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

Re: Pitch orientation-discriminating feature detectors?

Dear list,

two years ago, Jont and also me obeyed Al's suggestion to purify the list
from  cochlear and heretical nitty-gritty, respectively. This time, I feel
obliged to point you to recent work by Delgutte, see my M240. As far as I
understood, it might be an open question whether or not it is really always
justified to ascribe perceptual peculiarities like sluggishness, dynamic
asymmetry, and masking just to the big unnkown, i.e., the cortex. An
apparent bottleneck of hearing is the limited number of inner hair cells.
Typically ten auditory nerve fibers contact a single inner hair cell of
human. They differ from each other by spontaneous rate and threshold. Just
a small percentage of afferents travel to a slightly more basal position
and contact OHCs. So the response of auditory nerve to a change of
frequency is thought to largely be a Poisson process. It depends on random
stepwise neurotransmitter release. The corresponding delay increases
towards apex of cochlea. Perhaps, that's why we are able to even
distinguish between rarefaction-condensation-rarefaction and
condensation-rarefaction-condensation sequences. Goto my M130 and listen to
pnp.wav vs. npn.wav.


Frequency peaks (upward-then-downward frequency
>modulation) are more salient and produce much lower difference limens
>than frequency troughs (downward-then-upward FM).


>The next question is of course why would it be useful in the real
>world to process differently upward vs. downward frequency