LDC Forced Aligner

Xiaoyi Ma


Abstract
This paper describes the LDC forced aligner which was designed to align audio and transcripts. Unlike existing forced aligners, LDC forced aligner can align partially transcribed audio files, and also audio files with large chunks of non-speech segments, such as noise, music, silence etc, by inserting optional wildcard phoneme sequences between sentence or paragraph boundaries. Based on the HTK tool kit, LDC forced aligner can align audio and transcript on sentence or word level. This paper also reports its usage on English and Mandarin Chinese data.
Anthology ID:
L12-1636
Volume:
Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12)
Month:
May
Year:
2012
Address:
Istanbul, Turkey
Venue:
LREC
SIG:
Publisher:
European Language Resources Association (ELRA)
Note:
Pages:
3405–3408
Language:
URL:
http://www.lrec-conf.org/proceedings/lrec2012/pdf/1065_Paper.pdf
DOI:
Bibkey:
Cite (ACL):
Xiaoyi Ma. 2012. LDC Forced Aligner. In Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC'12), pages 3405–3408, Istanbul, Turkey. European Language Resources Association (ELRA).
Cite (Informal):
LDC Forced Aligner (Ma, LREC 2012)
Copy Citation:
PDF:
http://www.lrec-conf.org/proceedings/lrec2012/pdf/1065_Paper.pdf