Information Services banner Edinburgh Research Archive The University of Edinburgh crest

Edinburgh Research Archive >
Centre for Speech Technology Research >
CSTR publications >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1842/3820

This item has been viewed 19 times in the last year. View Statistics

Files in This Item:

File Description SizeFormat
garau-taslp08.pdf373.74 kBAdobe PDFView/Open
Title: Combining Spectral Representations for Large Vocabulary Continuous Speech Recognition
Authors: Garau, Giulia
Renals, Steve
Issue Date: 2008
Journal Title: IEEE Transactions on Audio, Speech and Language Processing
Volume: 16
Issue: 3
Page Numbers: 508--518
Abstract: In this paper we investigate the combination of complementary acoustic feature streams in large vocabulary continuous speech recognition (LVCSR). We have explored the use of acoustic features obtained using a pitch-synchronous analysis, STRAIGHT, in combination with conventional features such as mel frequency cepstral coefficients. Pitch-synchronous acoustic features are of particular interest when used with vocal tract length normalisation (VTLN) which is known to be affected by the fundamental frequency. We have combined these spectral representations directly at the acoustic feature level using heteroscedastic linear discriminant analysis (HLDA) and at the system level using ROVER. We evaluated this approach on three LVCSR tasks: dictated newspaper text (WSJCAM0), conversational telephone speech (CTS), and multiparty meeting transcription. The CTS and meeting transcription experiments were both evaluated using standard NIST test sets and evaluation protocols. Our results indicate that combining conventional and pitch-synchronous acoustic feature sets using HLDA results in a consistent, significant decrease in word error rate across all three tasks. Combining at the system level using ROVER resulted in a further significant decrease in word error rate.
URI: http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=4443886
http://hdl.handle.net/1842/3820
Appears in Collections:CSTR publications

Items in ERA are protected by copyright, with all rights reserved, unless otherwise indicated.

 

Valid XHTML 1.0! Unless explicitly stated otherwise, all material is copyright © The University of Edinburgh 2013, and/or the original authors. Privacy and Cookies Policy