Newsgroups: comp.dsp,comp.speech
Path: cantaloupe.srv.cs.cmu.edu!rochester!udel!gatech!howland.reston.ans.net!Germany.EU.net!EU.net!uknet!festival!news
From: jwaf@ee.ed.ac.uk (Justin W A Fackrell)
Subject: Question about Multi-band Excitation (MBE)
Message-ID: <D7LLt0.6Eq@festival.ed.ac.uk>
Sender: news@festival.ed.ac.uk (remote news read deamon)
Reply-To: jwaf@ee.ed.ac.uk
Organization: The University of Edinburgh, Electrical Engineering
Date: Tue, 25 Apr 1995 16:00:35 GMT
Keywords: Speech,Coding,HOS,MBE
Lines: 32
Xref: glinda.oz.cs.cmu.edu comp.dsp:18809 comp.speech:5400

Can anyone point me toward a reference for a type of speech coding known as
Multi-band excitation (MBE).  I understand that it is a technique in which use is
made of the fact that speech has a banded structure with different amounts of
deterministic/stochastic information in frequency bands.

In particular, I would like to know how is it possible to determine the level of
stochastic signal content in a particular frequency band ?    If you assume that
the signal in any band is an additive mixture of deterministic (ie predictable)
and stochastic (ie nonpredictable) components, then you must have some way of
deciding how much of each component is present in a given band.

I think I have a method for doing this, but I know little about how MBE 
techniques decide how much stochastic content there is in a frequency band.

Any references, suggestions, or discussions welcomed.

Thanks

Justin Fackrell

===================================================================
Justin Fackrell			email:	jwaf@ee.ed.ac.uk
Signals and Systems Group	email:	J.Fackrell@ed.ac.uk
Dept of Electrical Engineering	WWW  : http://www.ee.ed.ac.uk/~jwaf/
University of Edinburgh
King's Buildings	  	tel (office) : +44 [0]131 650 5655
Edinburgh EH9 3JL		tel (home)   : +44 [0]131 667 6156
Scotland, UK.
===================================================================



