Yet another temperament comparison

This for people who want to get serious about it. I don't who, if
anyone, will have the time and energy to pursue this, but here is my
entry:

http://bahamas.eshockhost.com/~xenharmo/mp3/bwv542/

This gives seven different versions of BWV 542, in 12-et, 31-et, grail,
Werck3, squiggle, Valotti-Young and Neidhard1. The ordering was chosen
by computer selection of a random permutation.

The point is not to guess which is which but to see what you like; I'd
be interested to hear about that if anyone is up to all the listening
this involves. In any event, I'll use these files to create a
temperament comparison page, with the names, later on.

Gene Ward Smith wrote:
> This for people who want to get serious about it. I don't who, if > anyone, will have the time and energy to pursue this, but here is my > entry:
> > http://bahamas.eshockhost.com/~xenharmo/mp3/bwv542/
> > This gives seven different versions of BWV 542, in 12-et, 31-et, grail, > Werck3, squiggle, Valotti-Young and Neidhard1. The ordering was chosen > by computer selection of a random permutation.
> > The point is not to guess which is which but to see what you like; I'd > be interested to hear about that if anyone is up to all the listening > this involves. In any event, I'll use these files to create a > temperament comparison page, with the names, later on.

Well, this is pretty long, so I'll just listen to the fantasia for now and leave the fugue for later.

A: This seems like an appropriate, powerful tuning for this music. Nothing really stands out, but it does sound like it might be a little unequal (or in any better than expected for ET). Now watch as this one turns out to be 12-ET! But I'm guessing that f.mp3 is the ET one.

B: Something about this that I can't quite identify seems a little less solid than the first one. Parts of it sound pretty good though. In any case, it sounds as appropriate as any of the others, and some of the key color variations are effective.

C: A pleasant tuning. The sequence of key colors around 4:00 has a nice progression effect; you can tell there's something going on. I'll tentatively identify this one as the one I like the best, but many of the others are about as good.

D: This one doesn't hold up as well as the others. It's an interesting effect, like playing Beethoven in 15-ET, but some parts really don't work at all. Parts sound like meantone, and other parts sound like meantone with misspelled notes.

E: This tuning has a generally pleasant effect, without having much of a distinctive flavor of its own. It does seem appropriate for the organ timbre. This would be my second favorite after C. Or maybe this one is better; it's really a pretty good fit.

F: Seems like a pretty generic tuning. I can't detect any key contrast in the 4:00-4:20 progression. I vote for this one as "most likely to be mistaken for ET".

G: I like this one, although after hearing all the others it's getting harder to point out anything distinctive about it. The key differences are pretty subtle, but it does seem as if some of the keys have a different quality from others.

--- In tuning@yahoogroups.com, Herman Miller <hmiller@...> wrote:

> Well, this is pretty long, so I'll just listen to the fantasia for
now
> and leave the fugue for later.

Very interesting; I'll respond but probably not for a while.

> This for people who want to get serious about it. I don't who,
> if anyone, will have the time and energy to pursue this, but
> here is my entry:
>
> http://bahamas.eshockhost.com/~xenharmo/mp3/bwv542/

Wow, long files. I'll just listen to the fantasia. One of
my favorite Bach pieces of all time.

Also, these are the best-sounding renderings I've ever heard
from you.

> This gives seven different versions of BWV 542, in 12-et,
> 31-et, grail, Werck3, squiggle, Valotti-Young and Neidhard1.
> The point is not to guess which is which but to see what
> you like; I'd be interested to hear about that if anyone is
> up to all the listening this involves.

d immediately sounds different than a-c. It's gotta be
grail. It works really well in the introduction with the
diminished chords and pedal point. It starts to sound
too far out somewhere around 1:15 (where there's an awkward
upward melodic leap in the tenor). And the run at 1:39
sounds wrong. It's still enjoyable and ear-opening to
listen to the piece, but it wouldn't be my choice if I
was performing it.

E is interesting. Probably my favorite, given the very
limited listening I'm doing.

I'm surprised 31 didn't jump out at me.

-Carl

Hey Gene,

It's late here and my mind and ears are tired. I wanted to thank you
for doing this.

Quickly, I want to listen more to this with more critical ears, but I
found that I liked e for it's attrctive consonance--I would guess off
the bat that this was 31-equal, but I could be wrong.

c sounded close to equal, at least for the beginning part that I
heard. It seemed that the interval sizes were pretty consistant.

I'm confident d was grail---it's out there.

I'll comment more as I get more into it...the files are long and I
only got partway in most of the time. I'm fried.

I wanted to say that I liked the performance---one of the more
realistic MIDI files in terms of expressive articulation and agogic
accent---it was nice to here a MIDI file that didn't march in strict
metronomic time, but had bouyancy...where did you find it.

I'm also, on the side, curious how you went about retuning to
31-equal---are you just using a subset of 31-equal, matched to the
gamut (ala 1/4-c. meantone?). What's your software and process?

Thanks,
Aaron.

--- In tuning@yahoogroups.com, "Gene Ward Smith" <genewardsmith@...>
wrote:
>
> This for people who want to get serious about it. I don't who, if
> anyone, will have the time and energy to pursue this, but here is my
> entry:
>
> http://bahamas.eshockhost.com/~xenharmo/mp3/bwv542/
>
> This gives seven different versions of BWV 542, in 12-et, 31-et, grail,
> Werck3, squiggle, Valotti-Young and Neidhard1. The ordering was chosen
> by computer selection of a random permutation.
>
> The point is not to guess which is which but to see what you like; I'd
> be interested to hear about that if anyone is up to all the listening
> this involves. In any event, I'll use these files to create a
> temperament comparison page, with the names, later on.
>

P.S. Gene, I'd love to hear your comments on my listening test!

-A.

--- In tuning@yahoogroups.com, "Gene Ward Smith" <genewardsmith@...>
wrote:
>
> This for people who want to get serious about it. I don't who, if
> anyone, will have the time and energy to pursue this, but here is my
> entry:
>
> http://bahamas.eshockhost.com/~xenharmo/mp3/bwv542/
>
> This gives seven different versions of BWV 542, in 12-et, 31-et, grail,
> Werck3, squiggle, Valotti-Young and Neidhard1. The ordering was chosen
> by computer selection of a random permutation.
>
> The point is not to guess which is which but to see what you like; I'd
> be interested to hear about that if anyone is up to all the listening
> this involves. In any event, I'll use these files to create a
> temperament comparison page, with the names, later on.
>

--- In tuning@yahoogroups.com, "Aaron Krister Johnson" <aaron@...>
wrote:

> I wanted to say that I liked the performance---one of the more
> realistic MIDI files in terms of expressive articulation and agogic
> accent---it was nice to here a MIDI file that didn't march in strict
> metronomic time, but had bouyancy...where did you find it.

Kunst der Fugue, I think. It's recorded from a live performance,
which gets away from the mechanical sound.

> I'm also, on the side, curious how you went about retuning to
> 31-equal---are you just using a subset of 31-equal, matched to the
> gamut (ala 1/4-c. meantone?). What's your software and process?

No, it's extended meantone.

(1) I take a midi file and convert it to a seq file, in the
sharps/flats, not numerical, notation.

(2) I run that through a sed script which produces another file,
where the notation is now numerical in terms of generators--that is,
F is -1, C is 0, G is 1, D is 2 and so forth.

(3) This is read into Maple as a Maple file, and then I have Maple
programs analyze it. An important thing I look for is the meantone
span of each chord--that is, the farthest up the circle of fifths
minus the farthest down. Hence a G major triad, {1,2,5}, would have a
span of 4. If the span of every chord in the piece is 10 or less, I
have a good starting point, because then every interval is either a
dissonance no matter how you adjust it enharmonically by a diesis up
or down, is a 5-limit consonance, or is a 7 or 9-limit consonace.
That doesn't mean I'm done, but it does mean things are not going to
get totally out of hand.

(4) I now use the information Maple provides me to edit the seq file,
replacing notes with enharmonic equivalents--ie, adjusting up or down
by a diesis. That is, I swap an Ab for a G# and so forth.

(5) Then I feed this into the sed script again and iterate the
process.

Writing a program to automate this process would be possible, and
quite an achievement for someone.