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

Re: Robust method of fundamental frequency estimation.

Zitat von "Richard F. Lyon" <DickLyon@xxxxxxx>:

> You [Arturo] say "Although some of these algorithms were
> initially proposed using a time-domain approach,
> all of them can also be formulated using the
> spectrum of the signal, and that is the approach
> we took."  That may be true, but there are other
> good time-domain correlation-based pitch models
> that can NOT be expressed in terms of the
> spectrum.  For example, the Meddis & Hewitt or
> Meddis & O'Mard models, or Slaney & Lyon models,
> derived from Licklider's duplex theory, which do
> the ACF after what the cochlea model does, which
> is a separation into filter channels and a
> half-wave rectification.

Kaernbach and D. provided psychophysical evidence against these models.
Moreover, there is still scant physiological evidence for the required neural
connections. Corresponding delay lines are missing. Cross-channel links are too
short within midbrain. Furthermore, it is hard to imagine how a sophisticated
structure might have developed step by step. Therefore my favoured explantion
differs from yours. I agree in that cochlear frequency analysis is just the
first step. However, something similar to autocorrelation or cepstrum can be
easily performed via further quite simple neural frequency analysis.

You are quite right in that purely frequency based models cannot account for all
audible features of sound.

> Your poster says that the spectra were estimated
> using FFT, and the next sentence says using a
> gammatone filterbank.  Which is it?  Or both?
> Oh, I see, one says the algorithm and the other
> the model.  Why would you choose an algorithm
> that doesn't match the model?  Why treat these as
> conceptually different things?  An algorithm is a
> computational model, is it not?

While I support this caveat, I do not expect a simple mathematical solution at
all, because the multipolars within CN do not synchronously respond to the
frequency which stimulated the IHCs. Neurons are generally too slow as to
directly convey all audible frequencies. Chopper frequencies in the kHz range
are impossible due to refractory time. So auditory nerve and cochlear nucleus
perform something like downsampling. So harmony and in particular octave unison
are quite natural phonomena. We need not look for their 'learned' basis.