[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: harmonic vs. inharmonic sounds (one last time)

Dear Martin and List,

As a complete amateur when it comes to the physiology of the auditory
system, I am still baffled by the connection between the notions of
"tonotopic registration of frequencies" and a "harmonic template or
sieve".  I don't see how the latter implies the former.

Take the case of two voices on different pitches heard at the same
time.  To create a separate representation for each one, the auditory
system would have to detect two fundamental frequencies (or
periodicities).  Of course an autocorrelation function might show two
peaks, but it's not sure that the human nervous system can do
autocorrelation per se.  The alternative would be the periodicity
detectors that you referred to.

However, there is a periodicity for every harmonic that is
sufficiently intense in the two voices; so the system will register
many periodicities.  To decide which fundamentals are present, it has
to determine some set of underlying "fundamental" periodicities that
would account for the set of registered ones.  In other words we need
a harmonic template or sieve operating on the output of the
periodicity detectors, and it would very much resemble the process
that would have to operate on a place-based analysis of the signal.

So we are really left with only two logically distinct methods of
detecting the fundamental: (1) a harmonic sieve operating on either
place or periodicity information, or  (2) an autocorrelation device
operating on the global signal and picking up the biggest peaks in the
autocorrelation function.

I hope that the signal-processing mavens and physiology experts among
us will correct any errors that crept into my message.  But as I see
it, the problem of fundamental detection is _logically_ the same for
either place or periodicity information.  However, it may be the case
that it would easier to implement such a template physiologically by
working with periodicity detection.


Albert S. Bregman, Emeritus Professor
Psychology Department, McGill University
1205 Docteur Penfield Avenue
Montreal, QC, Canada H3A 1B1.
    Tel: (514) 398-6103
    Fax: (514) 398-4896

On 3/16/07, Martin Braun <nombraun@xxxxxxxxx> wrote:
Dear Al and others,

Al Bregman wrote:

> I have framed the explanation in terms of repetition rate because it
> was easy to do, but it might not be correct.  For example a "harmonic
> sieve" or template, might derive pitch.

We have detailed data on pitch shift caused by the medical drug
carbamazepine. These data exclude the possibility of pitch extraction
via a ' "harmonic sieve" or template '.

These data are compatible, however, with the concept of pitch extraction via
tuned periodicity detectors.

Chaloupka, V., Mitchell, S., Muirhead, R., 1992. Observation of a medication-induced change in pitch perception. J. Acoust. Soc. Am. 91, 2436-2437.

Chaloupka, V., Mitchell, S., Muirhead, R., 1994. Observation of a
reversible, medication-induced change in pitch perception. J. Acoust. Soc.
Am. 96, 145-149.

Braun, M., Chaloupka, V., 2005. Carbamazepine induced pitch shift and octave
space representation. Hear. Res. 210, 85-92.


Martin Braun
Neuroscience of Music
S-671 95 Klässbol
web site: http://w1.570.telia.com/~u57011259/index.htm