Verituner Update

[Dave Carpenter]

--- Greg Newell <gnewell@ameritech.net> wrote:

> Mr. Carpenter,
>     How is this possible? If notes are measured and
> then tuned "on the fly" is
> not the octave stretching changed also "on the fly"?
> If this is true then some
> notes will be stretched more than others and you
> will have a very uneven scale
> throughout because the tempering is in a constant
> state of flux. Further, does
> it solve for A-440 or is that subject to change also
> based on what it hears?
> Could you please explain?


Greg,

You need to consider the order in which notes are usually tuned. For
example:

1. Tune A4 to Partial 1 = 440 Hz, as in aural tuning. Simultaneously, you
measure the inharmonicity of A4.

2. Since you know the inharmonicity of A4, you also know much of the
criteria for tuning A5; you know at least what it is going to take to get A5
to produce your desired single octave width.

3. At this point you also know something about the inharmonicity of the rest
of the piano (In fact, the SAT 1 used to base its entire tuning from just
F4, not too far away from A4).

4. Next you tune A#4. Knowing where you might tune A5, you can choose an
appropriate amount of stretch for A#4 that will smoothly progress on its way
to your idea of the target for A5.

5. You continue tuning upwards. As you do, more inharmonicity measurements
are made. This refines your idea of the inharmonicity of the rest of the
piano, because now you also see at what rate the inharmonicity is
increasing.

6. You go back to A4 and start tuning downwards. Same methods are used.

As each note is tuned and measured, its inharmonicity measurements are added
into a sort of forecasting algorithm that continually refines its idea of
the inharmonicity for the entire piano. This algorithm considers the
inharmonicity levels measured, the rate of changes from one note to the
next, and the quality factor of the measurements taken.

After each note, target tunings for ALL remaining untuned notes are
recalculated. This tuning is based off the latest inharmonicity forecast for
the ENTIRE piano, plus the actual pitches of the notes that have already
been tuned.  Octave widths are selected based on goals set by user selection
of a style, and individual inharmonicity measurements are used to smooth out
the progression of interval beat rates.

There are a multitude of other note sequences that could be used that may
give the tuning calculator a better idea of the inharmonicity earlier on in
the tuning process. One easy example might be: A4, A3, A3-up, A3-down. Of
course the extreme case is to tune the piano twice. Many tuners regularly do
this anyway. In this case, the entire inharmonicity picture is known before
you start, for every note, and an incredibly smooth tuning can be
calculated. Since the Verituner can save these measurements for later use,
you can recall these measurements on repeat tunings and get a similar
effect.

Having explained all that, it is important to put this all into context. The
One-Step Tuning feature was designed as a TIME SAVING TOOL which allows you
to just start tuning (without a separate measurement step) and still get
good results, comparable to or better than alternative ETD methods. In my
experience, the benefits of measuring the inharmonicity of each note
outweigh the benefits of measuring, say, a span of a few isolated notes and
not knowing whether these readings are representative of their surrounding
notes. Then in the cases where you want to go further and get the best
possible tuning, you can use alternate tuning sequences, or tune the piano
twice, or recall a previous tuning.

Additionally consider this: if you tune outward from A4, the first few notes
will have had the least benefit of knowing the span of inharmonicity in the
piano. However, these are also the notes that are the LEAST AFFECTED by this
lack of knowledge. That is, A#4 would only move very slightly if you changed
your mind about the amount of overall stretch, because it is close to A4
which is the one note that never moves. (If you really wanted to retune A#4
later, there is a ReTune function that selects a new target for any
previously tuned note).

So, the short answer to the original question is that yes, the octave
stretching is changed somewhat "on the fly", but the results of this are
minimized because:

1. Earlier tuned notes close to A4 are the least affected
2. All notes are tuned based on actual pitches of previously tuned notes
3. The general octave stretch is based on the trend of all inharmonicity
measurements taken, not each note individually


Dave Carpenter, RPT
Veritune, Inc.