Re: Masking release due to the pitch-based grouping? (Joby Joseph )

Subject: Re: Masking release due to the pitch-based grouping?
From:    Joby Joseph  <joby(at)dsplab.ece.iisc.ernet.in>
Date:    Fri, 16 Jan 2004 15:20:20 +0530

HI,
  I think these results can possibly explain monoaural source segragation of
many natural sounds including speech at low frequencies  because of the
following reasons.

a) Most of the natural sounds are generated by vibration of physical bodies
with finite dimensions.
b) So the generated sound complex from a particular source  will have to be in
phase because of the boundary conditions.
c)So if there is a mechanism which phase locks to some time syncronised phase
points  of individual frequency components can distinguish frequency
components from the same source. The previous threads seem to sugges that
there are such mechanisms.
d) Statistically there are often large number of frequency components of a
given source that dont overlab with other sources. (Unless it is a very
difficult scenario in which case we are any way not able to listen to except
the strong source.) Therefore assoiciating frequency components of a source
should be doing the perception equivalent(?).

       joby

--
Joby Joseph
SSP Lab, Acoustic Bldg.      Phone : +91 80 293 2280
Department of ECE            Email : joby(at)dsplab.ece.iisc.ernet.in
Indian Institute of Science          badjoby(at)yahoo.co.uk
Bangalore 560 012            URL   : http://www.dsp.ece.iisc.ernet.in/~joby/
India