Skip to main navigation Skip to search Skip to main content

Improved Mandarin speech recognition by lattice rescoring with enhanced tone models

  • Huanliang Wang*
  • , Yao Qian
  • , Frank Soong
  • , Jian Lai Zhou
  • , Jiqing Han
  • *Corresponding author for this work
  • Harbin Institute of Technology
  • Microsoft USA

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Tone plays an important lexical role in spoken tonal languages like Mandarin Chinese. In this paper we propose a two-pass search strategy for improving tonal syllable recognition performance. In the first pass, instantaneous F0 information is employed along with corresponding cepstral information in a 2-stream HMM based decoding. The F0 stream, which incorporates both discrete voiced/unvoiced information and continuous F0 contour, is modeled with a multi-space distribution. With just the first-pass decoding, we recently reported a relative improvement of 24% reduction of tonal syllable recognition errors on a Mandarin Chinese database [5]. In the second pass, F0 information over a horizontal, longer time span is used to build explicit tone models for rescoring the lattice generated in the first pass. Experimental results on the same Mandarin database show that an additional 8% relative error reduction of tonal syllable recognition is obtained by the second-pass search, lattice rescoring with enhanced tone models.

Original languageEnglish
Title of host publicationChinese Spoken Language Processing - 5th International Symposium, ISCSLP 2006, Proceedings
Pages445-453
Number of pages9
DOIs
StatePublished - 2006
Event5th International Symposium on Chinese Spoken Language Processing, ISCSLP 2006 - Singapore, Singapore
Duration: 13 Dec 200616 Dec 2006

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume4274 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference5th International Symposium on Chinese Spoken Language Processing, ISCSLP 2006
Country/TerritorySingapore
CitySingapore
Period13/12/0616/12/06

Keywords

  • Lattice rescoring
  • Supra-tone units
  • Tonal syllable recognition
  • Tone modeling

Fingerprint

Dive into the research topics of 'Improved Mandarin speech recognition by lattice rescoring with enhanced tone models'. Together they form a unique fingerprint.

Cite this