Tuning Acrosonics

[Dean L. Reyburn, RPT]

Like Kent Swafford, I also found Jim Coleman Sr.'s post on tuning
Acrosonics fascinating, and hate to disagree with someone as smart as
Jim, but I'm afraid I'll have to respectfully take another viewpoint.

I spent some time analyzing Jim's Acrosonic data in a spreadsheet.
The results were very interesting, and I've put them on the web at:

 www.reyburn.com/acrosonic.html

The data is there in downloadable MS-Excel format, dBase (DBF), and
Claris Works format. 

Reyburn CyberTuner gives the tuner a number of choices for tuning
partials. In the low bass partials 6, 7, 8, 10 and 12 are available. In
the mid-bass partial 4, 5 and 6 are the choices and in the midrange (A2
to G#4 for RCT) partials 3 and 4 are the choices. Above that partials 1
and 2 for an octave, then only fundamental above A5.

So what guidelines does the tuner use to decide on which partials to
tune with? In RCT the default partials (6-6-3-2-1) always work well, but
how does one determine the optimum tuning partial?

Here are the guidelines I used with RCT in setting up the default
partials for the midrange:

1. The partial must be strong, certainly strong enough to give a good
   visual display.
2. The partial should be in the middle of the range of strong partials 
   to which an aural tuner listens.

So what's wrong with the fourth partial in the midrange? It doesn't meet
either of these criteria on most smaller pianos. On many small pianos
such as the Acrosonic the 4th partial is very weak in the upper 4th
octave, about G4 ~ B4. The 4th is almost always weaker than the third
partial for most of the temperament range, and often gives a weak
pattern on the visual display.

Is the 4th partial the most consistent across the breaks? Jim's own data
says it's not, and that's my experience also. If you do the math, the
jumps in the 3rd partial curve at the bass/tenor and tenor/treble breaks
are the smallest with the 3rd partial (1.26 and 0.90 cents respectively).
The runner up is the *2nd* partial with jumps at the breaks of just
under two cents. The IH (inharmonicity) jumps at the 4th partial are a
distant third place at 4.67 and 3.43 cents at the two breaks. Take a
look at the graphs of the partial ladders and the stability of the third
partial (the yellow line) is even more striking than the numbers.

Even if you convert the cents to beats, the 3rd partial is still more
than twice as consistent as the 4th across the breaks on this Acrosonic.

On this piano, tuning a smooth curve across the two breaks at the 3rd
partial will give the best compromise possible. The intervals involving
the 2nd partial will be off in one direction, and those involving the
4th partial will be slightly in error in the opposite direction by about
the same amount. Certainly some tweaking could be done by ear on this
tuning even if it's tuned to a 3rd partial smooth curve across the
breaks, but it's unlikely any changes will be more that 0.5 cents, and
these would have to be done carefully, checking *all* the aural
intervals or it's easy to make one interval better at the expense of
another.

Jim also states that:

"In this chart you as see the irregularity of the lower partials
especially"

Maybe I'm looking at these numbers differently from Jim, but I'm seeing
the lower partials as most consistent, and the upper ones as less 
consistent. My spreadsheet analyzes how much each partial ladder from
C3 to A4 varies from the average, the 2nd and 3rd partial are both more
consistent than the 4th partial.

As Kent Swafford said in his January 1997 PTJournal article, the third
partial splits the difference between the 4:2 and 6:3 octaves, making
the best compromise possible on many or most small pianos. 
(See www.reyburn.com/innovations.html for the entire article)

Jim also recommends tuning this small piano using OTS 9, (Octave Tuning
Style) and a 2nd tuning record with a more conservative OTS in the bass.
In my opinion, this is very unnecessary with RCT. It has both way too
much stretch and adds complication where none is needed. 

This much stretch in the treble is way more than anyone that I know
would use on anything smaller than a 9' grand.

And I don't believe there's any need for the second tuning record. RCT
is very smart about custom fitting the tuning to the piano. On this
piano, if you do choose a high OTS such as 7, 8 or 9, RCT will smoothly
graduate the tuning so that it uses tons of stretch in the treble and is
very  conservative in the bass. If you use the 3rd partial both breaks
will be just about as smooth as you can get them since both breaks are
contained in the 3rd partial area (A2 to G#4).

Since Jim Coleman Sr., the long-time SAT teacher makes recommendations
on how to use RCT, I'll take the risk  of making a few recommendations
as to using the SAT: <g>

The two-page tuning technique Jim recommends does work well with the
SAT. And if you're using a SAT, then yes you should be using the 4th
partial in the midrange. That's what the SAT's FAC tuning calculator
produces and to use the 3rd partial would be a lot of work. Also, the
SAT displays motion relative to hertz only, so if you use a lower
partial with the SAT, your effective resolution goes down. For instance
if you choose a partial that's an octave lower, the display will move
half as fast, effectively cutting down sensitivity. Moving the
listening partial down from the 4th to the 3rd partial with a hertz
relative display will reduce the speed of the display by 25%.

I mention the above not as a criticism of the SAT at all, just to point
out that you have to keep in mind the specific capabilities of the
tuning device when you're talking about. What works best on one machine
may not be optimal on another. For instance, with RCT changing partials
has no effect on the sensitivity or speed of the spinner since it's
normally set to cents-relative (though you can choose either).

In RCT you normally get a tuning that is well fitted from low bass to
high treble with little need for anything but over-all stretch
adjustment to taste. Chameleon tuning calculator built into RCT has lots
of "smarts" and compensates well for short-scale pianos. And if none of
the standard nine OTS styles exactly match your taste in stretch, then
RCT has an advanced but easy-to-use tool called 'Custom Equalizer" which
lets you customize the stretch in each octave in aural tuning terms.
Custom EQ looks like a graphic equalizer and lets you micro-adjust and
predict a number of aural intervals with extreme precision. If you
should want to, you can tell Custom EQ to use OTS 9 in the treble and
OTS 4 in the bass, and it will even remember that as a "User Setting".

BTW if you're tuning a finely scaled piano with very consistent
inharmonicity from note to note, it usually won't matter which partial
you use, 3rd or 4th. The concert grands I tune regularly will target
blush (ie stay within a few tenths of a cent) both 3rd and 4th partial
RCT tunings, even across the breaks on most notes.

Sorry again to have to disagree with someone with a "Dr. of Pianology"
degree. :) I have a great longtime affection and admiration for Jim C.
Sr. having taken and benefited from and enjoyed his tuning classes at
PTG conventions over the years.

-Dean Reyburn


~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 Dean L. Reyburn, RPT      RPS, Inc.          email:  dean@reyburn.com
 2695 Indian Lakes Road                      web page: www.reyburn.com
 Cedar Springs, Michigan, 49319 USA
 1-888-SOFT-440 (or 616-696-1002)                    Fax: 616-696-8121