Department of Electrical Engineering - Columbia University

[SEAS logo]

ELEN E6820 - Spring 2009

SPEECH AND AUDIO PROCESSING AND RECOGNITION

Home page

Course outline

Matlab scripts

Problem sets

Projects

Columbia Courseworks

Matlab resources

This page links to the various Matlab resources associated with the course. Most of these resources are located in the course Matlab directory.

You will need access to a computer running Matlab to complete the course material. The Unix workstations in the AcIS Engineering Terrace (215 Mudd) are one place to do this, and the EE department "ILAB" machines (1235 Mudd) is another. You can also buy a fully-functional "student edition" of Matlab various places including directly from The Mathworks.

Matlab release R2008b is the current version. Any version from 6.5 onwards (release 13) should be fine for the course. The Columbia Bookstore may have some boxed copies of the student version.

2009-01-20:
Here is the diary of the Matlab examples used in today's class. The associated scripts and data can be downloaded from the course Matlab directory.

2009-01-27:
Here is the diary of the Matlab examples from today's class on acoustics and travelling waves, etc.

2009-02-03:
Today's diary demonstrates the use of Gaussian models for simple pattern classification, including generating the figures shown during class. See also this nice online demo of a range of classifier algorithms, and this online demo of fitting a GMM with EM (alternate version here).

2009-02-10:
There are some nice MATLAB cochlea models available from the Boston University EARlab (down as of 2009-02-14, but may reappear). Try downloading the Peterson and Bogert model along with the Basilar Membrane animation routine from their Matlab models page (local copies: PBmodel.m, BManimation.m). Running the model (for frequencies in the range 1000-8000 Hz) then running the animation gives a nice visualization of the behavior of travelling waves in the variable-resonance structure of the Basilar Membrane.
There's also a nice animated graphic that makes the same point on WADA lab auditory mechanics page from Tohoku University. And there's a very nice flash animation of spectral analysis in the cochlea from the Hudspeth Lab at Rockerfeller University.

2009-02-19:
Diary of LPC analysis/ synthesis examples relating to today's lecture. (The functions lpcfit.m and lpcsynth.m are in the Matlab directory).

2009-03-03:
The sinewave modeling stuff from today's class is all documented on my sinewave modeling page. Auditory icon examples can be found at the Sounding Object site. Under "Software" on that site you'll find some Matlab code for synthesizing these kinds of sounds.

2009-03-05:
Diary of Encoding and quantizing speech.

2009-04-07:
Diary of broad phonetic classification of speech frames using MFCC features.


Valid HTML 4.01! Dan Ellis <[email protected]>
Last updated: Tue Apr 07 10:40:09 AM EDT 2009