> Other harmonic-based tunings don’t
need this – or at least have enough other problems to not turn to
Stretch for their solution.
I think this might be the crux of the situation.? Equal
Temperament is very close to an ideal choice, but tuners noticed
that following the rules of tuning exactly didn't end up with the
best end results, at least, to the ear.
But the string issue is real:? the windings do not track the
center core perfectly, and the issue grows over the distance from
the reference string -- in our case, usually Middle C.? So, it
becomes a choice:? do we follow the mathematical model -- which
results in the same beating between intervals of the
*fundamentals* across the keyboard -- or do we focus on the
beating between harmonics by stretching the tuning?? Since a piano
is harmonic-rich, the answer seems to be the latter.
________
But I still wonder if the process applies to digital piano
samples; or to non-string instruments like the Rhodes.? It's
difficult to answer, because any reference piano is only going to
remain perfectly tuned for a very short time.? Within minutes,
string tension and temperature/humidity changes conspire to alter
the pitch of each string.? So, one cannot compare a grand piano
vs. a sample (or modeled) piano and make strong conclusions about
frequency relations, unless its possible to do all the analysis
within minutes of a piano tuner's work.
That said, sometimes adding a little imperfection improves the
overall sound, even if the results aren't 100% accurate.? As an
example:? I've found that the Mellotron samples of my Nord Electro
sound much more realistic if I apply the 'Vibrato' effect at a
very slow setting.? Now, if one examines what that effect does at
slow speed, it actually creates a non-linear pitch bend that is
mostly flat, with a slight 'whip' of FM at the end, kind of like
this --
?????? ______/\______/\______/\__
This is clearly *not* how the wow and flutter of a Mellotron's
tape mechanism occurs.? Nevertheless, the slow vibrato adds a
feeling that is missing when playing the dry samples -- even
though they were recorded on a real Mellotron, with its own wow
and flutter.? It may be that adding stretch values to a digital
piano better simulates the effect of tuning an acoustic piano,
then allowing the piano to settle a bit before use.? So, if
stretch tuning improves your Kurzweil's illusion, or the realism
of an FM piano, that's not a bad thing.? I just wouldn't stress
over finding the 'perfect' tuning rules, because it may be that
any small variation improves the sound.
Regards,
-BW
Bruce Wahler
Halfmoon-Switch.com?
bw@...
978.597.7008
PLEASE NOTE: ?The AshbySolutions.com domain will be closed soon.
?Please use the new
(wahler.us) email for all future correspondence. ?Thank you for
your understanding.
?
?
On 8/20/2021 7:05 PM, LarryS wrote:
toggle quoted message
Show quoted text
I have heard this
version before and I have heard other versions of WHY
stretch tuning.?
While I’m not
certain which explanation best encompasses all the physical
world, I do know the debate carries on in certain circles.?
What debate??
It’s not that the
strings or their makeup that have anything to do with it,
it’s our own imperfect, non-mechanical ears.? Of course, who
wants to blame listeners, and ultimately customers?? Poor
form.? So we blame the strings.
Trouble is, as we
get closer models, both synth and sample, stretch tuning
becomes more apparent – with nary a string involved.? My
go-to for piano is a Kurzweil and yes, it sounds better with
a few tweaks of stretch.
Stretch Tuning
seems to be an outgrowth of Equal Temperament tuning.? Other
harmonic-based tunings don’t need this – or at least have
enough other problems to not turn to Stretch for their
solution.
But like Bose
speakers and Aphex Bottoms, it’s not about what’s
mathematically correct, it’s what “sounds better”.? After
all, this is how stretch tuning came about.? The various
explanations came about much later.
For a couple of
decades, it’s been a given that anything that can bring more
animation into the sound of a DX7 is a good thing.? My guess
is this is no exception.? Lots of unexplored possibilities,
here.
BUT – this will
be at-best a very subtle change to a 6 op FM patch.
Heck, it’s a
pretty subtle change to an acoustic piano.
?
L.
?
?
?
Exactly. The Railsback stretch/effect is
due to the “inharmonicity” of the strings vibrating. To my
knowledge, not even Pianoteq goes so far as to model this. I
expect you’ll find no positive benefit from employing the
“stretch tuning” as a result. But do keep us posted; it’s an
interesting notion. It would never have occurred to me to
apply those offsets to virtually generated tones.
?
?
?
I suspect
that if stretch tuning benefits an FM piano, it will be
something of a 'happy accident.'? Stretch tuning is
necessary on an acoustic piano because of the way that a
wound string behaves -- i.e., the windings do not perfectly
follow the frequency of the core.? Stretch tuning is
actually a compromise between the 'correct' mathematical
pitch and one which sounds more pleasing to the listener.
If one attempts to stretch tune say, a Fender Rhodes
electric piano, the curve will be different, because tines
act differently than strings.? Nevertheless, a stretch-tuned
Rhodes sounds different than a equal-tempered Rhodes.? (I'm
not going to choose a winner here, because attempting to
copy acoustic piano idiosyncrasies on a Rhodes is a
debatable endeavor.? A piano is a piano, and a Rhodes is its
own instrument.)
Since an FM piano is closer to a mathematical model, I would
expect that the fundamental/harmonic ratio differences would
be minimal across the scale; in fact, any differences might
not even be linear.
________
This is, of course, quite different from the microtuning
changes used by Wendy Carlos and others.? In this case, the
composer or player is trying to better duplicate the
historical performance, which may use a tuning other than
equal temperament.
Regards,
-BW
Bruce Wahler
Halfmoon-Switch.com?
bw@...
978.597.7008
PLEASE NOTE:
?The AshbySolutions.com domain will be closed soon.
?Please use the new (wahler.us) email for all future
correspondence. ?Thank you for your understanding.
?
?
On 8/20/2021 7:27 AM, Daniel Forró via
groups.io wrote:
I have found in my archive another
table sent to me by George Chakiris, it uses only
small differences. I attached it.
The basic question is how stretch
tuning will sound with FM generated electronic piano
sound, which is far from the real acoustic piano
sound, and if it has any reason to use is. I don’t
think it can help to improve sound.
There was interesting discussion
many years ago in microtonal composers circles about
the relation between timbre and microtonal tunings. I
think the person behind the research is a microtonal
music composer William Sethares. Try to start here:
?
?
On Aug 20, 2021, at 3:26,
philfrei via <philfrei@...>
wrote:
?
Thanks for the nudge. I was
personally emailed a file with settings
that contain pitch adjustments for every
note, and the equation that generated the
numbers. It looks quite interesting, and
was developed by the person who contacted
me. I assume that the individual reads
this list and I'd like to encourage them
to share the chart and equation with the
full group.
The bottom line, though, is that I don't
know yet if there exists a chart on the
Railsback effect produced by Yamaha
engineers in any of their manuals. I've
been a bit more focused on learning
whether or not there is an industry-level
standard practice on the issue than in
testing how it sounds. So far, from the
various people who are closer to studio
recording tech than I, there doesn't seem
to be much use or awareness of these pitch
adjustments, which kind of surprises me.
Maybe it is related to the ability of an
acoustic musician to make adjustments on
the fly, as I learned to do when playing
the oboe in the studio.
So, unfortunately, the task of testing how
it actually sounds got back-burnered. A
related project, more front and center a
the moment, is a pilot test I have going
for an experiment on whether timbre
affects the perceived pitch of notes. I'm
collecting data from a handful of people,
and need to find a good data-analysis
person to ensure that what I'm collecting
via the test is suitable for
quantifying/modeling. (I'm more a
composer/programmer than a scientist.)
I haven't used my DX7S in the studio
much lately and figuring out the
logistics of inputting and using the
data is a bit nervous-making (due to
unfamiliarity), though probably doable.
I've been doing more with an FM synth I
programmed using Java that can produce
equivalent voices to those I developed
on the DX7. When I get a chance, the
plan is to work the Railsback Curve data
into the functions that convert MIDI
pitches to Hz and vice versa, and see
how it sounds. If I get this working,
I'll post links to a something that
compares the two intonations
back-to-back. But it could be another
month or more.
>From: Eb Myatt
>Date: Thu, 19 Aug 2021 02:19:50 PDT
>Good Morning Phil,
>
>We haven't received any feedback
from you since you last posted. So,
did you manage to resolve the
>"Railsback" tuning issue that you
discussed? Please let us know if you
have the table for the DX7s >or any
other Yamaha synth ?
>
>Best Regards,
>E?
?
?
?
?
?
?
?
|
