Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!agate!howland.reston.ans.net!newsserver.jvnc.net!news.edu.tw!news!debbie!ccsun20!u8013536
From: u8013536@cc.nctu.edu.tw ()
Subject: Weighted speech signal
Message-ID: <1993Mar19.122443.14857@debbie.cc.nctu.edu.tw>
Sender: usenet@debbie.cc.nctu.edu.tw
Organization: National Chiao Tung University
X-Newsreader: TIN [version 1.1 PL6]
Date: Fri, 19 Mar 1993 12:24:43 GMT
Lines: 20

	Hello, everyone

	  Could any one tell me how could I choose a weighting window in 
	speech recognition ?

	  Usually, we could get better result in speech recognition task
	if we could give each frame of speech signal a weight in computing 
	the vaule of the total distance (DTW) or likelihood function (HMM).

	  But if I can only use a weighting window, what will meet my need ?
	Or, could I use a adjustable window function (with some parameters) ?
	Then What is it ? Could FIR or IIR filter do this ?

	  I need some information !! Thank you for your help !!

					liao,
					u8013536@cc.nctu.edu.tw


	 
