/* ---------- "Cochlear Modelling in Mathematica" ---------- */
I'd like to announce that the following technical report is now available:
"Lyon's Cochlear Model," by Malcolm Slaney, Apple Technical Report #13.
I'm making an electronic announcement because this report is being published
on paper and more importantly as a Mathematica notebook. A Mathematica
notebook is an electronic document that includes equations like any other
symbolic math package but also explanatory text, graphs and pictures. Since
most of the value is in the notebook this paper doesn't really fit into the
normal technical publishing scheme. Thus it is a technical report.
I think a number of audiences will find this report interesting.
1) Speech Researchers and Brain Modelers - This is the main audience.
We are hoping that publishing this report will allow more
people to incorporate these ideas into their own research.
(We also distribute the C/Fortran source code to implement
the models in unix-like environments.)
2) Signal Processors - I think this report illustrates a good way
to write, document and teach signal processing algorithms.
3) Mathematicans - Mathematica finally provides something I've
been waiting a long time for; the ability to create a truly
interactive mathematical report. While the equations are a
long way from looking as pretty as Troff or TeX it makes this
document come alive. If a reader doesn't understand something
then it is easy to plot a new graph or change an existing
equation to fit their own viewpoint.
4) Mac Fans (and NeXT too) - I think this is another form of
HyperMedia. Unlike Hypercard where buttons are just a
navigation tool here the equations and graphs really mean
something and users are encouraged to play with the models
and see how things change.
5) Mathematica Users - As far as I know this is the first report to
be published using Mathematica. I often describe Mathematica
as my favorite word processor (great for the shock effect.)
This report describes how sound is translated from acoustic pressure waves
in the cochlea (inner ear) into nerve firings. There nerve firings are then
used by higher level of the brains for speech recognition and other things.
This model pretty much treats the cochlea as a black box. We assume a very
simple model for the outer and middle ear and then only model the input/output
characteristics of the cochlea. We (and other researchers throughout the
world) are working on more detailed models so this report should definitely
not be taken as the final word.
This report is available (for free) either from me or the Apple Corporate
Library. We can be reached at:
Malcolm Slaney malcolm@apple.com
or
Corporate Library corp.lib1@applelink.apple.com (408) 974-2400
The report includes a printed copy of the report (64 pages) and a Macintosh
floppy. The floppy contains the full text of the report as a Mathematica
notebook, a sample Mathematica notebook reader and the C/Fortran source code
to implement this model on a Unix-like machine. The notebook reader was
graciously provided by Wolfram Research as a means for people without
Mathematica to be able to read the notebook on a Macintosh but does not
allow the reader to do any mathematics or change any graphs.
If you would like a copy of this report please drop me a line. For my
own curiosity please indicate which of the interest groups above you fit.
Numbers 1 and 2 (speech and signal processing) make me feel good and 3 through
5 make my bosses happy. Be sure to send a paper-mail address since this
report is much too large to send through the email systems (>600kbytes
compressed just for this notebook.)
Enjoy.
Malcolm Slaney
malcolm@apple.com
Research Scientist
Advanced Technology Group
Apple Computer
20525 Mariani Ave, MS 22-Y
Cupertino, CA 95014