ETD Question

[Brian Trout]

Hi Richard,

I'm a little late to respond in this thread, and I use TuneLab with David
Porritt's "Calculate" program instead of Cybertuner, but I'll speak a bit on
what I've observed with that.

First of all, David Porritt or somebody with a similar background could
probably give a much more technically accurate account of what I've seen, so
take what I say with a grain of salt.  :-)

When I take inharmonicity measurements with the Tunelab / Calculate program,
after taking the individual measurement, I get a little box which has a
series of numbers representing those inharmonicity levels and the partials
associated with them.  (Sorry I can't be a bit more technically accurate.)
I do pay attention to the numbers that are given.  It'll give a list such
as:

Fund:  0.00
1st:  0.04
2nd:  0.45
3rd:  2.79
4th:  4.12
5th:  8.49
6th:  14.84
7th:  19.09
8th:   28.66
and so on, up to the 12th partial or so.  That's only an example from
memory, and only intended to be an illustration.  Those kinds of numbers are
what I'm expecting to see.  They usually progress more quickly as you move
higher in the scale of the piano, and it's harder to get good measurements
for the higher partials as you get higher as well.

If you get a reading like this:

Fund:  0.00
2nd:  0.43
3rd:  1.81
4th:  24.80
5th:  58.97
6th:  2.76
7th:  11.52
8th:  12.84

or:

Fund:  0.00
2nd:  1.04
3rd:  6.93
4th:  26.81
5th:  11.77
6th:  0.44
7th:  4.01
etc.

...the program usually doesn't know what to do with those numbers to make a
decent tuning.

When I've collected all of the IH readings, and performed the 'calculate'
function, I immediately go to a function called "Graphically Edit Tuning" or
something similar.  This gives me a screen which graphically displays the
tuning.  Just from seeing tuning curves from previous tunings, one can get
pretty good at telling whether a particular tuning will 'work' or not.  When
the individual numbers line up in a reasonable sequence, usually, they will
work well.  If they don't, they won't normally give as predictable a tuning
curve.  (Also, if I just can't get it to do what I want with the
inharmonicity readings and the Calculate program, I can manually tell it
what I want by moving sections of the curve or individual notes to my
liking.)

Some things that seem to matter are, how hard one strikes the string,
whether the hammers are rock hard, (on a grand shifting the action can
help<hint, hint>) what kind of scale the piano has, overall condition, room
noises (air conditioners, kids)... there's a lot of stuff that might affect
those readings.

With the TuneLab / Calculate setup, if I really can't get a good reading
with a particular note that's needed, I will try a neighboring note to see
if one of them will give a more reasonable harmonic layout.  This
programming will allow those numbers to be substituted for the note I
couldn't get.

I think TuneLab and Calculate are excellent programs, and I commend Rob and
Dave both for their efforts.  But by the same token, they are tools, and are
subject to the skills of the operator.  I often tell my customers, the
computer makes a great tool, but a terrible master.

There's some food for thought.  Perhaps Dave will have more technical input
if anyone is interested.

Best wishes,

Brian Trout
Quarryville, PA
btrout@desupernet.net