Bach in non 12/keyboard tuning

Wendy Carlos's "Switched On Bach 2000" uses a number of non 12, circulating temperaments...and Richard Egarr's recent recording of the Goldberg Variations uses the infamous Lehman "squiggles" tuning. I find it fascinating, and a bit amazing, that someone could get a very practical, useable well temp out of those squiggles if they were just random doodles...Egarr was convinced enough to do a whole CD using that system, I think that says a lot.
And, I really like Ross Duffin's new book "How Equal Temperament Ruined Harmony;" one of the cool things he does in the book is to not focus so much on just keyboard tunings. He quotes famous violinists, flutists, and vocal teachers who talk, pretty consistently, about the whole step being divided into 9 commas (which, he says, implies 55 tones/octave, which Telemann mentioned)...old Leopold Mozart also mentions how flat notes and sharp notes (G#-Ab) are separated by a comma...it explains a lot of basic concepts really well to boot, I recommend it to my students who are interested in tunings.
And, lastly, a quote from the great Pablo Casals from the book: "Do not be afraid to be out of tune with the piano. It is the piano that is out of tune. The piano with it's tempered scale is a compromise in intonation." Amen. Maybe other cultures who used fretless instruments (the oud for example), or axes with moveable tuners (the zheng, or the tar, with it's moveable frets) were onto something...best...Hstick
microstick.net agenturadellarte.net guitar9.com

Hi Neil,

As your list of non-12 Bach points out, these Bach performances are not
WIII. My position is that Bach used WIII and that one needs to hear Bach in
WIIII to appreciate what I am addressing. Wendy Carlos and I have talked about
this for years, quite positively I might add.

Pablo Casals did have that intuitive ability to anticipate intonationally
where WIII would be: The D minor cello suite has a sharp minor third to ET while
the C minor cello suite has a flat minor third to equal. This mirrors
beautifully what is contained in WIII and many other WTs.

You said you had heard Bach in WIII. May I ask what you heard? Here's the
complete List of WIII on CD:
Igor Kipnis, harpsichord and clavichord
Ton Koopman, conductor, keyboard
Rebecca Pechefsky, harpschord
PITCH Early CD, Brandenberg's, WTC selections, Werckmeister Christmas
Cantata, JM Bach Concerto

There was an attempt to say suggest the musical market is saturated with
WIII. I thought it might be so in Europe since it was Daniel who said so. Um,
nothing comes up. I really think I am talking in part about an experiential
thing. No harm there.

best, Johnny

--- In tuning@yahoogroups.com, <microstick@...> wrote:
> Richard Egarr's recent recording of the Goldberg Variations
> uses the infamous Lehman "squiggles" tuning. I find it
> fascinating, and a bit amazing, that someone could get a very
> practical, useable well temp out of those squiggles if they
> were just random doodles...

I refer you to:

http://en.wikipedia.org/wiki/Runamo

-Carl

To those interested in the difference between equal,
WIII, and another well temperament in a recording of
a Bach WTC piece, may like to listen to this:

http://lumma.org/stuff/bach-tuningA.mp3

http://lumma.org/stuff/bach-tuningB.mp3

http://lumma.org/stuff/bach-tuningC.mp3

See if you can guess which tuning is Equal temperament,
Werck III, Young. It isn't impossible, but I'd be surprised
if half the people participating in this thread can do it
correctly.

To control for performance differences, these were all
rendered from the same MIDi file using the same samples.
I suppose some will say the samples offend their ear so
they can't judge the tuning: a potentially valid, if
suspect, excuse.

-Carl

The first sounds like Young, the second, 12-eq, the third, Werckmeister III.

Oz.

----- Original Message -----
From: "Carl Lumma" <clumma@yahoo.com>
To: <tuning@yahoogroups.com>
Sent: 15 Ocak 2007 Pazartesi 18:57
Subject: [tuning] Re: Bach in non 12/keyboard tuning

> To those interested in the difference between equal,
> WIII, and another well temperament in a recording of
> a Bach WTC piece, may like to listen to this:
>
> http://lumma.org/stuff/bach-tuningA.mp3
>
> http://lumma.org/stuff/bach-tuningB.mp3
>
> http://lumma.org/stuff/bach-tuningC.mp3
>
> See if you can guess which tuning is Equal temperament,
> Werck III, Young. It isn't impossible, but I'd be surprised
> if half the people participating in this thread can do it
> correctly.
>
> To control for performance differences, these were all
> rendered from the same MIDi file using the same samples.
> I suppose some will say the samples offend their ear so
> they can't judge the tuning: a potentially valid, if
> suspect, excuse.
>
> -Carl
>
>
>

The first sounds like Young, the second, 12-eq, the third, Werckmeister III.

Oz.

Yesaroni. Young sounds even and predictable. ET sounds more rinky-dink. WIII, which could have started anywhere for a different basic key sound, sounds quite expressive.

Johnny

________________________________________________________________________
Check out the new AOL. Most comprehensive set of free safety and security tools, free access to millions of high-quality videos from across the web, free AOL Mail and more.

I take it you agree with my observations. But what's with all the veneration about WIII?

Oz.
----- Original Message -----
From: Afmmjr@aol.com
To: tuning@yahoogroups.com
Sent: 15 Ocak 2007 Pazartesi 19:33
Subject: Re: [tuning] Re: Bach in non 12/keyboard tuning

The first sounds like Young, the second, 12-eq, the third, Werckmeister III.

Oz.

Yesaroni. Young sounds even and predictable. ET sounds more rinky-dink. WIII, which could have started anywhere for a different basic key sound, sounds quite expressive.

Johnny

I'll give the answers on Sunday night for anyone else
who wants to try.

-Carl

> The first sounds like Young, the second, 12-eq, the third,
> Werckmeister III.
>
> Oz.
>
> ----- Original Message -----
> From: "Carl Lumma" <clumma@...>
> To: <tuning@yahoogroups.com>
> Sent: 15 Ocak 2007 Pazartesi 18:57
> Subject: [tuning] Re: Bach in non 12/keyboard tuning
>
> > To those interested in the difference between equal,
> > WIII, and another well temperament in a recording of
> > a Bach WTC piece, may like to listen to this:
> >
> > http://lumma.org/stuff/bach-tuningA.mp3
> >
> > http://lumma.org/stuff/bach-tuningB.mp3
> >
> > http://lumma.org/stuff/bach-tuningC.mp3
> >
> > See if you can guess which tuning is Equal temperament,
> > Werck III, Young. It isn't impossible, but I'd be surprised
> > if half the people participating in this thread can do it
> > correctly.
> >
> > To control for performance differences, these were all
> > rendered from the same MIDi file using the same samples.
> > I suppose some will say the samples offend their ear so
> > they can't judge the tuning: a potentially valid, if
> > suspect, excuse.
> >
> > -Carl

Carl Lumma wrote:
> To those interested in the difference between equal,
> WIII, and another well temperament in a recording of
> a Bach WTC piece, may like to listen to this:
> > http://lumma.org/stuff/bach-tuningA.mp3

I'm guessing this is Young. It sounds a little unpleasant.

> http://lumma.org/stuff/bach-tuningB.mp3

It seems a bit more pleasant. Werckmeister? Ah, but this is in A major. Now I don't remember which keys are better in which temperaments, but I don't recall A major being one of the better ones in Werckmeister. Whatever this is, it sounds better than the other one. I'm sticking with my first answer, since I generally tend to like Werckmeister better than Vallotti or Young.

> http://lumma.org/stuff/bach-tuningC.mp3

Equal temperament.

> See if you can guess which tuning is Equal temperament,
> Werck III, Young. It isn't impossible, but I'd be surprised
> if half the people participating in this thread can do it
> correctly.

I'd be interested in a Kirnberger III version for comparison, or something more exotic like Sorge 1758 or one of Neidhardt's.

> To control for performance differences, these were all
> rendered from the same MIDi file using the same samples.
> I suppose some will say the samples offend their ear so
> they can't judge the tuning: a potentially valid, if
> suspect, excuse.

Is that a sample of Peter Schickele playing the tromboon in the bass part? :-)

Is that the original key? My first impression was, it's in the wrong
key.

--- In tuning@yahoogroups.com, "Carl Lumma" <clumma@...> wrote:

> To control for performance differences, these were all
> rendered from the same MIDi file using the same samples.
> I suppose some will say the samples offend their ear so
> they can't judge the tuning: a potentially valid, if
> suspect, excuse.

The real problem isn't that the samples offend the ear, it's the
relationship between the pitch of the origianl samples and the
amount of pitchbend in the retuning. If the original samples are
pitched for 12-EDO, another tuning can sound worse simply because of
crappy resampling and subtle deformation of the formants.

This problem plagues "xenharmonic" music.

Anyway, I have to agree with Oz and Johnny on this one, but I
wouldn't be too surprised if the third one is Young and the first
WIII and the whole thing is in the "wrong key".

-Cameron Bobro

> Is that the original key? My first impression was, it's in the
> wrong key.

The piece is in A, the tunings are from the Scala archive,
and I believe I used C=262 or whatever.

> The real problem isn't that the samples offend the ear, it's the
> relationship between the pitch of the origianl samples and the
> amount of pitchbend in the retuning. If the original samples are
> pitched for 12-EDO, another tuning can sound worse simply because
> of crappy resampling and subtle deformation of the formants.

These samples aren't full-keyboard anyway, so they'd be stretched
far further in 12 than the difference between 12 and a
well temperament.

-Carl

--- In tuning@yahoogroups.com, "Cameron Bobro" <misterbobro@...> wrote:

> The real problem isn't that the samples offend the ear, it's the
> relationship between the pitch of the origianl samples and the
> amount of pitchbend in the retuning. If the original samples are
> pitched for 12-EDO, another tuning can sound worse simply because of
> crappy resampling and subtle deformation of the formants.
>
> This problem plagues "xenharmonic" music.

How finely do you think the octave would need to be sampled to prevent
this? Of course, rather than bending a single sample, other things
could be done, such as a sophisticated interpolation scheme.

--- In tuning@yahoogroups.com, "Gene Ward Smith" <genewardsmith@...>
wrote:
>
> --- In tuning@yahoogroups.com, "Cameron Bobro" <misterbobro@>
wrote:
>
> > The real problem isn't that the samples offend the ear, it's the
> > relationship between the pitch of the origianl samples and the
> > amount of pitchbend in the retuning. If the original samples are
> > pitched for 12-EDO, another tuning can sound worse simply
because of
> > crappy resampling and subtle deformation of the formants.
> >
> > This problem plagues "xenharmonic" music.
>
> How finely do you think the octave would need to be sampled to
prevent
> this? Of course, rather than bending a single sample, other things
> could be done, such as a sophisticated interpolation scheme.

It's a fairly complex thing, and not entirely predictable.

The very poor sampling instruments suffer the least, in a way- if
the whole instrument is just a high pitched squawk mickey-moused
down, it doesn't really matter much exactly where it's sampled to.
That should and often does sound like crap, but even that needs a
footnote, because there are some completely decent sampled sounds
that are done just that way, when you peak under the hood.

When you get a modern instrument, which has samples ranging from at
least every octave to multiple velocity layers on every single key,
like a Gigasampler instrument, it can be just straight playback,
assuming that the sounds have been carefully pitched beforehand and
there's no difference between the pitch of the sample and its
playback pitch. IE, when you play A-440, the sample at that key is
pitched at a-440 and there's no resampling or other pitchshifting
going on when you hit the key.

Which sounds great- if you want 12-EDO at A-440. As soon as you tune
away from this, even with great interpolation, you're still moving
the formants along with the sample. Even with formant filtering at
the end of the chain, if it's a real acoustic instrument, the
formant information is in the original sample, and it's never going
to be right.

Throw in tunings with more than 12 pitches per octave, unless you go
in yourself and assign samples to notes, and it can get seasick very
quickly.

So the best way is to sample instruments at the pitch they'll be
played back at, and if that's not possible, fine-tune each
individual sample in a quality DAW.

Since that's not usually practical, especially if you're trying many
tunings, what I do is try not to task the sampler too much. So place
a sample and have its range go down a few keys, definitely a
different sample for different registers, at least one for each
string, limited ranges, etc. The classic tricks.

VSampler is around 150 dollars, has several interpolation settings,
for realtime and superb rendering, reads and writes all kinds of
formats, and takes Scala files. Awesome!

-Cameron Bobro

--- In tuning@yahoogroups.com, "Carl Lumma" <clumma@...> wrote:
>
> > Is that the original key? My first impression was, it's in the
> > wrong key.
>
> The piece is in A, the tunings are from the Scala archive,
> and I believe I used C=262 or whatever.
>
> > The real problem isn't that the samples offend the ear, it's the
> > relationship between the pitch of the origianl samples and the
> > amount of pitchbend in the retuning. If the original samples are
> > pitched for 12-EDO, another tuning can sound worse simply because
> > of crappy resampling and subtle deformation of the formants.
>
> These samples aren't full-keyboard anyway, so they'd be stretched
> far further in 12 than the difference between 12 and a
> well temperament.
>
> -Carl
>

Yes, see my post to Gene. I was speaking in general. Anyway in this
case it would be no excuse.

-Cameron

--- In tuning@yahoogroups.com, "Gene Ward Smith" <genewardsmith@...>
wrote:
>
> --- In tuning@yahoogroups.com, "Cameron Bobro" <misterbobro@> wrote:
>
> > The real problem isn't that the samples offend the ear, it's the
> > relationship between the pitch of the origianl samples and the
> > amount of pitchbend in the retuning. If the original samples are
> > pitched for 12-EDO, another tuning can sound worse simply because of
> > crappy resampling and subtle deformation of the formants.
> >
> > This problem plagues "xenharmonic" music.
>
> How finely do you think the octave would need to be sampled to prevent
> this? Of course, rather than bending a single sample, other things
> could be done, such as a sophisticated interpolation scheme.

Of course, the obvious solution is to dispense with sample-based
synthesis entirely for tuning comparisons. Additive synthesis is
inherently indifferent to tuning systems, and can theoretically match
the performance of any sample-based synth using any sample set. A
project for someone on the list, perhaps? It's been on my to-do list
for quite some time, but I haven't got around to it yet.

Peace,
Patty