CSLU: 22 Languages Corpus

Item Name: CSLU: 22 Languages Corpus
Author(s): T. Lander
LDC Catalog No.: LDC2005S26
ISBN: 1-58563-356-9
DOI: https://doi.org/10.35111/zkn2-5x88
Release Date: November 29, 2005
Member Year(s): 2005
DCMI Type(s): Sound, Text
Sample Type: ulaw
Sample Rate: 8000
Data Source(s): telephone speech
Application(s): language identification, language modeling, language teaching, machine translation
Language(s): Yue Chinese, Vietnamese, Tamil, Swedish, Russian, Portuguese, Polish, Korean, Japanese, Indonesian, Hindi, English, German, Arabic, Swahili, Spanish, Mandarin Chinese, Italian, Hungarian, Persian, Czech
Language ID(s): yue, vie, tam, swe, rus, por, pol, kor, jpn, ind, hin, eng, deu, ara, swa, spa, cmn, ita, hun, fas, ces
License(s): CSLU Agreement
Online Documentation: LDC2005S26 Documents
Licensing Instructions: Subscription & Standard Members, and Non-Members
Citation: Lander, T.. CSLU: 22 Languages Corpus LDC2005S26. Web Download. Philadelphia: Linguistic Data Consortium, 2005.
Related Works: View


CSLU: 22 Languages v 1.2 was developed by the Center for Spoken Language Understanding (CSLU) and contains approximately 84 hours of fixed vocabulary and fluent continuous telephone speech in 21 languages and orthographic transcriptions for a subset of the utterances.

The corpus is distributed by the Linguistic Data Consortium and includes the following languages: Eastern Arabic, Cantonese, Czech, Farsi, German, Hindi, Hungarian, Japanese, Korean, Malay, Mandarin, Italian, Polish, Portuguese, Russian, Spanish, Swedish, Swahili, Tamil, Vietnamese, and English.


All of the data in this corpus were collected over digital telephone lines. The files were recorded with the CSLU T1 digital data collection system, with an 8-bit signal and an 8 kHz sample rate, stored as ulaw files. All files are stored in standad 16-bit linear RIFF format.

Each of the 50,191 utterances is verified by a native speaker to determine if the caller followed instructions when answering the prompts. For this release, approximately 19,758 utterances have corresponding orthographic transcriptions in all the above languages except Eastern Arabic, Farsi, Korean, Russian, and Italian.


For an exampe of the data in this corpus, please listen to these Arabic (WAV) and English (WAV) samples.


None at this time.

Available Media

View Fees

Login for the applicable fee