Excitation-pattern models (reply to Peter Meijer)

[SOREN BUUS <BUUS@xxxxxxx>]

Dear Peter,
  You probably considered excitation-pattern models already.  If not,
I'd like to put in a plug for them.  They might be useful and I think=
 that
at least a handful of people might believe in them.

  Look in Zwicker and Fastl (1990) for the fundamentals and for calcu=
lations
of virtual pitch.  For jnds, look in Florentine and Buus (JASA, 1981)
for how to get jnds.  Buus (1997; Encyclopedia of Acoustics) shows ho=
w
to use an auditory-filter model (which is just a type of excitation-p=
attern
model) to provide calculations of frequency effects and signal durati=
on
effects in simultaneous masking.

  In that chapter you will also find an empirical formula for calcula=
tion of
forward masking, but that doesn't qualify as a model.  You could look=
 to
Zwicker (JASA 1976 or 1977; Zwicker and Fastl, 1990) for a more model=
-based
approach to forward masking.  Last, but not least, Brian Moore and co=
-workers
(mostly Brian Glasberg) have published several papers on their versio=
n of the
excitation-pattern model.=20

  I'd suggest that you use the auditory-filter model to produce excit=
ation
patterns and then apply the decision models described in the various =
papers
to obtain the results that you are looking for.

  I am not aware of any work on excitation-pattern models for auditor=
y-profile
discrimination, but there is no reason why it shouldn't work quite we=
ll.  The
Ito, Braida, Durlach (later used by Green) model is essentially an ex=
citation
pattern model, except that it looks only at the physical stimulus--as=
 opposed
to the excitation patterns.  I don't think that the results would be =
much
different if it were to use excitation patterns--and for my money, it=
 would
make the model a lot more compelling.=20

Best wishes,

S=BFren Buus