Speed limit: 4.7 cents/sec.

This quote may be of interest to John deLaubenfels and other
"autotuners":

"In a formal listening test with professional musicians pieces by
Purcell, Gesualdo and the contemporary Swedisch composer Bäck were
presented in pure (beat-free) tuning, in equally tempered tuning and
in the melodic charge tuning presented above (1). The results showed
that musicians tended to dislike all these alternatives, least
though, equally tempered tuning. However, a computer controlled
synthesizer does not need to have a preset tuning, provided a better
alternative is available.
The best solution as yet tried, though not yet formally tested, is
to tuning all notes according to melodic intonation from the onset
and change the tuning toward just intonation at a rate of 4.7
cents/s. This rate, being rather critical, conveys the idea that the
ensemble members seemingly pay attention to salient beats and is at
the same time slow enough to convey the impression of a stable
pitch."

(1) Sundberg, Johan, Anders Friberg and Lars Frydén. "Rules for
automated performance of ensemble music", _Contemporary Music
Review_ vol. 3, 1989, pp. 89-109.

This melodic charge tuning involves playing augmented tones sharper
and diminished tones flatter than 12-tET. So it's like Pythagorean,
but the deviations from the equally tempered scale are smaller.

The quote is from the article
Sundberg, Johan, Anders Friberg and Lars Frydén. "Common Secrets of
Musicians and Listeners: An analysis-by-synthesis Study of Musical
Performance", _Representing Musical Structure_, Peter Howell (ed.),
Academic Press, London, 1991, pp. 161-197.

So John, is this easy to try for you? Start tones at 12-tET and move
them to just (or your computational optimum) at that rate? Perhaps
soft tones can be changed a bit faster.

Manuel

--- In tuning@y..., <manuel.op.de.coul@e...> wrote:
>
> This quote may be of interest to John deLaubenfels and other
> "autotuners":
>
> "In a formal listening test with professional musicians pieces by
> Purcell, Gesualdo and the contemporary Swedisch composer Bäck were
> presented in pure (beat-free) tuning, in equally tempered tuning and
> in the melodic charge tuning presented above (1).

If they had a chance to hear the results of John's program applied to
these pieces, they might have liked that even better than any of
these alternatives. How were commas handles in the "pure (beat-free)
tuning" in this experiment?

[Manuel Op de Coul wrote:]
>This quote may be of interest to John deLaubenfels and other
>"autotuners":

>"In a formal listening test with professional musicians pieces by
>Purcell, Gesualdo and the contemporary Swedisch composer B„ck were
>presented in pure (beat-free) tuning, in equally tempered tuning and
>in the melodic charge tuning presented above (1). The results showed
>that musicians tended to dislike all these alternatives, least
>though, equally tempered tuning. However, a computer controlled
>synthesizer does not need to have a preset tuning, provided a better
>alternative is available.

>The best solution as yet tried, though not yet formally tested, is
>to tuning all notes according to melodic intonation from the onset
>and change the tuning toward just intonation at a rate of 4.7
>cents/s. This rate, being rather critical, conveys the idea that the
>ensemble members seemingly pay attention to salient beats and is at
>the same time slow enough to convey the impression of a stable
>pitch."

>(1) Sundberg, Johan, Anders Friberg and Lars Fryd‚n. "Rules for
>automated performance of ensemble music", _Contemporary Music
>Review_ vol. 3, 1989, pp. 89-109.

>This melodic charge tuning involves playing augmented tones sharper
>and diminished tones flatter than 12-tET. So it's like Pythagorean,
>but the deviations from the equally tempered scale are smaller.

>The quote is from the article
> Sundberg, Johan, Anders Friberg and Lars Fryd‚n. "Common Secrets of
>Musicians and Listeners: An analysis-by-synthesis Study of Musical
>Performance", _Representing Musical Structure_, Peter Howell (ed.),
>Academic Press, London, 1991, pp. 161-197.

>So John, is this easy to try for you? Start tones at 12-tET and move
>them to just (or your computational optimum) at that rate? Perhaps
>soft tones can be changed a bit faster.

Thanks for the reference, Manuel! My NeXT real-time software had an
adjustable maximum rate of tuning change, and JI Relay does as well, but
in my leisure retuning, I've got horizontal motion under pretty tight
control, so, to my ear, it hasn't proven necessary to gliss tuning
motion. Do you disagree, for any of the sequences on my web page?

JdL

--- In tuning@y..., "John A. deLaubenfels" <jdl@a...> wrote:

> Thanks for the reference, Manuel! My NeXT real-time software had an
> adjustable maximum rate of tuning change, and JI Relay does as
well, but
> in my leisure retuning, I've got horizontal motion under pretty
tight
> control, so, to my ear, it hasn't proven necessary to gliss tuning
> motion. Do you disagree, for any of the sequences on my web page?

A very good point! J. Sundberg should be made aware of deLaubenfels'
advanced work in this area.

John deL. wrote:
>I've got horizontal motion under pretty tight
>control, so, to my ear, it hasn't proven necessary to gliss tuning
>motion. Do you disagree, for any of the sequences on my web page?

I haven't listened to any lately. I shall do that. But the example
retuning by William Sethares of a fragment of a Scarlatti sonata
on the CD with his book sounds convincing and pleasing to me.
He used glissandi but I don't know at which rate. Have you heard it?
That's the only example I have heard and I'd really like to be
able to compare the two methods with the same piece.
The way you do it now, is putting two compromises on top of each
other, minimising horizontal motion and maximising harmony.
With slow glissandi you could disconnect these two. With short tones,
the melodic aspect is more important, accommodating short term pitch
memory. Longer tones and chords would slowly get in tune.
I see lots of ways for improvement here. One could experiment with
the change rate depending on beat rates, volume, beats strength,
number of tones changing simultaneously, direction of change, etc.

Manuel

Paul wrote:
>If they had a chance to hear the results of John's program applied to
>these pieces, they might have liked that even better than any of
>these alternatives. How were commas handled in the "pure (beat-free)
>tuning" in this experiment?

Indeed. It didn't say. I assume that comma shifts, if any, were left
audible.

Manuel

[I wrote:]
>>I've got horizontal motion under pretty tight
>>control, so, to my ear, it hasn't proven necessary to gliss tuning
>>motion. Do you disagree, for any of the sequences on my web page?

[Manuel Op de Coul:]
>I haven't listened to any lately. I shall do that. But the example
>retuning by William Sethares of a fragment of a Scarlatti sonata
>on the CD with his book sounds convincing and pleasing to me.
>He used glissandi but I don't know at which rate. Have you heard it?
>That's the only example I have heard and I'd really like to be
>able to compare the two methods with the same piece.
>The way you do it now, is putting two compromises on top of each
>other, minimising horizontal motion and maximising harmony.
>With slow glissandi you could disconnect these two. With short tones,
>the melodic aspect is more important, accommodating short term pitch
>memory. Longer tones and chords would slowly get in tune.
>I see lots of ways for improvement here. One could experiment with
>the change rate depending on beat rates, volume, beats strength,
>number of tones changing simultaneously, direction of change, etc.

Yes, this may have merit. I need to find the Sethares CD and listen to
the piece! Actually I've been thinking of making a change to my program
so that, even when notes don't change for a longer period of time,
there would be horizontal springs, say 16 per second or so, which could
accommodate gradual tuning motion. If I jump off the linear spring
restriction, motion could be rigidly restricted to 4.7 cents/sec, or
any other value.

JdL

Manuel Op de Coul quoted Johan Sundberg, Anders Friberg and Lars
Frydén:
...
> >The best solution as yet tried, though not yet formally tested, is
> >to tuning all notes according to melodic intonation from the onset
> >and change the tuning toward just intonation at a rate of 4.7
> >cents/s. This rate, being rather critical, conveys the idea that
the
> >ensemble members seemingly pay attention to salient beats and is at
> >the same time slow enough to convey the impression of a stable
> >pitch."
...
> >This melodic charge tuning involves playing augmented tones sharper
> >and diminished tones flatter than 12-tET. So it's like Pythagorean,
> >but the deviations from the equally tempered scale are smaller.
...

Thanks Manuel. This is brilliant. It beautifully eliminates the
apparent conflict between Pythagorean and 5-limit-JI views of "ideal"
tuning.

JdL:
...
> in my leisure retuning, I've got horizontal motion under pretty
tight
> control, so, to my ear, it hasn't proven necessary to gliss tuning
> motion. Do you disagree, for any of the sequences on my web page?

I personally can't hear the difference between most of your retunings
and 12-tET, except on sustained chords. But doesn't the above suggest
that even with a piece like Pachelbel's Canon with no comma problems
(so it doesn't need adaptive tuning at all and could stay in
5-limit-JI), musicians would prefer the initial pitch of every note to
sound closer to Pythagorean and only go to 5-limit-JI if it is
sustained long enough?

BTW, I want to thank Herman Miller publically (I thanked him
privately some days ago) for his brilliant Warped Canons page!

-- Dave Keenan

[Manuel Op de Coul quoted Johan Sundberg, Anders Friberg and Lars
Fryd‚n:]
>>>The best solution as yet tried, though not yet formally tested, is
>>>to tuning all notes according to melodic intonation from the onset
>>>and change the tuning toward just intonation at a rate of 4.7
>>>cents/s. This rate, being rather critical, conveys the idea that the
>>>ensemble members seemingly pay attention to salient beats and is at
>>>the same time slow enough to convey the impression of a stable
>>>pitch."

>>>This melodic charge tuning involves playing augmented tones sharper
>>>and diminished tones flatter than 12-tET. So it's like Pythagorean,
>>>but the deviations from the equally tempered scale are smaller.
...

[Dave Keenan wrote:]
>Thanks Manuel. This is brilliant. It beautifully eliminates the
>apparent conflict between Pythagorean and 5-limit-JI views of "ideal"
>tuning.

[I wrote:]
>>in my leisure retuning, I've got horizontal motion under pretty tight
>>control, so, to my ear, it hasn't proven necessary to gliss tuning
>>motion. Do you disagree, for any of the sequences on my web page?

[Dave:]
>I personally can't hear the difference between most of your retunings
>and 12-tET, except on sustained chords. But doesn't the above suggest
>that even with a piece like Pachelbel's Canon with no comma problems
>(so it doesn't need adaptive tuning at all and could stay in
>5-limit-JI), musicians would prefer the initial pitch of every note to
>sound closer to Pythagorean and only go to 5-limit-JI if it is
>sustained long enough?

Well, as I posted a couple of days ago,

/tuning/topicId_24689.html#24750
/tuning/topicId_24689.html#24891
/tuning-math/message/207

the Pachelbel Canon, despite lack of comma problems, still benefits
greatly (in achieving consonant intervals) from adaptive tuning. Quite
a few people have said that my softly sprung tunings sound almost
identical to 12-tET to their ears; the "solution" would be to rigidify
the vertical springs until the consonance is clearly audible and/or
resultant horizontal motion becomes objectionable.

I would not want to speak for individual musicians, much less musicians
as a whole, regarding what intonational practices would sound best to
their ears. To _my_ ears, although I agree that some tuning motion
after the start of a note may help the overall adaptive practice, I do
not consider that deliberately targeting a starting tuning different
from what one is aiming for in the long run is a good idea.

[Dave:]
>BTW, I want to thank Herman Miller publically (I thanked him
>privately some days ago) for his brilliant Warped Canons page!

Ditto!! I'm still working my way through them. Love the 5-limit JI
minor version, even with minor dominant; it's a whole 'nother piece!
And 34-tET sounds particularly vivid to me, though I don't know why.

JdL

I listened to some of your files yesterday. The Mozart
sonata sounded very good. It would probably still sound in
meantone. The Fantaisie Impromptu was over the edge I
thought, not an improvement of the original, although in
the beginning it sounded almost the same.

I don't remember that you ever mentioned what your 5-limit
target scale is. Is it Malcolm's monochord? Ellis's duodene?

>To _my_ ears, although I agree that some tuning motion
>after the start of a note may help the overall adaptive practice,
>I do not consider that deliberately targeting a starting tuning
>different from what one is aiming for in the long run is a good
>idea.

Why? A strange sentence. Would you explain? Do you think 4.7
cents/sec. is too fast?

Manuel

[Manuel wrote:]
>I listened to some of your files yesterday. The Mozart
>sonata sounded very good.

Kyool...

>It would probably still sound in meantone.

Sound good? Yes, I'm sure you're right. I show 1/4 comma meantone
(either Gb to B or Db to F#) to be almost as good as the best COFT I
can come up with. There is fairly good pain reduction from that to
adaptive tuning, though.

>The Fantaisie Impromptu was over the edge I thought, not an improvement
>of the original, although in the beginning it sounded almost the same.

Did you happen to notice whether uneven melodic step size was part of
your objection? As I've reported on the tuning-math list, I've got
preliminary results from a new type of spring (melodic) to partially
hold this variation in check.

>I don't remember that you ever mentioned what your 5-limit target scale
>is. Is it Malcolm's monochord? Ellis's duodene?

Sorry to say I'm not familiar with either of these. The "es2" tunings
simply target JI intervals for thirds, fourths, fifths, and sixths, and
let seconds and sevenths float as they like. A dominant 7th thus has
the 7th degree pushed to 6:5 above 3:2, or 9:5 from root. The "cs5"
tunings target 7th degree as 16:9 of root, slightly flatter (and closer
to 12-tET in that regard).

[JdL:]
>>To _my_ ears, although I agree that some tuning motion
>>after the start of a note may help the overall adaptive practice,
>>I do not consider that deliberately targeting a starting tuning
>>different from what one is aiming for in the long run is a good
>>idea.

>Why? A strange sentence. Would you explain? Do you think 4.7
>cents/sec. is too fast?

No, I think 4.7 cents/sec is pretty slow. My "strange" reasoning is
that the price of any such technique is additional horizontal motion,
the bane of adaptive tuning, IMHO. There are already so many
conflicting forces, and on top of them we're going to add deliberate
motion? That's not something I'd want to do, unless the reasons were
_very_ compelling. At 4.7 cents/sec, it'd take LONG sounding notes to
achieve close to JI vertical sounds (depending upon the degree of
adjustment to the pitch at the start of the note, which isn't clear to
me), so there are significant vertical prices to pay as well.

I don't want to discount the results of actual listening tests, though!
There may well be benefits I don't yet understand...

JdL

>Sound good?

Yes, it was a typo.

>Did you happen to notice whether uneven melodic step size was part of
>your objection?

It was either that or my 12-tET disposition towards Chopin.

>Sorry to say I'm not familiar with either of these. The "es2" tunings
>simply target JI intervals for thirds, fourths, fifths, and sixths, and
>let seconds and sevenths float as they like.

Ah, ok so you're not using a full 12-tone scale, that's understandable.

>My "strange" reasoning is
>that the price of any such technique is additional horizontal motion,
>the bane of adaptive tuning, IMHO.

But you said "to my ears". Is it your ears, your intellect or a dogma?
Should the adaptation be restricted to the onset of the tones?

>There are already so many
>conflicting forces, and on top of them we're going to add deliberate
>motion?

I don't follow you. It's an extra degree of freedom!

Manuel

[I wrote:]
>>The "es2" tunings simply target JI intervals for thirds, fourths,
>>fifths, and sixths, and let seconds and sevenths float as they like.

[Manuel:]
>Ah, ok so you're not using a full 12-tone scale, that's understandable.

Right, I'm only considering the actual vertical intervals present at any
given moment.

[JdL:]
>>My "strange" reasoning is that the price of any such technique is
>>additional horizontal motion, the bane of adaptive tuning, IMHO.

[Manuel:]
>But you said "to my ears". Is it your ears, your intellect or a dogma?
>Should the adaptation be restricted to the onset of the tones?

I'm simply saying that horizontal motion is painful (to my ears, not my
"intellect" or "dogma"). I have already said that some adaptation
after onset of the tones may be desirable, because it means less at the
note transition. Perhaps there is something wonderful to the other
suggestions in this study (deliberately modifying the intonation during
a note when it would not otherwise be necessary) that will be clear once
I've heard it, but as measured from the parameters I'm already aware of
as being important, deliberately adding horizontal motion, and beginning
vertical intervals farther from JI than necessary, represent steps
backward.

[JdL:]
>>There are already so many conflicting forces, and on top of them we're
>>going to add deliberate motion?

[Manuel:]
>I don't follow you. It's an extra degree of freedom!

Yes... Of course, _anything_ we might do to the tuning could be
considered "freedom". We could, for example, begin each note a full
semitone sharper than its target and gliss it downward. Why shouldn't
we be free, after all? It's all a question of what's gonna sound nice.
Do you know if the people who did this study left behind any audio
examples?

JdL

--- In tuning@y..., <manuel.op.de.coul@e...> wrote:
>
> I don't remember that you ever mentioned what your 5-limit
> target scale is. Is it Malcolm's monochord? Ellis's duodene?

Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
scale, as any such scale has comma problems. Instead, each vertical
third and sixth is targeted toward the corresponding 5-odd-limit JI
ratio, while other harmonic intervals may or may not have harmonic
targets depending on the version of John's program in question.
Retuning motion is normally kept to within a small fraction of a
comma, by the horizontal springs in John's matrix model.

So what's Malcolm's monochord?

> >To _my_ ears, although I agree that some tuning motion
> >after the start of a note may help the overall adaptive practice,
> >I do not consider that deliberately targeting a starting tuning
> >different from what one is aiming for in the long run is a good
> >idea.
>
> Why? A strange sentence. Would you explain? Do you think 4.7
> cents/sec. is too fast?
>
I would think it's too slow given a 12-tET or Pythagorean starting
point for each note -- 4.7 cents/sec. would result in very little "JI-
ness" for a piece that moved along at a reasonable pace.
Unfortunately, motion faster than 4.7 cents/sec. would sound like
audible sliding, not desirable.

> Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
> scale, as any such scale has comma problems.

The version of bge that I have certainly -- and certainly
unwisely -- does!

> Instead, each vertical third and sixth is targeted toward the
> corresponding 5-odd-limit JI ratio, while other harmonic
> intervals may or may not have harmonic targets depending on the
> version of John's program in question.

When did all this happen!?

-Carl

--- In tuning@y..., carl@l... wrote:
> > Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
> > scale, as any such scale has comma problems.
>
> The version of bge that I have certainly -- and certainly
> unwisely -- does!

What's bge?
>
> > Instead, each vertical third and sixth is targeted toward the
> > corresponding 5-odd-limit JI ratio, while other harmonic
> > intervals may or may not have harmonic targets depending on the
> > version of John's program in question.
>
> When did all this happen!?
>
Starting about 2 years ago, I believe.

[Paul E wrote:]
>>Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
>>scale, as any such scale has comma problems.

[Carl L:]
>The version of bge that I have certainly -- and certainly
>unwisely -- does!

Uhhh... I'm not sure whether I want to ask you what "certainly unwisely"
means! ;-> The version you've got doesn't support what Paul is talking
about... [Paul, bge is the name of my executable that does leisure
retuning. I sent you one too - do you remember?].

[Paul:]
>>Instead, each vertical third and sixth is targeted toward the
>>corresponding 5-odd-limit JI ratio, while other harmonic
>>intervals may or may not have harmonic targets depending on the
>>version of John's program in question.

[Carl:]
>When did all this happen!?

While you were away, about six months ago. Every interval gets JI
springs based not upon reference to a template in a tuning file, but
directly from the number of semitones in the 12-tET source file. Among
other things, this supports non-self-consistent tunings, such as
chain of fifth chords, augmented triads, full diminished tetrads, etc.
The downsides are: no 7-limit tuning support, and dominant 7ths have
rather sharp 7th degrees, which I'm becoming more used to over time.

JdL

--- In tuning@y..., "John A. deLaubenfels" <jdl@a...> wrote:
> [Paul E wrote:]
> >>Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
> >>scale, as any such scale has comma problems.
>
> [Carl L:]
> >The version of bge that I have certainly -- and certainly
> >unwisely -- does!
>
> Uhhh... I'm not sure whether I want to ask you what "certainly
unwisely"
> means! ;-> The version you've got doesn't support what Paul is
talking
> about... [Paul, bge is the name of my executable that does leisure
> retuning. I sent you one too - do you remember?].

No. I thought the leisure retuning one was called JI Relay, and
centered around 12-tET rather than around a fixed 5-prime-limit JI
scale . . . so I'm confused as to what Carl is referring to.

> The downsides are: no 7-limit tuning support,

Carl might get upset about this if you don't qualify it . . . you
still do support 7-limit retunings, but these require "tuning files"
with chords of at least three notes, rather than merely intervals,
being assigned various "target" tunings . . . correct?

>>When did all this happen!?
>
>While you were away, about six months ago. Every interval gets
>JI springs based not upon reference to a template in a tuning
>file, but directly from the number of semitones in the 12-tET
>source file. Among other things, this supports non-self-
>consistent tunings, such as chain of fifth chords, augmented
>triads, full diminished tetrads, etc.

Great! I've been asking for this since the beginning! At this
rate, you'll arrive at my wish-list adaptive tuning in 18 months.

>The downsides are: no 7-limit tuning support, and dominant 7ths
>have rather sharp 7th degrees, which I'm becoming more used to
>over time.

Don't understand these downsides, why they would exist. The
retuning of intervals rather than scales should make it easier
than ever to get good 7-limit tunings, for example.

Do you think it's time to publish an update for JI Relay? I'd
certainly like to update the bge'd versions of my stuff. Hope
there's still a 'pure vertical JI' option...

-Carl

--- In tuning@y..., carl@l... wrote:

> Don't understand these downsides, why they would exist. The
> retuning of intervals rather than scales should make it easier
> than ever to get good 7-limit tunings, for example.
>
If you look solely at intervals, rather than larger groupings, there
would be no way to get good 7-limit tunings, because, for example,
you'd have no way of targeting one minor third to 6:5 while
simultaneously targeting another minor third to 5:4.

Anyway, Carl, it's a shame you were away for so long, John's program
has been discussed in great detail . . . I don't recall it ever using
a fixed 5-prime-limit JI target, though . . . perhaps you're thinking
of something else?

(It just occured to me that "leisure" is the opposite of "real-
time" . . . so sorry, John, about the last message where I said I
thought the leisure one was JI Relay . . . I got confused and know
that JI Relay is the real-time one.)

I wrote:

> you'd have no way of targeting one minor third to 6:5 while
> simultaneously targeting another minor third to 5:4.

Oops, I meant 7:6, not 5:4, of course!

Sorry!

> If you look solely at intervals, rather than larger groupings,
> there would be no way to get good 7-limit tunings, because, for
> example, you'd have no way of targeting one minor third to 6:5
> while simultaneously targeting another minor third to 5:4.

Who said you wouldn't look at larger groupings?

> Anyway, Carl, it's a shame you were away for so long, John's
> program has been discussed in great detail . . . I don't recall
> it ever using a fixed 5-prime-limit JI target, though . . .
> perhaps you're thinking of something else?

It used to map the meantone diatonic scale to the classic
5-prime-limit just diatonic.

In my experience, the great length of discussions isn't the
part the passerby cares about.

> (It just occured to me that "leisure" is the opposite of "real-
> time" . . . so sorry, John, about the last message where I said I
> thought the leisure one was JI Relay . . . I got confused and know
> that JI Relay is the real-time one.)

Oops. In that case, make my last post say, "Do you think it's time
for an upgrade to bge?"

-Carl

On Fri, 15 Jun 2001 06:14:54 -0600, "John A. deLaubenfels"
<jdl@adaptune.com> wrote:

>Ditto!! I'm still working my way through them. Love the 5-limit JI
>minor version, even with minor dominant; it's a whole 'nother piece!
>And 34-tET sounds particularly vivid to me, though I don't know why.

Check out the new Slendro, Pelog, and Chopi scale versions I just put up!
My approach for these is a bit different from the previous ones; these are
tuned to an approximate 5-7 note scale that repeats at the octave (a
slightly stretched octave, that is). They lose some of the character of the
real tunings, which vary slightly from one octave to the next, but I still
think the results are interesting. The Slendro version uses the same
orchestration as the 5-TET version, for easy comparison; the other two have
completely new voice assignments.

--
see my music page ---> ---<http://www.io.com/~hmiller/music/index.html>--
hmiller (Herman Miller) "If all Printers were determin'd not to print any
@io.com email password: thing till they were sure it would offend no body,
\ "Subject: teamouse" / there would be very little printed." -Ben Franklin

--- In tuning@y..., carl@l... wrote:
> > If you look solely at intervals, rather than larger groupings,
> > there would be no way to get good 7-limit tunings, because, for
> > example, you'd have no way of targeting one minor third to 6:5
> > while simultaneously targeting another minor third to 5:4.
>
> Who said you wouldn't look at larger groupings?

John _does_ look at larger groupings in many cases, including the 7-limit cases. These
larger groupings are what the "tuning files" are for.
>
> > Anyway, Carl, it's a shame you were away for so long, John's
> > program has been discussed in great detail . . . I don't recall
> > it ever using a fixed 5-prime-limit JI target, though . . .
> > perhaps you're thinking of something else?
>
> It used to map the meantone diatonic scale to the classic
> 5-prime-limit just diatonic.

John deL . . . is this true? I don't remember this ever being your method.

--- In tuning@y..., carl@l... wrote:

> In my experience, the great length of discussions isn't the
> part the passerby cares about.

Huh?

I've just got to swear off sleep! The tuning list is a 24-hour
"reality TV", or maybe "unreality TV".

[I wrote:]
>>The downsides are: no 7-limit tuning support,

[Paul E wrote:]
>Carl might get upset about this if you don't qualify it . . . you
>still do support 7-limit retunings, but these require "tuning files"
>with chords of at least three notes, rather than merely intervals,
>being assigned various "target" tunings . . . correct?

Correct.

[Carl L:]
>Great! I've been asking for this since the beginning! At this
>rate, you'll arrive at my wish-list adaptive tuning in 18 months.

That'd be nice! Your request for infinitely rigid vertical springs is
still undone, and that's been _more_ than 18 months ago... Note, too,
that exact JI can not be applied to non-self-consistent tuning sets,
such as chain-of-fifths chords.

[JdL:]
>>>The downsides are: no 7-limit tuning support, and dominant 7ths
>>>have rather sharp 7th degrees, which I'm becoming more used to
>>>over time.

[Carl:]
>>Don't understand these downsides, why they would exist. The
>>retuning of intervals rather than scales should make it easier
>>than ever to get good 7-limit tunings, for example.

[Paul:]
>If you look solely at intervals, rather than larger groupings, there
>would be no way to get good 7-limit tunings, because, for example,
>you'd have no way of targeting one minor third to 6:5 while
>simultaneously targeting another minor third to 7:6.

Right.

[Carl:]
>>Who said you wouldn't look at larger groupings?

[Paul:]
>John _does_ look at larger groupings in many cases, including the
>7-limit cases. These larger groupings are what the "tuning files" are
>for.

Correct.

[Carl:]
>>It used to map the meantone diatonic scale to the classic
>>5-prime-limit just diatonic.

[Paul:]
>John deL . . . is this true? I don't remember this ever being your
>method.

It's not true, if I'm understanding Carl. Each tuning file _does_
specify a self-consistent tuning, but as the piece modulates, the tuning
file is re-applied in any or all of the 12 keys. I don't think there
has ever been a forced adherence to any fixed scale in any of my work.

JdL

>>>It used to map the meantone diatonic scale to the classic
>>>5-prime-limit just diatonic.
>>
>>
>>John deL . . . is this true? I don't remember this ever being your
>>method.
>
>It's not true, if I'm understanding Carl. Each tuning file _does_
>specify a self-consistent tuning, but as the piece modulates, the
>tuning file is re-applied in any or all of the 12 keys. I don't
>think there has ever been a forced adherence to any fixed scale in
>any of my work.

I'm having a hard time understanding why this wasn't clear. I've
participated in thousands of lines of conversation on this topic,
and used the software for many hours. I know you used a key-
guessing algorithm, John. Does this or does this not change the
fact that you were treating scale positions as the fundamental
object of tuning? In any case, I'm glad to learn you've developed
an interval-based approach. The chopin is killer.

-Carl

>> In my experience, the great length of discussions isn't the
>> part the passerby cares about.
>
> Huh?

The great length is often required to reach a conclusion.
I want just the conclusion.

As someone who's spent many hours in long threads, I know what
it's like to have somebody come along and ask for a summary.
"Check the archives" is what I've told those people many times.
But I knew even then that "check the archives" is useless.
The chance that the end user is going to be able to find, read,
decipher all the unique language developed for, and understand
a thread is nil. Especially with the caveman-primitive
threading and message-handling tools used for these lists.

There are two things which will play a key role in wether this
community survives its growing pains and continues to be
productive:

() We need a FAQ. I'll volunteer 40 hours of my time to it.
Who's with me?

() We should look at moving to a more advanced, thread-based
forum. There'd be one list, with a web site. A database on the
site tracks threads by topic (i.e. harmonic entropy, practical
microtonality...), subject (i.e. 4.7 cents per sec.), contributing
authors, and activity. The reader can sort by these fields, have
the database show all threads with any of his favorite authors,
etc. Threads could be linked by a "see also" field. E-mail could
be turned on or off by thread, topic, author, or the entire forum.
I've used forums like this, but I've never seen anything quite
this powerful. Does anybody know if it exists?

-Carl

[Carl L wrote:]
>>>>It used to map the meantone diatonic scale to the classic
>>>>5-prime-limit just diatonic.

[Paul E:]
>>>John deL . . . is this true? I don't remember this ever being your
>>>method.

[I wrote:]
>>It's not true, if I'm understanding Carl. Each tuning file _does_
>>specify a self-consistent tuning, but as the piece modulates, the
>>tuning file is re-applied in any or all of the 12 keys. I don't
>>think there has ever been a forced adherence to any fixed scale in
>>any of my work.

[Carl:]
>I'm having a hard time understanding why this wasn't clear. I've
>participated in thousands of lines of conversation on this topic,
>and used the software for many hours. I know you used a key-
>guessing algorithm, John. Does this or does this not change the
>fact that you were treating scale positions as the fundamental
>object of tuning? In any case, I'm glad to learn you've developed
>an interval-based approach. The chopin is killer.

Glad you like it! I've got a new version, as reported on tuning-math,
that isn't uploaded yet. Carl, I'd describe my methods (when using
tuning file(s)) as chord-matching more than scale-position matching.
Each tuning file provides one or more templates for well-tuned chords
(and of course, a whole host of poorly-tuned chords, noted in the
"goodness" measurements for the 66 intervals in each key).

Let me approach our confusion with this question: what do you think was
(or is) "certainly unwise" about my past methods that is "not as unwise"
with the newer tuning file free methods?

JdL

>Let me approach our confusion with this question: what do you think
>was (or is) "certainly unwise" about my past methods that is "not
>as unwise" with the newer tuning file free methods?

What confusion? I was agreeing with Paul, and presumably you as well...
you switched, didn't you!?

-Carl

[I wrote:]
>>Let me approach our confusion with this question: what do you think
>>was (or is) "certainly unwise" about my past methods that is "not
>>as unwise" with the newer tuning file free methods?

[Carl L:]
>What confusion? I was agreeing with Paul, and presumably you as
>well... you switched, didn't you!?

LOL! No, I'm still tuning as before, targeting intervals, not scale
degrees.

JdL

[JdL:]
>>>My "strange" reasoning is that the price of any such technique is
>>>additional horizontal motion, the bane of adaptive tuning, IMHO.

[Manuel:]
>>But you said "to my ears". Is it your ears, your intellect or a dogma?
>>Should the adaptation be restricted to the onset of the tones?

[JdL:]
>I'm simply saying that horizontal motion is painful (to my ears, not my
>"intellect" or "dogma").

It should be done in a musical way. Vibrato is horizontal motion too.
I don't see a reason to be so fundamentally opposed to it as to call
it the bane of adaptive tuning. Provided it's done in a tasteful way,
I think it can make an improvement. For instruments with a fixed tuning
like the harpsichord one should of course be extra careful with it.
But the rest you write shows that you perhaps didn't mean it so strongly.

>I have already said that some adaptation after onset of the tones may
>be desirable, because it means less at the note transition.

Exactly. Sethares made the adaptation of already sounding notes ten times
slower than that of the note attack. So this slow adaption would
probably be in the same range as the 4.7 cents/sec.
Because we know that there's a lot the ear doesn't notice (the success
of the MP3-format is based on it) I think there's good reason to do
more experiments with this kind of adaptive tuning before we dismiss it
or not.

>Do you know if the people who did this study left behind any audio
>examples?

No, I don't have any contact with the authors.

Manuel

Paul wrote:
>Manuel, John _very_ wisely does _not_ target a 5-prime-limit JI
>scale, as any such scale has comma problems.

Yeah, I realised that after I posed the question. I read that he
tried two different minor sevenths, which made me wonder about
the other intervals.

>So what's Malcolm's monochord?

I don't know much about its history. Invented by a Malcolm in
1721. It's in my DX7 as the just major scale:
16/15 9/8 6/5 5/4 4/3 45/32 3/2 8/5 5/3 16/9 15/8 2/1

Manuel

[I wrote:]
>>I'm simply saying that horizontal motion is painful (to my ears, not
>>my "intellect" or "dogma").

[Manuel Op de Coul:]
>It should be done in a musical way. Vibrato is horizontal motion too.

Well, that's a good point! I'm not a huge fan of vibrato, but it's not
as painful as some horizontal motion, to be sure.

>I don't see a reason to be so fundamentally opposed to it as to call
>it the bane of adaptive tuning. Provided it's done in a tasteful way,
>I think it can make an improvement. For instruments with a fixed tuning
>like the harpsichord one should of course be extra careful with it.
>But the rest you write shows that you perhaps didn't mean it so
>strongly.

Everything I say of course reflects my particular focus in the way I
tune. I do also try to make clear that I know there are other forces
and other possibilities which I do _not_ yet take into account; learning
about them is an ongoing process.

[JdL:]
>>I have already said that some adaptation after onset of the tones may
>>be desirable, because it means less at the note transition.

[Manuel:]
>Exactly. Sethares made the adaptation of already sounding notes ten
>times slower than that of the note attack. So this slow adaption would
>probably be in the same range as the 4.7 cents/sec.
>Because we know that there's a lot the ear doesn't notice (the success
>of the MP3-format is based on it) I think there's good reason to do
>more experiments with this kind of adaptive tuning before we dismiss it
>or not.

Oh, I agree! Might be a bit tough to graft onto my stuff, especially
since I'm still a bit vague on the details.

>>Do you know if the people who did this study left behind any audio
>>examples?

>No, I don't have any contact with the authors.

OK, then we all just need to keep our ears open for more info. Thanks
again for bringing all this up!

JdL

--- In tuning@y..., Carl Lumma <carl@l...> wrote:
> >Let me approach our confusion with this question: what do you think
> >was (or is) "certainly unwise" about my past methods that is "not
> >as unwise" with the newer tuning file free methods?
>
> What confusion? I was agreeing with Paul, and presumably you as
well...
> you switched, didn't you!?
>
> -Carl

I don't think, and presumably John doesn't either, that there was
anything "certainly unwise" about the past methods, at least nothing
nearly as "certainly unwise" as targeting Ellis' Duodene, as Manuel
was orignally suggesting in this thread.