A grapheme-based method for automatic alignment of speech and text data

A. Stan, P. Bell, S. King

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper introduces a method for automatic alignment of speech data with unsynchronised, imperfect transcripts, for a domain where no initial acoustic models are available. Using grapheme-based acoustic models, word skip networks and orthographic speech transcripts, we are able to harvest 55% of the speech with a 93% utterance-level accuracy and 99% word accuracy for the produced transcriptions. The work is based on the assumption that there is a high degree of correspondence between the speech and text, and that a full transcription of all of the speech is not required. The method is language independent and the only prior knowledge and resources required are the speech and text transcripts, and a few minor user interventions.
Original languageEnglish
Title of host publicationSpoken Language Technology Workshop (SLT), 2012 IEEE
PublisherInstitute of Electrical and Electronics Engineers (IEEE)
Pages286-290
Number of pages5
ISBN (Electronic)978-1-4673-5125-6
ISBN (Print)978-1-4673-5124-9
DOIs
Publication statusPublished - 2012

Keywords

  • acoustic signal processing
  • natural language processing
  • speech synthesis
  • text analysis
  • word processing
  • automatic speech data alignment
  • automatic text data alignment
  • grapheme-based acoustic models
  • language independent method
  • orthographic speech transcripts
  • text transcription
  • unsynchronised-imperfect transcripts
  • utterance-level accuracy
  • word accuracy
  • word skip networks
  • Acoustics
  • Data models
  • Error analysis
  • Hidden Markov models
  • Speech
  • Speech recognition
  • Training
  • grapheme-based models
  • imperfect transcripts
  • speech alignment
  • word networks

Fingerprint Dive into the research topics of 'A grapheme-based method for automatic alignment of speech and text data'. Together they form a unique fingerprint.

Cite this