Re: maximum 'tatum' speed perception ("John K. Bates" )

Subject: Re: maximum 'tatum' speed perception
From:    "John K. Bates"  <jkbates(at)COMPUTER.NET>
Date:    Tue, 9 Apr 2002 17:42:59 -0400

In a similar vein, when I hear announcers doing their fast-talking "fine
print" on radio advertising commercials I have wondered about the limits of
perceiving intelligible speech.  There must be enough intelligibility in
this kind of speech to satisfy legal requirements, at least if a listener
wants to pay attention. I think that this also presents interesting
technical problems.
  What does the time spectrograph look like? Is there a useful formant
pattern? What about the stop consonants? Is there a proposed cochlear or
CASA model that could show how to make sense out of this kind of speech?
  The human ability to understand rapid-fire speech seems to indicate that
a processing method for extracting linguistic meaning must use a temporal
rather than a spectral approach.

- John Bates