Re: There is an easy solution to this problem....

Hi Joseph, Carl & PageWizard

Actually, two or three measures was just a way of speaking. Could
be much more than that.

One could start a repeating section with a note ambiguous between
Ab and G#, then develop from it, going to remoter parts of
lattice, and never visiting the nearer regions for quite a while,
so that for a considerable length of time one just doesn't have
any clues to show which of the two the note is. One could play
with this ambiguity too.

E.g. this plan for a short piece:

diatonic intro in C major.

Now play a single sustained G# (which will come as a "surprise"),
and follow with diatonic section in C# major.

Or indeed, follow with pentatonic section played entirely on the black notes.
Through that entire section, one doesn't yet know how the new notes
fit in with the intro.

Then at the end of the first repeat, play the G# with an E (maybe
as part of E major triad?), and maybe drop down to C as well after
that to really anchor it and make it clear it was a G# all along.
Maybe linger for a while with the C and E, and some diatonic C major,
and recall the intro there.

Then, repeat the section, but this time first note played is Ab,
and entire section needs to be a comma sharper. I.e., on
second repeat it has been transposed up by a comma to Db major,
or poss. the pentatonic scale on Db.

At end of second repeat, play the Ab in unison with C to anchor it
in that way. Here the C will perhaps come as a surprise again as one will
be expecting to get to it via the E as before, rather than directly. Now go off
into C minor at this point, playing with the original diatonic tune in minor
key (I've just tried this out improvising, and it sounds a quite nice
succession of keys). This could just be beginning of a larger piece perhaps.
Or indeed could sound nice just to end it in C minor and fade away I think.

Suppose there are no comma pumps involved.

Then the leisure time adaptive tuning has a pretty clear job of it
I'd have thought.

The real time one has to make a decision. No reason in advance to expect the
repeat to be any different from first time; that is just the composer's
idea, and another composer might do it differently, and may go
somewhere else altogether at the end of the section.

So supposing it makes the G# choice both times, then second time
round it has to pitch slide when playing the C at the end of second
repeat. If it chooses Ab both times, then it pitch slides at end
of first repeat.

Okay, if one accepts pitch slides, it doesn't matter. But then it is
no longer optimal adaptive tuning, and I think one can see that
the leisure time retuning always has an advantage. In this case,
even if one can look many bars ahead, but not to the end of the
entire piece.

I think prob. one will most likely prefer the leisure time tuning
of this particular piece than the real time one with a pitch slide of the
G# to an Ab with the introduction of the C at the end.

Here one would have to look ahead for the entire duration of the repeat,
and we have no idea how long that will be, so the leisure retuning
is the only real option.

I'm sure one can come up with more elaborate ones, and certainly
ones that are more compositionally polished, but this is
a very simple kind of example which does work, and it's easy
to see exactly what is happening.

(I might even write this piece once I've got the root channel
thing programmed; kind of promising somehow).

Robert