From: malcolm@interval.com (Malcolm Slaney)
Date: Mon, 13 Feb 1995 14:31:34 -0800
Subject: AM/FM at the Hearing Seminar
Message-Id: <v02110104ab658a6565dd@[192.203.7.70]>

This week at the CCRMA Hearing Seminar we are fortunate to have Brent
Edwards, a  doctoral graduate from University of Minnesota, talking about
his work on encoding AM and FM.  He'll be visiting Northern California
later this week and he'll be presenting his work at CCRMA.   Brent is
trained as both a psychoacoustician and an engineer under Neal Viemeister.

Amplitude and frequency modulation have been a subject of much discussion
at the Hearing Seminar.  Amplitude Modulation shows up as both pitch (when
it's done at >50Hz) and as onsets for slower rates.  FM has been a subject
of much debate at the Hearing Seminar.  Are there separate detectors for
frequency modulation, or is it another example of AM?  These questions get
at how the perceptual system encodes the auditory signal, and which cues
are important for sound separation systems.

        Who:    Brent Edwards
        What:   Perceptual Encoding of AM and FM
        When:   Thursday, February 16 at 11AM
        Where:  CCRMA Library (Top Floor of the Knoll)

See you at CCRMA!

-- Malcolm
P.S.  Coming on Feb. 23, Dan Salzman will be reviewing work he's done in
Neurophysiology on whether directions are encoded as winner-take-all, or
via averaging.

P.P.S.  Last week's ARO conference was amazing: 800 papers on all aspects
of hearing.  Things that stuck in my mind include 1) everybody acknowledges
that the ear is non-linear, although whether there is actual energy being
added to the signal, or whether the impedance gets lower is a subject of
many arguments. 2) Lots more interest in both hearing development in
infants, and how our hearing degrades as we get older (especially with
respect to the cocktail party effect.)  3) Lots of paper about time!!!



Perceptual Encoding Mechanisms for Amplitude and Frequency Modulation

The perception of amplitude and frequency modulation (AM and FM) is
important for many simple acoustic tasks, such as understanding speech and
identifying musical instruments. Any narrowband signal can be defined as a
combined AM and FM signal and much research has focused on detection and
discrimination thresholds using AM, FM and narrowband stimuli. Still,
there is considerable debate over how these signals are peripherally
encoded and subsequently perceived. Using results from psychoacoustic
tasks involving single- and multi-component AM and FM signals,
similarities and differences between their encoding mechanisms are
discussed. A simple vector space model can account for much of the results
that have eluded previous encoding theories. In light of these results,
previous ways of thinking about AM and FM signals may need to be revised.