Frame rate for auditory models

[Christian Spevak <christian@xxxxxxxxx>]

Dear list,

I am using auditory models as a pre-processing stage for self-organizing
timbre maps. Because I am dealing with non-stationary sounds (evolutions of
timbre and/or pitch), I have to devide the sound signal into short frames.
The auditory models, however, give an output signal for each input sample,
corresponding to the probability of auditory nerve responses in the
particular frequency band.

What would be an appropriate value for a frame rate concerning the
perception of timbre (20~25 Hz?), and how
can I achieve the recuction from 22 or 44 kHz without losing crucial
information (windowing/averaging?) ?

Thanks in advance for any suggestions,

Christian

---
Christian Spevak
University of Hertfordshire, UK
Music Department
christian@spevak.de
http://www.spevak.de/