Re:[tuning] Kyma. How is it ?

Graham wrote:-

>That looks sufficiently flexible that it might work. I did set this up on
>Kyma list night. With high Q and resonance, moving the filter by
>semitones made no obvious difference to the sound. But there's also a
>"feedback" mode, which gives a whine at the cuttoff frequency when >turned
>to extremes. That means you can play tunes in the feedback: the only
>rewarding part of the experiment. It appears the usual keyboard range
>isn't enough for the effects I get by twisting the cutoff knob. Getting
>portamento to work might help.

I have been thinking about getting a Kyma system as it seems to be easy to
implement tunings and scale based timbres a la Sethares as well as all the
other nifty features.

I have just one hesitation though. I am interested in implementing some
kind of adaptive tuning in the vein of Sethares' model [featured in his
book 'Tuning, Timbre, Scale'] or at second best a JI only model like
Justonic Tunings 'pitch palette' or JI relay. Does Kyma allow one to do
this ?
It would require I suppose control of tuning to individual notes so that
any tuning chage would affect only those pitches required to make just
chords [playing [sliding] the comma so to speak].

If it could do this I don't think I would need any other synth or module or
software synth. It seem to do it all.

Justin White

This email and any attachments are confidential and are intended solely for
the recipient. If you are not the intended recipient, you must not disclose
or use the information contained in it. If you have received this email in
error, please tell us immediately by return email and delete the document.

The contents and any attachments are the opinion of the sender and not
necessarily that of David Jones Limited.

--- In tuning@egroups.com, "Justin White" <justin.white@d...> wrote:
>
> I have just one hesitation though. I am interested in implementing some
> kind of adaptive tuning in the vein of Sethares' model [featured in his
> book 'Tuning, Timbre, Scale'] or at second best a JI only model like
> Justonic Tunings 'pitch palette' or JI relay. Does Kyma allow one to do
> this ?

Since you're interested in adaptive tuning, I suggest you familiarize yourself with John
deLaubenfels' latest work . . . I believe JI Relay is somewhat outdated . . . John?

In-Reply-To: <4A256991.00168227.00@dj307.davidjones.com.au>
Justin White wrote:

> I have been thinking about getting a Kyma system as it seems to be easy
> to
> implement tunings and scale based timbres a la Sethares as well as all
> the
> other nifty features.

It's not as easy as it should be to implement tunings, but I'm working on
simplifying that.

> I have just one hesitation though. I am interested in implementing some
> kind of adaptive tuning in the vein of Sethares' model [featured in his
> book 'Tuning, Timbre, Scale'] or at second best a JI only model like
> Justonic Tunings 'pitch palette' or JI relay. Does Kyma allow one to do
> this ?
> It would require I suppose control of tuning to individual notes so that
> any tuning chage would affect only those pitches required to make just
> chords [playing [sliding] the comma so to speak].

This is something I've got in the back of my mind. You can tune notes by
arbitrary expressions, that's not a problem. What may be is getting a
function that can listen to all notes, and remember past ones. The way
synthesizers work, each degree of polyphony lives in its own world, and
the way expressions work they don't have memory. The first problem could
be fixed provided you tune all notes to the same scale. The second one
I'm looking at for other reasons, and I expect I'll fix it with a bit of
poking.

> If it could do this I don't think I would need any other synth or
> module or
> software synth. It seem to do it all.

Oh yes, it can do everything, but not all at the same time. At least not
with the base system. So you either need non-real time hackery, or some
other synths. It's also based around low-level components, and there
aren't many presets to start you off with. And it doesn't have any of
these new-fangled physical modelling algorithms. So it's still useful to
have something else around.

Note though, on the positive side, that it's external *effects* you won't
need as well.

It works well with external sound sources, and can be used to process
sample files. I'm hoping there'll be a way of dumping the samples into a
retuned SoundFont to play through the soundcard, but that's still blue
skies stuff.

Also, as it uses external DSPs, you can use it alongside software synths.

Graham.

[Justin White:]
>>I have just one hesitation though. I am interested in implementing
>>some kind of adaptive tuning in the vein of Sethares' model [featured
>>in his book 'Tuning, Timbre, Scale'] or at second best a JI only model
>>like Justonic Tunings 'pitch palette' or JI relay. Does Kyma allow one
>>to do this ?

[Paul E:]
>Since you're interested in adaptive tuning, I suggest you familiarize
>yourself with John deLaubenfels' latest work . . . I believe JI Relay
>is somewhat outdated . . . John?

Yes. I had better real-time adaptive tuning going on the NeXT (though
still very far from ideal). JI Relay is very primitive in its tuning
choices, and tends to change the tuning too often. I tossed JI Relay
off in the first two months after joining the list in March 1999, and
since then have focused on retuning complete sequence files in
non-real-time.

The problems of real-time adaptive tuning are daunting, for the simple
reason that a new chord tends to start sounding just before an old
chord ends, but the new notes don't have "knowledge" of which existing
notes are about to switch off, so tuning decisions become bad. Starting
a note at one tuning, then moving it later, is painful, but what else is
there to do?

I can imagine a very sophisticated real-time adaptive tuning program
that "learns" what transitions to expect and makes tuning decisions on
that basis. But I'm deep into leisure (full knowledge) retuning and
will probably continue to be so for the foreseeable future.

To the best of my knowledge, there is no really good real-time adaptive
tuning program available today. Other readers who have tried the
Justonic software haven't seemed to be too thrilled with its adaptive
capabilities. It may be better than JI Relay, however.

JdL

For Renaissance, Baroque, and early Classical music, a Vicentino-like
approach would be an excellent starting point for a real-time adaptive
tuning algorithm . . .

[Paul E:]
>For Renaissance, Baroque, and early Classical music, a Vicentino-like
>approach would be an excellent starting point for a real-time adaptive
>tuning algorithm . . .

For 5-limit, yes. Come to think of it, JI Relay probably does a pretty
fair job at adaptive 5-limit tuning. Got any good ideas for 7?

JdL

I wrote,

>>For Renaissance, Baroque, and early Classical music, a Vicentino-like
>>approach would be an excellent starting point for a real-time adaptive
>>tuning algorithm . . .

>For 5-limit, yes.

As you know, I don't like 7-limit retunings of this music.

John deLaubenfels wrote,

>Come to think of it, JI Relay probably does a pretty
>fair job at adaptive 5-limit tuning.

What does it do to I-vi-ii-V-I and I-vi-IV-V-I progressions?

[I wrote:]
>>Come to think of it, JI Relay probably does a pretty fair job at
>>adaptive 5-limit tuning.

[Paul E:]
>What does it do to I-vi-ii-V-I and I-vi-IV-V-I progressions?

Don't you remember when Brett Barbaro, Paul E, and I went thru all this
around 04-15-99? Brett and Paul used to engage in wild debate with each
other, while the rest of the list could only stand back and watch in
stunned amazement. Imagine our surprise when Brett and Paul turned out
to be the same person!

(Sorry - Paul and Brett ARE the same person, but it was never proved
that the two engaged in actual debate on the list ;->).

JI Relay's tuning decisions for the "comma pump" and other sequences
of chords are described in:

http://www.egroups.com/message/tuning/2275 (JdL)
http://www.egroups.com/message/tuning/2289 (Brett)
http://www.egroups.com/message/tuning/2302 (JdL)
http://www.egroups.com/message/tuning/2325 (Brett)
http://www.egroups.com/message/tuning/2333 (JdL - a good summary)
http://www.egroups.com/message/tuning/2364 (Brett)
http://www.egroups.com/message/tuning/2373 (Brett)
http://www.egroups.com/message/tuning/2379 (JdL)

JdL

John deLaubenfels wrote,

>Come to think of it, JI Relay probably does a pretty fair job at
>adaptive 5-limit tuning.

OK, I guess "pretty fair" is right -- not "pretty good", but "fair".

[I wrote:]
>>Come to think of it, JI Relay probably does a pretty fair job at
>>adaptive 5-limit tuning.

[Paul E:]
>OK, I guess "pretty fair" is right -- not "pretty good", but "fair".

I've always made it clear that I consider it lacking, and consider the
whole process of real-time adaptive tuning to be extremely challenging.
The Vicentino method has been shown to have significant problems for
common chords such as A,C,E,G, and in any case, as far as I know, has
only existed on drawing boards all these centuries. My experience in
twenty years of programming has consistently been that when something
moves from the drawing board to reality, many more difficulties
invariably come up. When/if I see Vicentino demonstrated in real-life,
I'll be much more confident of its capabilities, such as they may prove
to be.

Five-limit real-time adaptive tuning is not nearly as daunting as
7-limit, because nothing moves from 12-tET by all that much, so wrong
decisions at chord transitions can be fixed without a lot of motion. My
own interest, as you know, is in seven, where several common 12-tET
intervals are highly ambiguous as to their proper JI interpretation, and
where, therefore, wrong decisions become quite painful.

JdL

>The Vicentino method has been shown to have significant problems for
>common chords such as A,C,E,G

Not really -- why not just tune it halfway between where (A,C,E) would
suggest and where (C,E,G) would suggest?

>When/if I see Vicentino demonstrated in real-life,
>I'll be much more confident of its capabilities, such as they may prove
>to be.

I'm sorry I haven't been able to make clear to you the proof of how it
works, but the fact is, if no enharmonic equivalents are used in the music,
it works excellently. It is a matter of seeing what the maximum errors of
meantone are and how they are distributed. It should be rather easy to
program if the music is restricted to a 12-tone meantone compass -- and a
bit more challenging if it exceeds that but a 12-tone MIDI files is all
you're provided with.

[I wrote:]
>>When/if I see Vicentino demonstrated in real-life, I'll be much more
>>confident of its capabilities, such as they may prove to be.

[Paul E:]
>I'm sorry I haven't been able to make clear to you the proof of how it
>works, but the fact is, if no enharmonic equivalents are used in the
>music, it works excellently. It is a matter of seeing what the maximum
>errors of meantone are and how they are distributed. It should be
>rather easy to program if the music is restricted to a 12-tone meantone
>compass -- and a bit more challenging if it exceeds that but a 12-tone
>MIDI files is all you're provided with.

Yes, well, my skepticism may be completely without merit, but skeptic I
am! I definitely have a "show me" attitude, which experience has
continuously reinforced. "Proof". A proof laden with qualifications.
And how many real sequences trigger the exceptions? Only when it's
fired up will we know. I'm prepared to be delighted - if it truly
passes muster.

JdL