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

Re: HC selectivity ... was Re: Physiological models of cochlea activity - alternatives to the travelling wave

At 8:01 PM +0200 10/4/07, Eckard.Blumschein@xxxxxxxxxxxxxxxxxxxxxxxxxx wrote:
Look at what I am offering, the purely mathematical cosine spectrogram. It can be imagined like an ideal so called filter bank, a continuum of local resonators, and surprizingly, it shows in response to a click epiphenomenal traveling waves, a propagating one and stationary ones rolling on the spot at the harmonics, and phase accumulation.

Eckard, this is true, and a point I've often made myself. You can't really distinguish a filterbank model (causal filters like your cosine spectrogram) from a traveling-wave model, if they both have similar magnitude and phase transfer functions that model the cochlea.

However, I can't interpret that as you do, as surprising, or as evidence against the traveling wave. Rather, I take it as an opportunity to make a simplified model, a model that describes the response at each location as the output of a linear parallel filterbank with common input from the stapes. The next step is then to incorporate nonlinearities in a way that describes the real nonlinear behavior of the cochlea; it's here that the physics of the traveling-wave system can lead to cascade-filter models that incorporate nonlinearities in a natural way to give good fits to a wider range of data and phenomena.

Perhaps I'm too closed-minded about this stuff that I thought was settled a decade or two ago. Maybe some day I'll have to do the work to do detailed comparisons of a good nonlinear cascade model against all the latest data of various sorts. Or maybe some grad student wants to do that and I can help and advise...