So introducing "artificial" latencies that are unfamiliar can make a
mess of an individual musician's integrated experience. However this does
not mean that a new integration can not be achieved, with "new latencies",
but just that experiential adjustments are needed when things change.
However, an experienced musician who has perfected his/her art may well not
appreciate having an additional term thrown into the equation.

Now having said all that, if you were to alter a well-tuned performance
situation by inserting an artificial latency of 1ms into it somewhere, I can
well believe it might have an effect on the music and even on the musician's
ability to perform. A tenth of a ms? Harder to believe. Or rather the
effect that a fraction of a ms latency might have would probably disappear
into the kinds of things that will be different in every room, and every
time the same instruments are set up in the same room. Ok, I admit it,
every difference *is* a difference. Which ones matter? Which ones are just
part of the quality of the moment, whether "right" or "wrong"? I can't
really claim to know. I've given it the above kind of thought, but then my
inner scientist bows to the truth of the musical moment.

Respectfully,
Kurt

> Cheers,
> Jon

🔗Jon Szanto <JSZANTO@ADNC.COM>

🔗gooseplex <cfaah@eiu.edu>

--- In tuning@yahoogroups.com, Kurt Bigler <kkb@b...> wrote:
> on 4/3/04 5:48 AM, Alexandros Papadopoulos
<Alexmoog@o...> wrote:
>
> > Hello
> > I think some people from the list are very interested about
this
> > generalized keyboard
> > that was mentioned a while ago.
> > I personally can't stand it anymore not being able to play
other
> > tunings besides what my piano offers , and I cannot find
> > any other serious solutions.
> > I could have bought the Starrlabs microzone but I don't know
exactly
> > what will the new keyboard offer , and
> > how will be different from the microzone.
> >
> > I know that some people here are involved in this project , so
please ,
> > can we have some information , news , updates on the
> > progress of the keyboard making?
> >
> > I am sorry if I am asking too much , but this is torturing me.
> > Thanks for your time
>
> Don't forget this keyboard which was discussed some in early
march:
>
> on 3/4/04 3:22 PM, gooseplex <cfaah@e...> wrote:
> > http://www.HakenAudio.com/Continuum/
>
> One thing that I don't know whether people realize about this
keyboard is
> that with suitable software filtering it could be mapped to have
any desired
> "key" configuration you like. The documentation doesn't make
this obvious
> and the feature is kind of hidden, but you don't have to use the
standard
> left-to-right pitch information. However to do something else
with it you
> have to be able to write some software unless you can find
some already
> written that does what you need.
>
> However: if anyone buys one of these keyboards and wants to
lend it to me
> for a little while, maybe I could write the necessary software to
implement
> a hexagonal-matrix key arrangement (for example). The open
question is how
> to control the tuning and I don't know whether a single
"canonical" system
> of finite scope has been worked out that would give say
everyone on this
> list everything they want without requiring infinite complexity in
control.
> On the other hand it could be implemented in Max or PD and
ongoing tweaking
> would then be easy for the end user to do, with a little help.
>
> What I don't know is whether the playing surface permits an
overlay on which
> key markings could be printed. I will post an email to the
manufacturer
> about that. Even without it, the existing markings combined
with a small
> number of front-to-back rows might be adequate for some
purposes.
>
> -Kurt

Hello Kurt!

Maybe you have heard back from Haken Audio about this
already, but since you mention it, the idea of mapping tunings
onto the Continuum keyboard is something I discussed recently
with the inventor, Lippold Haken. In fact, during my visit, Lippold
wrote some code in Kyma which mapped a 19 tone grid to the
surface.

You are correct in saying that software of this kind could be
written in MAX to achieve discrete step microtunings through
MIDI or digital audio. I suggested that a cross-platform software
package should come with the instrument to allow the user this
kind of control. Since I have experience with programming for my
own microtonal instruments, I offered to develop this software for
the Continuum. So, at least the idea has a possibility of being
fully developed.

One thing to keep in mind if you decide to write your own
software is that the keyboard does not handle 'unisons', or
front-to-back vertically aligned pressure points. Also, according
to the inventor, the response of the front-to-back sensing is such
that tuning grids will be limited to three rows of virtual keys.

The surface does not lend itself to overlays, although we
discussed possibly swapping out the playing surface. The
problem there is that swapping the surface requires
manufacturer service; it is not a recommended user-servicable
operation.

On the issue of latency, having played the instrument myself I
can say it is _not a problem in the least. The Continuum can
control very complex synthesis as well as tuning without any
difficulty at all. Four milliseconds is fast; remember that the
perceptual threshold of annoyance is really around 30 ms
anyway.

Best regards,
Aaron

🔗Kurt Bigler <kkb@breathsense.com>

on 4/4/04 5:34 PM, gooseplex <cfaah@eiu.edu> wrote:

> Hello Kurt!
>
> Maybe you have heard back from Haken Audio about this
> already, but since you mention it, the idea of mapping tunings
> onto the Continuum keyboard is something I discussed recently
> with the inventor, Lippold Haken. In fact, during my visit, Lippold
> wrote some code in Kyma which mapped a 19 tone grid to the
> surface.
>
> You are correct in saying that software of this kind could be
> written in MAX to achieve discrete step microtunings through
> MIDI or digital audio. I suggested that a cross-platform software
> package should come with the instrument to allow the user this
> kind of control. Since I have experience with programming for my
> own microtonal instruments, I offered to develop this software for
> the Continuum. So, at least the idea has a possibility of being
> fully developed.

Hah! I just suggested the same thing to them in my last message, though as
a third-party development.

> One thing to keep in mind if you decide to write your own
> software is that the keyboard does not handle 'unisons', or
> front-to-back vertically aligned pressure points. Also, according
> to the inventor, the response of the front-to-back sensing is such
> that tuning grids will be limited to three rows of virtual keys.

Oh, well that kind of shoots it a bit in the foot. Though there are still
possibilities, it seems like there are a growing number of points against
it, including the overlay problem...

> The surface does not lend itself to overlays, although we
> discussed possibly swapping out the playing surface. The
> problem there is that swapping the surface requires
> manufacturer service; it is not a recommended user-servicable
> operation.
>
> On the issue of latency, having played the instrument myself I
> can say it is _not a problem in the least. The Continuum can
> control very complex synthesis as well as tuning without any
> difficulty at all. Four milliseconds is fast; remember that the
> perceptual threshold of annoyance is really around 30 ms
> anyway.

See my previous messages. That threshold may apply for latency in limited
contexts, but certainly not for latency quantization, which is what scan
rate relates to. This is pretty easy to demonstrate and I have been through
it many times.

Thanks for chiming in here. I'm in good company.

-Kurt

> Best regards,
> Aaron

🔗Jon Szanto <JSZANTO@ADNC.COM>

🔗Alexandros Papadopoulos <Alexmoog@otenet.gr>

Thank you very for replying
One thing I did not understand :
You said that there will be an one octave prototype , and the final
model will have five.
The virtual octaves above these five will be on the prototype or the
finished model?

On Apr 3, 2004, at 11:38 PM, Afmmjr@aol.com wrote:

>
> Things are still on schedule. The first octave is now being built,
> before building all 5 physical octaves. There will be virtual octaves
> above and below those 5 octaves. This keyboard will work with SCALA,
> designed by Manuel op de Coul. Once the prototype has been turned
> over to Joel Mandelbaum, another 20 will be built for sale at a
> reasonable cost. When those are sold, we will build another 20.
>
> The physical details in the modeling of the keyboard are beyond my
> complete purview, involving toolings, materials, dimensions. Once
> there is an okay on the single octave by the New York group, the full
> prototype will be built in Toronto by Dylan Horvath. There is a team
> of people making this keyboard happen.
>
> The design by Sieman Terpstra based on some Bosanquet pioneering makes
> it a more natural feel for tunings, thus escaping the entrapment of
> the conventional 12-tone Halberstadt keyboard. The Terpstra keyboard
> will make it quite natural to explore different tunings. I expect the
> ability to macro quick changes between tunings. Manuel has seen the
> key arrangement and has agreed to make his free Internet tuning
> software available for the Terpstra keyboard.
>
> We all await progress on this front at its soonest.
>
> best, Johnny Reinhard
> AFMM, Director
>

🔗gooseplex <cfaah@eiu.edu>

--- In tuning@yahoogroups.com, "Jon Szanto" <JSZANTO@A...>
wrote:
> Aaron,
>
> --- In tuning@yahoogroups.com, "gooseplex" <cfaah@e...>
wrote:
> > Four milliseconds is fast; remember that the
> > perceptual threshold of annoyance is really around 30 ms
> > anyway.
>
> I respectfully disagree. "Annoyance" is hardly a defineable
place, and I'd love to see two musicians attempt to play
synchonous music, one with a responsive instrument (acoustic
or not) and one saddled with a 30 ms latency delay. Having
played a lifetime of rhythmic music (for instance, Reich's
"Drumming"), I can attest that my annoyance level is lower than
that!
>
> Cheers,
> Jon

Hi Jon,

I doubt it. I'm also a percussionist, and have also performed
pieces by Steve Reich such as Music for Pieces of Wood. So,
one drummer to another, I respectfully disagree.

Look, machines do not have 0 ms accuracy and real people
certainly do not have 0 ms accuracy. Your claim to 0 ms accuracy
is unreal. Quantized measurements of your own playing would
reveal offsets in the order of tens of milliseconds all over the
place. Granted, you are talking about response and I am talking
about 'rhythmic feel', but don't you see how the two are
inextricably related?

Say you are playing a ride cymbal with the butt end of a wood
stick, and someone else is playing a concert tom with a large
soft mallet. In that case your instrument has a sharp, crisp
attack, and his doesn't. In fact, with a soft mallet, there is latency
in the attack. But, as you know, in a situation like this you both
will adjust and perform together just fine.

Or think about playing a tam tam. You have to warm it up for it to
even speak. Lots of latency there - but you will hit it to speak on
cue. The point is that whatever the response characteristics of
the instrument are, you adjust, and, trust me, a few ms is
_nothing. We can't even perceive a millisecond.

But I guess the main point is this: I'll say again what I've said
already; I played the instrument and this latency thing is simply
not a problem, no matter what you believe about milliseconds
schmilliseconds.

Regards,
Aaron Hunt

🔗Jon Szanto <JSZANTO@ADNC.COM>

Aaron,

--- In tuning@yahoogroups.com, "gooseplex" <cfaah@e...> wrote:
> So, one drummer to another, I respectfully disagree.

That's OK, because besides our similarities we also have our differences, and part of that may simply be in how exacting we like things to line up.

> Look, machines do not have 0 ms accuracy and real people
> certainly do not have 0 ms accuracy. Your claim to 0 ms accuracy
> is unreal.

I may have phrased that poorly, but what I was getting at is very simple: when the head of my malled touches the bar of the marimba, the sound is made. There is *no* latency there. And I was relating this to the frustration of my first efforts, going back a couple of years, to finding tools for recording/electronic music that _approached_ that level of latency.

When I first tried using softsynths/samplers, the delay between when a key was pressed and when the sound was played was, to put it mildly, unacceptable. It is only now possible for me, with improvements in both software engineering and better (and faster) hardware that it is coming to be workable. I am sure that the latency on some of the first work I was trying to do was in the 30ms range, and it was impossible to record one track and play in sync with it.

Part of this might also be a fluxuation, where polyphony would induce further latency issues. All I can say is that from a lifetime (at least since I was *very* young) of having a sound come When I Want It, any delays, to the degree that is delayed, is not something I relish.

And I certainly acknowledge that there is some point between 0 and x ms that is a cutoff point, whereby any smaller delay wouldn't affect either performing or listening, and certainly not every note I've ever played is accurate to a 1 ms grid. The beauty of being human, I guess! :)

> Granted, you are talking about response and I am talking
> about 'rhythmic feel', but don't you see how the two are
> inextricably related?

Hopefully, in the above explanation I gave, you can see that I was very much focusing on response. And if one plays with a given feel, you certainly don't want the input device to rob you of letting you put notes Right Where You Want Them.

> In fact, with a soft mallet, there is latency in the attack.

No, there is a difference in the attack envelope, but if both stick and mallet hit at the same time, they hit at the same time.

> But, as you know, in a situation like this you both
> will adjust and perform together just fine.

Oh, I have to adjust all the time. Years of being the furthest from the podium, and having to play ahead of any visual beat, as well as ahead of (aurally) where the front of an orchstra is placing their notes, simply to come out in unison (rhythmically) at the lip of the stage - I've had that in spades. And it's awful, to tell the truth.

I've often mentioned to others in the group "what if you had to play progressively sharper the further back in the ensemble, just so it would be in tune at the front?". It seems ludicrous, but that is very much the (rhythmic) scenario I face daily. It varies according to hall and studio acoustics, but distance is distance.

> The point is that whatever the response characteristics of
> the instrument are, you adjust, and, trust me, a few ms is
> _nothing.

Trust? Wellll, I'll trust my experience too. But as for adjusting, sure - but adjusting is a compromise.

> I played the instrument and this latency thing is simply
> not a problem, no matter what you believe about milliseconds
> schmilliseconds.

On that, then, I will take your word. I'm curious - without the benefit of having it there to work with:

1. would it work in situations where you had to play exactly in time, say, in a multitrack recording situation?

2. would the latency accumulate as one became more polyphonic in performance?

Thanks for the good thoughts, and maybe I've strayed too far from tuning at this point. If so, we can talk over at metatuning...

Cheers,
Jon

🔗Afmmjr@aol.com

🔗Kurt Bigler <kkb@breathsense.com>

on 4/5/04 9:46 AM, gooseplex <cfaah@eiu.edu> wrote:

>>> Four milliseconds is fast; remember that the
>>> perceptual threshold of annoyance is really around 30 ms
>>> anyway.
>> I respectfully disagree. "Annoyance" is hardly a defineable
>> place, and I'd love to see two musicians attempt to play
>> synchonous music, one with a responsive instrument (acoustic
>> or not) and one saddled with a 30 ms latency delay. Having
>> played a lifetime of rhythmic music (for instance, Reich's
>> "Drumming"), I can attest that my annoyance level is lower than
>> that!
>
> Hi Jon,
>
> I doubt it. I'm also a percussionist, and have also performed
> pieces by Steve Reich such as Music for Pieces of Wood. So,
> one drummer to another, I respectfully disagree.
>
> Look, machines do not have 0 ms accuracy and real people
> certainly do not have 0 ms accuracy. Your claim to 0 ms accuracy
> is unreal. Quantized measurements of your own playing would
> reveal offsets in the order of tens of milliseconds all over the
> place. Granted, you are talking about response and I am talking
> about 'rhythmic feel', but don't you see how the two are
> inextricably related?
>
> Say you are playing a ride cymbal with the butt end of a wood
> stick, and someone else is playing a concert tom with a large
> soft mallet. In that case your instrument has a sharp, crisp
> attack, and his doesn't. In fact, with a soft mallet, there is latency
> in the attack. But, as you know, in a situation like this you both
> will adjust and perform together just fine.
>
> Or think about playing a tam tam. You have to warm it up for it to
> even speak. Lots of latency there - but you will hit it to speak on
> cue. The point is that whatever the response characteristics of
> the instrument are, you adjust, and, trust me, a few ms is
> _nothing. We can't even perceive a millisecond.
>
> But I guess the main point is this: I'll say again what I've said
> already; I played the instrument and this latency thing is simply
> not a problem, no matter what you believe about milliseconds
> schmilliseconds.

I'd like to clarify the issue about latency versus what I was calling
"quantization of latency" and the related issue of inconsistent latency.
The distinction between a certain amount of latency and a certain amount of
potential inconsistency in latency is a distinction that has been pretty
much ignored, making some aspects of the discussions that ensued possibly
not that meaningful.

As I have said even a full note-value's worth of latency can be assimilated
by an organist, and this is a *lot* of milliseconds. That is very clear.
My main experience with this probably involved a distance of about 15 feet
from the pipes, IIRC, which would translate to about 15 ms. This situation
involves no "quantization" of latency, just latency itself. Many organists
are able to play with a much larger distance involved. This is particularly
true when there is an antiphonal (if I have the term right) division located
at the opposite end of the building but playable from the same console,
which is *not* generally located in the middle. (Indeed the antiphonal
should not be played simultaneously with the main organ or major latency
problems occur for the listeners.) Cathedrals can be pretty big, and her we
are talking about latencies of 100 ms or more being quite ordinary.

Years ago, I had a software situation in which key events were acted on only
20 times per second (thus 50 ms passes between the times when key events are
sampled). This is not just a matter of latency, but quantization of
latency. With such a system, if you happen to play one note a nanosecond
after the system happens to check for key events the latency in this case is
50 ms more than when you happen to hit a key a nanosecond before the system
checks for key events. I am calling this a "latency quantization" of 50 ms.
Believe me this was intolerable--it was very hard to play at all, and I had
*plenty* of opportunity to learn to adjust to this and I simply could not.

When the technology got better I was able to sample key events 43 times per
second, creating a latency quantization of about 23 ms. Although this was
better, with fast pieces I had exactly the same problem I had with the 50 ms
quantization: It became *extremely* confusing to play because the latency
varied from one note to the next. The issue here is not latency but
inconsistent latency caused by key presses being sampled too infrequently.

So I'm pretty sure that even if 30 ms latency might be ok in some
situations, that does not say anything at all to the question of whether 30
ms worth of inconsistency in latency is ok. Two different questions.

When we are talking about the Continuum fingerboard and the 4 ms "scan
interval" we are not talking about a 4 ms latency. We are talking about a
quantization of latency which can result in latencies on subsequent notes
being inconsistent by as much as 4ms. A 4ms scan interval will result in an
average scanning latency of 2ms but the latency will vary from 0ms to 4ms,
and as far as the musician is concerned, it probably appear to vary
"randomly". The question of whether 4ms of latency is ok is a completely
different question to whether a 4ms scan interval is ok. Note that if the
pressure sensitivity of the Continuum is used directly to create the attack
of the note, the attack thus being very slow as a result, in that case a 4ms
scan interval will be no issue at all, I'd certainly believe. But if the
same device is used to create a percussive attack that is another thing
entirely.

I hope that clarifying the distinction between simple latency and
inconsistency of latency might help avoid some confusion. No point in
talking about two different things by the same name.

-Kurt

🔗gooseplex <cfaah@eiu.edu>

🔗Kurt Bigler <kkb@breathsense.com>

🔗Jon Szanto <JSZANTO@ADNC.COM>

🔗Kurt Bigler <kkb@breathsense.com>

on 4/6/04 8:05 AM, gooseplex <cfaah@eiu.edu> wrote:

> --- In tuning@yahoogroups.com, Kurt Bigler <kkb@b...> wrote:
>> on 4/5/04 9:46 AM, gooseplex <cfaah@e...> wrote:
> ....
>> I hope that clarifying the distinction between simple latency and
>> inconsistency of latency might help avoid some confusion. No
> point in
>> talking about two different things by the same name.
>>
>> -Kurt
>
>
> Hi Kurt!
>
> Turns out I had missed your previous postings about this; we
> were indeed discussing two (or three) different kinds of latency.
> Might it be better to refer to this type of 'inconsistency latency' in
> terms of statistics? Either 'precision' or 'accuracy' or perhaps
> both, in the quantization, would seem to apply. Or perhaps
> confining the language to 'quantization resolution' would reduce
> confusion? The latter seems more direct to me.

Yes well we are all informed by what has come before and when it comes down
to it we just have to pick the words that seem best! I think "quantization
resolution" confuses me by introducing some slight conceptual redundancy.
It sure would be nice to have simple words for all this but that is probably
a luxury for something which only a small population is talking about.
Language is clumsy on the frontier, you might say, and we can refine it to
the degree that it becomes important to do so. So we'll see.

By the way if we continue on latency, let's start a new thread. I'd like to
take the thread with this title back to the issue of the Continuum. At some
point I will probably have the chance to try one of these devices and I want
to be prepared to do sensible things with any alternate mappings. The
limitation of the device which I might call the front-to-back-monophonic
limitation which you described and which Lippold Haken also told me about is
something I'd like to understand well enough to understand the conclusions
you drew.

Given that limitation, I came up with another mapping idea. It is a little
disappointing perhaps compared to a nice hexagonal layout, but here it is.
Divide the Continuum left-to-right into discrete semitones (corresponding to
the visual cues that already exist), with no left-to-right pitch variability
within the semitone. Then create a front-to-back range of variability of
one or two semitones in each of those slots. This would permit transparent
transposition by simple hand movements and would also offer in effect a
"longer" pitch strip so that it would be eaiser to get close to a target
pitch visually. And if I understand the limitation right, this pretty much
eliminates any problem with chording unless you need several notes within a
semitone. What do you think?

I'd like to work out some useful mappings on this instrument so that when I
get my hands on one (which I think I can arrange some day) I can invite a
bunch of tuning group people over to try it out.

-Kurt

🔗Mark Smart <marksmartus@yahoo.com>

Hi all.
I just joined the group because I heard that the Continuum was being
discussed here. I am the guy that did a lot of the audio demos on the
Continuum web site and helped demonstrate it at the NAMM show, so
take my opinion with that in mind. But FWIW, I will say that I'm no
salesman and I try to be objective about these things. I would make a
lousy salesman because I would tell people everything that's wrong
with the gear I was trying to sell them.

The respose time on the Continuum feels instantaneous to me; very
very fast. I tend to be pretty picky about these things, especially
when it comes to guitar synths (I am mainly a guitarist). The delay
on most guitar synths drives me totally nuts. That's why I am so
fascinated by the old Roland GR-300 and have spent a lot of time
studying it...it has no percepitble delay on the high strings. To me
the Continuum feels like this when I am playing a hardware MIDI synth
and not a software synth on my slow old computer.

The Continuum's biggest weakness is that it is not nearly as accurate
in the Y direction (front-back) as it is in the X (left-right)
direction. Lippold is still working on this, but it's a very
difficult problem due to the mechanics involved.

I was hoping I could divide the Y direction into six zones, and have
the synth change pitch by a 4th every time you switched into a new
zone. This would make it like playing single note lines on the melody
half of the Chapman Stick. The Y measurement is not accurate enough
on the Continuum to do this.

Fortunately, if you are not trying to do zone mapping for pitch, the
Y direction is the one that is the least important in terms of
expressiveness. Lippold likes to compare it to bowing on a violin. If
you bow at the bridge, it sounds very different than when you bow
near the fingerboard, but if you move the bow only 2mm, you will not
hear a difference.

Response speed on it is simply not a problem for me, even when
playing highly rhythmic music like the funky demos I did. And the
dynamic response in the Z direction (up/down--how hard you are
pushing) is excellent as well, especially since he started using a
thicker sheet of neoprene (wetsuit material) for the playing surface.
Check out my "Summertime" mp3 on the Haken Audio site to hear what
can be done with the dynamics.

I hope this helps.

Mark Smart
www.marksmart.net

🔗gooseplex <cfaah@eiu.edu>

> Or perhaps
> > confining the language to 'quantization resolution' would
reduce
> > confusion? The latter seems more direct to me.
>
> Yes well we are all informed by what has come before and
when it comes down
> to it we just have to pick the words that seem best! I think
"quantization
> resolution" confuses me by introducing some slight
conceptual redundancy.

Quantization is the process of assigning a discrete value, and
resolution refers to the number of bits used in the value, so it
isn't redundant to talk about quantization resolution, although I
see why conceptually it may seem so, and maybe it's not an
appropriate term for what you are talking about.

> By the way if we continue on latency, let's start a new thread.

You might try discussing this issue on the makemicromusic list,
since it's not really a tuning issue

I'd like to
> take the thread with this title back to the issue of the
Continuum. At some
> point I will probably have the chance to try one of these devices
and I want
> to be prepared to do sensible things with any alternate
mappings. The
> limitation of the device which I might call the
front-to-back-monophonic
> limitation which you described and which Lippold Haken also
told me about is
> something I'd like to understand well enough to understand
the conclusions
> you drew.

As I understand it (Mark please correct me!) it's a differential
calculation based on the values of two sensors at the 'front' and
'back' under the playing surface. The website offers info on how
that works.

128 front-to-back values are possible and even demonstrable,
but the problem is simple physics; because there are so many
variables involved:

1) the physical parameters of the finger impact will vary
2) friction of the rods on the posts along with the spring
response, and the parabolic caluculation of the sensors will vary
3) taking the difference between them which accumulates the
variances for both the front and the back

> Given that limitation, I came up with another mapping idea. It
is a little
> disappointing perhaps compared to a nice hexagonal layout,
but here it is.
> Divide the Continuum left-to-right into discrete semitones
(corresponding to
> the visual cues that already exist), with no left-to-right pitch
variability
> within the semitone. Then create a front-to-back range of
variability of
> one or two semitones in each of those slots. This would
permit transparent
> transposition by simple hand movements and would also offer
in effect a
> "longer" pitch strip so that it would be eaiser to get close to a
target
> pitch visually. And if I understand the limitation right, this pretty
much
> eliminates any problem with chording unless you need several
notes within a
> semitone. What do you think?

Sounds interesting, as long as the front-to-back is 2 or 3 zones
only. Left-to-right discrimination is also limited at the halfstep
distance marked on the surface. Mark Smart can probably supply
more insight into these limitations.

Regards,
Aaron Hunt

🔗Jon Szanto <JSZANTO@ADNC.COM>

🔗Mark Smart <marksmartus@yahoo.com>

--- In tuning@yahoogroups.com, "gooseplex" <cfaah@e...> wrote:
> > Given that limitation, I came up with another mapping idea. It
> is a little
> > disappointing perhaps compared to a nice hexagonal layout,
> but here it is.
> > Divide the Continuum left-to-right into discrete semitones
> (corresponding to
> > the visual cues that already exist), with no left-to-right pitch
> variability
> > within the semitone. Then create a front-to-back range of
> variability of
> > one or two semitones in each of those slots. This would
> permit transparent
> > transposition by simple hand movements and would also offer
> in effect a
> > "longer" pitch strip so that it would be eaiser to get close to a
> target
> > pitch visually. And if I understand the limitation right, this
pretty
> much
> > eliminates any problem with chording unless you need several
> notes within a
> > semitone. What do you think?

Given the Continuum's lack of accuracy and repeatability in the Y
direction, I don't think this would work very well. Fortunately, I
don't think it's necessary. Round Initial Pitches mode is probably a
better solution.

If you have a sophisticated, configurable synth on the other end,
like Kyma or Reaktor, you could even set it up so that the synth
would play just pitches when it received Note On messages with 0
pitch bend. Then when you use Round Initial Pitch mode, you would
always get just pitches when you first attack the notes, then you
could bend and slide them from there.

Dang, I have to try that out with Reaktor...

Or how about this...you define an octave or two at the bottom of the
Continuum that controls what key the just intonation is in. If you
hit a C in that octave, the synth with round everything to the
nearest just pitch in C. Then you hit an F# in the low octave when
you want to play an F# major chord without it sounding totally awful.

(You guys probably talk about this on here all the time)

Of course...you could do this on any keyboard synth and you don't
need a Continuum. Definitely the easy way out. The only thing the
Continuum gains you is bending, sliding, and vibrato.

If you were a REAL hardcore microtonalist, you would practice the
Continuum for several years until you could do it with no
rounding...then act really condescending to all the "rounding wimps".

Mark Smart
www.marksmart.net

🔗Mark Smart <marksmartus@yahoo.com>

--- In tuning@yahoogroups.com, "gooseplex" <cfaah@e...> wrote:
> As I understand it (Mark please correct me!) it's a differential
> calculation based on the values of two sensors at the 'front' and
> 'back' under the playing surface. The website offers info on how
> that works.
>
> 128 front-to-back values are possible and even demonstrable,
> but the problem is simple physics; because there are so many
> variables involved:
>
> 1) the physical parameters of the finger impact will vary
> 2) friction of the rods on the posts along with the spring
> response, and the parabolic caluculation of the sensors will vary
> 3) taking the difference between them which accumulates the
> variances for both the front and the back

You've pretty much got it. Actually there are two separate issues
here:

1. You can't play two different notes with the same X value, i.e. no
unisons. This is a result of the fact that there are rods running
from front to back, and the computer has no way to tell the
difference between a rod with one finger on it and a rod with two
fingers on it.

2. The accuracy and repeatability of the Y measurement of a single
note being played on a rod is limited. This is because of the factors
you mentioned above. There is the problem of the rods binding on the
springs and posts, and other problems related to the fact that the
fields of the magnets on the ends of the rods interact somewhat; the
bar ends repel each other. Also, each magnet's field has a slight
effect on its neighbor's Hall Effect sensor. Lippold has done very
well working around these things, but they are big problems which are
not likey to go away. The result is that you get some variation in Y
value even when you hit the exact same spot repeatedly with your
finger. This is still not very noticeable unless you are controlling
some very sensitive parameter like pitch with the Y value.

> Sounds interesting, as long as the front-to-back is 2 or 3 zones
> only. Left-to-right discrimination is also limited at the halfstep
> distance marked on the surface.

No, this is totally untrue. The X-direction measurement is EXTREMELY
accurate. Lippold says it's accurate to 0.1mm and I believe him. Left-
to-right is so good that you will never get any unexpected results or
hear discrete steps when you slide notes. What's amazing is that this
remains true even if you set it up for a MIDI Pitch Bend range of 96
half steps. Lippold uses all 14-bits of the Pitch Bend message in the
MIDI spec and sends the data at a very high rate, so that you NEVER
EVER hear discrete steps, even when the pitch bend range stretches
across the whole full-size Continuum. You can hit a note at the very
bottom of the keyboard and slide it very quickly all the way to the
top. The Continuum won't send a new Note On value and will do all the
pitch adjustment with Pitch Bend Messages, and you absolutely cannot
hear the steps, no matter how fast you slide. It's totally amazing;
I've never seen such accurate use of MIDI Pitch Bend before.

If you have a synth that won't do a 96-half-step pitch bend range,
then you set up the Continuum for whatever bend range you have, and
the notes will stick at one pitch when you reach the end of the pitch
bend range while sliding. When I was demoing with the Sound Canvas,
its maximum pitch bend range is 24 half steps, so the notes stick at
one pitch when you slide more than two octaves without re-attacking.

The reason you hear steps on the pitch bend wheels of so many synths
is that they do not fully implement what the MIDI spec allows. They
either don't send the data fast enough, or they don't use a fine
enough A/D converter to read the wheel, or both. The Pitch Bend spec
is very good when implemented the way Lippold did it.

The reason that the Round Initial Pitches mode is provided is that it
makes it easier to play. When you are not Rounding, it's very
difficult to play chords in tune. This isn't a technical probem with
the device at all, it's just the nature of the instrument. Lippold
likes to call it a "fretless piano". Without rounding, it is as hard
to play in tune as a violin, i.e. very hard. But if you want a
continuous playing surface, you have to deal with how hard it is to
play. When playing the thing, you are not going to be saying "This
thing won't track my blazingly fast technique." You instead will be
saying "How many years am I going to have to practice before I can
hit a chord in tune at a fast tempo without rounding??"

Notice that on the "Drone and Solo" mp3 I am playing very slowly so I
can slide the notes in tune.

Round Initial Pitches is cool because it rounds the notes to the
nearest half step only on the attack, then references all subsequent
bending and sliding from the spot where your finger first went down.
This makes it possible to hit chords in tune quickly (like on my
ContinuFunk and Sound Canvas Rhodes mp3s), but you can still do
slides and vibrato. This is a really great thing.

Playing justly intonated triads is easier in this mode, too, because
you can hit a chord, then bend the third down slightly to the just
third. The root and fifth get rounded to their equally tempered
values, which are so close to the just values that I can't hear it at
all. This is much easier than trying to hit all three notes in tune
with no rounding.

The lines on the playing surface do not indicate the limits of the X
measurement; they are there to help the player hit the notes in tune,
like the lines on the fingerboards of some fretless electric basses.
The X measurement it at least 100 times as fine as the lines.

Hope this helps.

Mark Smart

🔗Kurt Bigler <kkb@breathsense.com>

on 4/7/04 11:09 AM, Mark Smart <marksmartus@yahoo.com> wrote:

> Given the Continuum's lack of accuracy and repeatability in the Y
> direction, I don't think this would work very well. Fortunately, I
> don't think it's necessary. Round Initial Pitches mode is probably a
> better solution.

Not if it rounds to 12et and I'm trying to get some 4:7 and 11:8 intervals
right in a 5- to 7-note chord. But by combining the rouding with some other
adaptive just smarts and then we might really have something with some
microtonal potential. Or are you really saying that you can adjust so
quickly that no 12et pitch is ever heard? Of course this is all guesswork
for me not having played the thing, so I'm really out of my league here.

> If you have a sophisticated, configurable synth on the other end,
> like Kyma or Reaktor, you could even set it up so that the synth
> would play just pitches when it received Note On messages with 0
> pitch bend. Then when you use Round Initial Pitch mode, you would
> always get just pitches when you first attack the notes, then you
> could bend and slide them from there.

Yes exactly. Of course this group has talked a lot about how best to do
this kind of thing algorithmically without being restricted to a fixed
12-tone scale of any kind. Off the top, you have:

Werner Mohrlok's "Hermode Tuning"

John deLaubenfels' adaptive tuning system (adaptune.com)

The thread "Hermode Tuning" started by Werner on 11/27/03 covers both of
these systems and does some comparison/contrast of the two.

> Dang, I have to try that out with Reaktor...
>
> Or how about this...you define an octave or two at the bottom of the
> Continuum that controls what key the just intonation is in. If you
> hit a C in that octave, the synth with round everything to the
> nearest just pitch in C. Then you hit an F# in the low octave when
> you want to play an F# major chord without it sounding totally awful.
>
> (You guys probably talk about this on here all the time)

Yes several of us have talked about and worked on aproaches based on
explicit modulation cues (as opposed to algorithmically inferred modulation)
and it is a long-term work-in-progress for me to develop this into something
that can be used effectively in keyboard/pedalboard performance. If you
search the archives for "XMW" you will find out about some of this. Or
better, look for posts by Robert Walker, Peter Frazer, Carl Lumma, and
myself in these threads:

re-tuning methods
re-tuning methods (was Re: odeion1-003)
work in progress on XMW [was: odeion1-003]
odeion1-003

A good place to start is perhaps with Peter Frazer's post "Re: Re:
odeion1-003" on 12/12/03 or possibly with his originating post on that
thread on 12/9/03. There may have been other contributors I didn't list
above.

[If we reawaken some of these old topics, let's start new thread subjects so
we don't make this "Continuum" thread into a big mish-mash of separate
topics. The ability to manage a relationship to archived messages of this
list over years and years depends on people's help in maintaining some
organization, although it will of course never be perfect. Note that this
list has been relatively quiet lately, whereas at times I think we have had
well over 50 messages per day, IIRC. If you have the old message around or
can find it in the archive and forward itself to yourself from the web
interface then you can even continue the old threads with their existing
subjects, which is a great thing if the subject was well-chosen.]

Changing keys is not just changing keys. There are a lot of issues
regarding horizontal versus vertical pitch relationships, and how horizontal
accuracy can be traded off to prevent horizontal drift problems. A
*non-rounding* or even an auto-rouding control over the key that also allows
subsequent bending (of the key root pitch and thus simultaneously of all the
notes in the chords) would throw a whole new interesting twist into this
effort.

> Of course...you could do this on any keyboard synth and you don't
> need a Continuum. Definitely the easy way out. The only thing the
> Continuum gains you is bending, sliding, and vibrato.
>
> If you were a REAL hardcore microtonalist, you would practice the
> Continuum for several years until you could do it with no
> rounding...then act really condescending to all the "rounding wimps".

This is definitely a potentially "life changing" kind of instrument and I
think is worth some investment of time to discover its potential even if
this means facing a lot of problems. I definitely want one. To the degree
I can learn it I will be able to break the barrier that prevents me as a
keyboardist from talking intelligently about tonallity to a cellist!

-Kurt

🔗Mark Smart <marksmartus@yahoo.com>

--- In tuning@yahoogroups.com, Kurt Bigler <kkb@b...> wrote:
> on 4/7/04 11:09 AM, Mark Smart <marksmartus@y...> wrote:
> Werner Mohrlok's "Hermode Tuning"
>
> John deLaubenfels' adaptive tuning system (adaptune.com)
>
> The thread "Hermode Tuning" started by Werner on 11/27/03 covers
both of
> these systems and does some comparison/contrast of the two.

Cool, I will check those out.

> > (You guys probably talk about this on here all the time)
>
> thread on 12/9/03. There may have been other contributors I didn't
list
> above.

This is all pretty interesting to me since I haven't studied
microtonality much beyond messing with just intonation a little on
different computers.

> Changing keys is not just changing keys. There are a lot of issues
> regarding horizontal versus vertical pitch relationships, and how
horizontal
> accuracy can be traded off to prevent horizontal drift problems. A
> *non-rounding* or even an auto-rouding control over the key that
also allows
> subsequent bending (of the key root pitch and thus simultaneously of
all the
> notes in the chords) would throw a whole new interesting twist into
this
> effort.

Yes, a very difficult problem. This is only somewhat related, but some
of you might be interested in some of the auto-harmonizer stuff I've
done with Reaktor to synthesize big band horns from a wind controller
input (you've probably already looked at this if you followed the link
from the Continuum example page):

http://www.marksmart.net/sounddesign/windsounds/re
aktorbigband/reaktorbigband.html

The way I implemented the harmonizer using Event Tables that control
the intervals for each harmony voice on each type of chord might be
adaptable for microtuning. (Actually, I thought about trying this but
never got around to it...making it so that the thirds in the chords
always got bent down to their just values. It could do this because it
has enough information to figure out where the third is at all times)

I have one version where it gets the chord progression from a
sequencer, and worked on another where it would try to figure out what
chords a keyboardist was playing from MIDI input. It kind of worked as
long as you always played the root on the bottom, and the chords all
had a third and a seventh/sixth to determine chord quality. Fun to
play with. Reaktor is great for this kind of thing. I suppose a lot of
you are already way into Reaktor or Kyma.

> This is definitely a potentially "life changing" kind of instrument
and I
> think is worth some investment of time to discover its potential
even if
> this means facing a lot of problems. I definitely want one. To the
degree
> I can learn it I will be able to break the barrier that prevents me
as a
> keyboardist from talking intelligently about tonallity to a cellist!
>
> -Kurt

I have come up with some cool demos, but still I feel like I'm just a
beginner on the thing. It would be possible to spend your whole life
working on the technique. It's a totally new instrument; having
regular keyboard skills doesn't help much. A lot of keyboardists at
the NAMM show complained to us because they couldn't just sit down and
start playing it. Some told us it should have bumps on it where piano
keys would be so they could feel for the keys without looking the way
they do on a regular keyboard. All we could say was, "This is a new
thing, it's not a piano!"

I just wanted to say one more thing, which is that the Continuum is
still very expressive even if you are not using the "Y" information at
all. On my "Summertime" example, there is no "Y" control, and Z is
controlling both volume and brightness. This makes it like a wind
controller, which only has two parameters, but is very expressive.
Continuum with no Y is like a polyphonic wind controller.

This seems like a great group. I don't know how involved I am going to
be able to get, because I am already pretty overloaded with projects
(lately I've been woodshedding a lot on Stick), but thanks for a very
stimulating discussion.

Mark Smart

🔗Eduardo Sabat-Garibaldi <ESABAT@ADINET.COM.UY>

🔗Jeff Olliff <jolliff@dslnorthwest.net>

The point of the Continuum is it's implementation of the pitch
continuum, so you can hit any pitch. Another interesting feature is
that harmonic intervals of any species are a constant span
throughout all registers, not the case on the fingerboards of string
instruments. Adaptive tuning is an endlessly fascinating idea for
keyboards and computer music. This Fingerboard does an end run
around adaptive tuning by returning intonation to the purview of the
performer, should s/he so dare. I wrote a suggestion for practicing
intonation in an earlier thread:

/tuning/topicId_52763.html#52833

but in ignorance of the real instrument, and the difficulty of
sewing soft frets on the surface. Some interesting comments in that
regard lately. Chopping the thing up into hex tiles or disabling
the continuum misses the point.

🔗Afmmjr@aol.com

🔗Kurt Bigler <kkb@breathsense.com>

on 4/7/04 6:48 PM, Jeff Olliff <jolliff@dslnorthwest.net> wrote:

>> Of course...you could do this on any keyboard synth and you don't
>> need a Continuum. Definitely the easy way out. The only thing the
>> Continuum gains you is bending, sliding, and vibrato.
>>
>> If you were a REAL hardcore microtonalist, you would practice the
>> Continuum for several years until you could do it with no
>> rounding...then act really condescending to all the "rounding
> wimps".
>>
>> Mark Smart
>> www.marksmart.net
>
> The point of the Continuum is it's implementation of the pitch
> continuum, so you can hit any pitch. Another interesting feature is
> that harmonic intervals of any species are a constant span
> throughout all registers, not the case on the fingerboards of string
> instruments. Adaptive tuning is an endlessly fascinating idea for
> keyboards and computer music. This Fingerboard does an end run
> around adaptive tuning by returning intonation to the purview of the
> performer, should s/he so dare. I wrote a suggestion for practicing
> intonation in an earlier thread:
>
> /tuning/topicId_52763.html#52833
>
> but in ignorance of the real instrument, and the difficulty of
> sewing soft frets on the surface. Some interesting comments in that
> regard lately. Chopping the thing up into hex tiles or disabling
> the continuum misses the point.

I'm not quite sure of what you intended in saying "misses the point", but
the following responses came up for me:

Part of the point of the Continnum is that you can do whatever you want with
it. All the X,Y,Z coordinate information is contantly being sent, even if
you do nothing with it. But it is there for those who want to do something
with it, which includes the possibility that the generated pitch information
can be ignored and other mappings substituted.

It turns out that there are some problems with the Y direction that dashed
my hopes a bit. However, the Continnum is here now, and costs considerably
less than the Starr Labs keyboard (which I'm not putting down, mind you), so
finding out all the details of what the Continnum might and might not do in
a discrete pitch mode is for me an important topic to pursue. But also when
I think of discrete pitch mappings with the continuum, I realize it can have
discrete pitches "with a little bending" around each note available,
something that no discrete-key board can do except by using the Z direction
for pitch, which has its limitations. I have heard that some keyboards have
tried to use X-direction wiggle for modulation purposes, but I don't think
it was a success.

So I don't think I'm missing the point when thinking of the things that
basically nothing *besides* the Continuum could do. Even with discrete
pitches and no other interesting modulation or polyphonic aftertouch going
on, the Continuum is the only surface that within its limitations lets you
map it however you want to map it, and remap it at a whim (with appropriate
software support, of course).

The flexibility of how the Continnum can be used is part of its potential
value for me. If I did not also want a continuous pitch mode then perhaps
my money would not be well-spent. On the other hand, David Wessel (CNMAT)
for example has used the Continnum X dimension very successfully for things
totally unrelated to pitch, from what I have heard.

In short, its got a lot of things going for it besides continuous pitch that
nothing else out there can touch.

-Kurt

🔗gooseplex <cfaah@eiu.edu>

> The point of the Continuum is it's implementation of the pitch
> continuum, so you can hit any pitch.
....
Chopping the thing up into hex tiles or disabling
> the continuum misses the point.

Perhaps it does not coincide with the aims of the inventor, but I
don't think the idea of mapping a tuning "misses the point"...

On the contrary, mapping an unconventional fixed tuning of your
choice to the flat playing surface in a _pattern of your choice with
any interval spans you want per whatever distance, etc. is
something worth exploring!

No amount of practice will result in being able to nail all the
pitches of a specific tuning like 22ET on the Continuum as if you
are playing a keyboard with one-to-one correspondences. And
yes, it wasn't built for that. But the fact remains that you can
theoretically map that tuning to any pattern you want, and there
you have it. I understand that it was not designed to do this, but I
think this idea of mapping exploits a very strong potential of the
Continuum rather than misses the point.

Regards,
Aaron Hunt

🔗gooseplex <cfaah@eiu.edu>

🔗Mark Smart <marksmartus@yahoo.com>

🔗Robert Walker <robertwalker@ntlworld.com>

HI there,

Here is an idea for another way you could map the continuum keyboard that
could be itneresting for exploring alternative tunings.

The idea is to explre something in between the discrete and the
continuous. At one extreme, each pitch is steady
for each "key" then abruptly changes - the discrete keyboard.
At the other extreme it is completely continous as it is now,
and in between you have "key"s that are fairly uniform
in pitch. changing continuously, but very gradually, then in the
cracks between the "keys" the pitch varies much more rapidly.

So, e.g. if your current tuning has a 7/6 in it, then a region around
the 7/6 "key", will play a 7/6 to within a few cents.

Then the pitch changes rapidly until you get to the next key,
maybe tuned to a 6/5. Then there's another broad patch around the 6/5 where
relatively large movements of the finger cause small pitch
adjustments above or below a pure 6/5 and so on.

Then the user can easily hit the exact pitches, and
can also explore subtle nuances of the pitches.

I can add this as a feature to FTS if any FTS user
has the keyboard and wants to use it in this way
with FTS. FTS could look at the pitch received from
the keyboard just before the note gets played and
move it closer to the nearest note in the currently
selected tuning.

Indeed, come to think of it, maybe I'll add that
in anyway - make it so that you can have a non linear
pitch bend adjustment with the pitch bend wheel
as a way to help deal with the low resolution
of some pitch bend wheels by allowing more
shades of pitch in the immediate vicinity of
every note in the tuning.

N.B. if you use it in its current form,
with the round off mode switched on,
and retuning to another tuuning in FTS, it
will work jsut like the ET tuning except that
the initial finger position gets rounded off
to a pitch in the currently selected tuning
in FTS rather than 12-et. So, you could
make it round off to WIII, quarter comma meantone,
or whatever you want. This won't need any re-programming
- FTS already does it.

Another thought, if you have two or more fingers on the
same part of the keyboard, but not exactly in line, then
there will be some overlap, but they won't both touch exactly the
same rods. Maybe there is one rod that is only touched
by one of them to one side, another only touched by
another to the other side. So, perhaps you get a kind
of skewed "footprint" which maybe software could learn to recognise
and use to work out what the separate finger positions
were and how many fingers were pressed, don't
know if it would work, just suggesting it as an idea.

Thanks,

Robert

http://www.robertinventor.com

🔗Jeff Olliff <jolliff@dslnorthwest.net>

Kurt and Aaron (and with thanks to Mark for contributing so much
hands-on experience),

My apologies for the remark about missing the point. Your rebuttals
are well taken. I was in too big a hurry.

I'm very happy to see this discussion happening. I hadn't
considered the problem of the granularity of half steps, although it
should have been obvious from the implementation. My very old
fashioned mean tone orientation uses diatonic half steps, wider than
12tet, but not the narrower chromatic ones. Maybe the diatonic
variety would work on the continuum, for the most part, but I'd have
to try it to find out. I'm shooting for the same flexibility the
17th century Italians were, a practical distinction between sharps
and flats, basically because temperament is bad for polyphony.

I also hadn't realized the limitation of resolution on the y axis,
front to back. Using the z axis (key depth) for fine tuning is
analogous to the clavichord's subtle and hazardous intonational
control, and is a very interesting idea. I jumped to the conclusion
we had to move tuning to the x axis, a momentary mental straight
jacket.

Another quick thought: a run stepping through chromatic half steps
(F-F#, etc) would require a slightly detached technique, just as on
a fretted clavichord where those notes share a string choir, and
very unmusical ciphers occur trying to play them together.

Jeff O

🔗Alexandros Papadopoulos <Alexmoog@otenet.gr>

Ooops! sorry , wrong email recipient!
Although not too off topic...
On May 15, 2004, at 10:35 PM, Alexandros Papadopoulos wrote:

> Hello
> I am Alexandros Papadopoulos from Greece , we have talked about this
> keyboard some time before.
> You then told me that 20 keyboards will be build , and when these are
> sold, more will come.
> Is there a list of people waiting for these?
> If yes , can I be on the waiting list ? It is very important for me .
> I only need to see a picture of the instrument and a few technical
> details in order to buy.
>
> Thank you very much
>
>
>
> ------------------------ Yahoo! Groups Sponsor > ---------------------~-->
> Check out Music Videos, Internet Radio, Artist Photos, Music News!
> LAUNCH Music on Yahoo!
> http://us.click.yahoo.com/Wkrq7C/JARHAA/n1hLAA/RrLolB/TM
> ---------------------------------------------------------------------> ~->
>
> You can configure your subscription by sending an empty email to one
> of these addresses (from the address at which you receive the list):
> tuning-subscribe@yahoogroups.com - join the tuning group.
> tuning-unsubscribe@yahoogroups.com - leave the group.
> tuning-nomail@yahoogroups.com - turn off mail from the group.
> tuning-digest@yahoogroups.com - set group to send daily digests.
> tuning-normal@yahoogroups.com - set group to send individual emails.
> tuning-help@yahoogroups.com - receive general help information.
>
> Yahoo! Groups Links
>
>
>
>
>
>

🔗Afmmjr@aol.com

🔗kraig grady <kraiggrady@anaphoria.com>

🔗Afmmjr@aol.com

🔗Alexandros Papadopoulos <Alexmoog@otenet.gr>

Hello ,
Any news on the keyboard?
I hope I am not bothering you by asking.
I have sent an email to Dylan Horvath some time before , only because I
saw in a design oriented website (www.coroflot.com) that they are
making the keyboard , but I got no reply.
Thank you

On 17 Μαϊ 2004, at 6:06 μμ, Afmmjr@aol.com wrote:

> Hello Alexandros Papadopoulos,
>
> The Terpstra keyboard will have its prototype this summer, in a
> couple of months. At that time another 20 will be built, hopefully
> learning something significant from the experience of working with the
> prototype. We are keeping a list of those interested in one of the 20
> and I am happy to include you on this list.
>
> best, Johnny Reinhard
>
> You can configure your subscription by sending an empty email to one
> of these addresses (from the address at which you receive the list):
> tuning-subscribe@yahoogroups.com - join the tuning group.
> tuning-unsubscribe@yahoogroups.com - leave the group.
> tuning-nomail@yahoogroups.com - turn off mail from the group.
> tuning-digest@yahoogroups.com - set group to send daily digests.
> tuning-normal@yahoogroups.com - set group to send individual emails.
> tuning-help@yahoogroups.com - receive general help information.
>
>
>
>
>
> Yahoo! Groups Sponsor
>
> ADVERTISEMENT
> <122603_LREC_bkup_BR_Ashanti.jpg>
> <l.gif>
>
> Yahoo! Groups Links
>
> • To visit your group on the web, go to:
> /tuning/
>
> • To unsubscribe from this group, send an email to:
> tuning-unsubscribe@yahoogroups.com
>
> • Your use of Yahoo! Groups is subject to the Yahoo! Terms of
> Service.
>
>

🔗Alexandros Papadopoulos <Alexmoog@otenet.gr>

Oh no!
please don't read the message , I made a mistake by sending it here!

I am very sorry
On 27 Απρ 2016, at 3:14 πμ, Alexandros Papadopoulos wrote:

> Hello ,
> Any news on the keyboard?
> I hope I am not bothering you by asking.
> I have sent an email to Dylan Horvath some time before , only because
> I saw in a design oriented website (www.coroflot.com) that they are
> making the keyboard , but I got no reply.
> Thank you
>
>
> On 17 Μαϊ 2004, at 6:06 μμ, Afmmjr@aol.com wrote:
>
>> Hello Alexandros Papadopoulos,
>>
>> The Terpstra keyboard will have its prototype this summer, in a
>> couple of months. At that time another 20 will be built, hopefully
>> learning something significant from the experience of working with
>> the prototype. We are keeping a list of those interested in one of
>> the 20 and I am happy to include you on this list.
>>
>> best, Johnny Reinhard
>>
>> You can configure your subscription by sending an empty email to one
>> of these addresses (from the address at which you receive the list):
>> tuning-subscribe@yahoogroups.com - join the tuning group.
>> tuning-unsubscribe@yahoogroups.com - leave the group.
>> tuning-nomail@yahoogroups.com - turn off mail from the group.
>> tuning-digest@yahoogroups.com - set group to send daily digests.
>> tuning-normal@yahoogroups.com - set group to send individual
>> emails.
>> tuning-help@yahoogroups.com - receive general help information.
>>
>>
>>
>>
>>
>> Yahoo! Groups Sponsor
>>
>> ADVERTISEMENT
>>
>> <122603_LREC_bkup_BR_Ashanti.jpg>
>> <l.gif>
>>
>> Yahoo! Groups Links
>>
>> • To visit your group on the web, go to:
>> /tuning/
>>
>> • To unsubscribe from this group, send an email to:
>> tuning-unsubscribe@yahoogroups.com
>>
>> • Your use of Yahoo! Groups is subject to the Yahoo! Terms of
>> Service.
>>
>>