Re: Effect of duration on pitch perception

[Martin Braun <nombraun@xxxxxxxxx>]

Dear Christian and list,

> ........  Imagine that a pure tone of 500 Hz is perceived with stimuli as
> short as 15 ms, and a pure tone of 600 Hz within, say, 12 ms. A place
> theory would then conclude that after 15 ms there is enough information
> available to go for a pitch of 100 Hz.

This is correct. But then the place theory would in fact be a temporal
theory. Let me explain. The data of Patterson et al. (1983) actually provide
the scenario that you suggested. Their minimum duration times for pure-tone
pitch were:

> 80 ms for 100 Hz
< 40 ms for 200 Hz
< 20 ms for 400 Hz
~ 10 ms for 900 Hz

So indeed, after ca 15 ms the partials of 500 and 600 Hz from their complex
tone with the partials 200, 300, 400, 500, 600 Hz could have excited two
positions on a "harmonic grid" (Erik), which could have been sufficient
information for a neural place-pattern comparison for f0 detection.

The point, however, is that such large dispersion of minimum duration cannot
be explained by differences in cochlear and neural transport times. These
are well-known, and they are much too small between 100 and 900 Hz .

The table above shows that close to eight cycles are necessary for pure-tone
perception. Such an effect is typical for filtering by resonance. In other
words: it is typical for periodicity filtering. It is here that the large
frequency dependence of latencies occurs (as referred to in my first post in
this thread).

In conclusion: a place-pattern theory of pitch that has to include
periodicity filtering should be considered as a dead theory, because the
periodicity filter can filter the periods of the partials and the period of
the fundamental at the same time by the same mechanism.


> I think that temporal coding plays an enormous role in hearing, and that
> this role is not yet appreciated by everybody. However, one should not be
> too fast in ruling out place theories. One point for place theories that I
> can hardly think of how to deal with is the perception of pitch of pure
> tones with very high frequencies, say above 4000 Hz.

I never disputed mapping of frequency. I only wanted to recall that our
data, for a long, long time now, have ruled out the extraction of f0 pitch
via a place-pattern comparison on such a frequency map.


Patterson, R.D., Peters, R.W., Milroy, R., 1983. Threshold duration for
melodic pitch. In: R. Klinke, W. Hartmann (Eds.), Hearing - Physiological
bases and Psychophysics, Springer-Verlag, Berlin, pp. 321-25.


Martin

----------------------------

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