Newsgroups: comp.speech
Path: lyra.csx.cam.ac.uk!warwick!slxsys!pipex!howland.reston.ans.net!spool.mu.edu!torn!watserv2.uwaterloo.ca!watserv1!crg4.uwaterloo.ca!jxwu
From: jxwu@crg4.uwaterloo.ca ((Jim) JianXiong WU)
Subject: N-best results on TIMIT database
Message-ID: <CsyIF0.AK5@watserv1.uwaterloo.ca>
Sender: news@watserv1.uwaterloo.ca
Organization: University of Waterloo
Date: Fri, 15 Jul 1994 00:52:11 GMT
Lines: 18

Hi, 

Has anyone generated N-best candadite sentences for the TIMIT speech database? 
In order to improve the recognition accuracy, we are currently building more 
sophisticated acoustic models to rescore N-best results. I'd like to know the recognition
performance of top N candadite sentences using either context-dependent 
models or context-independent models.

Thanks for any information.

Jim
-- 
---------------------------------------------------------------------------
* Jianxiong WU,                     |  E-mail: j.x.wu@ieee.org            *
* Dept. of Elec. & Comp. Eng.       |  Phone : (519) 885-1211 ext6988 (O) *
* University of Waterloo            |    Fax : (519) 746-3077             *
* Waterloo, Ontario, Canada N2L 3G1 |                                     *
---------------------------------------------------------------------------
