Re: excitalking

I wrote:

>> Be happy to. The basic idea is to take all the notes that are sounding
>> at any given time and to retune them so as to maximize a measure
>> of "sensory consonance," or perhaps more properly to minimize the
>> "sensory dissonance."

to which Gene replied:

>This is funny; I've been thinking about doing this very thing, and
>was going to ask over on tuning-math for easily computed smooth
>dissonance functions, and whether critical bands might work better
>for this purpose than plunging into the murky depths of harmonic
>entropy. I am wondering how you deal with a disonnace function whose
>derivative is not monotonic, or am I leaping to conclusions that you
>did things the way I was turning over in my mind? I also wonder if
>you assume a certain timbre in advance.

Well, hard for me to know what is in your mind... but basically I
parameterized the Plomp-Levelt dissonance curve for pure sine waves
as a difference between 2 exponentials (doing a least squares fit
to find the appropriate parameters). (There are other reasonable
parameterizations, too.) The sensory dissonance is then the sum of
all these exponentials. Since the Plomp-Levelt curves are smooth,
and since the derivative of a sum of smooth curves is smooth, the
derivative is always well defined. In the adaptive tuning, I do a
(stochastic) gradient minimization of this function.

In principle, each note can have its own spectrum, but in the
implementation,
I assume that all notes on a given MIDI channel have the same spectrum,
which I must prespecify.

The model itself is described in detail in
http://eceserv0.ece.wisc.edu/~sethares/paperspdf/consonance.pdf
while the algorithm is at:
http://eceserv0.ece.wisc.edu/~sethares/paperspdf/adaptun.pdf

--Bill Sethares

--- In tuning@yahoogroups.com, sethares@e... wrote:

> The model itself is described in detail in
> http://eceserv0.ece.wisc.edu/~sethares/paperspdf/consonance.pdf
> while the algorithm is at:
> http://eceserv0.ece.wisc.edu/~sethares/paperspdf/adaptun.pdf

I downloaded that, and from the bibliography of adaptun2002 I got the
idea that downloading and unzipping adaptun.zip would be of benefit,
so I did that also. On looking at the unzipped directory, I have one
urgent question--what the heck is it??

If either you or John has actual binaries which do retuning I would
be happy if I could use them.

Gene asked:

>I downloaded that, and from the bibliography of adaptun2002 I got the
>idea that downloading and unzipping adaptun.zip would be of benefit,
>so I did that also. On looking at the unzipped directory, I have one
>urgent question--what the heck is it??

>If either you or John has actual binaries which do retuning I would
>be happy if I could use them.

These are the Max source code, and hence will require Max
(a programming language) for use. I am trying to figure out
how to "compile" these so that others can easily use them.

So the bad news is that I don't yet have executables that you
can download. The good news is that this *should* now be possible
on the Mac, and since Max has just been ported to windows,
should soon be available there as well.

Of course, you realize that John's approach and mine are quite different.
They do happen to coincide (more or less) in the case
when the notes have harmonic spectra - and in this case they
also coincide (moreor less) with the hermode tuning as well.

--Bill Sethares

--- In tuning@yahoogroups.com, "Bill Sethares" <sethares@e...> wrote:

> Of course, you realize that John's approach and mine are quite
different.

I've been pondering ideas on retuning. Is this the place to discuss
them? I'm not sure you and John read tuning-math.

I wrote:

>>Of course, you realize that John's approach and mine are quite
different.

to which Gene replied:

>I've been pondering ideas on retuning. Is this the place to discuss
>them? I'm not sure you and John read tuning-math.

I havent been reading the math forum. As for what's appropriate,
if we get into detailed math-oriented discussions, then Id say
we should move over or continue off-line. But until then....
BTW - I'm going to be travelling for the next 3-4 days, so
I may not reply right away...

--Bill Sethares